THEROMANARMY AS A COMMUNITY Including papers of a conference held at Birkbeck College, University of London on 11-12 January, 1997 edited by Adrian Goldsworthy and Ian Haynes with contributions by C. E. P. Adams, L. Allason-Jones, R. Alston, M. C. Bishop, A. Goldsworthy, M. Hassall, I. Haynes, M. Henig, A. J. King, B. Rankov, E. Sauer, C. S. Sommer, T. J. Strickland, R. S. 0. Tomlin, and J. J. Wilkes Portsmouth, Rhode Island 1999 JOURNAL OF ROMAN ARCHAEOLOGY SUPPLEMENTARY SERIES NUMBER 34 AN INTERNATIONAL JOURNAL Addresses of contributors C. E. G. Adams, Department of Classics, University of Warwick, Gibbet Hill Road, CV4 7AL, England L. Allason-Jones, Museum of Antiquities, University of Newcastle-upon-Tyne NE1 7RU, England R. Alston. Department of Classics, Royal Holloway College, Egham, Surrey TW20 OEX. England M. C. Bishop, Braemar, Kirkgate, Chirnside, Duns, Berwickshire TDll3XL. Scotland I. Goldsworthy, 4 Church Road, Penarth, South Glamorgan CF64 IAE, Wales M. Hassall, Institute of Archaeology, 31-34 Gordon Square, London WClH OPY, England I. Haynes, Department of History, Birkbeck College, Malet Street, London WClE 7HX, England M. Henig, Institute of Archaeology, 36 Beaumont Street, Oxford OX1 2PG, England A. King, King Alfred's College, Winchester S022 4NR, England B. Rankov, Royal Holloway College, Egham, Surrey TW20 OEX, England E. Sauer, Keble College, Oxford OX1 3PG, England C. S. Somtner, LandesdenkmalamtBaden-Wiirttemberg, Silberburgstrage 193, D-70178 Stuttgart, Germany T. Strickland, Gifford and Partners, 20 Nicholas Street, Chester CH1 2NX, England R. S. 0.Tomlin, Wolfson College, Oxford OX2 6UD, England J. J. Wilkes, Institute of Archaeology, 31-34 Gordon Square, London WClH OPY,England ISBN 1-887829-34-2 ISSN 1063-4304 (for the supplementary series) Copyright O 1999 Journal of Roman Archaeology L.L.C. Printed by Cushing-Malloy, Ann Arbor, Michigan. This and other supplements to the Jolrrnal of Roman Archaeology may be ordered from: JRA, 95 Peleg Road, Portsmouth, RI 02871, U.S.A. Telephone (USA) 401 683 1955 telefax (USA) 401 683 1975 e-mail: jra@wsii.com Web site: JournalofRomanArch.com Introduction: the Roman army community Ian Haynes Introduction The character and ethos of the Roman army were primarily determined by its r61e as a fighting machine and an instrument of imperial power. To understand the army fully, however, it is necessary to go beyond a n investigation of formal military structbre and try to appreciate the army as a community or group of communities. The great range of evidence available makes this endeavour more feasible than for almost any other provincial group. Religious dedications and calendars, personal correspondence and legal texts, all illuminate the relationships, values, and social codes that underpinned the army. Archaeology reveals the ways in which, through dress, display, and the use of space, members of the military selectively emphasised their distinctiveness from the surrounding civilian society. This book will develop the growing recognition among students that the army was itself a community, the study of which illuminates the character of Roman society as a whole: exploring the nature of communities of soldiers will deepen our understanding of the Roman army as a force in provincial society. The history of the Roman world is partly the history of the Roman army. Military force carved out the empire and maintained it, constantly reinforcing the unequal peace of the pax romana. The army was the bedrock of imperial power and political control. All major studies of Roman history have acknowledged these realities, though only a few (e.g., Campbell 1984) have focused on the complex relationship between ruler and soldier under the empire. While the army's r81e and status were fundamentally linked to the office of emperor, military factions perceived their relationship to the person of the emperor and/or imperial claimants differently. The army was not a monolithic power block; it could and did break down into provincial, regional, and even regimental interests. Given the factionalism of the army's politics, it may seem inappropriate to discuss the arrny as a single entity. While senior personnel and even entire regiments could be transferred from one end of the empire to another, in practice this happened less and less frequently. Although the empire was bound together by the movement of people and ideas, only a relatively small proportion of soldiers and their equestrian commanders would have served in both east and west. As a result, armies developed regional or provincial attributes and allegiances. Similar processes were at work at the local level. Individual units drew local recruits and were stationed in one place for generations and were influenced by the community in which they were stationed. We should recognise diversity in the army; multiple local communities existed within the large military community. Attention to such details, however, should not blind us to those attributes that bound the army together across the empire. Such characteristics were not merely structural, limited to a particular form of organisation or ranking system; they also took on a ritualized form. From the Euphrates to the Tyne, soldiers celebrated the same festivals (Fink,Hoey and Snyder 1940: 2829) and swore the same oaths (Campbell 1984: 19-32). Through such rituals the entire army shared in the military ethos that originated in Rome herself. Recent scholarship and the Roman army Research on the Roman army has proliferated in recent years. Work on the frontiers alone is so extensive that a single scholar can hardly follow developments in more than a few regions. Academic specialisation in a particular area, or at least in one half of the empire, is common. Ian Haynes In the Roman west our knowledge of the army remains largely dependent on material excavated. Recently, study of the army in the east has taken on a new life, stimulating some of the most interesting debates in years. B. Isaac's The limits of empire (1990) and subsequent discussion1 introduced many scholars both to the region and to new sources of evidence. We may point, for example, to Isaac's use of the Talmud to illuminate tensions between the local populations and the Roman garrison. Our understanding of the army's operation in different provincial settings is significantly informed by the available evidence. In some areas scholarship is heavily dependent on papyrology; in others archaeological excavation and field-walking are the principal sources of data. These differences may make inter-regional comparisons and generalisations problematic, yet to some extent the differing types of evidence are capable of addressing the same questions. At a local level students can seek to examine the army's impact on provincial society. How important was the military in establishing the civilian infrastructure of newly conquered provinces? How significant was the army within local economies? How dependent was the army on the civilian community for food and supplies? Questions such as these are considered in this volume. The army has frequently been discussed by historians in the context of social changes but it is seldom appreciated as a dynamic society in its own right. Military historians and epigraphers have focused on its organisational attributes, since the army's structure is accessible even if the interpretation is seldom straightforward. For the archaeologist, organisational studies offer a framework for the study of types of military equipment or installations. However, too great a concentration on structure may b h d us to key aspects of the soldier's experience. Military organisation has a comforting familiarity; modern armies too are carefully structured, often in ways similar to the Roman army, yet in seizing upon such similarities one can fall prey to modernizing assumptions and attribute to Roman soldiers characteristics they would never have understood. On a daily basis, relationships are more important to people than organisational structures. Formal structures may influence patterns of social interaction but they do not necessarily determine the forms of interaction. Recent scholarship is moving beyond organisation, to illuminate soldiers' relationships with one another and with the local community. There has been a shift in emphasis away from institutions and towards underlying social tensions and ties. MacMullen (1984), for example, considered the legion as a society, not just as a fighting machine, though he emphasised that the two aspects were related. In exploring the relationships of civilians and soldiers, Shaw (1983) adopted another approach, arguing that a legionary garrison could be seen as a 'total institution' - a community socially and culturally isolated from the civilian population it o v e r s a ~ ;so ~ pronounced was this isolation that the legion was able to act unhesitatingly in support of the emperor, brutally suppressing local rebellion (1983, 148). More recently, Pollard (1996) employed the 'total institution' model to 1 2 Recent reviews of Isaac include: A. R. Birley, CR 61 (1991)411-13; J. Crow, Britannia 23 (1992) 335-39; A. Demant, Historische Zeitschrift 256 (1993) 439-41; M. Gawlikowski, Topoi 4 (1994) 371-76; M. Goodman, Jozirnalfor the Sfzidy ofJlidaisrn 22 (1991) 147-50; G. D. B. Jones, JRS 83 (1993) 241-44; Y. Le Bohec, Latomiis 51 (1992) 672-74; C. S. Lightfoot, The International Historical Review 24.1 (Feb. 1992) 111-14; S. T. Parker, JRA S (1992) 467-72; T. Rajak, JJS 44 (1993) 142-43;M. Sartre, REA 93 (1991) 199201; I. Shatzman, IEJ 44 (1994)129-35; M. P. Speidel, B f i 192 (1992)649-50; G. P. Verbrugghe, CW 35 (1991) 62-63; G. Von Biilow, Klio 74 (1992) 509-10; C. R. Whittaker, TLS, 22 March 1991, 23; G. Wirth, JbAC 34 (1991) 193-200; T. Wiedemann, GbR 38 (1991) 103-4. See also in general the papers in D. L. Kemedy (ed.),The Roman army in the East URA Suppl. 18,1996). The 'total institution' model is defined by Goffrnan (1961,xiii) as "a place of residence and work where a large number of like-situated individuals, cut off from a wider society for an appreciable period of time, together lead an enclosed, formally administered life". The Roman army as a community examine the auxiliary garrison of Dura-Europos and explore the social distance that existed between soldiers and civilians. By contrast, Alston (1995) has used papyrological evidence from Egypt to indicate more relaxed patterns of interaction between the two groups. While it is not necessary here to consider these divergent pictures in detail, it is important to note the introduction of new approaches to old evidence. The concept of community In the parlance of modem social science, all these approaches may. fall within the category of 'community studies'. Today the word 'community' occurs frequently in casual speech and in political 'sound-bites'. Even during the 1950s, some 94 different definitions of the concept were in academic use (Hillery 1955), and it may be as well that, in assembling this volume, the editors have not sought to impose a uniform definition of community upon the contributors. Yet because the term offers new insights into ancient societies in general and the Roman army in particular, some discussion is required. Many traditional views of community emphasise structure, defining it with a series of attributes. Tonnies' Gemeinschaft/Community, for example, depicts a society "which, being based upon consensus of wills, rests on harmony and is developed and ennobled by folkways, mores and religion" (Tonnies 1955,261); in other words community is partially defined through its opposition to other forms of communal living based upon convention and political legislation. A more popular and complex definition comes from the 'symbolic interactionists' who see community as something existing largely in the mind, rather than as a list of traits. Hamilton, in his preface to Cohen's seminal work, The symbolic construction of community (1985)' summarizes this definition: "symbolically constructed, as a system of values, norms and moral codes which provides a sense of identity within a rounded whole to its members". Symbols, be they ideas or things, become reference points to which members of a community relate whether or not they have identical meanings to all community members. A community need not have a specific geographic focus, and 'occupational communities' need not be associated with a particular place. Today their formation may be connected to a number of factors (Gerstl 1961). First, individuals in occupational communities are intensely involved in their work. Their work may be highly skilled and/or personally satisfying, or it may seem particularly responsible or dangerous to those who undertake it (Salaman 1974,29). Second, the job may carry a 'marginal status', meaning that its practitioners may see themselves worthy of membership in a higher status group to which, however, they are not fully accepted. Third, some occupations are both intentionally and accidentally inclusive. Today resources such as social centres, holidays, and creches can be provided for employees outside work so that they continue to mix with one another, giving the employer the benefits of a close-knit, dependent work force. Sometimes working conditions such as hours and geographic isolation lead to similar inclusivity, whether or not this is desired by the employer. It is not surprising that members of occupational communities frequently see themselves in terms of their occupational rBle and tend to socialize with one another rather than mixing with 'outsiders' (Salaman 1974, 21). Differing organisations conform to the model to varying extents. Factory work-forces are not considered occupational communities; police forces, however, are classic examples. Communities in the Roman army By reminding us of the close relationship between community and identity, these perspectives of the modern social sciences offer a valuable approach to Roman army studies. In his preface to a conference on 'Military equipment and the identity of Roman soldiers', Coulston noted that the theme could include the "personal identity of a man as a soldier in ancient society" (1988, X). While relatively little was made of this idea at the time, Bishop and Coulston have since developed it elsewhere (1993, 196-98, 206-9). Archaeology illuminates the Ian Haynes ways members of military communities expressed their collective identity, as in the bearing of specific types of equipment, notably the sword and sword-belt. So long as a man wore both, he was identifiably a soldier; if deprived of those items, he was temporarily suspended of his status and rendered -useless to his warrior c o m m ~ n i t yThe . ~ importance of this equipment is reflected in the decoration lavished on the belts themselves and in their careful depiction on monuments of soldiers not shown in battle-dress. Archaeology shows how the architecture of fortresses and forts, as in the special treatment of walls and gates, through whitewashing or other decorative treatment,4 could distinguish the military community from that around it. A further point emerging from the social sciences is that 'community' is not synonymous with 'institutional organisation'. A community may comprise a group of individuals within a category, such as soldiers from particular regions serving within cohors 11 Tungrorum equitata. In the mid 2nd c. groups of men of distinct origins erected three altars at the regiment's headquarters in Birrens. An altar to Mars and the Victoria Augusta (RIB 2100) was erected by men from Raetia; a dedication to the goddess Ricagambeda was put up by soldiers from the Vellavian district of Germania (RIB 2107); and an altar to Viradecthis (RIB 2108) was dedicated by the pagus Condustis militans in cohorte. In commissioning these dedications, these groups were commemorating their distinctiveness in a unit where they may have been in the minority. Other inscriptions, such as those referring to the collegia at Lambaesis, record associations which contained soldiers of different ranks or specialisations: one collegium (ClL V111 2553) appears to have had optiones valetudinarii, pequarii, a librarizis and discentes capsariorurn as its members, indicating that collegia could stretch across official professional categoriese5The Birrens inscriptions record informal groupings; the Lambaesis inscription attests a formally constituted association. Other communities, such as cult groups, soldiers' dependents, and the socially isolated families of commanding officers, cut across military and civilian boundaries. Even regiments and sub-units appear in a different light when viewed as communities. This can be demonstrated by the great Roman military symbol, the standard. The legal and sacred status of unit standards is well attested but their symbolic r81e within army communities might reward further analysis. For example, standards of ala Longiniana (CIL XI11 8094) and cohors I Gallica (ILS 9127) appear to have been adorned with animals popular in Gallic religious iconography - respectively, a three-horned bull (Green 1976, 13) and a small boar.6 Such animals almost certainly appeared on signa of the type referred to by Tacitus (Germ. 7; Hist. 4.22), borne by the pagi and tribal groupings of the Gauls (Royrnans 1990,19).They appear open to different symbolic interpretations: depending on the cultural perspective of the viewer, they could have been seen primarily as traditional Gallic tribal totems or as Roman military 3 4 5 6 Suetonius records that Augustus deprived soldiers of their swords and sword-belts as a punishment (Aug. 24.2) intended to shame the men. Conversely, Juvenal (16.45-SO),writing in jest, observes that a sword-belt was one of the things a soldier needed to wear to ensure that he enjoyed the legal privileges to which he was entitled. For the exterior decoration of forts and fortresses in Britain and abroad, see Bidwell (1996, 11-29). Excavated fragments of stonework and sculpture from Housesteads indicate that the fort gates were richly adorned. Recent work on material from the site suggests that figures of Victory and Mars flanked a dedicatory inscription on the fort's east gate (Crow 1995,33).Both deities were particularly dear to the military community. Our attention has been drawn to this text by the research of A. Pegler (Birkbeck College, University of London). The tombstone of Vellaunus of the ala Longiniana depicts him carrying a vexillurn. The motif on the vexillurn is generally interpreted as the head of a three-horned bull. The only commentator to challenge this view was Webster (1986,112)who felt that the feature projecting from the centre of the bull's head looked more like a feather than a horn. But ILS 9127 refers to the celebration by cohors I Gallica of the natale apnrnctrlon~m,suggesting that the unit standard was itself styled on a small boar. The Roman army as a community regimental standards. In the early days of the auxilia a soldier's sense of the community of his unit may have owed more to the former than to the latter. Further insights into the community aspect of military life may be drawn from the study of the genii. The genii of certain organizations, ranks, or groups within the unit are attested more frequently than those of others. Could emphasis on particular spiritual guardians reflect the importance that soldiers themselves (rather than officials) placed on the communities? In this respect, the dedication of a statue of a genius by men from different legions of the Upper Pannonian garrison is particularly interesting (CIL I11 4452): the dedicants seem to have been speculatores celebrating their common professional links across unit boundaries. Consciously and sub-consciously, scholars have often drawn analogies between the Roman army and certain modern occupational communities. For parallels scholars most frequently have looked to modem fighting forces, but they also note the policing rBle of the Roman army. Recent sociological studies of police highlight many themes of interest for the ancient world. Both Banton (1964) and Cain (1971) note that colleagueship and friendship within police tend to develop along lines less common outside that profession. This process is clearly related to their particular social status, the pressures of their job, and their relative social isolation. Such factors may also have had a significant impact on ancient interaction between soldiers and civilians in personal, not professional, situations. Extensive socialising between the two groups does not follow from the employment of soldiers in 'civilian' contexts: collecting taxes, policing roads, purchasing supplies. Under certain circumstances, groups working in close proximity may mutually or unilaterally shun social contact, fearing stigma or the erosion of authority. Alston's analyses of Egyptian papyri (1995, and p. 175 ff below) is particularly important in this regard, for he argues that the formation of friendship regularly crossed military-civilian lines. However, the extent might have varied according to a range of factors: linguistic differences, endemic ethnic tensions, localized guerilla warfare, or a history of recent conquest might all have diminished the potential for the formation of military-civilian friendships, and under such circumstances 'mateship' within military units would have been of primary importance, reinforcing a strong sense of community. Evidence for the social interaction of Roman soldiers is not restricted to Egypt. Other documentary evidence survives, notably the Vindolanda tablets. Here, the condescension implicit in the reference to Brittunculi (Tab. Vindol. 11, 164.5) contrasts with the happy memories of cont~bernales.~ As the recent discovery of tablets at Carlisle demonstrates (Tomlin, 127 ff.) there is every reason to believe that archaeologists will recover more examples of soldiers' correspondence. Nor need the exploration of friendship be limited to tablets and papyri; epigraphic evidence reveals something of those closest to the soldiers - their heirs. This book As mentioned above, the individual contributors to this volume were free to define and use the concept of community as they wished. As a result, some papers employ it as a synonym for the army and its dependents, others as the key to understanding the distinctiveness of military organisations, and still others as a device for appreciating the dynamics of garrisons and campaigning forces. It is well accepted that the Roman army's relationships with provincial societies changed over time and space, and that these relationships varied in tandem with patterns of social 7 This term is open to various interpretationsbut is often translated as 'messrnates'. Vindolanda examples include Tab. Vindol. 11, 181.14, 310.2, 311.ii.2, 343.29, 346.ii.4, 349.ii. Ian Haynes interaction. These factors in turn determined the degree to which soldiers felt isolated and distinct from society at large. Central to an understanding of the military community is knowledge of its deployment. Identifying forts and fortresses obviously comes first, but determining the nature of their use is less straightforward than some discussions have implied, as Strickland reveals (105 ff.) in his analysis of the archaeology of Chester. Also, we know that individuals, detachments and regiments were not confined to narrow frontier zones or tightly delineated bases or forts; individual soldiers were frequently detached from their units. As Rankov demonstrates in his discussion of the oficium consularis (15 ff.), military personnel performed vital and diverse administrative tasks for governors. Even early imperial governors enjoyed the benefits of sizeable bureaucratic assistance from the army. The ordinary course of military duties required soldiers to travel far from their regiments and to meet a range of provincial subjects as they went. Such men appear repeatedly in papyri and writing tablets from the Roman world. The duty rosters from Dura and Vindolanda allow us to imagine the widespread movement necessitated by soldiers' working routines. Our understanding of soldier-civilian interaction in the immediate context of military installations is still limited but two contributions demonstrate the potential for future research. In an illuminating survey, Allason-Jones (41 ff.) examines the British evidence for women linked to the army by family and economic ties. Hassall too (35 ff.) explores the relationships of women to soldiers on the basis of recent work at forts, fortresses, and colonise; he observes that artefactual evidence for the presence of women and children in some forts suggests that they contained a larger civilian population than was previously recognised. The picture of army camps as all-male bastions is now obsolete. Concentrated in large numbers, Roman soldiers could have a significant impact on local settlement, affecting the interaction and character of both military and civilian communities for generations. Several papers argue that the army's movements significantly impacted subsequent development. Some regional studies illustrate the range of historical and cultural factors that determined the deployment of army units. Wilkes (95 ff.) considers their r81e in the Danubian lands, a region of fundamental importance to Rome's military history. His account charts several centuries of change, from the devastating impact of military action under Augustus to the growth of a vibrant frontier economy fostered by the army's presence. Development is also a key theme of Sommer's work on the vici (81 ff.). The rale of the military in developing local centres is well attested but Sommer's study of the Upper German material provides fresh insights. In addition to examining evidence that the army may actually have founded some vici, he explores the way in which the distribution of vici formed the core of later civilian development. At a more local level Sauer (52 ff.) identifies traces of the military impact upon settlement patterns: by examining the wealth of numismatic evidence from Bourbome-les-Bains, he argues that the site owed its early growth to the army. His study has important implications for our understanding of both the deployment of the army under Augustus and the history of spa sites. While agreeing that the dispersal of Roman forces could have significant consequences for civilian development, contributors do not necessarily agree on the factors which determined that dispersal. Bishop (111 ff.) draws upon literary, sub-literary and epigraphic evidence to offer a new view of Roman army deployment. He argues that dispersal was largely dictated by the need to feed large numbers of soldiers. His challenging ideas may require that popular concepts of distinct military and civilian zones be rethought. Questions of deployment also have important implications for discussions of military logistics, still of fundamental importance for understanding how the army worked as a wider community within wider society. As Adams demonstrates (119 ff.), a detailed reading of the papyrological evidence illuminates how the Egyptian bureaucracy oversaw supply. His analysis leaves little doubt The Roman army as a community that logistics involved significant interaction between soldiers and civilians. A different approach to supply is adopted by King (139 ff.) in his study of meat diets in civilian and military Britain, Gaul, and Germany. Careful statistical analysis of the fauna1 material from a variety of sites suggests patterns in the types of animals supplied to the army and in the dietary identities of military communities. All these findings help elucidate the impact of the army's basic needs on provkcial society. One of the tablets recently discovered at Carlisle, expertly interpreted by Tomlin (127 ff.), reminds us that questions of supply did not operate simply at an army or regimental level. "Making the machine work", as Tomlin describes it, involved officers and men right through the chain of command. The organisation of supply within regiments was a matter of routine to the military in war as in peace; the maintenance and replacement of weapons was vital, since without them the Roman soldier lost his public identity and became a useless cog in the machine. The theme of identity appears in several contexts. Community membership and professional affiliation, military or otherwise, played important r6les in the way individuals perceived themselves. This theme is evident in the study of Roman art, as Henig demonstrates in his study of artistic patronage in the army (150 ff.). Identity is also a central theme in Haynes's paper (165 ff.) as he explores the way in which service within the military community could actually transform the cultural identity of Roman soldiers, and argues that many auxiliaries underwent a form of Romanization. Without shared 'symbols', common cultural points of reference, the army simply could not function. The bonds holding the army together are the subject of analyses by Alston and Goldsworthy. In his discussion of the Egyptian papyrological evidence, Alston (175 ff.) seeks to trace group loyalties within the military and between soldiers and civilians. Recognising that loyalty and patronage helped bind army units together, he seeks to emphasise that soldiers could and did develop significant personal links with their host communities. While Alston explores army dynamics in peacetime, Goldsworthy seeks to understand them during war (197 ff.). The army was first and foremost a fighting machine and it was under the pressure of combat that some aspects of soldiers' collective identity became most obvious. Through a careful analysis of the accounts of Josephus he illuminates how these factors manifested themselves in the Roman army at the siege of Jerusalem. In compiling this book we have been anxious to underscore the vibrancy of the Roman army. We have sought to emphasise the ways in which soldiers' relationships with their colleagues and civilian counterparts extended beyond organisational structures. States of war and peace necessarily influenced these relationships at all levels, from the institutional to the individual. In aiming for this goal we hope, to paraphrase Sir Michael H o ~ a r d that , ~ we have not lost sight of what armies are for or what they spend most of their time doing. Birkbeck College, University of London References R. Alston, Soldier and society in Roman Egypt (London 1995). M. Banton, The policeman in the community (London 1964). P. Bidwell, "The exterior decoration of Roman buildings in Britain," in P. Johnsonwith I. Haynes (edd.)9-29. M. C. Bishop and J. C . N. Coulston, Roman military equipment (London 1993). M . Cain, "On the beat: interactions and relations in rural and urban police forces," in S. Cohen (ed.), Images of deviance (London 1971) 62-97. J. B. Campbell, The emperor and the Roman army 31 BC.-A.D. 235 (Oxford 1984). 8 Quoted in Keegan (1976,28). Ian Haynes A. P. Cohen, The symbolic construction of community (London 1985). J. C. N. Coulston (ed.) Military equipment and the identity of Roman soldiers. Proceedings of the Fourth Roman Military Equipment Conference (BARS394, Oxford 1988). J. Crow, Housesteads (London 1995). R. 0.Fink, A. S. Hoey, and W. F. Snyder, The Feriale Duranum (New Haven 1940). J. E. Gerstl, "Determinants of occupational community in high status occupations," Sociology Quarterly 2 (1961) 37-48. E. Goffman, Asylums (Chicago 1961). M. J. Green, A corpus of religious materialfrom the civilian areas of Roman Britain (BAR 52, Oxford 1976). G. A. Hillery, "Definitions of community: areas of agreement," Rural Sociology 20 (1955) 111-23. P. Johnson with I. Haynes (edd.), Architecture in Roman Britain (CBA Research Report 94,1996). D. L. Kennedy (ed.), The Roman army in the Easf (JRASuppl. 18,1996). J. Keegan, The face of ba ttle (London 1976). R. MacMullen, "The legion as society," Historia 33 (1984) 440-56. N. Pollard, "The Roman anny as 'total institution' in the Near East? Dura-Europos as a case study," in D. L. Kennedy (ed.) 211-27. N. Roymans, Tribal societies in Northern Gaul (Amsterdam 1990). G. Salaman, Community and occupation (Cambridge1974). B. Shaw, "Soldiers and society: the army in Nurnidia," Opus 2 (1983) 133-59. F. T o M ~ ~Community s, and association (London 1955). G. Webster, "Standards and standard-bearers in the alae," BJb 186 (1986) 105-16.