Skip to main content
Open AccessArticle

The Extramural Settlement at Vindolanda in the Early Second Century CE: Defining a Glocalized Environment on the Romano-British Frontier


Examining the Roman military settlement at Vindolanda, this article explores the archaeology of the northern frontier of the Roman empire in a glocalization framework, investigating the site during a specific occupation period to understand how the material culture found there operated within its particular local context. The soldiers and the extended military communities of auxiliary settlements that dominated the imperial frontiers make a complicated and intriguing case study because of their origins as subaltern and conquered subjects of imperial rule, followed by incorporation into the Roman army. A close examination of the extramural settlement outside the fort at Vindolanda in the site’s Period 4 (ca. 105–120 CE) allows the opportunity to apply a glocal lens to the architecture, foodways, literacy, and dress preserved in the material record. We are presented with a picture of adoption, adaptation, and retention that ultimately can be understood only as the result of ongoing change and creation in a multilayered imperial context. These spaces and their material culture are fully analyzed here, with careful consideration of the community present at Vindolanda, in order to tease out the unique and novel outcomes that this population created in their local context.1


This article applies a glocalization framework—examining the interrelation of global and local—to the particularly rich case-study site of Vindolanda on the Roman frontier in Britain to better understand the complex and dynamic space of an auxiliary military settlement with a diverse population.2 Employing a glocalization approach offers freedom to explore and interpret the differences in material culture packages, especially when considering their meanings in a specific localized context.3 We employ tools to better interpret the unexpected archaeological finds that in the past were forced into unhelpful binaries such as “Roman” or “native,” both terms that themselves have been found to carry little meaning in the evaluation of material culture.4 The evidence presented here has the potential to contribute a great deal to our understanding of cultural change and responses to conquest because the fort at Vindolanda was established only shortly after the conquest and consolidation of northern Britain. It offers the opportunity to explore the material signatures of a changing society that was very much a part both of the imperial machinations of the Roman army and of a highly localized context at the farthest edges of the empire.

The incorporation of glocalization and globalization frameworks into Roman archaeology holds great promise, but relatively few treatments focusing directly on glocalization as a concept have been applied to Roman and Greek material culture.5 Although theorized for more than 20 years, neither concept has made a major impact in the discipline of archaeology, in contrast to many other social science fields.6 As archaeological tools, especially in the context of Roman archaeology, concepts of both glocalization and globalization have emerged from inquiries about identity in the past and how it relates to material culture.7 Both approaches attempt to understand the processes of cultural change broadly, and they can help challenge and replace the now outdated binaries central to romanization debates.8

In this article, we work with a definition of glocalization that allows for interpretation of “real-life experiences and situations.”9 Providing a single definition is fraught because its use predominates in business and marketing fields and is in some ways an unapologetically modern concept;10 however, its migration to the social sciences and to historical disciplines has been steady, and the useful employment of both global and local elements in understanding historical processes has been evident for some time, if not necessarily by using the term glocal explicitly. Roudometof’s simple definition of glocalization that requires—but is separate from—a global element is key to the discussion here: “Glocality is defined as experiencing the global locally or through local lenses (which can include local power relations, geopolitical and geographical factors, cultural distinctiveness, and so on).”11 Part of understanding the hyperlocal environment in Roman provincial sites such as Vindolanda requires detecting the incorporation of newly adopted habits with traditional ones, which together form a novel material culture package for that new environment. Not every object used in a particular household will be something global that is experienced in a local way. Characteristic of frontier towns in northern Britain are the displacement of certain populations associated with the Roman auxiliary armies, their settlement in new locations, and the heterogeneity of those populations. We use a glocal lens to interpret this new environment, and its physical remnants, to illuminate the range of new and old influences active in shaping life in a new frontier community. This study considers the local environment at Vindolanda in depth and inserts that material culture carefully into a framework created by the known elements of that community—not into preconceived dichotomous categories.

We begin with an introduction to the archaeological context, the material remains, and the local environment at Vindolanda. We selected an area of the extramural settlement from the fourth phase of occupation at Vindolanda (Period 4) as our case study because its rich preservation in oxygen-reduced conditions has left a more complete record of activities in this space than is typical. This evidence is then interpreted using glocalization approaches to conceptualize life in frontier and provincial settlements in Britain.

Case Study: The Archaeology of the Period 4 Extramural Settlement at Vindolanda

Our case study examines a specific area of occupation outside the military fort at Vindolanda in the early second century CE. Vindolanda was located on the northwest frontier of the Roman empire in Britain and was one of the original forts on the early defensive line established in the 80s CE, now called the Stanegate frontier (fig. 1).12 We focus on a specific group of buildings from Period 4, dating to ca. 105–120 CE. These buildings are of particular importance for understanding the nature of the population and the settlement during this early phase of Roman occupation in the north because they are the only remains of the extramural settlements dating to this early period. The growth of a settlement, often called the vicus or canabae, outside the walls of a Roman fort was common and can be found at most military sites in the empire. However, no other remains of extramural settlements at Vindolanda or elsewhere on the northern frontier date to the period so close to conquest. For the most part, only the robust stone remains of the later second and third centuries have been explored in any depth.13 In addition, the first four periods of occupation at Vindolanda (ca. 85–120) are of particular interest because of the extensive preservation of organic materials, including architectural remains and objects made from wood, leather, and bone.14 The analysis, therefore, includes evidence that is simply invisible on most archaeological sites and results in an extraordinarily comprehensive and robust picture.

Fig. 1. 
Fig. 1. 

Map of Roman settlements and forts, including Vindolanda, along the Stanegate Road in what is now northern England. Gray lines represent roads or rivers, and the black line indicates Hadrian’s Wall (© Newcastle University, WallCAP).

Discovery and Excavation of the Period 4 Extramural Occupation

During the excavation seasons of 2012 and 2013, an area just outside the northern rampart of the Period 4 fort was excavated and thoroughly investigated (fig. 2).15 Two structures were found, preserved unusually well in anaerobic environments, both constructed using the same wattle-and-daub technique, one round in form and the other rectilinear. Both buildings had two phases of occupation, made clear by the construction developments and the slight shift in the footprints of both buildings. It is difficult to discern in some cases whether objects were associated with the earlier or later phase, but their association with the complex generally is evident. The life of these structures, including both phases, sits neatly within the Period 4 occupation, ca. 105–120 CE.16

Fig. 2. 
Fig. 2. 

Plan of the remains of the Vindolanda third-century fort (in black), and the Period 4 fort (in red) lying partly underneath and partly to the west of the third-century fort. The green line marks the outer edge (partly proposed) of the Period 4 fort walls. The preserved Period 4 extramural area (also in red) is to the north of the fort wall (drawing by A. Birley; © The Vindolanda Trust).

Both buildings appear to have been used primarily as domestic spaces. They contained typical domestic ceramic assemblages, including mortaria, cooking bowls, and serving dishes.17 The floor levels were built up in the manner of other timber-built domestic spaces in this period at Vindolanda, with a beaten-earth floor softened with a carpeting layer of bracken, straw, and other locally available organic material. This is the same flooring recorded in the barracks and other domestic spaces such as officers’ quarters inside the fort.18 Hearths were not found in either of the extramural buildings, but cooking facilities were located nearby, with a communal oven built into the clay bank 15 m to the northwest of the complex. This is a typical domestic arrangement at Vindolanda for both intramural and extramural spaces. Most domestic spaces inside the fort lacked a hearth, indicating a preferred emphasis on provisions for communal cooking either through rampart ovens, cook houses, or other shared facilities.19 This minimized fire hazards in timber buildings and may also have promoted communal activities. The domestic character of these spaces was also indicated by the discovery in each area of a personal hand-turned quern stone, of a type documented in other domestic spaces at Vindolanda, including barracks, presumably used for limited production of flour in the home.

Settlements found outside auxiliary forts in Britain were dominated by domestic structures.20 Shops were also commonly attached to the front of domestic spaces, but there was nothing about these Period 4 structures to suggest the occurrence of openly commercial activities. Nor can the structures be construed as temples or bathhouses, both often present in extramural settlements.21 Similarly, there was no evidence of metalworking or other activities associated with a fabrica, such as the presence of slag, heavy tools, or the storage of raw materials. Although arguably any space could have been used for storage, the structural diversity and the variety and abundance of daily-use objects found makes storage unlikely as a primary focus for the spaces discussed here. Many storage areas elsewhere on site from different periods produced few objects and even fewer quotidian personal items (e.g., brooches, ligula probes, hair combs).22

The stratigraphy of this area was clearly defined; the fill that leveled this part of the site to support later construction was a very thick layer (a meter in places) of gray boulder clay and rubble almost entirely devoid of artifacts. The clay had been deposited over the area from the nearby digging of the new western fort ditch in the later Period 5, and it is the upcast from this ditch, located only 4 m to the east of the Period 4 buildings, that entirely and quickly buried the remains of the timber buildings, preserving them in anaerobic conditions.23 This material sat above and separate from the occupational levels sealed below, which were distinct in both color and character. The considerable amount of material culture found in the occupation levels compared to the fill above was striking. Objects were predominantly found on floor surfaces and between layers of the bracken and straw carpeting mentioned above. The patterning of object deposition also indicates that these were primary occupational deposits. A small, wattle-and-daub, circular feature was found in the round building that may have been used as a pen or cage but was devoid of any objects or material culture.24 Similarly, in the rectilinear structure, a pit of hazelnuts was found, countersunk into the ground, with no other material culture inside. In both spaces, daily-use objects were strewn around the floors but respected these small internal contexts. If these layers represented random fill brought from elsewhere, they would have more likely covered the internal area of these structures uniformly, without regard for pits or other small features.

The archaeological contexts in the Period 4 extramural buildings and the details of the architecture and finds are described below. A list of the finds and their contexts can be consulted in the appendix. Under consideration here are the internal floor levels and occupational debris of the two structures, which must represent the final phases of Period 4 occupation before departure from the site and subsequent filling of the area for later construction. Outside the structures, a few contexts appear to belong to this same phase of occupation and represent general discard from the habitation of the buildings. These include the drains associated with the rectilinear structure, a ditch that demarcated the northern and western sides of the round structure, and a rubbish deposit between the two buildings. The material that came from these external spaces resembled the artifacts from the internal floors of the structures, leading the excavators to treat this assemblage together. The connection between the deposits was further confirmed by the presence of matching pairs of shoes with one deposited inside and one outside the structure. We treat the material from the floor surfaces of each internal space as certainly part of the assemblage from that specific building, while the objects from the external areas, especially the discard dump between the two structures, cannot be assigned a more specific context than having come from the complex generally. The fill in the shallow ditch adjacent to the round building may have been the debris associated with the first phase of occupation and filled during reconstruction, but this cannot be proven with certainty since the whole area was occupied for such a short period of time. The majority of objects, however, emerged from the internal contexts of both buildings (122 of 194 objects, or 64%).

Architectural Form and Construction Techniques

The two structures had similar features, but at the same time physical differences were immediately apparent. The rectilinear house on the south side of the area sat quite close to the roundhouse to its north, and both had at least two phases of construction. They were both, at least in their first phase, single-room dwellings, perhaps with a porch or vestibule appended. Regarding form, the obvious observation is that one is round (subcircular) and the other rectilinear (figs. 3, 4). The different forms suggest we are dealing with two different builders, owners, and occupants, though this was not necessarily true. Their proximity may indicate a complex with a single owner who, for reasons unknown, decided to build in two different forms. For the sake of analysis, however, we will proceed with the idea that there are two discrete structures occupied by different people or groups and two material assemblages.

Fig. 3. 
Fig. 3. 

Aerial view of the Period 4 extramural area excavation (full scale = 2 m) (A. Stanford; © The Vindolanda Trust).

Fig.  4. 
Fig. 4. 

Plan of the Period 4 extramural area on the north side of the complex, roundhouse at the right and rectilinear house at the left (drawing by A. Birley; © The Vindolanda Trust).

Both structures used the same wattle-and-daub construction technique, despite their obvious disparities in shape. Most internal structures of the contemporary fort were built of wattle-and-daub, but rectilinear architecture dominates. Perhaps more telling is that each employed a distinctive construction practice of filling between the wattle of the outer walls using scraps of leather from discarded tent panels. This construction technique has also been discovered in a roundhouse from the same period located farther west on the site, as well as in the rampart construction of Period 1 (ca. 85–90 CE), when a military unit with the same Tungrian origins (from Belgic Gaul) inhabited Vindolanda.25 This is a building technique that was likely used as some sort of water- or weatherproofing or perhaps as an insulation agent. Such techniques have not been observed in other periods at Vindolanda or at other sites in the region, although the preservation of oxygen-reduced levels in which wattle-and-daub can most easily survive remains relatively rare outside of Vindolanda. It is possible that the construction technique of insulating with leather was local knowledge at Vindolanda, perhaps derived from the Tungrian or northwest European origin of many of the inhabitants at this time.

Material Culture from Inside the Buildings

An interesting picture emerges from the material culture found inside the structures. As noted above, each building had a personal hand-turned quern stone found within to grind flour, but they were of two different styles. The roundhouse contained a beehive or bun quern. This type originated in Britain and was typical in Iron Age British contexts, but it is found only rarely elsewhere at Vindolanda or on other military sites from the Roman period.26 It is certainly a style that comes from a non-Roman tradition. In the rectilinear house next door was found a disc quern, a type that was made in Germany from lava stone and imported into Britain at the time of Roman conquest. The pattern of quern finds elsewhere at Vindolanda shows the predominance of disc querns. In this landscape, therefore, the beehive quern was atypical, especially as it sits within the round timber structure that also contradicts the dominant architectural style of rectilinear buildings elsewhere on site at this time. In contrast, the owners of the disc quern used a tool that was familiar elsewhere on site and at other military sites in the region.27

The rest of the objects in these two spaces consisted of household items of everyday use and objects of daily adornment. For the most part, the nonorganic objects from both the roundhouse and rectilinear house indicate that the inhabitants took part in daily practices that show participation in broad networks of connectivity in Britain and the northwest provinces.28 Both spaces produced knives with Roman-style bone handles, mortaria sherds with makers’ stamps known from Roman-period sites throughout Britain, as well as beads, rings, bracelets, brooches, and decorative studs that would be found on any second-century site in the region or the province. The brooches were all bow brooches, including two examples of the trumpet-type brooch that developed in Britain and became popular throughout the province in the late first and second centuries.29 One of the trumpet brooches revealed a plate soldered to its surface to repair the object, presumably with the intention to prolong its use rather than discard this object of adornment.30 The beads, as well, were consistent with objects found at Vindolanda and throughout the northwest empire, including melon beads and small glass beads. The roundhouse also produced personal items such as game counters, combs, and bronze ligulae (small probes for personal hygiene) that can be found on any site in Britain of the Roman period. The inhabitants of both structures made choices about the objects that surrounded them for daily use, whether for adornment, hygiene, or food preparation, and those choices, at first glance, look quite similar in the two houses. Equally important, however, is the allowance that even objects that appear thoroughly Roman, for instance mortaria vessels for food preparation, may have been utilized in new ways to continue old habits. Ghisleni has expertly shown such results for these vessels in Dorset in southwest England, where residue analysis indicates that the new pottery form was used to continue their traditional cooking practices, demonstrating how they participated in simultaneous processes of change and continuity.31

Other objects from these spaces were somewhat surprising. Both houses produced evidence for writing in the form of iron stylus pens and fragments of wooden stylus tablets. Writing implements themselves tie the actual act of writing to these spaces, not just participation in literate culture. The shallow ditch just outside the roundhouse produced a fragment of ink tablet and an iron ink pen nib. Vindolanda is famous for its ink writing tablets and wax stylus tablets that recorded everything from military supplies to letters between officers’ wives; however, writing is regularly associated with the Roman military and governance, not the community outside the fort.32 The organic preservation reveals other interesting evidence and potentially localized habits in a large stash of hazelnuts found stored inside a countersunk, clay-lined container in one end of the rectilinear house (see fig. 4, the larger gray circle in the rectilinear structure).33 The container appears to have been a store for foodstuffs, and its contents were important enough to require special care, with maintenance and insulation.34

The unique use of leather in the structural support of the buildings discussed above was complemented by everyday leather objects that reveal interesting patterns of potential ownership. A range of leather footwear was associated with these spaces and several interesting shoes were found in the occupational layers of debris associated with the life of the buildings. A pair of carbatinae (house shoes made from a single piece of thick leather) found in the dump between the two houses was of higher quality than most of the commonly found examples of this type, completed with rather decorative cutouts at the ankles. One of the finest shoes from the complex, found in the ditch outside the roundhouse, had a fishnet pattern of uppers that was probably costly to produce and purchase. The presence of this expensive shoe type in the context of the extramural settlement was a surprise. Its type has been hitherto recovered only from high-end spaces like the commanding officer’s residence inside the fort and has in the past been associated with the most elite residents of the settlement.35 At the same time, there were shoes that suggest a very different character to this space. Two examples of what appear to be homemade carbatinae were found, one in the floor debris inside the roundhouse and another in a discard deposit between the two structures. Altogether, these assemblages provide much for interpretation within a framework of glocalization.

What is “Local” At Vindolanda in the Second Century CE?

The evidence from Vindolanda presented above provides an ideal opportunity to examine glocalization approaches: on one level, the area presents one of the most basic and well-known binaries between Iron Age Britain and the Roman conquest—that the locals preferred roundhouses, while the Romans favored rectilinear construction.36 Not long ago, one might have concluded that the roundhouse was built by a member of the indigenous population moving closer to the fort, perhaps to take advantage of the new opportunities brought by a steady market. In this largely binary narrative, the square house was built by the Romans, importing their new construction and living habits to the provinces. In the context of this frontier auxiliary setting, however, the meaning of the terms local and Roman should be interrogated. Neither term sits comfortably, because they reify old, and largely false, binaries rather than the novel outcomes of the reinterpretation of material culture in a new environment.

A clear definition of terms is difficult to pin down, especially the meaning of the term “local” with its range of intended connotations. In the context of an auxiliary fort, an element simply labeled “local” might mean it came from the soldiers stationed there who originated from another part of the empire, or from the noncombatant population living at the fort, or from the geographical area directly around the fort. At the same time, it is difficult to banish entirely the term “local” (or conversely “Roman”) from an article about the Roman empire and the provinces. To balance this uncertainty, we employ a tight definition of “local” to describe the particular environment that was created at Vindolanda by all those who inhabited the site and exerted influence on the archaeological record in the early second century. Our interpretation of the archaeology stems from what we know of this particular settlement in its specific time period. We do not use “local” to mean broadly non-Roman, particularly because of the heterogeneity of an auxiliary fort and the different possible interpretations of the word.37 This approach requires serious consideration of the local environment at Vindolanda and its population to understand how this extramural complex was both familiar and different within its setting.

The northern frontier was guarded predominantly by auxiliary soldiers who were organized into units based on their region of recruitment, for example the First Cohort of Tungrians or the Ninth Cohort of Batavians, which were both units stationed at Vindolanda around the early second century. New recruits to these garrisons could also derive from population groups in Britain, as well as from elsewhere in the empire.38 The extended military community in residence at any of the forts would have included soldiers’ dependents, who joined them during their service,39 in addition to entrepreneurial merchants or other individuals who chose to take up residence near the fort. With an auxiliary unit in residence and with the whole of the frontier guarded by auxiliary units (as opposed to legionary), the likelihood that these individuals held Roman citizenship, or in fact identified themselves as culturally Roman in any way, is slim. To be sure, there would have been a diverse group living at any fort on the northern frontier in Britain, but focusing on what they have in common provides a better sense of what created their specific local setting and how it contrasted to the wider global context in which they existed. What most of the community shared at this point was probably their lack of Roman citizenship, their provincial origins, and their attachment, in varying ways, to the Roman army. Certain members of the group would have been linked by common social status, degrees of wealth, or other social identifiers that commonly unite and divide.40

To what extent inhabitants of the Roman provinces constructed their daily identity in terms of what they were not, as in noncitizen or non-Roman, is unknown,41 but for auxiliary military communities in the early second century, what they lacked could not be ignored. Their home was either in or dominated by the military fort that was the very symbol and physical presence of Rome’s subjugation of the region in which they now lived, and may have been a scene deeply reminiscent of where they originated. The central spaces of the fort would have reminded the inhabitants every day of the global context of their existence, particularly the principia (headquarters building) filled with and promoting the signs and symbols of empire—the so-called shrine of the standards, along with dedications to Jupiter Optimus Maximus, the genius of the unit, the emperor, or the empire.42 The commanders of many auxiliary units, though not all, were Romans of equestrian status.43 Regular visits could be expected from legionary officers or sometimes governors, or others involved with running the province and the defense of the frontiers, all of whom were of the highest status. Even the visitation of a vexillation of regular legionaries would mark a difference. They were the very soldiers that auxiliaries were not—citizen soldiers.

The global context was omnipresent in an auxiliary settlement, and they would have known where that placed them within a broader network of interactions and customs; but most of all they would have been conscious of the differences, the fact that they represented the broad category of non-Roman soldier. The positioning of the two houses under consideration here is also telling, as they sat directly next to the defensive ditch, in the shadow of the fort wall, reflecting a traditional power narrative about those who inhabited these spaces.44 Whether they strived for or hoped to become Roman—whatever that might have meant at the time—is difficult to say, and it is unlikely that every member of the community, family, or group held shared aspirations.45 All these elements were part of the local context at Vindolanda in the early second century, and it is this complexity that offers valid reasons to interpret the local in relation to the global here. Vindolanda and its specific setting only makes sense against the backdrop of the globalized context of the Roman army and the empire.46

The First Cohort of Tungrians in residence at this time was a unit originally raised in Belgic Gaul. Some number of soldiers and perhaps also their families are likely to have identified still as Tungrian, even amidst a somewhat mixed population. A variety of evidence supports this view, including a finely rendered inscription to a water goddess named Ahvardua recovered from a nearby section of the fort’s defensive ditch during the same excavation.47 This is a goddess who is unattested from elsewhere in the empire and is associated linguistically with the Ardennes mountains, the region from which the Tungrians hailed.48 Its dedication by the Tungrian cohort and its discovery in a Period 4 context, when the unit resided on site, point to the probability that there was a component who considered themselves culturally Tungrian and adhered to customs of their homeland.49 It is also telling that this inscription was found broken but with very little wear on the surface or at the break points, suggesting it had not stood in the elements for long before it was dismantled and deposited in the ditch. Its significance or relevance to the residents at Vindolanda diminished to such an extent that the dedication was deposited into the Period 4 ditch, presumably after the Tungrian unit and their dependents departed the site in the early 120s.50 There is also contemporary evidence from the site for the presence of a vexillation of Vardulli cavalry, who hailed from the Iberian Peninsula (Spain). The resulting mix of troops raised from different parts of the empire but serving during the same period at the same site offers a glimpse into a diverse population where a potentially localized koine existed. At Vindolanda in the early second century, there were probably very few individuals who considered themselves Roman in the inexact way that the term is often used in provincial contexts.51

Finally, the relationship between the extramural population and the army is an important consideration in a study of the local environment at Vindolanda. The population of extramural settlements associated with Roman forts were likely a mixed group with varied connections to the unit in residence. As at any other town in the empire, we could expect merchants, shopkeepers, and others who appreciated the commercial advantage offered by a military garrison. The families of soldiers were also housed in the settlement, which may have included extended family members and multigenerational family units.52 The shoes from the two houses under discussion here make it clear that children and adults lived in this space outside the fort. The identities of these individuals is less clear; they may have been the family of a soldier in residence who had little enough status that the fort was closed to them, or they may have been native to the region around the newly established frontier. In this sphere, we can see the role of subordination and potential violence in the everyday existence of these individuals living in a landscape of conquest and occupation. From the start, they constructed their buildings on land that was probably owned and administered by the Roman army, a space that presumably did not allow for private ownership or longevity of use beyond the occupation of the nearby fort, and there may have been restrictions on construction and occupation of extramural land.53 At the same time, some people living near and prospering from the nearby garrison would have been doing quite well indeed.54 Such a scenario would have required negotiations and renegotiations at a local level.

The potential complexities of life and interactions in this period are tantalizingly laid bare in contemporary texts from the site. Several of the Vindolanda writing tablets could be taken to imply that the military was engaged in building projects outside the walls of the early forts, which included buildings with revenue potential like cottages and residences, such as the hospitium referred to in tablet 156, or the bathhouse mentioned in tablet 155.55 We know that some soldiers and their dependents would have paid for their stays when visiting forts, and it may have been in a building such as this.56 An amount for those trips was set aside in the accounts of the fort for a visit to Catterick, listed in Vindolanda tablet 185.57 Under those circumstances it is not far-fetched to suggest that merchants, soldiers, their families, or enslaved persons visiting military bases could have found accommodation, at a price, in the extramural settlement at Vindolanda. Regular travel was a common feature of life on the frontier, and such activities may have provided one of many potential sources of income for the regiment.58 The substantial income from industry, rent, and other means may be revealed in tablet 178, which outlines the revenues of the fort at Vindolanda (the reditus castelli).59 This account refers to a few days in July when the garrison made more than 80 denarii, a princely sum at a time when you could buy a horse at Vindolanda for only two and a half denarii.

What remains debatable is how much the population at Vindolanda had the opportunity to realize the potential for making money from the activities and needs of the army. Kolbeck addresses the role of the “command economy” around military settlements, with particular reference to the land owned by forts as a location of economic revenue.60 He points out that the extramural settlements were a locale of exchange in the embedded economy of the region and concludes that they could have been quite prosperous, while admitting that the ability for any individual to be part of these local economic forces depended on diverse social factors such as rank, status, and other obligations.61 At the same time, while Roman coinage was commonplace in the forts and extramural settlements, its deposition and penetration into the wider rural landscape surrounding frontier forts is far less visible, so the picture may have looked quite different outside the military settlements of the frontier.62

The relationship between military communities and the local Britons also requires some discussion, especially considering that much of the extramural settlement would have comprised noncombatants from further afield.63 At the time the northern frontier and sites like Vindolanda were settled, the native peoples of the area were the Brigantes who, after various alliances and rebellions, were conquered under the Roman governors Petillius Cerialis and Julius Agricola (70s and 80s CE). From the Vindolanda tablets again, we know that the Tungrian unit and other garrisons directly engaged local Britons for goods and services, such as supplying wagons.64 Not all dealings and economic interactions were smooth or successful, as shown in a tablet recovered from the floor of a contemporary centurion’s apartment from the Period 4 fort.65 This document records a plea for mercy from a man who had been beaten with rods for a crime he claimed he did not commit. His defense includes the remark that his punishment was unacceptable because he was a “man from overseas,” implying that a beating might have been appropriate had he been a local Briton. In a conflict situation, the Britons, or even those who were simply not soldiers, were at a distinct disadvantage.

This evidence suggests, in no surprising way, that those living in the extramural settlements were in a subordinate position to the military administration, whether or not they benefited financially from the presence of the garrison. The fort and settlement, especially in the early phases of frontier occupation, were by their very nature temporary and often were entirely removed or remodeled depending on the movement of the units. Under such circumstances, long-term ownership was presumably impossible, and even permission to build or rent would have therefore been temporary. The materials of the fort and the extramural settlement beside it—wood, wattle, and clay rather than stone and brick—reflected this compromise and lack of permanence. The timber forts and their associated extramural buildings had temporary rather than permanent lives, with both settlement and fort remodeled extensively after periods of brief abandonment.66 When the garrison left, so too did the dwellers of the rectilinear house and roundhouse, and the upcast from a new fort ditch sealed the remains of their homes and belongings.67 All of these interactions between people, objects, architecture, landscape, and time at Vindolanda created the local environment in which the archaeology is interpreted here.

Glocalization and Globalization: Roman Material Culture in Local and Global Frameworks

A glocalization framework is one way to interpret this material culture and to help refine our understanding of the site and its people in the second century CE. As discussed above, glocalization concepts provide a fruitful way to understand the reality of life in antiquity.68 Roudometof’s recent work on glocalization has taken a prominent role in academic research in the humanities and social sciences. He leads with the notion that glocalization stands on its own as an independent concept but requires a global element for its proper meaning to take shape.69 Because of the detailed knowledge about Vindolanda available through a unique range of organic artifacts and remains, we can obtain a fine-grained picture of the site and its people that works well for applying a glocalized theoretical approach.

The terms and concepts surrounding glocalization and globalization require some discussion because of the varied interpretations and applications of both concepts.70 At times the similarity and interchangeability of the terms feel inevitable, and this characteristic is found in much of the principal research on the subject from the last decade. The present study adopts the position that the approaches are different, though closely related, and that they complement each other in useful ways. We do not intend to argue here whether the Roman empire was globalized, but rather we start from the premise that many of the tenets of globalization theory help produce fruitful insights into the imperial Roman world.71 At the same time, globalization does not mean homogenization,72 and it is in the nuances and differences in assemblages that glocalization approaches become most useful. Through a glocal lens the meaning of material culture is articulated better within the context of patterns across a site, between sites, and between regions and provinces, allowing a clearer view of different situational experiences in the empire. Such an approach is valuable when making sense of the cascade of data that often flows from a Roman-period frontier settlement, especially the complex locations of Roman military activity.

The last decade of research in Roman archaeology has seen a good deal of work with the goal of reorienting questions about identity in the empire, and it seems that both globalization and glocalization studies help in this regard. Versluys has offered a useful update on the romanization debate, focused closely on the concept of the “object turn” to interpret material culture.73 Together with work by Pitts, these ideas have recently been elaborated to focus closely on the notion of “objectscapes” to better understand the material record and its relationship to the people who created it.74 Static approaches to the significance of material culture are discarded in favor of using the now familiar concept that the objects themselves held agency.75 These arguments offer a starting point for our study, but the independent agency of objects does not dominate our analysis. Although we scrutinize a traditional assemblage of material from a single archaeological context, this is interpreted with the understanding that it was part of a larger landscape of meaning.76 The interpretation of the assemblage cannot be realized without a deep understanding of the overall context of the site and its inhabitants and, as Versluys urges, with a clear eye on how these objects operated in their specific contexts.77 New and multiple meanings were often at play in the use of material objects within the varied local and regional contexts of the Roman provinces and frontiers.78 Moreover, single objects themselves are usually only a fragment of the complete picture. The whole assemblage and its overarching context within a site and particular landscape give depth of meaning, which is missed when we divorce an object or assemblage from the wider context of its use.79

The relationship between the use of an object and the identity of the user is also important. The adoption and use of objects that were available throughout the Roman empire does not inherently mean the user identified as global or as someone we might call Roman.80 Owning and using an object, even on a daily basis, is very different from understanding its full origin in the manufacturing and commercial networks of the empire and internalizing that meaning as part of one’s identity. Presumably, very few people in antiquity could identify the origin of their daily goods or thought consciously about the difference between the previous significance of an object and its shift in meaning in a new setting. It is far more likely that people and objects existed in the reality of their immediate surroundings, and the meaning of an object was first and foremost tied to notions of value in that specific community. For instance, a localized network of value might rely more heavily on the material from which an object was made rather than its form or function. The reinterpretation of objects with novel meaning or new functionality did not necessarily carry with it the knowledge by its user that they shared taste or value with someone elsewhere in the empire.81 This shift in value is much more difficult to interpret from the archaeological record and requires careful attention in our interpretations.

A good example of the pitfalls of our assumptions is found in the debate about resistance by provincial communities.82 The absence of goods that are typically considered Roman in origin (whether that label is correct or not) or the adherence to traditional objects and habits was once interpreted as intentional resistance to romanization in a particular location. Jiménez has shown the danger in this reasoning and the discursive use of the past in Baetica (southern Iberian Peninsula) as one example of its problematic outcomes.83 Her investigation of burial contexts shows that the retention of traditional forms was not employed necessarily to display resistance, and the use of Roman material culture was not inevitably part of integration and acceptance. The transformations of meaning can best be understood within their context of use by having knowledge also of potential local hierarchies, social values, and power structures. A glocalized theoretical approach becomes a valuable addition to interpretation because of its prioritization of the specific place and time—and everything we know about the site and its population at that point—to explain the material remains in that particular context.84 The detailed picture we have of the community at Vindolanda described above puts us on relatively steady ground to explore ideas of cultural bilingualism and hybridity.85 There is still a place for individualism and situational constructions of identity that were malleable.

Remembering what is at stake in any specific context is particularly worthwhile in the archaeology of the Roman empire because the overt context of imperialism and subjugation would have deeply affected many of those living in the conquered territories under discussion.86 González-Ruibal warned of the danger that certain postcolonial applications foster when the softer side of colonialism, such as hybridity and negotiation, is emphasized, which give much promise to individual choice.87 It is important to consider agency, but this should be done without masking the power differentials that engender inequality and were prevalent in the Roman provinces.88 In a more direct indictment, Fernández-Götz et al. critiqued the ideas around “material agency” and the “object turn,” precisely because of the ease of ignoring the social hierarchies that produce extreme inequality.89

Power differentials on a Roman frontier, especially within an auxiliary military context such as existed at Vindolanda, are a key component to understanding life in these settlements.90 The community of soldiers and the extended military communities of auxiliary units that dominated the frontiers of the empire are a case in point.91 They were at once a diverse population of both the conquered and conquering elements of empire. Their origins as soldiers recruited from the subjugated provinces of the empire surely mark them as subaltern;92 at the same time, their incorporation into the Roman army centers them directly as primary agents of further conquest and violence.93 This situation created a complex overlay of identifying factors for soldiers and their dependents, many of whom joined them in military communities across the Roman frontiers. Inclusion in the Roman army might expose a soldier to new ways of expressing power and status, such as material objects that articulated belonging to the military and its inherent authority.94 This does not mean, however, that the same individual disavowed their origins or belonged to no other identity categories.95 This is where looking through a glocal lens offers a more realistic understanding of identities in the provinces, with frameworks that discard binaries and approach such communities on their own terms. For an auxiliary soldier or his family living at Vindolanda, the meaning of an object was created within that local context; some value may have been gained by its association with global ideas or conversely with something more akin to traditional customs, but not certainly. Objects may have taken on entirely different symbolic meanings in novel contexts.96

Moving between global and local scales and between different contexts of site, region, and province is critical for understanding the globalized Roman world. Highlighting the nuances of very localized contexts and understanding how they differ from patterns seen elsewhere is where a glocalization approach becomes useful. This is not to say that the meaning or appearance of a glocalized site will look the same everywhere. The material signatures of rural sites with no military connections located in the same region as Vindolanda look dramatically different, though they may each be considered globalized or glocalized in their own particular way. This variety and flexibility should be a hallmark of glocalization studies, not a deterrent to its fruitful application.

Interpreting the Local and Global at Vindolanda

The challenge is to interpret what this material culture from Vindolanda means within this very local context of time, place, and population. An interesting feature of the evidence is its variation. On a quick glance one might like to fall back on old paradigms: that is, the roundhouse dwellers built in a time-honored form because that is what they knew best; they had a local-style quern because foodways can be personal and traditional; but they adopted new and interesting goods such as fancy knives and flashy brooches. But this picture is too simplistic and probably false. A glocalization approach feels especially important at a site like Vindolanda, where the local context of this military community had some specific characteristics that are quite particular to this frontier and its inhabitants.

Starting with the construction style of the two spaces under study here, the basic binary that local Britons built roundhouses and Romans preferred rectilinear styles is of no real interpretive use. In the past few decades, it has become clear that the roundhouse concept did not disappear at the point of conquest but rather remained a dominant structural form, especially in rural communities.97 This same structural form is visible elsewhere at Vindolanda in the same period, but the archaeology is less robust and more difficult to understand. Outside the Period 4 fort defenses on the west end of the site, another round timber building stood next to a rectilinear timber structure. This complex was investigated in the 2010 excavations, but it was badly damaged by later Roman construction and cannot be discussed in detail.98 Inside the fort, excavations in 2004 revealed the ephemeral remains of a similar subcircular wattle-and-daub structure standing next to square buildings,99 suggesting this form was not simply a curious decision by a single household but may have been a recurring arrangement.

Even a full century after conquest of this frontier region, the roundhouse style persisted at Vindolanda. In the early third century, the inhabitants of the settlement constructed an intriguing field of round hut structures.100 Dozens of roundhouses were built with stone foundations and were lined up in neat rows of five huts back-to-back to make sets of ten (figs. 5, 6).101 The complex of roundhouses reused building materials—stones, opus signinum, bricks, and tiles—from the demolition of an earlier second-century fort. The assemblages from inside the third-century roundhouses are entirely different in both quantity and variety compared to the material recovered from the military areas of the site at this time. The Severan roundhouse complex at Vindolanda, with its neat rows, appears to have been a planned development and may have been encouraged or administered by the military garrison. At the same time, their form shows the builders were not following architectural norms and that the desire to build and live in roundhouses endured well into the Roman period. The onset of Roman conquest and occupation did not remove traditional building habits but rather situated them with a new backdrop.

Fig. 5. 
Fig. 5. 

Plan of the fort in Period 6B (ca. 200–212 CE) with a series of roundhouses constructed to the east (drawing by A. Birley; © The Vindolanda Trust).

Fig. 6. 
Fig. 6. 

Aerial view of the early third-century CE roundhouses as they sit below the later stone remains of the mid third-century fort (A. Stanford; © The Vindolanda Trust).

From this perspective, the roundhouse as a form is only different and apparently local in contrast to the rectilinear structure that stands next to it, the latter representing novel ideas introduced with Roman conquest.102 The persistence of roundhouse construction well after the point of conquest could in some sense be interpreted as an act of continuity of tradition, but can only be appreciated as such within the global context of an otherwise Roman setting.103 Once rectilinear structures are, at first, an option and then the dominant form, the meaning of a roundhouse necessarily changes from a normative dwelling to a divergent practice. In that way, the seemingly traditional element is not merely stagnant but instead is part of changing practices and evolving ideas. The local needs the global and the global needs the local for either one to receive the label. It is the whole context with both elements present that is a new formation of what a settlement can look like in this time and place.

At the same time, certain objects stand solidly outside the binaries of Roman and non-Roman categorization. For instance, much of the material evidence that represents everyday life in these houses is preserved in wood, such as tools, container lids, and barrel stoppers. Whether made of local or imported materials, these objects do not betray any affiliation with an origin or identity, providing a needed reminder that much of daily life was distinct from these binaries and ought to be approached from a neutral perspective from the start. Other objects from the Period 4 extramural complex tell a contrasting story. There were certainly plenty of items, such as brooches, hairpins, and jewelry, that indicated the inhabitants had access to and chose to participate in trends that were common to sites in the western Roman empire at this time. However, if we interpret these items through a glocalized lens, an alternative picture presents itself and raises the notion that globalization does not create homogenization.104 In a glocal framework, these objects attracted new meanings and more likely worked in nuanced ways within their new context.

A good example is the significant amount of jewelry found in these spaces, such as beads, bracelets, and rings, that align at first glance with a Roman-style material culture package. However, if we look elsewhere on the site to interpret the use of these types of objects, a different picture emerges. The carved gemstones (intaglios) that were set into many finger rings were part of a strong Graeco-Roman tradition that spread ubiquitously across the empire.105 It has been argued that in the context of the army, images of Graeco-Roman military heroes, such as Achilles, carved on a gemstone may have conjured notions of Homer’s Iliad and similar heroic stories.106 This may have been true for officers who were raised in Roman cultural traditions. From the perspective at Vindolanda, however, where even officers sometimes originated from the provinces,107 the image takes on different meaning. Certainly, by displaying a finger ring set with a gemstone, the wearer adopted a very Roman way of showing status, which had meaning in the hierarchical context of a Roman military community. However, rather than imagine that a Tungrian or Batavian soldier originally from northwestern Europe conjured an appreciation of Homer, it seems more probable that this image of Achilles carried personal connotations of general military strength, which we know was also a source of status in Batavian culture.108 Thus, the object itself articulated status in the hierarchical context of the Roman army, but the incised image may have resonated with the values from the soldier’s homeland. To value the object principally because of its Graeco-Roman artistic expression misses the point within the wider framework of its discovery. In this particular place and at that specific time, the object developed a new set of implications, one that may have been quite far removed from a straightforward classical model. The beads, rings, and bracelets from the extramural houses may have taken on similarly new symbolism in this context.

Some features of the finds assemblage that betray daily activity in the Period 4 houses were surprising, for instance evidence of literacy and written communication. Both houses produced iron stylus pens and fragments of wooden stylus tablets. Writing is not often associated with those who inhabited extramural frontier settlements, especially at this early date, but more often with the identity of an educated elite and especially the bureaucracy directly related to the army. Writing implements in the context of a military fort are generally considered a Roman influence and are part of debates about the Latinization of the Roman provinces.109 Implications of the presence of writing implements in this space are complex and are better understood through the lens of class, status, and social identity of the extended military community.110 Mullen has recently brought together a great deal of evidence that indicates literacy levels were higher than previously thought in provinces like Roman Britain. Her arguments surrounding what she terms “socio-literacy” shift our old view of the mostly nonliterate provincial population that was presented by Harris three decades ago.111

We may look to other provincial regions to hypothesize the meaning of writing culture for the inhabitants of these spaces. Derks and Roymans proposed that the Batavians in the Rhine delta adopted writing practices specifically to stay in touch with family members who were recruited into the military.112 This argument is intriguing for its explanation of the adoption of a new habit to suit the particular needs of this social group. Thousands of soldiers were removed from their homelands to serve in the Roman auxiliary army, and new skills and habits were needed to maintain traditional familial networks.113 At Vindolanda, we have several letters that suggest families stayed in touch and greeted each other by way of letter writing. Many tablets greet daughters, sisters, and other family members, while tablet 650 sends greetings to “Capito and all my fellow-countrymen and friends.”114 Perhaps similar reasons motivated those living outside the Period 4 fort to adopt writing culture. If indeed military families lived in these spaces, as was certainly true for a good part of any extramural military settlement, perhaps the adoption of communication tools was necessary here for the first time as their mobility and separation from their home communities increased.115 A glocal interpretation allows proposal of more realistic scenarios about the new habits created to cope with displacement and change.

The evidence for writing found in the Period 4 extramural houses makes clear that old expectations of who might be found in these settlements and what we infer from them are changing dramatically, moving us beyond simplistic categorizations of an illiterate rural Britain and the Latinizing force of the Roman army. Mullen and Tomlin have both avoided the temptation to estimate rates of literacy in Roman Britain, and this evidence supports such a rejection.116 Whether the inhabitants of these spaces were part of a 5, 10, or 15% minority of literate people in the province gets us no closer to a better understanding of their lived experience.

Other evidence from the extramural houses is intriguing because of its indication of potentially localized practices, such as the significant stash of hazelnuts found in the rectilinear structure. The Roman diet was no stranger to hazelnuts, but they are not usually found in such large quantities on Roman-period sites in the region, and in the context of the northern frontier in Britain they may have a unique interpretation.117 The pre-Roman inhabitants of Britain appear to have used hazelnuts as a staple, and large numbers are found on most Iron Age British sites.118 If we consider cuisine and food preparation as an important form of cultural expression, then we would more readily associate a large stash of hazelnuts with a non-Roman European identity.119 At the same time, the inhabitants of both of our extramural houses clearly adopted food preparation practices that were consistent with the networks of goods and ideas emerging from the Roman conquest of the northwest provinces. Sherds of mortaria, a vessel type introduced to Britain with the Roman conquest, and other wares that show connectivity with trade networks throughout the northwest provinces in this period, were found in the houses, together with objects with a probable British origin, such as the beehive quern.120 What resulted was a new discourse of daily life that allowed for both traditional and adopted habits to be part of a newly developing cultural milieu.

The recognition of two spheres of interaction—private versus public—is one way to think through this evidence.121 Van Driel-Murray has argued that we can see the maintenance of traditions brought from the homelands of soldiers in the domestic sphere and in the settlements around the forts specifically because women and families accompanied the men into the military environment.122 At Vindolanda, it has been shown clearly that at least some of the Tungrian soldiers in residence in Period 4 were accompanied by their families, some of whom even lived in the fort itself.123 In other spheres, both ancient and modern, it has been shown that women are often the conduits through which tradition is maintained.124 This can be apparent in domestic spheres through traditional ceramic assemblages used for food preparation, while the public-facing dinnerware—what guests see if invited to the home—may be a new form that advertises something different, perhaps participation in novel ways of expressing status with access to wider networks of interaction.125 The formula can also be inverted, where the public or display pieces retain tradition in order to uphold social customs and hospitality relationships when a group is displaced together.126 The ceramic assemblages of the roundhouse suggest the former scenario, with imported wares such as terra sigillata used as a public-facing feature in the home. The presence of terra sigillata in the roundhouse is especially interesting because it diverges from some of the other, more traditional material culture choices. The cooking wares were simple and nondistinct, but the beehive quern introduces further complexity. It was a British type that predominated in the Iron Age. Its use may reflect a local habit of food preparation that either continued or else was adopted from the region directly around Vindolanda and was not necessarily brought from the home region of the Tungrians. This object stands out against the choice to use terra sigillata for serving, which shows participation in broad imperial networks, at least subconsciously.

The shoes from the two houses also reveal a pattern of different choices simultaneously. On the one hand, both buildings produced evidence of the adoption of shoe styles available throughout the northwest empire. Shoes, in particular, took on a standardized form, apparent in their style of uppers and the nailing patterns on the outer sole, that clearly stemmed from Roman ideas about footwear manufacturing and Roman networks of trade.127 Shoes from the Roman period are very different from the footwear worn by Iron Age inhabitants of northwest Europe and there was consistent uptake of this advancement throughout the northwest provinces.128 A feature of both the rectilinear and roundhouse was the presence of good-quality footwear in the occupational deposits, showing that inhabitants took part in a shared style that was common throughout the western empire.129 On the other hand, there were two shoes, from different pairs, associated with the roundhouse that appear to have been homemade and might mimic a carbatina style of shoe, which was a Roman form made from a single piece of thick hide and lacked hobnails. These homemade shoes are quite curious because, although we can see that the inhabitants did indeed own shoes that were consistent with regional and interprovincial patterns, they also participated in a homespun attempt to imitate those same shoe styles.

Carbatinae are ubiquitous at those Roman-period sites throughout the northwest empire that produce leather objects. They are, however, also quite similar in form to the simple shoes manufactured from one piece of leather found in Iron Age European sites.130 Mould stresses, in her discussion of homemade shoes from the Medieval period, that this basic type of shoe has been worn from the earliest prehistoric period through the 20th century.131 She posits that the type might have been worn by those living in rural contexts and by those who may have lacked access to a full range of choices.132 Only a few homemade shoes are known from Roman contexts in northwest Europe and an interesting aspect of Mould’s discussion stresses that they are found alongside otherwise more sophisticated footwear, much like what we see in the Period 4 extramural complex. On the one hand, some footwear examples from the complex indicate participation in up-market trends, while others suggest a need to make do and mend. Another shoe from the complex shows the same trend to fix objects, with at least two patches added to the shoe uppers to extend the life of the item.133 Though this shoe eventually ended up in the discard pile between the two structures, it appears that there was some necessity to extend its use considerably. Indeed, the two shoes of this type found associated with the roundhouse were an attempt to manufacture footwear from a single piece of hide; but were they imitating a Roman carbatina or falling back on a familiar style from the owners’ homeland that was also easy to produce? At the same time, the very high-end shoes with fishnet uppers that were part of elite Roman-style adornment in other spheres at Vindolanda were found in the same space.134

If viewed through the lens of public and private spheres, some meaning in the footwear assemblage starts to take shape. The homemade carbatinae were simple and were meant to be worn at home, in the private sphere. Meanwhile, the footwear with fishnet uppers that sent a message of status and belonging were shoes to be worn in public, something to be seen in. If both were owned by the same family unit, both motivations make sense. In the local setting of this auxiliary fort, symbols such as attire and adornment can identify one in this new hierarchical space of the military, but customary habits allow for retention of that original identity that had meaning to the household or the cultural group displaced to the new location. In the end, the hybridity and the ownership of both items are what characterize the new context, one that is created by both old and new habits together. Perhaps the desire to fit into a new order, coupled with identity stress, contributed to the employment of objects of different origins together to provide meaning in this particular context. The specificity of the local context is highlighted even further if, as mentioned above, the origins of objects often were probably not known to their users.

We are left to ask, who were the inhabitants of these spaces and what spheres of interaction dictated the material culture in their lives? Both spaces were part of regional and interregional networks that brought novel goods and ideas that could be considered global, while at the same time demonstrating the endurance of traditional habits that were also part of an emerging cultural milieu at Vindolanda in this period. The pottery, hairpins, brooches, and other items that were part of wider consumer networks of the western empire demonstrate a shared desire or, at the least, the ability to advertise these associations. But how well understood were those associations by their owners? We have a good understanding of when, where, and how objects moved around the Roman world, but this knowledge was not necessarily shared by the second-century CE inhabitants of Vindolanda.135 The networks at Vindolanda in the early second century were also not exclusively economic or object-based. The evidence for writing discussed above revealed the potential that literacy and letter writing could have been a way that a displaced group, such as a Tungrian community, maintained connections to their homelands. This would represent a geographically circumscribed network that may have only included two nodes, northern Britain and the Tungrian homeland.

We must consider closely the choices, such as the roundhouse building form, that assert a divergent position within this specific community to think through the processes of dynamic change at this time. The roundhouse may have advertised difference amidst a sea of rectilinear construction, especially standing directly adjacent to the rectilinear structure, but at the same time, other features may have spoken louder.136 Though certainly not a dominant form, there were at least one or two other round structures in this community.137 Whether it was an intentional announcement of habitual practices from a Tungrian or Vardulli homeland, the continuation of practices by a Briton, or even the adoption of something new by anyone in this environment, we cannot say with certainty. We could look to the two other subcircular timber structures discussed above—each of which stood next to a rectilinear timber building in Period 4 phases—to suggest this form may have been a recurring phenomenon at this time. The round form, whatever its meaning, must have been a meaningful choice that was legible in the context of this early second-century environment.


In this article, we have presented a way to read some of the archaeology at Vindolanda through the lens of glocalization. The interpretation of this evidence is specific to the discrete context of this site—a settlement of individuals who had been dislocated from their original sphere by their entry into a new environment of the military and its community.138 The construction of a roundhouse in this place is only curious when the house next door and most of the nearby architecture is the opposite—not round. The roundhouse can only be an anomaly because it sits solidly within a globalized world: a Roman fort and settlement dominated by rectilinear architecture and populated by groups and individuals originating from different parts of a vast empire. The interpretation of these spaces requires both a local and a global lens. The roundhouse, surrounded by nearly all square buildings, preserved some form of non-Roman tradition and perhaps advertised that fact boldly, as it stood directly next to and in sight of the walls of a Roman fort. However, the result is neither strictly Roman nor strictly local in literal terms, but rather is a combination of elements that created something novel, something glocal.139 As we consider how everyday choices came together with ingenuity and knowledge applied to old customs, such as how to build a house, or to novel ones, such as that which came to represent status and prosperity, we may acquire a better understanding of this specific frontier context.

The result at Vindolanda may look very different from a rural site down the road. Each community tapped into its own networks (perhaps strengthened by written communication) and had different spheres of meaning and value. A rural site in the north of Britain was not as well connected to other parts of the empire as a military site in the same region, whether a legionary garrison full of Roman citizens or an auxiliary fort with a heterogeneous population from different parts of the empire. Glocalization approaches should always expect variation and diversity when interpreting objects or practices at different sites. In this study, a basic tenet of glocalization theory rings true—this scene makes no sense without both the global and the local elements playing a part. To what degree each predominates reflects the lives of the inhabitants of the site in antiquity.

Appendix.  List of Objects and Their Find Contexts from Vindolanda Period 4 Extramural Area

C. = coin, L = leather, SF = small find, W = wood, WT = writing tablet

Rectilinear Building Interior Floor Levels

Find no.Object identification
SF-17472Copper-alloy sheet
SF-17495Bone knife-handle
SF-17498Ceramic gaming counter
SF-17507Lead waster
SF-17530Mortarium sherd with maker’s stamp
SF-17551Stylus pen
SF-17552Knife with bone handle
SF-17563Horn handle
SF-17565Copper-alloy bracelet
SF-17603Iron plate
SF-17614Iron arrowhead
SF-17623Bone handle
SF-17628Glass bead
SF-17629Trumpet brooch
SF-17649Trumpet brooch with evidence of repair (riveted metal plate)
SF-17651Bone key
SF-17658Stylus pen
SF-17661Shaped bone
SF-17697Bone handle
SF-17698Basketry fragment
SF-17699Bone ring
SF-17700Bone hairpin
SF-17701Bone key
SF-17704Joiner’s dog
SF-17705Copper-alloy decorative stud
SF-17706Amber bead
No find no.Disc quern, complete
L-2013-3Scrap leather
L-2013-4Leather tent fragment
L-2013-5Scrap leather
L-2013-6Leather patch
L-2013-7Shoe scrap
L-2013-9Scrap leather
L-2013-18Leather thong through hole in rock
L-2013-21Scrap leather
L-2013-22Scrap leather
L-2013-26Incomplete shoe (child)
L-2013-27Incomplete shoe (child)
L-2013-28Scrap leather
L-2013-29Scrap leather
L-2013-32Scrap leather
L-2013-33Incomplete shoe sole (adult male)
L-2013-34Shoe scrap
L-2013-35Almost complete boot with uppers (adult male)
L-2013-42Leather tent fragment
L-2013-44Scrap leather
L-2013-47Scrap leather
L-2013-50Leather tent fragment
L-2013-51Incomplete shoe (unknown owner)
L-2013-52Incomplete shoe (female or adolescent)
L-2013-53Incomplete shoe insole (female or adolescent)
L-2013-55Scrap leather
L-2013-57Scrap leather
L-2013-58Scrap leather
L-2013-59Scrap leather
L-2013-60Scrap leather
W-2013-2Wooden bung
W-2013-3Wooden comb
W-2013-4Wooden key
W-2013-5Wooden comb
W-2013-7Wooden bung
W-2013-8Wooden mallet
W-2013-9Wooden comb
W-2013-11Wooden lid (from small barrel)
W-2013-15Wooden bung
W-2013-16Wooden lid
W-2013-17Wooden lid
View Table Image: 1 | 2

Roundhouse Interior Floor Levels

Find no.Object identification
SF-16939Maker’s stamp on terra sigillata sherd
SF-16940Melon bead
SF-16942Lead ring
SF-16943Armor fitment
SF-16944Armor fitment
SF-16947Ligula probe
SF-16948Ligula probe
SF-16950Copper-alloy stud
SF-16951Bolt head
SF-16952Iron needle
SF-16956Graffito on terra sigillata sherd
SF-16958Copper-alloy brooch
SF-16959Joiner’s dog (iron)
SF-16960Bone handle
SF-16965Copper-alloy stud
SF-16966Serrated iron saw blade
SF-16967Maker’s stamp on terra sigillata sherd
SF-17570Graffito X+ on terra sigillata sherd
SF-17571Stylus pen
SF-17600Knife blade
SF-17601Copper-alloy needle
SF-17612Black glass melon bead
SF-17622Copper-alloy strip
SF-17626Copper-alloy finger ring
SF-17627Copper-alloy strip with rivets
SF-17702Bone handle
No find no.Bun quern, complete
L-2012-20Incomplete shoe (female)
L-2012-21Incomplete shoe (adult male)
L-2012-25Incomplete shoe (adult male)
L-2012-27Shoe scrap
L-2012-28Incomplete small shoe (child)
L-2012-30Incomplete shoe (adult male)
L-2012-31Incomplete small shoe (child)
L-2012-32Shoe scrap
L-2012-35Scrap leather
L-2012-42Carbatina shoe (female or adolescent)
L-2012-46Shoe (adult male)
L-2012-56Scrap leather
L-2012-57Scrap leather
L-2013-14Carbatina shoe, homemade
L-2013-24Complete shoe sole (adult male)
L-2013-30Leather tent fragment
L-2013-49Leather tent fragment
W-2012-8Barrel bung with central hole
W-2012-9Wooden pommel
W-2012-10Wooden comb
W-2012-11Wooden bung
W-2012-12Wooden comb
W-2012-13Wooden comb
W-2013-6Wooden comb
W-2013-10Wooden comb
WT-2012-4Stylus tablet fragment
View Table Image: 1 | 2

Rectilinear Building Drains

Find no.Object identification
SF-17551Stylus pen
SF-17552Knife with bone handle
SF-17564Bone whistle
SF-17603Iron plate
SF-17661Shaped bone
L-2013-19Complete sandal (child)
L-2013-20Decorated carbatina shoe uppers
L-2013-23Incomplete shoe (adult male)
L-2013-48Leather tent fragment
W-2013-5Wooden comb

Roundhouse Outer Ditch

Find no.Object identification
SF-17466Copper-alloy stud
SF-17469Iron pin
SF-17470Glass bottle
SF-17630Mortarium sherd with maker’s stamp
SF-17646Mirror fragment
SF-17647Mortarium sherd with maker’s stamp
SF-17648Iron hook
SF-17650Iron lynchpin
SF-17652Knife with bone handle
SF-17653Iron ink pen nib
SF-17655Copper-alloy stud
SF-17656Copper-alloy bracelet
SF-17657Bone gaming counter
SF-17659Ligula probe
SF-17662Ballista bolt head (iron)
SF-17663Copper-alloy disc
C.2484Sestertius of Nerva (97 CE)
L-2013-1Scrap leather
L-2013-2Leather strap
L-2013-11Scrap leather
L-2013-25Incomplete shoe (unknown owner)
L-2013-36Incomplete shoe (female or adolescent) with fishnet-pattern uppers
L-2013-37Incomplete shoe (female or adolescent)
L-2013-41Incomplete shoe (child)
L-2013-43Scrap leather
L-2013-54Scrap leather
L-2013-56Scrap leather
W-2013-12Wooden comb
W-2013-13Wooden bung
WT-2013-2Fragment of ink writing tablet
View Table Image: 1 | 2

Rubbish Dump Between Buildings

Find no.bject identification
SF-16954Copper-alloy needle
SF-16957Graffito on terra sigillata sherd
SF-16961Iron lance head
SF-16964Maker’s stamp on terra sigillata sherd
SF-16968Ligula probe
SF-17624Iron latch
L-2012-24Shoe scrap
L-2012-26Carbatina shoe, homemade (child)
L-2012-29Shoe scrap
L-2012-33Incomplete shoe (child)
L-2012-34Shoe scrap
L-2012-36Incomplete shoe (adult male)
L-2012-37Incomplete shoe (child)
L-2012-38Incomplete shoe (unknown owner)
L-2012-39Complete carbatina shoe with decorative uppers (adult male)
L-2012-40Leather tent fragment
L-2012-41Leather tent fragment
L-2012-43Incomplete shoe (female or adolescent)
L-2012-44Leather patch, trapezoidal with large stitch holes
L-2012-45Shoe uppers only
L-2012-47Incomplete carbatina shoe (female or adolescent); pair with L-2012-52
L-2012-48Shoe sole (adult male)
L-2012-49Incomplete shoe (adult male)
L-2012-50Incomplete shoe (adult male)
L-2012-51Leather tent fragment
L-2012-52Incomplete carbatina shoe (female or adolescent); pair with L-2012-47
L-2012-54Incomplete shoe with uppers, evidence of patches and repair (child)
L-2012-55Shoe uppers only
L-2012-58Scrap leather
L-2013-45Leather tent fragment
W-2012-14Partial wooden box lid with rectangular hole
WT-2012-3Fragment of wax stylus tablet
View Table Image: 1 | 2


1 We are grateful for early comments on this research by Florence Liard and Emilie Colpaint, the organizers of a glocalization conference at Université Saint-Louis Bruxelles in 2021, and the anonymous reviewers who offered excellent help in improving the article in many ways. We would like to thank everyone at the Vindolanda Trust and its archaeological team for support and helpful discussion of the research. Financial support for this research project was provided by the Canada Research Chairs Program (Award: CRC-2019-00113), the Social Science Humanities Research Council Canada, the University of Western Ontario, and The Vindolanda Trust.

This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0), which permits noncommercial reuse of the work with attribution.

2 The auxiliary army included the units originally composed of noncitizen soldiers, who were recruited from the conquered provinces of the empire. They supported the legionary units in battle and guarding the empire’s borders.

3 Pitts (2010, 125) stresses the need to concentrate closely on the relationship between objects and the very specific local contexts in which they operated. See also Versluys 2014.

4 Woolf 1997; 1998; Mattingly 2004, 6; Ghisleni 2018;Pitts and Versluys 2021.

5 Recent case studies can be found in Pitts 2007; 2008; papers in Hales and Hodos 2010; van Alten 2017; Diers 2018; Kolbeck 2018. Gardner (2013) provides a good discussion of recent theoretical approaches used in Roman archaeology, including glocalization.

6 Fine and Thompson 2018, 11. For an overview of globalization theory generally, see Nederveen Pieterse 2020. For the notion that there is no distinct theory of glocalization across the social sciences, see Roudometof 2016a. For original introductions to glocal concepts in social science inquiry, see, e.g., Robertson 1994; 1995. Theoretical approaches to the local and global in Roman archaeology have been developing for some time and include, e.g., Woolf 1995; 2003; Witcher 2000; Mattingly 2004; Hingley 2005; Pitts 2008; 2010; Hodos 2010; 2015; Gardner 2013; Laurence and Trifiló 2015; Pitts and Versluys 2015.

7 Hodos 2015, 250. For a recent overview of the development of identity studies, see Hodos 2010; see also Allason-Jones 2001; Hill 2001; Díaz-Andreu et al. 2005; Pitts 2007; Gardner 2007; 2013.

8 Globalization theories have been fruitfully applied to Mediterranean and specifically Roman contexts and material culture, e.g.: Hingley 2005; papers in Pitts and Versluys 2015; papers in Hodos 2017, esp. Pitts 2017, Versluys 2017a, Witcher 2017. Witcher (2000) includes important warnings that globalization does not mean cultural homogenization. For a useful reminder of the problems of globalization theory in Roman archaeology, see Gardner 2013, 8–9.

9 Roudometof 2016b, 1.

10 Roudometof 2016b, 4.

11 Roudometof 2016a, 401. Another statement from Roudometof (2016a, 399) is also useful: “glocalization is globalization refracted through the local.”

12 It has been debated whether this defensive line can be defined formally as a frontier, and it is sometimes referred to as the Stanegate “system.” See Hodgson (2000) for arguments and summary of the past debate. Note that all dates herein are CE unless otherwise stated.

13 On the vicus at Vindolanda, see R. Birley 2009, 162–68. Some extramural occupation is understood at Housesteads, a Hadrian’s Wall fort a few miles east of Vindolanda; see Rushworth 2009; Allason-Jones 2013. For a full review of extramural settlements in Britain, see Sommer 1984.

14 van Driel-Murray et al. 1993; R. Birley 1994; Greene and B. Birley 2024. On environmental conditions, see Seaward 1976; Hefford 2005.

15 Blake 2014, 107–11.

16 Some evidence may point toward initial habitation in the late first century, but the evidence is slight, and the bulk of dateable material points to predominantly Period 4 occupation; Blake 2014, 108.

17 Blake 2014, 107.

18 R. Birley 1994, 92–112.

19 Later examples are better preserved archaeologically; see, e.g., A. Birley and Blake 2007, 38–41.

20 For plans and discussion of vici in Britain, see Sommer 1984.

21 On military buildings located in the vicus at Vindolanda, see R. Birley 2009, 159–62.

22 E.g., the third-century granaries in the fort discussed by A. Birley (2013) or the storage spaces in the North Field at Vindolanda discussed by Greene and Meyer (2017, 236–44).

23 Blake 2014, 86–88.

24 Blake 2014, 107.

25 Blake (2014, 109) reports on the extramural structure. On the Period 1 rampart, see Russell et al. 2022, 186–88.

26 Heslop (2008) presents a typology of beehive querns in Britain; see also Taylor 2015, 85–92.

27 A similar pattern is seen in the northwest empire on the continent where Roman-style querns came to predominate very quickly but rare examples of Iron Age types continued in use; Wenzel 2020.

28 Although Pitts’ (2008, 495) discussion of the subordinate economic role of a province such as Britain amid wider global economic trends might also be taken into consideration when discussing broad networks.

29 Bayley and Butcher 2004, 160–64. Vindolanda find nos. SF-16958, SF-17629, SF-17649.

30 Vindolanda find no. SF-17649.

31 Ghisleni 2018, 149, with further discussion in the response by Versluys on 159.

32 Bowman and Thomas 1994; 2003; Bowman 1998; A.R. Birley 2002; Greene 2013a.

33 Blake 2014.

34 Greene and B. Birley 2024, for further discussion of the remains in these structures.

35 Greene 2014a.

36 Hingley 1999; 2017, 102; Mattingly 2006, 367–78.

37 When referring to material culture from, e.g., the homeland of the soldiers stationed at the site, we employ different and precise language, not simply “local.”

38 For in-depth treatment of the practice of local recruitment, see most recently Haynes 2013, 121–34. For a specific case study of local recruitment using the Batavians, see van Driel-Murray 2012.

39 van Driel-Murray 1995; 1997; 1998; Allason-Jones 1999; James 2006; 2018; Allison 2011; 2013; Greene 2013b; 2016; 2020. On the civilians at Vindolanda, see R. Birley 2000.

40 Woolf (2003, 134–35) considered an extensive list of identifiers that separated provincial society beyond the simplistic Roman and native dichotomy. Also see Jones 1997; Woolf 1997; Mattingly 2004, 8–9; Hodos 2010, 5.

41 Mattingly 2004, 9–13.

42 Gardner (2013, 11–12) provides a good discussion of the physical manifestation of the military hierarchy in the layout of a Roman fort.

43 The Ninth Cohort of Batavians stationed at Vindolanda in Period 3 were very likely commanded by one of their own elite (A.R. Birley 2001, 14–17).

44 The extramural settlements were historically described in pejorative terms as “shanty towns” and the inhabitants, termed “camp followers,” were a group negatively portrayed as the rabble; Breeze and Dobson 1987, 183. For discussion of this patently incorrect view of this group, see James 2001, 80; 2018, 37. The rehabilitation of this population as a legitimate community began in the 1990s with the growing recognition that women and children were regular components of military settlements; van Driel-Murray 1995; 1997; 1998; Wells 1997; Allason-Jones 1999.

45 Perhaps in the case of military personnel, thinking more along the lines of their becoming a Roman soldier is a better approach; Haynes 1999; James 1999; 2014.

46 Woolf (2003, 137–38) argues similarly about the extreme localism in matronae cults in the northwest provinces having meaning only within an imperial context and with the knowledge of how unique local expressions fit into the larger context of empire. He argues that the situation of empire forces the drive to localism and, therefore, can only be understood with the background of both local and global trends; see also van Alten 2017, 12.

47 A.R. Birley et al. 2013, 289–91, for the archaeological context and discovery.

48 A.R. Birley (in A.R. Birley et al. 2013, 287–88) points out the very rare formulation of putting the word for god or goddess after the deity’s name but notes a connection to Gaul, especially Aquitania, for this formulation. For a complete linguistic discussion of the origin of the name Ahvardua, see A.R. Birley et al. 2013, 297–300.

49 For a full review of the Tungrians, see A.R. Birley et al. 2013, 291–93. The tribal unit known as the Tungri were first part of Gallia Belgica but were later included with Germania Inferior after Domitianic redistricting.

50 A comprehensive discussion of the movements of this unit from Vindolanda to Hadrian’s Wall and farther north in the province is provided in A.R. Birley et al. 2013, 293–96.

51 Mattingly (2004, 11–12) provides a convenient breakdown of the number of elite Roman individuals (senatorial and equestrian status) in Britain, conjecturing there were around 300 elites in total at the top of the hierarchy at any one time.

52 An excellent example of a nonnuclear family in residence in an extramural settlement comes from Lussonium in Pannonia Inferior (Danube region). Military diplomas from a house in the vicus suggest that one soldier married the sister of his comrade, and they all cohabited in the home of the siblings’ father; Greene 2015, 142–44. Visy (1982, 60–65; 1987, 96–98) discusses the archaeology of the building and its finds.

53 The most recent and thorough treatment of the economy of military forts and the role of extramural settlements has been Kolbeck 2018. Also see Evers 2011, with reference specifically to the Vindolanda economy.

54 Kolbeck 2018.

55 Tab.Vindol. 156: “… sent with Marcus, the medical orderly, to build the residence [hospitium], builders, number 30, to burn stone, number 19 (?), to produce clay for the wattle fences of the camp”; trans. Bowman and Thomas 1994, 100–1. The bathhouse mentioned in tablet 155 may be the one located by excavation outside the southern walls of the Period 4 fort (A. Birley 2001).

56 The term hospitium generally refers to the custom of offering accommodation and hospitality. In a military context it often means the residence of an officer, or a soldier billeted in civilian accommodation. Bowman and Thomas (1994, 100–1) further suggest that in this case at Vindolanda it may refer to a simple guest house. They indicate the structure was inside the fort, but there is no evidence for where it may have been located.

57 Tab.Vindol. 185: “at Cataractonium, for accommodation (?), denarii 1/2”; trans. Bowman and Thomas 1994, 141–45.

58 Good examples are the merchants Octavius and Candidus. Tab.Vindol. 343; see Bowman and Thomas 1994, 321–29. The identity of Octavius as a soldier or a civilian is debated, but see Kolbeck 2018, 4–7, for a sustained discussion of this tablet and his identity as an “entrepreneurial contractor” (6).

59 Tab.Vindol. 178. It is difficult to know exactly to what this list of revenue refers, but it is logical to imagine it derived at least in part from the local population in some way. Bowman and Thomas (1994, 120–21) speculate the source could have been commercial interactions with the surrounding populace in the form of food or other items.

60 Kolbeck 2018; see also Evers 2011.

61 Kolbeck 2018.

62 Shotter 1993, 14. For recent consideration of the trends of coinage use in Roman Britain, see Walton 2012; 2015.

63 Kolbeck (2018) discusses this point generally. It is worth noting that no material culture from the Vindolanda fort or settlement can be associated strictly with pre-Roman occupation (R. Birley 2009, 39–40).

64 Tab. Vindol. 649: “You will receive out of the Britons’ carts … from Rac..romaucus (?) three hundred and eighty-one modii of … grain”; trans. Bowman and Thomas 2003, 106–8. See also Mitchell 1976 on the practice of requisitioning carts and goods from locals. Tacitus (Agr. 19) reports another practice whereby the Britons were forced to sell their grain to the state to supply the army. Whether this was true or more for rhetorical flourish can be debated.

65 Tab. Vindol. 344: “As befits an honest man (?) I implore your majesty not to allow me, an innocent man, to have been beaten with rods and, my lord, inasmuch as (?) I was unable to complain to the prefect because he was detained by ill-health I have complained in vain (?) to the beneficiarius and the rest (?) of the centurions of his (?) unit. Accordingly (?) I implore your mercifulness not to allow me, a man from overseas and an innocent one, about whose good faith you may inquire, to have been bloodied by rods as if I had committed some crime”; trans. Bowman and Thomas 1994, 329–44.

66 A. Birley 2016, 149–50.

67 Blake 2014, 86–88.

68 Roudometof 2016b, 1.

69 As Roudometof (2016a, 401) has said, quoted above, glocality is the experience of the global locally or through local lenses.

70 Hodos 2017, 3–4.

71 Recent work by Witcher (2017, 634–35), among others, makes clear that we can consider the Roman empire a globalized world. Jennings (2011, 3) similarly contends that a period of intense interregional interaction that prompts extensive social change, such as can be argued for Roman conquest, constitutes an era of globalization.

72 Witcher 2017, 639, 643; Jennings 2017, 15; Witcher (2000, 220) states: “Not only should we be suspicious that the widespread adoption of particular material culture was accompanied by a new homogeneous identity, but we might expect a proliferation of alternative identities with locally-diverse and context-dependent relationships with that material culture.”

73 Versluys 2014.

74 Versluys 2017b; Pitts and Versluys 2021.

75 Pitts and Versluys 2021, 367. See also, e.g., Gosden 2005; Hicks 2010; Jones and Boivin 2010; Hodder 2012; Antczak and Beaudry 2019; Pitts 2019.

76 Pitts and Versluys (2021, 368) encourage movement away from thinking in terms of individual objects and assemblages.

77 Versluys (2014, 17) argues: “We should not so much focus on what things with their stylistic and material properties would represent—or to what historical narrative they testify—but on what things do in a certain context.” For similar views, see Jiménez 2020.

78 Hodos 2015, 242.

79 Pitts and Versluys 2021.

80 Jiménez 2008, 16–17.

81 Witcher 2000, 220; 2017, 639.

82 On resistance to Rome, see Bénabou 1976; discussions in Bhabha 1994; Alcock 1997; Hingley 1997; Webster 2003; Van Dommelen 2007; Jiménez 2008.

83 Jiménez 2008.

84 Versluys 2014, 17.

85 Among others to discuss cultural bilingualism and hybridity, see Mattingly 2006; Jiménez 2008. The related idea of code-switching was put forward first by Wallace-Hadrill 1998; 2000.

86 Hodos 2010, 10–11; van Alten 2017, 2.

87 González-Ruibal 2010, 46; see also Van Dommelen 2006, 107; Gardner 2013, 5–6.

88 González-Ruibal 2010, 46; Gardner 2013, 6.

89 Fernández-Götz et al. 2020, 1631.

90 This article does not claim to tackle subjects of violence and domination broadly but rather addresses the role of inequality and hierarchy in understanding the local social context of a frontier military community.

91 Haynes 2013.

92 Gardner 2013, 12–14.

93 The Vindolanda tablets offer a glimpse of military violence, e.g., Tab.Vindol. 344: Bowman and Thomas 1994, 329–34. See also Mattingly 2004, 15–16. This picture is also presented in Roman literature, e.g., Apul., Met. 9.39–42.

94 Haynes 1999; James 1999; 2014. For a similar viewpoint from Italy, see Witcher 2000.

95 Meyer (2020) focuses on the creation of smaller ethnic groupings within auxiliary units.

96 See also Hodos 2010; van Alten 2017, 2.

97 Hingley 1989; 2017, 102; Fulford and Holbrook 2011, 337. Also see discussion in Gosden 2005; Ghisleni 2018.

98 Blake 2014, 107–11.

99 A. Birley and Blake 2005, 35.

100 Blake 2001; A. Birley 2003, 52–65; 2013, 65–71; A. Birley and Blake 2007, 27–30.

101 For discussion, see A. Birley 2013, 65–71, 67 for plan.

102 In Gosden’s (2005) discussion of the shift from round to rectilinear construction in Britain, he chooses to look instead at the similarities in use, and therefore continuity of ideas, rather than exact construction style, and points to the concept of central open space as a similarity.

103 We hesitate to call it resistance, which would put clear intention and purpose, that may not have originally existed, on the action of building in a local style.

104 Jennings 2017, 15; Witcher 2017, 639, 643.

105 B. Birley and Greene 2006; Greene 2014b. Cf. Marshman 2015.

106 Henig 1970.

107 A.R. Birley 2001, 14–17.

108 Roymans 1996, esp. 13–20.

109 Mullen 2021, 372, and on 376: “[literacy] was a core part of the Roman cultural package.”

110 Hodos 2010, 5.

111 Harris (1989) presents an overarching picture of minimal literacy in the provinces: roughly 5% of the population. Mullen (2021, 365–74) collects archaeological and textual data from recent projects to show that this number is probably far too low. Mullen (2021, 368) states: “New data enables us to think about what we might call ‘differential literacy’ and to move away from stark provincial percentages to pictures of social, chronological and geographical complexity”; see also Hanson and Conolly 2002; Häussler and Pearce 2007.

112 Derks and Roymans 2002.

113 On the effects of military recruitment on the Lower Rhine region, see van Driel-Murray 2008; 2009; 2012.

114 Tab.Vindol. 650; trans. Bowman and Thomas 2003, 109–11. On the community represented by the corpus of tablets, see Greene 2013a.

115 Mullen (2021, 371–73) points out that the military had influence on literacy in military settlements, as well as in soldiers’ home communities upon their return (e.g., as described in Derks and Roymans 2002 for the Batavian homeland) but not perhaps as much influence in the rural regions where forts were located. She also argues that these patterns may, however, be an oversimplification and require scrutiny; see also Raybould 1999, 159–60.

116 Tomlin 2011, 134; 2018; Mullen 2021, 374–76.

117 Pliny (HN 23.78) attributes medicinal properties to them. On hazelnut cultivation in the Roman world, see Alcock 2006, 48, 178.

118 López-Dóriga 2015, 39. Hazelnuts have been found associated with prehistoric European sites from the Mesolithic through the Iron Age.

119 On Roman identity and food consumption at Vindolanda, see Pearce 2002; see also Taylor 2015. For broad Roman studies, see Van der Veen 2008; papers in Broekaert et al. 2016; Banducci 2021.

120 Taylor 2015.

121 For further treatment of this theme, see Mikl Curk 1990, 134; Galsterer and Galsterer 1992, 377.

122 Van Driel-Murray (2009) uses four case studies, including the Tungrians at Vindolanda in Period 4, to show the presence of families and maintenance of homeland traditions; see also van Driel-Murray 2012, 116–17.

123 van Driel-Murray 1995; 1997; contra Hodgson 2014, who questions this conclusion; see also Le Bohec 2017.

124 Rothe 2009, 69–72; 2012. For a modern example from Mexico, see Nadig 1986.

125 Okun (1989, 136) observed this on the Upper Rhine frontier and argues that “assimilation only occurs among those segments of society which will profit from adopting new cultural processes.”

126 Van Driel-Murray (2009, 816–20) presents a perfect case study for glocalization approaches in which Housesteads Ware, a ceramics form found on Hadrian’s Wall in Britain but mimicking forms and technology from the Netherlands, was altered slightly to have specific meaning in the new context in Britain. In this case, the formal wares are retained as a way to maintain old symbols of hospitality and social relationships, and, she argues, the quotidian vessels could have been replaced easily with new styles.

127 van Driel-Murray 2007.

128 On Iron Age footwear, see Groenman-van Waateringe 2007.

129 For an overview of the styles present between the first and fourth centuries CE, see van Driel-Murray 2007, 343.

130 Van Driel-Murray (2009, 815) points to the shoes from sites such as the Saalburg in Germany, where non-Roman-style shoes found throughout the settlement betray the presence of northern Germanic troops and their families.

131 Mould 2014, 125–29.

132 Mould 2014, 130–31.

133 Vindolanda find no. L-2012-54.

134 Greene 2014a.

135 Gosden 2005, 198–99, 207–8.

136 Ghisleni 2018, 152–53.

137 A. Birley and Blake 2005, 35; Blake 2014, 107–11.

138 The particularity of individual sites is highlighted by Gardner (2013, 16) in a discussion of Claydon Pike in the British Cotswalds, where food consumption and architecture took on new forms, but the small finds assemblage appears traditional. This is the exact opposite of the situation seen in the Period 4 extramural settlement at Vindolanda.

139 Roudometof’s (2016a, 397–402) “analytical autonomy” of glocalization theory. This hybrid result is labeled the “Third Culture” in some approaches, e.g., Hodos 2010, 22; see also Rothe 2012 for a Roman case study using the “third way” to understand a new articulation of status through female dress by the Treveri in Gaul.

Works Cited