Archaeol. J., 159 (2002), 1–58 Writing the Legions: The Development and Future of Roman Military Studies in Britain By Simon James The province of Britannia and the Roman world as a whole were largely created and maintained by martial means. However, while Roman military studies in Britain achieved much during the last century, it is argued here that they have become isolated, theoretically stagnant and increasingly marginalized. Current archaeological discourse on the Roman world tends to concentrate excessively on social elites and the supposedly peaceful civil core of the empire. The historical reasons for this situation are discussed, the need for proper theorization of Roman military archaeology is argued, and its potential contributions to central debates in Roman archaeology are explored. Some possible routes towards reintegrating the martial aspect of Roman antiquity into mainstream archaeology are sketched out. INTRODUCTION Few would deny that the Roman empire was in large measure created and maintained by the threat, and the use, of armed force, albeit alongside other ‘social glues’ which bound it together (Terrenato 1998a; 1998b; 2001; Keay and Terrenato 2001). While, in considering Rome or any other past society, contemporary archaeological scholarship pays great attention to factors like politics, religion, and the means of production in the development of human societies, the impact of the ‘means of destruction’ or violence, its control and deployment, are frequently downplayed or even ignored, partly no doubt due to distaste at the inescapable central theme of bloodshed. Yet many societies have been shaped as much by armed force as by any other factor (Ehrenreich 1997, 143; Giddens 1985, 2, 18). Eighteenth-century Prussia, for example, ‘was not so much a State which possessed an army as an army which possessed a State’ (Howard 1991, 52; Ehrenreich 1997, 182; Anderson 1998, 157–58, 167–69). With reason, Rome is remembered as one of the most bellicose states in history. Surely, the roles of violence in general, and of warfare and the institutions which facilitate its waging in particular, merit more attention from archaeologists than they customarily receive. In the case of Rome, the result is an unfortunate relative neglect of one of the richest data sets from antiquity, a body of evidence highly relevant to some of the concerns of current archaeological theory such as identity and resistance. Yet in Britain at least, military studies have long been mostly the preserve of a small community of specialists. In recent decades, the dominant discourse on Roman archaeology in Britain has focused on civil society, especially the Romanization of the ‘provincial core’. This effort, led by people such as my former teacher, Richard Reece (e.g. 1988), and my 2 writing the legions friend and contemporary, Martin Millett (e.g. 1990), has largely shaped my own understanding of the Roman world. However, having a long-standing interest in the military side, I have long felt that this discourse was imbalanced, and has unduly avoided dealing with martial aspects, or indeed the wider spectrum of social conflict and violence. Although there are now encouraging signs of change, real advance demands a fundamental review of the subject, especially when almost all other fields of archaeology and related disciplines, including those studying the Roman world, have long been engaged in self-examination and theoretical overhaul. From my own observations it seems that, beyond emotional revulsion against violence, there are also deep structural reasons for what amounts to the virtual ‘ghettoization’ of military studies. The field has something of an academic image problem. The present paper is an attempt to understand and address this situation, and to advocate the reintegration of military studies as a central and vital part of mainstream research on Roman antiquity (for an alternative view, see Reece 1997). It seeks to provide an overview of what seem to me to be the key achievements and shortcomings of past Roman military studies conducted by British scholars, to draw lessons from them, to highlight important current developments in and around the field, and to suggest directions for future research. It does not seek to present a comprehensive historiography covering all the major scholars in the field, which would require years to complete and a weighty monograph to present. Neither does it seek to determine a detailed programme for the future, which would be both presumptuous and premature; the aim is to highlight issues and promising alternative approaches, and to encourage thorough-going debate. Crucial to understanding the discourse of any discipline, and fundamental to its direction and achievements, are its historical context and its internal social and political dynamics. These dynamics comprise a field of largely informal interaction on sites, in offices, common rooms and bars which is arguably as important as the formal academic discourse in lecture-rooms and publications. Such matters are usually deemed inappropriate or simply too dangerous to discuss in print, but because they so strongly shaped the field of Roman military studies and its attainments, my discussion necessarily deals not just with data-sets, methods and theoretical approaches, but also with personalities, perceptions and academic politics. THE DEVELOPMENT AND ACHIEVEMENTS OF ROMAN MILITARY STUDIES IN BRITAIN British scholars working on the Roman military achieved much during the course of the twentieth century. Systematizing existing traditions of fort excavation, scholars like F. Gerald Simpson and his successors elucidated the shape and the chronology of the Roman imperial presence on their home turf, the northernmost of frontier zones; it was Simpson, for example, who finally proved that the Wall really was of Hadrianic date (Birley 1961, 65–66). This resulted in perhaps the most detailed understanding of any Roman frontier, and a data-set of world value. Fieldwork outside Britain has been on a more modest scale, although significant, notably in the Middle East, exemplified by Ian Richmond at Masada (1962), and David Kennedy and Derek Riley on aerial writing the legions 3 photography of military installations (1990). More widely, British scholars have been prominent in international efforts to elucidate the shape of the Roman military as a whole, painstakingly mapping out an organization of bewildering complexity from an enormous body of epigraphy and other fragmentary texts. To name but a few, this scholarly community had included figures such as James Curle at the beginning of the century, Ian Richmond a few decades later, and John Mann (1996), Charles Daniels, Lawrence Keppie and Bill Hanson towards its end. Within it, the most lastingly-influential grouping, at home or overseas, has been the ‘Durham School’, led by Prof. Eric Birley (1906–95), who was active from the late 1920s until after his retirement in 1971. Birley’s scholarly reputation among military specialists in Britain, and equally in continental Europe, was very high; his Gedenkschrift was jointly edited by continental and British scholars, and was published in Germany (Alföldy et al. 2000). Brian Dobson has written: ‘In Roman army studies he was recognised as the successor to Alfred von Domaszewski, as the foremost scholar of his time, with papers of enduring influence’ (Dobson 1998, 232). In his Foreword to Birley’s collected papers, Michael Speidel noted the continuing value of works written a century ago by men such as Domaszewski and Ritterling; . . .their dedication to the sources is still our Genius. Eric Birley, more than anyone else, has carried on and refined their grand tradition with such skill and steadfastness, that today scholars throughout the world look to him as their Altmeister. . . and like those giants of the past, his papers are sure to teach and to inspire readers for a long time to come (Birley 1988a, v). In such contexts panegyric is permissible, but this does not much exaggerate Birley’s standing among many of his peers. His death at the age of 89 in 1995 marked the end of an era (Breeze 1996; Dobson 1995; 1998; see also Alföldy 1995). Birley and his School pursued a set of aims and strategies unusual in the overtness of their articulation, which epitomized, and to a large degree set the agenda for, British work on the Roman military for much of the twentieth century (Speidel 1992b, 18–19). Consequently, appreciation and critique of their work are central to much of what follows, for the Durham School shared with the rest of the scholarly community working on the Roman military a closely bounded tradition of discourse, defining what was studied, and how; equally important, as will be seen, it also implicitly determined what was not studied. DEFINING THE FIELD OF STUDY I believe the field of Roman military studies needs to be defined in the broadest terms, to include examination not just of the Roman armies and the military institutions of the state, but also the context of these within Roman society, culture and politics, and their interrelations — in peace as well as war — with societies beyond the frontiers. I would argue that such a contextual approach is absolutely essential to a proper understanding of the Roman military in particular, and of the Roman world as a whole. This broad project should be thoroughly interdisciplinary, exploiting impartially the full potential of all available sources of evidence, textual, representational, and not least material; the archaeological testimony is exceptionally rich. It should see an 4 writing the legions open and balanced interchange between the various disciplines and methodologies which have developed to deal with these diverse sources of information, from epigraphy and ancient history to archaeology. The desirability of this may seem self-evident, and indeed it is a century since the great military historian Hans Delbrück wrote that, ‘The history of the art of war is a single strand in the braid of universal history and begins with the latter’ (Delbrück 1975, 27; Keegan 1988, 48–49). However, it appears to me that, in Britain especially, few military scholars have conceived their research in relation to such a broad context, or to a truly interdisciplinary effort. The military aspects have usually been far more narrowly defined, in terms of two intimately-related fields: 1. ‘Roman Army Studies’ (e.g. Speidel 1992a), dominated by epigraphy and consisting primarily of research into the Roman military’s Order of Battle and prosopography (the study of careers through texts, especially inscriptions). 2. Roman Frontier Studies (Limesforschung), which concentrates on the archaeology of the military infrastructure of the frontier provinces. As we shall see, while dealing primarily with archaeological evidence, this has long been studied within a framework derived from ancient texts, which archaeological evidence was largely used simply to illustrate and elucidate. There has been relatively little acceptance that material culture may be able to tell us more or very different things about the Roman military. Outside these sharply defined foci, broader questions of research into military-related aspects of the Roman past, where studied at all, have largely been investigated by scholars based in other fields (see below). In my view, this narrow definition of what constitutes ‘proper’ scholarly discourse on Roman military affairs creates a serious structural problem, which prevents development of the kind of contextual, integrated, interdisciplinary approach advocated above. In Roman archaeology especially, there remains a significant gulf between civil and military studies. While prominent archaeologists have worked on both military and civil sites (e.g. R. E. M. Wheeler, I. A. Richmond and S. S. Frere), it seems to me that at an early stage different aims and rules were set for military and civil archaeology. I suggest that this early bipartite division of the discipline arose in part from the perception of distinct civil and military zones in the Roman Britain, and partly from the influence of particular prominent figures in setting academic agendas. THE PLACE OF ROMAN MILITARY STUDIES IN BRITAIN While research into the Roman military is a long-standing sub-discipline in Britain, it lacks any formal establishment. Although, of course, some scholars strongly specialize in it, there are no dedicated posts in the field. It also lacks a dedicated journal, although military subjects have been well represented in Britannia, the Romano-British archaeological journal of the Society for the Promotion of Roman Studies, and there is now a specialist Journal of Roman Military Equipment Studies. writing the legions 5 When I came into archaeology, in London in the late 1970s, study of ‘the Roman army and the frontiers’ was still mainstream in Britain, but already had a distinctive separate identity and seemed to have little real interchange with the rest of Roman archaeology practised in the United Kingdom. At that time, a group of scholars based mostly in the universities and museums of northern England and Scotland, centred on the members of the recently retired Eric Birley’s ‘Durham School’, were the dominant force in the field. By then, Roman military studies was already falling into disrepute among much of the wider scholarly community. Within the larger field of Roman scholarship in Britain, a conservative world not noted for self-critique, army studies were seen as one of the most conservative subdisciplines. As well as being morally or ideologically suspect to many, due to the inherent violence of their subject-matter, military scholars apparently had nothing of much interest to say to those working in neighbouring areas. Their work was and still is widely seen as narrow, unimaginative, unconcerned with theory and stagnant in methodology. A long-standing criticism is that it is too introverted, too focused on the anatomy of the military and its installations (Reece 1997, 3). The periodic international Congresses of Roman Frontier Studies ( Limeskongressen), in which British military scholars have been highly prominent, have often appeared as little more than catalogues of fort excavations (Elton 1996, viii). Many see the field as simply repetitive, publishing yet another tombstone or barrack excavation, dotting ‘i’s and crossing ‘t’s (e.g. Freeman 1996, 465). David Breeze has pointed out ( pers. comm.) that there have certainly been exceptions to this (e.g. Groenmann-van Waateringe et al. 1997, 217–60), but overall the charge remains well-founded. Major excavations of Roman military sites have continued in Britain and will surely go on, as the last two decades have seen (e.g. from Elginhaugh and Newstead in Scotland, to South Shields, Wallsend, Vindolanda and Birdoswald in the Hadrian’s Table 1. Numbers and proportions of archaeological sites of differing types excavated by five-year period from 1921 to 1995 (after Hingley 2000, 150–51, table 10.3) Year 1921–25 1926–30 1931–35 1936–40 1941–45 1946–50 1951–55 1956–60 1961–65 1966–70 1971–75 1976–80 1981–85 1986–90 1991–95 Military Military Major Major Villa % town town % 51 69 65 53 15 52 86 103 92 143 148 177 161 179 154 38 44 48 36 33 39 51 43 33 32 26 35 38 34 36 16 27 30 31 14 28 33 39 43 52 61 64 80 137 92 12 17 22 21 30 21 19 16 15 12 11 13 18 26 22 34 28 18 27 14 31 22 46 70 107 132 91 76 88 42 Villa % Small Small nontown town % villa 25 18 13 18 30 23 13 19 25 24 25 18 17 17 10 24 28 13 20 1 15 15 36 51 85 95 86 70 61 40 18 18 10 14 2 11 9 15 18 19 18 17 16 12 9 10 6 9 15 2 6 12 18 20 62 102 90 57 64 96 Nonvilla % 7 4 7 10 4 5 7 7 7 14 19 18 13 12 23 6 writing the legions Illus. 1 Numbers of excavations on different categories of Romano-British sites over time, showing the perennial dominance of military sites (from data in Hingley 2000, table 10.3) Wall Zone, to the current work at Alchester in Oxfordshire, not to mention legionary bases); indeed, military sites continued in their historical position as the most excavated category of Romano-British sites right to the end of the twentieth century (Table 1 and Illus. 1). However, for some considerable time, the academic study of the Roman military has been widely ignored as an irrelevant static backwater, while a wave of exciting innovation and change has been pouring into the rest of Roman archaeology and history. With exceptions, these developments, adopted with varying degrees of enthusiasm by those working on civil aspects of the Roman world, largely by-passed military scholars ( particularly archaeologists) and to date have been little applied to military data. This is apparently by mutual consent of two scholarly communities which have drawn ever further apart. But what are the root causes of this divergence? I believe that these are to be found deep in the history of scholarship. writing the legions 7 TOWARDS A HISTORIOGRAPHY OF ROMAN MILITARY STUDIES The genealogy of ideas is essential to explanation of the origins of prevailing attitudes and assumptions. Military historiography in general is an embryonic and neglected field (Keegan 1988, 47); that of Roman military studies is especially interesting, but as yet hardly documented. What follows is a provisional outline of what seem to me to be the main trends. Over the last twenty centuries, the power and prestige of the imperial Roman armies have never been forgotten. During the Renaissance, scholars and princes began using classical sources on the Roman military — selectively — as authorities and models for what a contemporary army should be like. Machiavelli’s Art of War was among the most important texts. ‘Perhaps Machiavelli’s most. . . significant contribution to the development of modern military science resulted from his revival of the idea of Roman legion organization, and all that. . . entailed. . .’ (Wood 1965, xxxi). Concerned at Florence’s dependence on foreign mercenaries, Machiavelli advocated a citizen militia, proposing an ideal army of 24,000 men modelled on a middle republican Roman consular army of two legions plus auxiliaries. In 1534 Francis I organized the French provincial militia into such a legion; apparently not very successful, nonetheless it formed the basis for some French units in the Wars of Religion (Wood 1965, xxxi). Seventeenth-century commanders like Maurice of Nassau and Gustavus Adolphus were also inspired by legionary organization and especially Roman notions of discipline; Maurice applied these ideas to the training of his armies, with great effect (Keegan 1993, 327; Wood 1965, xxxi). ‘Discipline’ was meant in both its senses; the idea of instant obedience to orders, but also training, in practices inculcated through drill and exercise, controlled sequences of movement in space. It thus also connoted rigorous method (c.f. ‘academic discipline’). Commanders like Maurice and their eighteenth-century successors made extensive use of translations of Polybius (especially Justus Lipsius’ adaptation of 1596: Grafton 1997, 179). Caesar, Vegetius and other writers also had a major influence, providing models for the disciplined manoeuvres which made for effective musket warfare (Wood 1965, xxxiii-iv). For example, in 1732, for his own amusement during a period of illness, de Saxe, the celebrated commander and military theorist who served under Marlborough and later led French armies, designed an ideal new legion-based army (de Saxe 1757; Phillips 1985, 276–83). It is interesting that de Saxe quotes examples of good military practice only from the preceding few generations, except for Roman ones. He took Rome as the acme of military achievement, except in fortification, the only area where he felt modern military science had truly surpassed the ancients (Phillips 1985, 259). For de Saxe, and for many others before and since, the memory of the legions was both so great and so distant in time that it had become a kind of Platonic ideal of an army, against which current military forces could be judged. Vegetius, composing his famous handbook of the military art in the late Roman period, was already creating a timeless, idealized legion which did not exist in his own day but which should serve as a model for contemporary practice (Vegetius, Epitoma rei militaris: Milner 1993). This 8 writing the legions imagined legion was something intangible, yet revered, which contemporary commanders might aspire to emulate. Vegetius, like earlier classical sources, continued to emphasize the centrality to military success of questions of morale, and the possibility of fear, rout, disaffection and mutiny among troops (Vegetius, Epitoma rei militaris, 2.28; 3.4, 25). It is important to note that writers from Machiavelli onwards discuss solely structures and techniques, not these less comfortable matters, which were omitted, either because these were inconvenient (since Rome was being used as a paragon) or, in the case of professionals like de Saxe, because such issues could be taken as read: it was the technicalities which interested him. From the seventeenth to the nineteenth centuries, a new metaphor for armies tended to encourage neglect of the idea that Roman (or other) soldiers were prone to emotional forces demanding close attention to morale and leadership. Officers and military historians (many of whom were or had been officers themselves), increasingly spoke of military formations and individual soldiers in terms appropriate to inanimate instruments. Armies were often conceived in terms of ‘chess-piece’-like units, an approach made feasible by the reconceptualization of the individual soldier as virtually an automaton, a component in a mighty, irresistible machine. Automata were a craze of eighteenth-century European courts: ‘Frederick II, the meticulous king of small machines, well-trained regiments and long exercises, was obsessed with them’ (Foucault 1979, 136). As James Boswell wrote after watching a Prussian regiment in 1764, ‘Machines are surer instruments than men’ (quoted in Anderson 1998, 168). A machine does exactly what it is told, and overwhelms human frailty. Such a view of soldiers fitted the power aspirations of generals and kings (Foucault 1979, 169). It was doubtless largely inspired, and apparently justified, by the beginnings of the actual mechanization of war, with the introduction of the flintlock musket, increasingly complex artillery and eventually other machine weapons. These were most effectively operated if the soldiers serving them ‘moved like clockwork’, with extreme precision and synchronization of movement and timing, achieved by rigorous drilling and brutal discipline (Foucault 1979, 151–55). Nonetheless, the urge to emulate Rome in politics, culture and military excellence continued, perhaps reaching its zenith with the French revolutionary armies and their overt Roman republican echoes (Schama 1989, 169–74), and with Napoleon and his imperial eagles, wonderfully encapsulated in J.-L. David’s vast oil-painting, Le Serment de l’Armée fait à l’Empereur après la Distribution des Aigles au Champ-de-Mars le 5 Décembre 1804 (Musée du Louvre 1989, 454–55, no. 187). The eighteenth and nineteenth centuries also saw wide equation of Britain with Rome, not least in civic and martial values. Edward John Poynter’s painting Faithful unto Death (1865) depicts a soldier staying steadfastly at his post in Pompeii as Vesuvius erupts. It embodies a Romanderived model of approved Victorian values and behaviour, social and military, especially unflinching dedication to duty and self-sacrifice. The painting is both a moral exemplar and, ostensibly, a kind of archaeological interpretation; Poynter believed (wrongly) that he was reconstructing an actual find from the city (Liversidge and Edwards 1996, 126, no. 43). writing the legions 9 Eighteenth- and nineteenth-century European armies, then, were in part attempts to imitate an idealized image of the Roman army. But as the Roman military became a source of rigorous academic study in the nineteenth century, I believe that this relationship of past and present subtly altered, and eventually reversed, along these lines: ‘if we, the descendants/successors of the Romans, are like them and equally successful in our imperial civilization, then surely the Romans must have been like us’. Here is a flip-over from emulation of a glorious mythologized past, to back-projection of our values and assumptions on to a Roman past which ‘must have’ been essentially like our present. For modern Western societies — at least, for the people writing the books — the best soldiers were the most automaton-like, the best formations the most like machines (seen in rigid formations and precision display-drill on modern parade grounds); these qualities were therefore expected of the much-vaunted Roman archetypes. This close identification and back-projection was reflected in Victorian popular culture, for example in Kipling’s Puck of Pook’s Hill, especially in the story ‘The Winged Hats’: for Hadrian’s Wall read the North-West Frontier of India, for Roman forts read British blockhouses, and for catapults read Maxims and mountain guns (Kipling 1906; Whittaker 1994, 2; Hingley 2000). Over the last century or more, the transition from emulation to back-projection has been completed. It is the British or Prussian armies of the last three centuries which have become idealized in memories still expressed in the ‘toy soldier’ regularity and glitter of Trooping the Colour. In recent decades it is the ghostly forms of these recent imperial armies which lie behind many of our assumptions regarding the organization, nature and practices of the Roman military — to which the modern term ‘war machine’, is indeed often applied, with its implication of superhuman invincibility. In his Roman War Machine, Peddie (1994, 1) quotes Slim’s analogizing of an army as like a clock; while he also discusses the central importance of morale, this remains very much an officer’s eye view, in which men and their morale are problems to be managed. Soldiers are tools to be husbanded and used for military objectives and are not the central subject of study (e.g. Peddie 1994, 9). Eric Birley also employed the machine metaphor for the structure of the Roman army (1988b, 5). This common conception was recently well expressed, and effectively challenged, by Goldsworthy: The popular image of the Roman army is of an incredibly modern force, highly organized and rigidly disciplined. When the army fought a war, it operated in a very methodical way. It advanced slowly and cautiously, and at the end of each day’s march it constructed a large camp, identical to the ones it had built, and would continue to build, on every other night of the march. . . In battle the army was just as slow and methodical. In all situations a legion would apply a carefully rehearsed battle-drill, the legionaries moving like components of some huge machine. Individual soldiers were mere automata, reduced by the army’s brutal discipline to a point where they were incapable of taking independent action. . . The Roman military machine was so perfect that it had no need for the overall direction of commanders. All they had to do was deploy the machine and point it in the right direction (Goldsworthy 1996, 283). This situation arises in no small measure from the role of soldiers — more specifically, officers — in Roman military scholarship. 10 writing the legions SOLDIERS STUDYING THE ROMAN MILITARY It is hardly surprising that professional soldiers, active or retired, have always played a prominent role in the field, from General William Roy’s surveys of Roman forts in the eighteenth century (Roy 1793), to the work of French, and especially German and Austrian officers, like Col Stoffel who dug for Napoleon III at Alesia (Anon. 1989), Schramm at the Saalburg (Schallmayer 1997), Groller von Mildensee at Carnuntum and Lauriacum (Niegl 1980, 160–61; see especially the early volumes of Der Römische Limes in Österreich) or du Mesnil, who excavated the famous siegeworks at DuraEuropos in Syria (du Mesnil du Buisson 1936; 1944). The great campaign of exploration of the German limes in the early years of the twentieth century was a joint civil and military enterprise, conceived by Mommsen with the active support of Helmut von Moltke, Prussian Chief of the General Staff, and was conducted in large measure by German army officers (highly controversial at the time, and involving the army in bitter political and academic infighting: Marchand 1996, 173–74). Archaeologists who were basically civilians have also applied their experiences of temporary peace- or war-time service, as did Wheeler to the siege of Maiden Castle (Wheeler 1943). Much earlier, Gibbon served as a captain of the Hampshire Militia from 1760–62, and rather endearingly remarked that ‘the discipline and evolutions of a modern battalion gave me a clearer notion of the phalanx and the legion; and the captain of the Hampshire grenadiers (the reader may smile) has not been useless to the historian of the Roman empire’ (Saunders 1981, 8). The Roman military continues to be written about by ex-soldiers (e.g. Peddie 1994). This is, to use an over-apt metaphor, a double-edged sword. Some aspects of military service are indeed virtually unchanging across time and space such as, for instance, the basics of survival in the field. Modern experience of service and battle can obviously give vital insights inaccessible to the armchair scholar, but may also limit views; for there remains the problem of archaeologists or historians who are exsoldiers projecting backwards their own experiences and expectations, which arose from one specific historical context, on to a very different one: ‘armies are always like this, and I should know’. For example, Dietwulf Baatz, ex-director of the Saalburg Museum, presenting a seminar on Roman catapults at the University of Newcastle Department of Archaeology on 1 November 1996, was quite specific that he considered that he had nothing to learn from others about Roman torsion artillery, because he had crewed 88-mm Flak guns in Berlin in 1945, so he knew how artillery was used, both recently and, therefore, in Roman times. Peddie projects modern military organization on to Rome to such a degree that he virtually invents a central ‘ministry for war’ (1994, 4; Campbell 1997, 480). Such long-held assumptions, largely implicit, of the basic similarity of Roman and modern professional armies have continued to be a feature of Roman military studies (e.g. Peddie 1994, xiii). It seems to me to have been a fundamental underlying characteristic of the approach evinced by Eric Birley and the Durham School. As we discover more direct evidence of the real Roman military, it seems we have continued to try to force it into our ideal, composed of largely implicit assumptions of what all regular armies must have been like (i.e. Frederick II’s army or the British writing the legions 11 Brigade of Guards in ceremonial parade order). But was the Roman military really closely analogous to modern armed forces? THE DANGERS OF ANACHRONISM Because historically our societies, and our armies, have to a significant degree modelled themselves on what we have thought Rome was like, it does not follow, as is widely and implicitly believed, that the Romans and their institutions must actually have been basically like ours. The Roman empire was of course not a modern industrial nationstate or trade-based maritime empire like that of Britain, and for that reason alone its military may be expected to have been fundamentally different in how it thought, and how it was structured and maintained (Giddens 1985, 1). Alston, from the viewpoint of an ancient historian, has put it well: Mainstream Roman history has itself adopted a position of broad consensus in the last twenty years, a consensus that is fundamentally opposed to that in Roman military history. Historians have come to lay great stress on the ancient aspects of the ancient world. This is no tautology. . . Life in the ancient world was very different from modern life and the institutions which shaped that life also differed greatly. There is, then, an incongruity between mainstream ancient history and military history. . . (Alston 1995, 4–5). Construing the Roman military as essentially like modern armies precludes the possibility of it being different, strange, or surprising — unless evidence to the contrary is overwhelming. Even then, ingrained resistance may be strong, as in the reaction to the shoes belonging to women and children found in an early imperial barrack block at Vindolanda (Van Driel-Murray 1995); the implication that women and children lived in barracks has proved hard to swallow for some, because ‘the Roman army couldn’t possibly have allowed this’. I have myself heard colleagues, not least from Germany, say this out loud, and it has long been asserted in print (e.g. Salway 1967, 23; Sommer 1984, 4, 30–31; see James 2001a, 83–84). Such responses are based on no specific evidence, but extrapolated from the attested legal ban on Roman soldiers being married (Watson 1969, 133–34), combined with a visceral conviction that such things were surely unthinkable in the Roman army, since women would never be found in any modern British or German barracks (something not true in practice anyway: Beevor 1991, 38). But Carol van Driel-Murray, who studied the shoes and suggested their interpretation, has not been constrained by this set of assumptions, because she was open to a wider range of permissible explanations, in this case drawn from the very different colonial military traditions of her adopted homeland, the Netherlands; for the Dutch colonial army in the East Indies did permit families to live in the barracks, a parallel which makes a similar practice in Roman times easier to envisage and accept (Van Driel-Murray 1995, 12–15). That women were indeed often, if not always, to be found in forts, is now gaining wider acceptance (AllasonJones 1999; Hassall 1999). Discourse about the Roman military often continues to be anachronistic, projecting too much of a modern framework on to the past (Alston 1995, 3–4; Luttwak 1976, xixii). In my view this is because Roman military studies, in both their documentary and 12 writing the legions archaeological aspects, have not interacted sufficiently with the rest of Roman studies and, in terms of theoretical development, have generally been left in a backwater. Why is this? In the case of Britain, I think it arises from the connections which army scholarship has neglected — with mainstream civil Roman studies in Britain and beyond — and those which it has emphasized; especially the link with German scholarship on the Roman army which, it is argued, shares basically the same ‘modernizing’ (or better, ‘presentist’: Matthew Johnson, pers. comm.) approach to the subject. Indeed it will be argued that British scholarship on the Roman military is and always has been essentially a major branch of continental, especially German scholarly tradition, while civil Roman studies have followed different trajectories. Other factors, not least the geography, sociology and politics of Roman studies as a whole, have reinforced this division. But before discussing these, it is useful to examine the role of wider social attitudes to military-historical scholarship in creating the present state of affairs. WIDER ATTITUDES TO STUDYING THE MILITARY AND WAR IN THE TWENTIETH CENTURY During the course of the twentieth century there has been a profound shift in general attitude regarding military affairs among historians and archaeologists. Suspicion towards, and disapproval of military studies has become widespread, as part of a general Western public rejection of things military since the World Wars, compounded in Britain by widespread cultural reaction against our own imperial past. This shift is palpable but rarely discussed. The change in basic attitude towards armies and warfare, from romanticized interest to suspicion and disapproval, can be seen happening in the carnage of the First World War. The deep integration of Classics into the education, ethos and ideology of the ruling classes of eighteenth- and nineteenth-century Europe, not least in Britain, resulted in great emphasis on values of patriotism, willingness to self-sacrifice, and the idea of glory. These values became even more ingrained with the growth of nationalism from the French Revolutionary and Napoleonic Wars (Best 1998b, 198), to the decades of rampant militarism following the Franco-Prussian War (Bond 1998, 32–33). In this environment, even if ordinary soldiers as a class were often despised, as Kipling’s poem, ‘Tommy’, makes graphically clear (Stallworthy 1984, 143, no. 81), publics generally approved of contemporary military forces, and study of past armies was highly respectable. This was the age of great military writers like the Napoleonic veteran, Carl von Clausewitz, whose seminal On War (Vom Kriege) first appeared in 1832, and Hans Delbrück, who wrote History of the Art of War within the Framework of Political History (1975), of which volume one, on Antiquity, was originally published in 1900. In Britain, Fifteen Decisive Battles of the World by Sir Edward Creasy, ‘England’s own philosopher of war’, went through forty editions between 1851 and 1894 (Creasy 1894; Keegan 1988, 49–50). In 1914 the young officers of the British Expeditionary Force went to war with their hearts filled with attitudes to battle drawn directly from Livy, Horace, and Sir Edward Creasy. But the unanticipated horrors of the First World War led to a writing the legions 13 profound and lasting change in public attitudes to warfare, exemplified by two of the Great War’s most famous poems. Rupert Brooke, who died in 1915, early in the war, is probably best known for his poem, ‘The Soldier’. It is expressed in the classical spirit of noble self-sacrifice and patriotism (Stallworthy 1984, 163, no. 103), a romantic vision of noble death which contrasts jarringly with the stark images of agonized ignoble suffering in Wilfred Owen’s ‘Dulce et Decorum Est’ written later, when poison gas had brought a new dimension of horror (Stallworthy 1984, 188–89, no. 140). This poem takes as its bitterly ironic title and closing words a line from Horace, which means ‘It is sweet and proper to die for one’s country’ (Odes, 3, 2). The contrast reflects the bitter disillusionment of a society. But another important change is also exemplified in the backgrounds of these two poets. Brooke was a product of the Classics-based elite education system of the time (Turner 1992, 7). Owen aspired to such an education but was denied it. The son of a railway clerk, he represents the voice of the lower classes striving for access to academic and cultural capital (Hibberd 1992, 2). The coming of full democracy with the end of the Great War, subsequent changes in the political landscape and improvements in state education of the masses led, over succeeding decades, to a profound shift in the make-up of the groups leading public opinion in general, and of those studying and writing about the Roman past in university departments. As a result of the First World War, study of armies and warfare lost credibility, becoming widely perceived as intellectually and morally suspect (Luttwak 1993, 4–5). While Freud saw warfare as a ‘natural’, if dark and frightening, aspect of the ‘normal’ psyche (1968, 76), many came to see it as a psychopathological state of society (Lesser 1968, 101). Ever since, there has been a widespread tendency to avoid things military, as they are regarded as inherently bad, if not evil. With the turning away of Western societies from imperialism, the particular association of the military with the creation of recent empires, now widely seen as an uncomfortable if not shameful aspect of our past, has added to the opprobrium. In the United Kingdom, another specific factor in the ‘alienation’ of martial affairs was the abolition of compulsory military service in the early 1960s; experience of armies from the inside has become the preserve of a dwindling number of professionals (thanks to Ian Haynes for pointing this out). In Britain, as in many other western societies, the twentieth century has also seen a more general ideological rejection of violence, albeit with mixed success in practice. However, the ‘denormalization’ of violence has resulted in, for example, its general abolition from the judicial system, and growing awareness, and condemnation, of domestic violence. These heightened sensitivities compound the generation of intense emotional reactions with regard to the subject matter of martial domination and violence, even regarding antiquity: few other subjects are today so difficult to discuss dispassionately. Acquired psychological attitudes and anxieties profoundly shape of the histories we write ( James 1993). Military studies, then, have long been academically embattled. Yet even before the carnage of 1914–18, the established tradition of military historiography already exhibited a marked ambivalence towards dealing with the bloody realities of warfare. Keegan has outlined the ways in which military historians generally shied away from dealing with the actual violence of war, beyond stylized ‘battle-pieces’. Military 14 writing the legions history at large has tended to be about generalship and strategy in war, or about institutional analyses of armies in peacetime. Little was written about the experience of the ordinary soldier until Keegan’s classic Face of Battle (1976). Much the same criticism may be made of Roman army studies, for much the same reason, but also because the individual’s experience is harder to reach directly through the available data (Campbell 1997, 479). Nevertheless, in effect military history of all periods had already voluntarily and implicitly demarcated some key potential areas of study as ‘offlimits’. In the social sciences as a whole, the general ‘turn from violence’ has been in large measure responsible for academic neglect of subjects like armies and warfare. Anthropology, for example, largely avoided such ground for most of the century (until the appearance of works like Fried et al. 1968; Ferguson 1984, 5–7). This neglect was significantly encouraged by longer-term theoretical trends in the social sciences. Giddens has emphasized that social scientists have long tended to ignore subjects like state violence in particular because of theoretical assumptions about the nature of social progress. Already in Victorian times, social thinkers such as Durkheim and Marx, who were to have a profound effect on the later development of so many disciplines including archaeology, were edging the study of violence and warfare out of the mainstream. Durkheim believed war was dwindling, and Marx in particular believed that armies and warfare would inevitably vanish from human societies in time, and so by implication study of these aspects in the past was not of great relevance (Giddens 1985, 22–30). More recently, processual or systems-based approaches to the study of present and past human societies have proved unable readily to accommodate or explain social conflict ( Johnson 1999, 77–78), while in archaeology many postprocessual approaches have exhibited no greater keenness to address issues of conflict and violence, despite their professed interest in power ( perhaps because, as Andrew Gardner notes, their interest is more in ‘power to’ than ‘power over’: pers. comm.). The study of a subject such as the Roman army, then, has for a lifetime been the subject of suspicion, disdain or even hostility from a substantial proportion of society in general, and of academia in particular. Within the context of Roman military studies in modern Britain, there were a number of special additional factors which also helped to shape the development of the subject. PERSONALITIES, POLITICS AND PRACTICALITIES While research into Roman Britain was largely pioneered by lone scholars and industrious local archaeological societies (Birley 1958, 3), over the last half-century intellectual leadership has lain with university archaeology departments. The distinction in the archaeology of the Roman province between the ‘civil zone’ of the south and east, and the ‘military zone’ of the north and west, came to be reflected in the geographical distribution of university departments and practitioners in the field. Until the 1960s, there was a physical gap, and it seems a growing gulf of comprehension, between the Roman scholars of Oxford, Cambridge and London in the South, and those of the northern schools — in Scotland, Durham and Newcastle. The divergence of North and South was reinforced by differing philosophies among practitioners, and writing the legions 15 personality clashes. At the time archaeology was a very small field in which strong individuals had great influence. In the North, for many in Britain and overseas the most prominent figure among a substantial group of talented scholars was Eric Birley. Within Durham University Birley was a controversial figure whose personality and style generated strong reactions. During his Mastership of Durham University’s Hatfield College (1949–56: Moyes 1996, 205–42), and his development of the Department of Archaeology of which he was first Professor from 1956, many of his contemporaries saw him as cantankerous. While they admitted that his research was of good quality, it was considered to be very narrow; and he ‘couldn’t write books’. It is a matter of opinion whether this is a fair or meaningful criticism; Birley published prolifically via papers and articles (Wilkes 1975). Of at least equal significance, he was also a generous and inspirational teacher who developed a circle of highly talented pupils (Dobson 1998, 220–21, 226–28) — now known as the ‘Durham School’ — while a far wider network of like-minded scholars across Europe were greatly influenced by him. Breeze has highlighted the strong bond of loyalty and sense of comradeship which existed between Birley and his circle, a major characteristic of the Durham School (Breeze 1996, xiii). As will be seen, it was no accident that this grouping of a professor and tightly-defined group of student-disciples reflected the situation typical of peer departments in German universities (Härke 1995, 52). Members of Birley’s circle I have spoken to express a personal loyalty to him fiercer than that I have encountered in the case of any other scholar in Britain; it truly has the characteristics of militarystyle esprit de corps, although this bond of loyalty can appear defensive, conveying a sense of ranks closed against expectation of outside criticism. This was, in part I suspect, both a cause and an effect of the tendency for Birley, and so his circle, to find themselves in conflict with others within Durham, and also in the context of ‘North versus South’. Tensions between scholars based in the North and those in the South were exacerbated by the physical practicalities of communications in Britain. Until as late as the early 1970s, when high-speed trains were introduced, trips from Durham, let alone Scotland, to London were considerable expeditions; day-trips for meetings were impracticable (thanks to Jeremy Taylor and Martin Millett for highlighting the significance of this). Such geographical distance helped foster the perception, and to a considerable degree the reality, of the isolation of the northern schools from the core of archaeological power and politics in the capital, Oxford and Cambridge. Those in the North increasingly saw the southern universities and London-based institutions as a ‘closed shop’. The latter were already inclined to see the Roman North as military and therefore inherently not very interesting. These dynamics helped to breed mutual suspicions. As late as the early 1980s, when I was a research student in London, (the then retired) Birley was still perceived as being at the heart of a ‘Wall Zone Mafia’ which guarded its academic and geographical territory jealously. One young excavator of a northern Roman fort told me in 1979 that, having committed some heinous breech of etiquette, he was advised by a (now-deceased) prominent frontier archaeologist that ‘if you want to play the Wall Game, boy, you’d better learn the rules first.’ There is a significant double entendre here: the ‘Wall Game’ is a traditional sport, 16 writing the legions little more than a violent brawl in the mud, played at a top English public school, Eton. This cultural reference is telling, with clear implications for the nature of Wall archaeology and its internal politics. Since the 1960s the proliferation and more even geographical distribution of archaeology departments have changed the dynamics; individual centres and powerful individuals have less impact on the field as a whole. Nevertheless, talking to archaeologists based in northern Britain in the later 1990s, I was struck by the evident bitterness which some still feel at the perceived virtual monopoly of academic power among the southern English universities. There was a further fundamental reason for the divergence of the military and civil schools of Roman studies in Britain: the intellectual roots, and continuing affinities, of the two wings were significantly different. Durham and other northern British universities had long tended to look across the North Sea for intellectual affiliations, and this made especially good sense for Roman military studies. From an early date, those studying the Roman army in Britain looked to Germany, where very similar archaeological and epigraphic evidence led to the early development of specialist studies in the field. Intimate co-operation and convergence of approach was the result, as will be seen. But meanwhile, those working in the civil zone of Roman Britain found many of their best archaeological parallels, and the most compatible communities of fellow-scholars, in other regions. Their divergence from the military community has accelerated in recent decades as civil archaeology has looked increasingly to other disciplines for its theoretical and philosophical underpinnings, not least sociology. The major intellectual connections of leading scholars in civil archaeology are at present to be found in the wider Anglo-Saxon world, and to varying degrees with Holland, France and Italy. Military researchers have to a considerable degree maintained their primary connection with Germany. THE GERMAN CONNECTION From the earliest days of the development of modern archaeology the similarities between the limites of Britain and Germany encouraged close, often cordial, links between frontier scholars, going back to Collingwood Bruce’s friendship with Emil Hübner (Bruce 1875, vi: information from David Breeze). The massive campaign of work on the Roman forts of Germany around the start of the twentieth century, conducted by the mixed civil- and military-run Reichs-Limeskommission (Marchand 1996, 173–74; Sommer 1993) made a profound impression in Britain (Freeman 1996, 465); the RLK was the personal brainchild of Theodor Mommsen, who was invited by Francis Haverfield to edit the inscriptions from Roman Britain — a strongly military body of material (Birley 1958, 4). It was this initial liaison with Mommsen over epigraphy which led Haverfield to take a decisive interest in Roman Britain, becoming the leading scholar of his day in the field; and so, unsurprisingly, Haverfield ensured that research on Hadrian’s Wall should take a comparative approach, particularly with regard to the frontiers of Germany (Birley 1958, 13; Freeman 1997). The tradition of Roman military studies established in Britain, then, is directly derived from the German tradition inaugurated by Mommsen, of study led by epigraphy and writing the legions 17 fort archaeology. Dominated by these aspects almost ever since, this has remained, to a notable extent, an Anglo-German co-operative venture. In the acknowledgements to his famous volume on Newstead, in addition to British luminaries (including Francis Haverfield, Reginald Smith and George Macdonald), James Curle notes the help of Heinrich Jacobi and other German colleagues — the only foreign scholars listed (Curle 1911, viii: thanks to Carol Van Driel-Murray for drawing my attention to this). Active co-operation continued after the First World War, when in 1928 Robin Collingwood introduced F. Gerald Simpson, Director of the Durham University Excavation Committee, to Ernst Fabricius, ‘the acknowledged leader in Limesforschung’. Fabricius visited the Wall to study Simpson’s methods, and was awarded an honorary doctorate by Durham (Birley 1958, 14, 16). The comparative emphasis was to be a particular focus of the work of the new Lecturer in RomanoBritish History and Archaeology, the post which in 1931 went to Eric Birley. It led eventually to the international congresses of Roman frontier studies ( Limeskongressen), Birley’s brainchild, the first of which he was to organize at Newcastle in 1949 (Birley 1952; Dobson 1998, 225). Strong academic and personal links with German scholarship have been maintained ever since by British scholars of the Roman military, down through the career of Birley himself, who spent much time on archaeological visits to Germany during the 1930s (Breeze 1996, xii), to, for example, Brian Dobson’s achievement in revising Domaszewski’s seminal Rangordnung des römischen Heeres (Domaszewski 1967) and his own Die Primipilares (Dobson 1978). Eric Birley’s son Anthony has long held a professorial chair in Düsseldorf, while German scholars have continued to work on Roman military archaeology in Britain (Sebastian Sommer on military vici: 1984; Eberhard Sauer at Alchester: 2000). This intimate engagement with German scholarship (Breeze 1996, xiv) as the most important outside link is perhaps the key reason for the divergence of British Roman military studies from the rest of British classical archaeology and history, and probably for its disinclination to partake of innovations in Roman civil archaeology. German archaeological scholarship in the decades following the Second World War was not noted for its own theoretical development (e.g. Härke 1995; 2000). Across the divide, the ‘Germanic’ flavour of British Roman military scholarship made (and makes) it seem alien to many civil-orientated Roman archaeologists in Britain. But what exactly have been the philosophy and aims of Roman military studies in Britain? These were set out most explicitly by Eric Birley, and pursued by the ‘school’ he established. THE WORK OF THE DURHAM SCHOOL The ‘Durham School’ flourished during the post-war decades, down to and beyond Birley’s retirement. In the introduction to the volume of his collected papers edited by his friend, Michael Speidel, he listed many of the major figures among his pupils and peers: . . . as a university teacher, I had the opportunity of guiding a number of my Durham pupils into the same area of research; and it is a pleasure to recall the published work and the co-operation of several of them: D. J. Breeze, R. W. Davies, Brian Dobson, M. G. Jarrett, Valerie Maxfield, 18 writing the legions G. R. Watson and J. J. Wilkes; and, away from Durham, I have been able to be of some help to Margaret Roxan in London, and Hubert Devijver in Belgium. For more than half a century, Sir Ronald Syme has been both friend and mentor; and in recent years I owe much to Géza Alföldy, and to Michael Speidel. . . One friend is alas no longer alive to receive my thanks: I dedicate this book to the memory of Hans-Georg Pflaum (1902–1979). . . (Birley 1984 unpag.). Birley himself succinctly summarized the foundations of the study of the Roman military on which he, his pupils, and like-minded scholars set out to build: Methodical study of the Roman army, founded by Mommsen in the nineteenth century, was continued by other German scholars trained in his school — Ritterling on the legions, Cichorius on the auxilia, Domaszewski on the ranks, careers and religion of the army (among other topics). For British readers, Cheesman’s monograph on the auxilia serves to show what a loss to such studies was his death in action in World War I. Meanwhile, French scholars had investigated the Roman armies of North Africa (Cagnat) and Egypt (Lesquier); and various English and American writers have contributed papers on one aspect or another of the Roman armed forces. But it has been left to a new generation of scholars to renew the task which those giants of the past set about. . . (Birley 1984 unpag.). As one would expect given these intellectual roots, the Durham School required a thorough grounding in textual sources, especially the epigraphy, and detailed knowledge of the archaeology of frontier systems. This Anglo-German-centred tradition of rigorous scholarship has, since its genesis in Victorian times, brought order and coherence to the fragmentary evidence, providing a framework for thinking about the military. The methods employed, and their formidable power, are readily seen in the works of Birley and many of his principal followers, many of which have been conveniently assembled in volumes of collected papers (among others: Birley 1953a; 1988a; Breeze and Dobson 1993; Speidel 1984; 1992a). Their nature and value can be exemplified by a couple of Birley’s own papers. The first, on Q. Veranius, briefly governor of Britain under Nero, illustrates how the study of a single career can give insight and understanding into the structures of the Roman imperial system, and the interactions and behaviour of the elite (Birley 1953b). Another, on the origins of equestrian officers, explains how prosopographical method can illuminate broader aspects of military organization — in this case an entire class of officers and their geographical and social connections (Birley 1953c).The tradition has clearly been very successful in continuing the work of Domaszewski in unravelling the structure of the Roman armed forces, and also in anatomizing its installations, and elucidating their development. Birley’s own career was really launched by his appointment in 1931 to teach Roman Britain at Durham University (first based in that university’s Armstrong College in Newcastle upon Tyne). Birley outlined his own intellectual history in the introduction to his collected papers. I quote these extensively because of their later relevance: My interest in the Roman army began when I was still at school, and then in Oxford I discovered the fascination of epigraphy, the most copious source for that subject, and the attractiveness of Roman nomenclature, hence also the study of prosopography. But from Oxford I was sent by R. G. Collingwood to work on Hadrian’s Wall, and up to the outbreak of World War II much writing the legions 19 of my time was taken up with excavations and research on and near the Wall, though I had begun to travel abroad, mainly to Germany and Switzerland, to study Roman frontiers and important military sites. It is for that reason that only three of the papers in the present volume were written before the caesura of nearly six and a half years in Military Intelligence, studying the German army and, incidentally, learning much about the methods of other modern armies (Birley 1988a, vii). His wartime intelligence achievements, in mapping the structure of the Wehrmacht, were of great importance, and have yet to be properly documented (Breeze 1996, xii; Shulman 1995, 30; Slusser 1993, 74; Dobson 1998, 221–22). An example of the product of his section of Military Intelligence (War Office 1944) was recently reprinted. It is clear that this labour helped shape his thinking, and also contributed to his reputation as a military scholar. When I returned to Durham, in 1946, I was able to resume my direct interest in the Roman army, and I found myself asking questions about its organization and methods which I would probably never have thought of, but for my practical experience of the ways in which modern armies work. Many of the papers here reproduced will serve to show the amount of my debt to those years spent away from my chosen field. . . (Birley 1988a, vii). Perhaps the clearest presentation of his philosophy, aims and methodology was made at a conference of classical epigraphers in Paris in 1952 (Birley 1988b). This is a key document. The approach and programme for the study of the military which he set out on that occasion is also taken as the starting point by his disciple Michael Speidel, in the latter’s 1989 review of past progress and future possibilities (Speidel 1992b, 13). Birley wrote: The study of the Roman army really involves an exercise in Military Intelligence (to use the British Army’s term for what the Americans, French and Germans call the work of their G-2, IIe Bureau and Ic respectively). . . The principal task of Military Intelligence is to discover and assess the strength, Order of Battle, organization, equipment and value for war of an enemy’s army (Birley 1988b, 3). This was a slightly whimsical address, showing how much his thinking was shaped by his activities in wartime Military Intelligence (which in 1952 will have chimed with the recent experiences of many of his audience). Yet these are telling passages, which, in my view, display both a clarity of purpose, and at the same time a narrowness of aim, which were reflected in the subsequent work of the Durham School. They reveal a fundamentally officer-class mindset, the outlook Birley acquired at school and university, sharpened in a particular way by his war experience. The principal tasks in studying the army, on this view, do not include consideration of the context of wider society and the economy, religion or, apparently, domestic politics. They do not include interest in the lives (as opposed to the careers) of soldiers, whose families and communities do not rate a mention. The realities of campaigning, and battle itself, are likewise ignored. Nor is there any direct interest in the question of, for example, logistics (Erdkamp 1998; Roth 1999; Adams 2001); the focus is on structures of fighting units, not even really on operations. Given Birley’s emphasis on the Order of Battle, this may seem strange, but such was not the task he set himself. Yet concentration on structures rather than war is in line with the wider traditions of 20 writing the legions military historiography outlined by Keegan (1988, 27–28). However, this is the kind of narrow mindset which Americans, at least, believed caused Rommel to lose North Africa; considerations like logistical factors, such as the harbour capacities which limited the rate of supply of the Afrika Korps, were outside that general’s aristocratic field of vision (Luttwak 1993, 7). Pursuing his Military Intelligence analogy further, it was clear that such a programme of study was to be led by epigraphy, supplemented by archaeology as a second-best tool: . . .its main sources of information, apart from what is derived from contact in the field and from interrogation of prisoners, are the reports of agents and the enemy’s own documents. In the case of the Roman army, it is the archaeologist who can provide contact in the field, by uncovering and interpreting its forts and frontier-works, and from them deducing something as to its value for war, or by digging up the sculptures or even some of the very weapons, which enable us to assess the quality and variety of its equipment. (Birley 1988b, 3). We are proceeding from the similar values of the Roman and modern European aristocratic and officer classes to investigate the great matters which have interested them in martial affairs, such as fellow-officers’ careers and the structures and articulation of military power. The interrogation of Roman prisoners, alas, is no longer possible; the reports of agents and observers — the surviving historians and other writers of the Graeco-Roman world — often have much to tell us, indeed, but seldom as much as we could wish, on the subjects with they deal, and there is a host of questions which occur to us and on which we look in vain for an answer. . . There remain the Roman army’s own documents. It is a mere accident that our Congress is only concerned with those of them which were inscribed on stone or metal or wood, or stamped on tiles; a full Military Intelligence survey must take into account not only the epigraphy but also the papyrology of the Roman army, for inscriptions and papyri alike represent, or at least reproduce, that army’s original documents, and if we can learn to read them aright, they have a very great deal to teach us about its strength, its Order of Battle, its organization — and how it actually worked. (Birley 1988b, 3). He then goes on to make a survey of what has been achieved, and what in his view remains to be done. He notes especially Mommsen’s diverse works, Domaszewski on Rangordnung, Cichorius and Ritterling on Order of Battle, and Cagnat and Lesquier, ‘who brought [the Roman army] to life as a working organism by their profound studies of the Roman armies of Africa and of Egypt.’ (Birley 1988b, 3). Among some other aims, Birley identifies the need for a study of legionary vexillations, a corpus of imperial military inscriptions, a monograph on the auxilia, another on the numeri, a study of soldiers’ cognomina, prosopographical studies of senatorial and equestrian officers, centurions, and work on the primipilares. He saw research on recruitment as the ‘largest problem of all’ (Birley 1988b, 4–10). As Speidel has detailed, this programme was in large measure successfully implemented, largely by Birley himself and his immediate circle (Speidel 1992b, 13). In its own terms, then, the Durham School set out to strive towards the objectives identified by Birley, and was substantially successful in achieving its aims, for which effort Birley and his circle earned the respect of their peers on the international stage. writing the legions 21 Yet, of course, all such programmes inevitably contain weaknesses, which usually look greater with lengthening hindsight, and the project of the Durham School is no exception. The tasks he set out do not, in practice, even cover so broad a range of activities as the already much-restricted definition with which he opened his survey; it is clear that in practice ‘study of the army’ amounts almost entirely to ‘study of the organization of the army and careers of soldiers’. It may be objected that in 1952 he was setting out epigraphic tasks to epigraphers, but there can be more even to military epigraphy than organizational studies; yet it is clear from the concentration on, and prestige given, to such work in ‘Roman army studies’ that they were considered to be the main task, as they still are to scholars like Speidel. It is interesting to see how Birley in the 1950s, and Speidel in the 1980s, both tended to elide ‘study of the organization of the army’ to become ‘the study of the army’; all else vanishes (Birley 1988b; Speidel 1992b). Birley’s characterization of the aims in studying the Roman military, then, are all ostensibly about the army as a ‘fighting machine’ and, above all, as an organizational system. Such an approach contains serious dangers. Construing his task as though it is a modern-style military intelligence analysis of the structure and internal workings of a potential enemy army, Birley’s approach automatically renders almost invisible Rome’s actual enemies, and the contexts in which they and the Romans interacted, both on the battlefield and in more pacific arenas. There is no indication that, to understand the military, he felt it necessary to look at the enemies it faced, or to pay much attention to its tactical, strategic, political, social or economic environment. Like an ant colony in a glass tank, the Roman army therefore tends to be seen in isolation from its real, working context (Freeman 1996, 468; Hanson 1997, 373). While scholars like Breeze and Hanson have emphasized the importance of including indigenous societies in work on the army (Hanson and Breeze 1991, 71–73; see also Breeze 1985), this situation still largely persists in the field. Perhaps Birley’s most important original contribution to Roman military studies was to draw overt methodological inspiration from outside the field. He had taken epigraphy and prosopography into the service of Military Intelligence, and now applied the techniques of Military Intelligence to Roman armies. Paradoxically, in this sense he was ahead of his time, foreshadowing the practices of later generations of archaeologists. The problem was the narrowness of application, and lack of reflection on the assumptions which this brought in its train. It is curious that, although the study of modern armies was deliberately chosen as a methodological model, specific modern military parallels are rarely cited in the research arising from this approach: apart from the ‘Military Intelligence approach’ itself, and the emphasis on interprovincial comparisons within the Roman empire, the enormous potential of comparative studies reaching beyond the Roman world itself remained off-limits. In contrast with common current practice in archaeology in general, it remains characteristic of traditionalist treatments that they rarely cite parallel cases from other periods or contexts, or draw overtly on the literature of other disciplines for ideas and inspiration. Among Roman military scholars the omission is not discussed or justified, in contrast to Millar’s explicit and reasoned rejection of such comparative practices for ancient history (Millar 1977, xii; thanks to Andrew Gardner for the reference). 22 writing the legions The primacy assigned to textual sources in Birley’s philosophy is evident. Birley was distinctly cool about the capabilities of archaeology, which was of less value than epigraphy, because it was less detailed and precise; and such value as it may have was again with respect to the war-fighting capability of the army. Other aspects are not of great significance. Birley never asked: ‘what can archaeology tell us about the Roman army in general?’ Projects like the excavation of forts and of vici are of interest only really insofar as they illustrate the main, predetermined aims of the study. Indeed, Birley was quite open about his attitude to archaeology; his title was ‘Professor of Romano-British History and Archaeology’, in that order, and he was not persuaded that archaeology had serious potential of its own. ‘Archaeology, by its very nature, cannot well be treated as a full and free-standing academic subject in its own right; it is essentially the handmaid of history, aiming at supplementing the historical record by observing, describing, classifying and explaining the physical remains of antiquity’ (Birley 1958, 18). It was not his primary personal interest anyway: ‘I am bound to confess that my own first academic interest is in Roman history. . . and that for me archaeology has always been a diversion, and not the real string to my bow.’ (Birley 1958, 19). Birley, then, did not even see an undeveloped potential in archaeology. Was this not an unduly conservative view, even in the context of the 1950s? In his defence, it is as well to remember that creation of autonomous archaeology departments, and the teaching of the subject as a first degree in its own right, was barely beginning. The perspective of scholars like Birley is quite comprehensible when the contemporary background is considered. The modern study of the Roman world began in the Renaissance as primarily a text-based field, and epigraphy was developed as a discipline before scientific archaeology, not least in military studies; already established, textbased scholarship long guided the agenda and constrained, even governed the development of archaeological approaches (Moreland 2001), since those who first came to Roman archaeology came to it from the Classics. Birley never taught a first degree in archaeology: he taught Roman Britain as a special subject to students who were first historians, classicists or others, some of whom went on to research in archaeology (Breeze 1996, xiii). However, with the growth and maturation of archaeology over the last thirty years, such views have long since ceased to be tenable. Despite the manifestly rich material record, prejudice towards texts, and against ‘mute’ evidence, meant that archaeology was not given as much intellectual attention as epigraphic studies; and insofar as it was used, the questions asked of it, being driven by a text-based historical approach, unsurprisingly did not always give especially satisfactory results. For these were often questions inappropriate to the nature of archaeological evidence — such as, for example, trying to trace on the ground the year-by-year progress of military campaigns and patterns of troop deployment during the conquest of Britain (e.g. Dudley and Webster 1965, Frere 1987, 49–125 Webster 1980) which requires chronological data much more precise than that available from archaeological evidence, except rarely from dendrochronology (e.g. Sauer 2001). Because it was not especially good at answering the central questions arising from the text-driven agenda of Roman Army Studies and Limesforschung, Birley’s dismissal of it to a secondary, supporting role seemed justified. But this was, then, something of a writing the legions 23 Table 2. A comparison of the numbers of excavations on civil and military sites in counties in the Military Zone Cumbria Durham Northumberland Lancashire Military Civil 9 4 18 20 20 18 35 7 Information from county sites and monuments records provided by Dr Jeremy Taylor. Data for Tyne and Wear not available. self-fulfilling prophesy. Reece has emphasized the very different rules of archaeology and documentary-based disciplines, and criticizes Roman military studies for having used the structures of the latter to frame their understanding, ‘while archaeology fills in the decorative details’ (Reece 1997, 3–4). That archaeology can make a leading contribution to military studies, when used to ask appropriate questions, is argued below. Although, elsewhere at least, archaeology has recently developed enormously in sophistication of theory and method, in Roman Army Studies it has never shaken off the subordinate position in relation to text-based research assigned to it by Birley, his peers and his pupils. In combination with the relatively low priority assigned by the Durham School to questions of the army’s role in wider society, or to its relations with surrounding communities, this has had serious consequences. For it resulted in a huge bias in the archaeology of the ‘military zone’ in Britain, especially in the North, where the effort expended on exploring the contemporary context of Hadrian’s Wall and other installations was dwarfed by the attention lavished on the minutiae of the installations themselves; by comparison with the number of scholars active on the military archaeology, George Jobey, who specialized in the exploration of the native settlement pattern, ploughed a pretty lonely furrow (Miket and Burgess 1984, 396–97). In Birley’s Research on Hadrian’s Wall (1961), it is notable how very little there is on the context of the Wall, on the landscape, or native settlement, society, or land use. The purpose of the Wall is almost entirely assessed from internal evidence of the forts and other structures (including to some extent vici: Birley 1935; Sommer 1984), and comparison with other frontiers, notably Upper Germany, Raetia and North Africa (Birley 1961, 269–73; for early work on Roman/native interaction in North Britain largely by scholars outside Birley’s circle, see Richmond 1958). The material remains from such sites are often very sparse, so it is understandable that people should be attracted to the more rewarding remains of the military installations, but even these were not explored evenly, with much less attention being paid to vici and cemeteries, for example, than to fort interiors. The result, a deep imbalance in our knowledge of the Roman landscape in these regions, is still with us. Table 1 and Illus.1 show that, even looking at the whole country, excavations of military sites have always outnumbered those of any other site category. The situation in the military zone is even more exaggerated. Table 2, based on data abstracted from sites and monuments records, shows Roman sites which have undergone scientific excavation in the English 24 writing the legions counties around the Wall Zone. Given their relatively small numbers in the total settlement landscape, the powerful bias towards military sites is apparent, and would be far more exaggerated if the number and scale of interventions on each site were also to be taken into account. (Information kindly provided by Jeremy Taylor, pers. comm.). However, here a note of caution is in order: it is important to emphasize that the narrow focus of purpose of the Durham School has by no means prevented its members, or of course other scholars, from working on a range of aspects beyond the ‘core tasks’ set out above, and increasingly so in recent years. To cite but a few examples, Birley himself conducted work on the Housesteads vicus (1935), and published on ceramics, particularly Samian (terra sigillata: Dobson 1998, 217, 223). He was instrumental in the production of Central Gaulish Potters (Stanfield and Simpson 1958, vii, ix; thanks to Vivien Swann for information), and also conducted important work on early antiquaries (e.g. Birley 1961, 1–24; Dobson 1998, 228). David Breeze has written on logistics, and presented a wide-ranging analysis of the Roman failure to conquer Scotland (Breeze 1993a; 1993b). Valerie Maxfield has produced an extremely valuable contribution on the relations between soldier and civilian (Maxfield 1995). Numerous other examples could be cited. Nonetheless, the overarching priorities of traditional military studies have always constrained effort in such areas, and approaches to them. Perhaps the best illustrations of this constraint are the works of two important figures in Birley’s circle, G. R. Watson and R. W. Davies, both of whom worked directly on the subject of the ordinary soldier. Watson’s Roman Soldier purports to be from the men’s viewpoint, yet it seems to me to be a work from which the soldier as a living, breathing person is conspicuously absent; this is a book about the system within which the miles was trained, and in which he lived and served. It is about laws, rules, customs and numbers. There is very little about the soldier’s own views or feelings, and, for example, no entries on women, children or families appear in the index; the fairly substantial section on marriage is mostly about legal matters and official attitudes towards such unions (Watson 1969 133–38). Watson does cite MacMullen’s Soldier and Civilian in the Later Roman Empire (1963) as a reason for paying so little attention to such aspects (Watson 1969, 9–10), but this seems inadequate justification, especially as MacMullen did not cover the early empire, on which Watson concentrates. Similarly, Roy Davies’ ‘Daily life of the Roman soldier’ (1989) is actually about the range of duties soldiers are known to have been detailed to perform; there is just a page on ‘offduty’ activities. The paper is on what soldiers are made to do, and on what is done to them, not on what they do for themselves; the men are ciphers, drones carrying out orders. Both these works, while fascinating and highly valuable, nonetheless represent, so to speak, the view of soldiers from the colonel’s office or the sergeants’ mess, not from the barrack room or married quarters. But very different approaches to the military and to soldiers can be taken, and have been by scholars in other disciplines, and other countries; for examples, among historians, in France Carrié (1993) has written of soldiers as social beings, while in the United States MacMullen has considered the legion as a social organization rather than simply a military formation (1984). Both these bring us far closer to soldiers as people. writing the legions 25 For the most part, even work such as that of Watson and Davies has remained firmly peripheral to the philosophy and main effort of Roman Army Studies. Breeze and Dobson’s own volume of collected papers includes, for example, work on the impact of the military on the indigenous population (originally published as Breeze 1985), but its title, Roman OYcers and Frontiers, confirms where the emphasis of the tradition has always lain (Breeze and Dobson 1993). It is clear that, even today, for many of the key figures the framework and aims have really changed little from those outlined by Birley in 1952. The epigrapher Michael Speidel, one of Birley’s chief intellectual legatees, still sees Roman Army Studies as essentially a text-based sub-discipline, above all based on epigraphy and papyrology (e.g. 1992b). The organization of the armed forces remains the supreme focus. At heart the field has exhibited little theoretical or methodological change, not moving beyond the essential groundwork of understanding the structure of the study matter, and its internal workings. Perhaps the development of the thriving sub-discipline of Roman military equipment studies has been one of the few, relatively modest, innovations since the 1960s, although, significantly, it was largely initiated by an outsider, H. Russell Robinson, Keeper of the Tower Armouries (Robinson 1975; Bishop and Coulston 1993; see also Connolly and Van Driel-Murray 1991, on reconstruction of equestrian harness). The style of scholarship exemplified by the Durham School has justly prided itself on the rigour of its method, yet there is always the danger that rigour can ossify into a strait-jacketed preoccupation with detail, myopic to the broader view. Of course it is important that detail should be mastered before wider statements can be made — but when there is so much information, as more and more is published, and syntheses may rarely be made at all, alternative approaches are left unconsidered. It seems to me that this situation has long prevailed in the military field. For example, there has been a dearth of synthetic works on the subject of the military as a whole, an area Speidel has identified as demanding attention (1992b, 18). The best-known recent and stillavailable general works are Graham Webster’s The Roman Imperial Army of the First and Second Centuries A.D. (1969; third edition 1985) and Yann le Bohec’s L’Armée Romaine sous le Haut-Empire (1989; translated as The Imperial Roman Army: 1994), are both limited to the early empire. Grant’s Army of the Caesars covers a slightly longer period (1974), while Keppie’s Making of the Roman Army covers the republic and JulioClaudian era (1984). Clearly, in writing synthetic books on the Roman military, the scale of the task is immense; yet synthetic works on other, equally massive subjects, such as the Roman East, are more often essayed (Millar 1993; Ball 2000; Sartre 2001), while it is widely agreed that there are actually too many surveys of Roman Britain, the most exhaustively explored province of all; and surely the Roman army is at least as attractive a subject for publishers. Apart from the few syntheses mentioned, we have, at most, works on particular aspects of the military, or individual frontiers (notably, that in Britain itself: Hanson and Breeze 1991, 59). This appears symptomatic of an over-specialized discipline short of ideas, in which knowledge of detail can be raised almost to a fetish, so that scholars are afraid to synthesize broadly, perhaps not least for fear of peer criticism. 26 writing the legions In summary, it seems to me that the agenda laid out for Roman Army Studies by Birley soon after the Second World War was sound, clear, in important ways bold and imaginative, and it has made major achievements which remain vital to the study of the field; but the research programme failed to develop or to keep pace with wider changes, and has become more and more outmoded since. Even accepting its demonstrable success within its own frame of reference, the traditional approach to studying the military remains a very partial one. As we have seen, Birley’s approach renders many major aspects peripheral and devalued, or even completely invisible, effectively excluded from discourse. As a result it has been introverted, and impoverished by its failure fully to contextualize the army. To use an appropriate figure, Birley led a highly effective advance, but it was on a front too narrow to sustain, and has run out of steam. The result became something of an impasse in Britain, of stale army studies and Limesforschung, with the rest of Roman scholarship becoming largely ‘military-sceptic’. However, for some time changes have been afoot in a number of areas, which have started to open up great potential for advances in new directions. These changes, perhaps inevitably, have come from outside the existing world of Roman military studies, notably from Roman historians, and from the wider social sciences where renewed interest in the study of violence and warfare is generating wholly new agendas. This movement arises from changes in society at large, and in the nature of the scholarly community itself. CHANGING PERSPECTIVES IN A CHANGING WORLD During the middle decades of the twentieth century, the twin traditions of epigraphy/ prosopography and Limesforschung were highly appropriate to the study of the Roman military; they followed a sound style of scholarship established in Victorian times, putting in the essential painstaking labour required to elucidate the structure of the subject matter. This effort was closely linked to the contemporary structure of academia, and the value system of the dominant social classes. Now, in a greatly changed society, a very different community of scholars is asking new questions of the past. As part of this, we need to explore new frameworks for thinking about the Roman military. As we have seen, one major change is the decline in military experience among scholars. This is, in different ways, both a loss and an advantage in studying the Roman military; it is certainly an important change. Of even greater significance is the profound demographic change in recent decades in the general archaeological community. This has shifted on more than one axis. There has of course been a major alteration in the gender balance; although women have long been active in Roman military studies in Britain (for example, Valerie Maxfield, Anne Robertson, Margaret Roxan and Grace Simpson), the increasing numbers and activity of women in the discipline as a whole, and the increasing importance of feminist perspectives, are changing the essential priorities of archaeology, and understandings of the shape of the past. writing the legions 27 There have also been fundamental shifts in the class and educational background of the archaeological community. Instead of coming from fairly privileged backgrounds with a private, classical education, students and practitioners have increasingly been drawn from middle-class, and to some extent working-class backgrounds, products of state education where the central role of the Classics has been replaced by the study of English. It has become increasingly rare for them to have studied any classical languages, literature or history at all before university. The result has been a fundamental change in the general tone of discourse about the Roman past since 1945. Few now enter Roman archaeology through the traditional route of classical education, coming to it via language and elite-framed primary texts, which tended to inculcate a ‘top-down’ view of society, from the centre of the empire. Today many, like me, have had little or no contact with the classical world at school, but started in general archaeology, and have approached Rome primarily through material culture. This tends to result in a more ‘bottom-up’ view of the Roman world, emphasizing provincial perspectives, and also an interest in the quotidian — and in the context of the army, the ordinary soldier, his family, and the communities they dwelt amongst. (My own perception of armies is very much from the viewpoint of the ranks. My paternal grandfather, Ernest James, whose experiences were instrumental in the foundation of my own interest in things military, was a regular Army Service Corps baker in the British Expeditionary Force that went to France in 1914 before serving as a lance corporal in the Durham Light Infantry.) These same decades have seen another vast transformation of Britain as a whole, to which, in my view, many pay insufficient attention; that Britain has changed from being a world-class military power with a relatively homogeneous domestic culture and a huge overseas empire to one of several, medium-sized European states, with substantial, recently arrived ethnic and religious minorities. The remarkably sudden dissolution of the British and other European colonial systems has seen the development of post-colonial literature and critique, and a sea-change in domestic British attitudes towards our imperial past and our multi-ethnic and globalizing present, at least among some sections of society — notably academia itself. All this has inevitably affected views of the Roman world, and there has recently been a race to deconstruct our received attitudes towards, and assumptions about the nature of Roman imperialism (e.g. Webster and Cooper 1996; Mattingly 1997; Hingley 2000). As a result of these shifts in philosophy, an interesting range of new views has arisen regarding the Roman world, drawing on many differing perspectives, ancient and modern. The broad demographic and socio-cultural shifts in academia mean that we have begun to escape the narrow confines of traditional discourse on Rome and its military aspect, not least by an increasing willingness to draw inspiration from developments in other disciplines. FUTURE POTENTIAL: NEW APPROACHES TO STUDYING VIOLENCE, WARFARE AND ARMIES Other fields of the social and historical sciences now have an immense amount to offer the study of the military aspect of the Roman world, in terms of philosophical 28 writing the legions approaches, and potential comparative frameworks. Over recent decades, scholars in a number of disciplines have worked beyond the earlier blanket rejection of dealings with the realm of violence, and have expressed the need to deal more directly and critically with the military aspect of human societies. Some major innovations have been made in the study of martial and other violence, and also of military organizations. These have direct relevance for the study of the Roman military, and a number are already being exploited, as will be seen. The following is not an exhaustive list, but highlights examples of work which suggest potentially fruitful future directions, by broadening the agenda, placing the military firmly back into its wide range of original contexts and, not least, by a critical reappraisal of basic concepts about its nature and workings. Since the Vietnam War, and especially since 1980, interest in warfare and violence has blossomed among anthropologists, and has produced a host of insights with obvious relevance to understanding the Roman case (Ferguson and Whitehead 1992a, xi; Riches 1986; Schmidt and Schröder 2001). A good example is work on the effects of exposure to contact with imperial societies on the war-making practices of nonstate societies; increased militarization and intensification of warfare result in many cases, suggesting that the existence and behaviour of states like Rome may have actually helped generate the bellicosity it encountered among its ‘barbarian’ neighbours (Service 1968; Ferguson and Whitehead 1992b), although this may be exaggerated (Keeley 1996). Anthropological understandings of the nature of war have also been under scrutiny. Lesser (1968, 101), for example, challenged the view that peace and war are normal and aberrant respectively, instead suggesting they be seen as alternate systemic states of society — a conception perhaps especially applicable to the case of Rome. The view of war-as-pathology (Lesser 1968, 101) may then be seen as a recent, specifically Western cultural judgement deriving from experiences in the twentieth century which, although probably valuable to our survival in the present, may seriously mislead and hamper understanding when projected inappropriately on to past societies and their conflicts. Among sociologists, Anthony Giddens has pioneered an interest in the role of violence in the service of the state, primarily the modern nation state, but he also has much of interest to say regarding ‘traditional states’ such as Rome (Giddens 1985). Among a host of other observations of relevant to the Roman world, his far-reaching survey emphasizes the centrality of ‘control of the means of violence’ to the very existence of state systems which lacked the means modern nation-states possess to control populations and enforce their will: especially efficient, bureaucratized systems of surveillance (Giddens 1985, 2, 18, 64). Among historians, it is now a generation since the ‘new social history’, with its tendency to focus on the mass of ordinary people and daily life rather than elites, was extended to the study of armies, soldiers and warfare. In modern history during the 1970s, a perceived gap in traditional historiographical treatment of war — how armed forces and war related to the societies involved — led to the development of the field of ‘war and society’, or the ‘new military history’ (Best 1998a, 5). Landmarks during the 1970s were Arthur Marwick’s ‘War and society’ course for the Open University (Marwick 1974; Best 1998a, 6), and a series of books treating the subject period by writing the legions 29 period (recently reissued: e.g. Anderson 1998). Around the same time, as we have seen, Keegan’s classic Face of Battle largely pioneered the study of the soldier’s view of the realities of battle itself (1976), now a well-explored genre (e.g. Ellis 1990; Hynes 1998). Such works reveal both the horrors of war, and aspects of its reality which are in some ways even harder to deal with, intellectually and emotionally. New historical studies exploring the complex responses of men — and women — to the act of killing in recent wars, strongly suggest that the idea that only psychopaths could enjoy battle, and not be traumatized by it, is mistaken. While millions during the twentieth century were psychologically scarred or destroyed by war (Holden 1998), it is now argued that many other service personnel found campaigning and combat the most intense, memorable and, if unpleasant, still largely positive experience of their lives; people who were quite ‘normal’ in peacetime, before and after war, could even exult in killing during combat (Bourke 1999). The very positive emotions often associated with joint action against a perceived common enemy, and the ambivalent feelings of terror, ecstatic excitement, revulsion and — harder to comprehend — the sense of fulfilment and even piety which are sometimes associated with war and violent death are explored in Barbara Ehrenreich’s Blood Rites (1997). This marks a major advance in understanding human attitudes and emotional responses to war, especially their intimate connection with religious behaviour. Her theme of the sacralization of war is especially relevant to Rome, where so much blood was shed not just on the battlefield and the altar, but in the arena. It is a fine example of innovative thinking on violence, and a new willingness to deal with it head-on. Perhaps in consequence of such recent work, studies, conferences and publications on violence and warfare, if not so much on military institutions, are once more becoming almost fashionable among archaeologists too, particularly prehistorians (Keeley 1996; Carman 1997; Carman and Harding 1999; Osgood et al. 2000; Guilaine and Zammit 2001; James forthcoming a; Parker Pearson forthcoming). Within Roman studies themselves, there have of course always been alternative views of military affairs available, not least in disciplines such as ancient history, and in other national traditions. We have already seen examples of social-historical approaches to Roman soldiers (MacMullen 1963; Carrié 1993). However, the first really effective challenge to the status quo, at least with regard to Frontier Studies, was Luttwak’s Grand Strategy of the Roman Empire (1976). A modern US strategist, his anachronistic application of a modern military systems approach was fundamentally flawed, and was based more on assumptions relevant to defending the West against the then-perceived Soviet threat than on concepts relevant to Rome (Mann 1979; Whittaker 1994, 62–66). Yet it is a fine example of a fruitful error, which has stimulated much valuable discussion and has permanently changed the way the military is written about: Luttwak put the question of frontier strategy firmly on the agenda (Freeman 1996, 465). Luttwak was very much an outsider, and it continues to be the case that innovation in thinking about the military facet of the Roman world has come primarily from those outside the communities of established scholars working in Roman Army Studies or Limesforschung, particularly ancient historians like Keith Hopkins (1978), William Harris (1979) and Brian Campbell (2002). Another good example of this is Rich and 30 writing the legions Shipley’s War and Society in the Roman World (1993), many of the papers in which are by scholars who are not particularly military specialists, who tackle a variety of pertinent issues, not least the sources, scale and nature of violence in Roman society and imperialism. This places emphasis on the context in which the Romans developed their military and political systems. Italy in the mid-first millennium b.c. was a very violent place, both between and within societies; Campanians, not Romans, invented gladiators. Roman society itself remained very violent. Woolf ’s contribution in particular provides an overdue challenge to the view, accepted since Gibbon, that the pax Romana meant general peace and prosperity within the empire; in fact it meant an absence of formal war, but contrast of peace and war was often more apparent than real in antiquity. High levels of domestic and juridical violence were accompanied by a lot of other unsanctioned mayhem; whether the latter is termed crime and brigandage or resistance and guerrilla warfare is a matter of context and political orientation (Woolf 1993). Also highly influential has been Whittaker’s Frontiers of the Roman Empire, a social and economic study of the subject, which starts with profound insights into the nature and impact of structures of Roman thought regarding cosmology, boundaries and war (Whittaker 1994, 10–29). The development of a post-colonial world view has also resulted in more critique of assumptions about what the Roman military in particular was doing. Isaac, notably, has criticized the assumption of its function as being mainly frontier protection: ‘Modern Europeans. . . who disapprove of foreign conquest and the suppression of liberty, still find much to praise in the principate because it is assumed to be a period of the defence of the provinces against barbarian invaders, since modern states may legitimately defend their borders against attack. . .’ (Isaac 1992, 2). He rightly emphasizes the fact that ‘the Roman army was in many areas and periods an army of occupation or an internal police force and [Roman historians] should try to realize the consequences for those who suffered occupation’ (Isaac 1992, 2). In contrast to the traditionally (at least implicitly) pro-army view, he focuses attention on how the Roman military was perceived by outsiders. The tendency to assume that, after initial conquest, Romanization normally followed ‘naturally’, and constituted progress, or at least a process, generally desired or accepted by the colonized is now regarded as highly suspect. Additionally, there is growing awareness of the importance of the perspectives and roles of Rome’s neighbours and enemies in shaping the development of Rome and therefore her military. How, for example, did the societies of the Parthian and Sasanian empires see Rome, and their dealings and conflicts with her? Because, hitherto, these societies have been seen as an amorphous ‘threat’, effort has been largely concentrated on the Roman military ‘response’, not on examining the confrontation from the other side, to discover what we can about the real nature of, in this case, Partho-Sasanian military power, ideology, and actual political and martial ambitions. Kennedy has made a refreshing attempt to address this issue, emphasizing that the history of the first century b.c. to the third century a.d. saw no real threat from the East to Rome’s power in the Levant, but was a chronicle of episodic Roman aggression which gradually undermined the Parthian state and was a major factor in its collapse (Kennedy 1996b). Wells has sought to view Roman interaction with the northern ‘barbarians’ writing the legions 31 from the perspective of the latter, and outlining their active contribution to the development of provinces and frontier zones (Wells 1999). Of particular interest is the work of scholars like Terrenato, which emphasizes the key role of negotiation between the Roman and other Italian elites in the unification of Italy, which then becomes seen as a process of federalization under Rome as much as a conquest (Terrenato 1998a; 1998b; 2001). In this persuasive view, military power was not the cause of Rome’s rise to the status of Mediterranean superpower, but a consequence of its political skill in uniting many equally or more bellicose Italian peoples. ‘Decentring’ Rome like this, by seeing it as one part of far broader political and cultural contexts, results in much more sophisticated conceptions of what it was doing and why. No longer can we justify explaining Roman military power purely in terms of the internal workings of Roman society, ideology and military institutions. The idea that the imperial military was simply ‘defending the empire’ against often remarkably vaguely-defined ‘external threats’ begins to look naı̈ve, indeed politically disingenuous, representing implicit identification with the imperial regime, and the way its ‘spin-doctors’ (like Virgil) would have liked Roman power to be portrayed. An interesting example of a challenge to traditional assumptions is provided by Drinkwater’s review of the ‘Germanic threat’ on the Rhine in the fourth century; he argued that it was greatly exaggerated, a myth fostered for their own mutual benefit by the ‘western emperors, generals, administrators and local aristocrats’ to validate their power and control of wealth (Drinkwater 1996, 27–28); the soldiers should surely also be added to this list. The value of overtly including the army’s antagonists when considering how it achieved its aims in war is to be seen in Millett’s Romanization of Britain (Millett 1990). Millett (a ‘military-sceptical’ archaeologist) presented a strategic appreciation of the Roman conquest of Britain, rooted in clear conceptions of how both Roman and late iron age British societies worked, and how they interacted; he was able to relate the military archaeology of the early conquest period not merely to the topography of the island, but to the ideology and mentality of the invaders ( politically and strategically Romans dealt with rulers and peoples, and places linked by routes, rather than blocks of territory: Woolf 1993, 179) and to the complexity of the multiple late iron age British polities (Millett 1990, 1–64, especially 44). In another macro-scale overview of military-civil interaction in a particular region, David Mattingly has looked at the Roman army against a detailed consideration of the archaeology and history of Roman North Africa (Mattingly 1992). He argues that the Romans went to great lengths to avoid war in the early imperial Maghreb, and showed great skill in doing so for prolonged periods. These constitute good examples of the potential of a more holistic contextual approach, taking the military out of its glass case, and putting it among living societies, an approach Hanson and Breeze have advocated for Roman Scotland (1991, 72–75). At the opposite end of the scale from the interaction of entire societies is the individual soldier in battle. Building on the approach pioneered by Keegan for medieval and later warfare (1976), and applied by Hanson to war in classical Greece (1989), Adrian Goldsworthy has examined the realities of battle, and especially the Roman soldier’s experience of combat, during the late republic and early empire 32 writing the legions (Goldsworthy 1996; see also Lee 1996; Sabin 2000). As we have seen, he has attacked the ‘machine analogy’, arguing that the Roman armed forces at war were vastly more ramshackle, ad hoc, and above all flexible and adaptable than is usually assumed. He demonstrates that Roman armies did not always move methodically and slowly, but were motivated by a doctrine of constant, often immediate attack, which not infrequently went disastrously wrong. Goldsworthy also powerfully emphasizes the fundamental importance of leadership and morale, as opposed to brutal discipline alone (Goldsworthy 1996). Such work has led the way in at last tackling directly the personal experience of the individual soldier in ancient war, not least among the horrors of the battlefield. If the human debris of battle is rarely found by archaeologists, when it appears, most often in the ruins of sacked cities, it retains the power to shock, as a salutary reminder of the full horror of war in general, and of Roman ferocity in particular, a characteristic and a tool of their style of war-making (Keegan 1993, 265–66; on the special horrors of the sacking of cities, see Ziolkowski 1993). The mutilated human remains found at Valencia, apparently victims of Pompey’s sack of the town, are a gruesome illustration, and the fact of their publication in graphic detail may itself be symptomatic of a new willingness to confront these issues (Ribera i Lacomba and Calvo Galvez 1995). Roman archaeology has, of course, also produced valuable new insights into the military off the battlefield as well. A few examples may be cited, which highlight the potential of studies of spatial layout of sites, of artefacts and ecofacts, of the contextualization of military archaeology against the wider social landscape, and of work drawing jointly on material, documentary and representational data. There are, for example, interesting new approaches even to the study of things as familiar as fort plans. The layout of well-explored military sites permits sophisticated analysis of the use of space (Hoffmann 1995), and especially their employment as settings constructed for surveillance and the exercise of power, inwardly as well as outwardly; in my view early imperial military base design was more to do with the control and containment of the soldiers inside, than with protecting them from enemy attack. With regard to studies of material recovered from sites, Van Driel-Murray’s work on the shoes from Vindolanda, discussed above, has been significant in crystallizing doubts over long-held assumptions regarding the social lives of early imperial soldiers, the contemporary attitude of the military towards soldiers’ dependants, and to the rules governing social access to space in forts. There have also been environmental studies, of individual sites (e.g. the plant material from Mons Claudianus: Van der Veen 1998), and regional surveys, for example of plant and vertebrate remains (Huntley 1999; Huntley and Stallibrass 1995; Dobney 2001) which have been highly informative regarding diet, ‘foodways’ and supply practices. These indicate that we have hardly yet begun to explore the potential of such data to tell us about fundamental aspects of life, military or civilian, from cultivation and procurement to processing of materials, foodstuffs and dietary practices, all of which may be highly culture-specific (e.g. patterns of meat consumption: King 2001). Among the most promising approaches are studies of multiple material culture categories in spatial context. Lindsay Allason-Jones has conducted pioneering and writing the legions 33 revealing work on ‘small-finds’ assemblages from Hadrian’s Wall turrets (AllasonJones 1988), while at York legionary base, Hilary Cool has been developing methodologies to study even wider ranges of material, including glass and ceramics as well as small-finds, in their archaeological contexts, in search of complex patternings and associations (Cool et al. 1995). We will return to this promising area below. I am convinced of the great potential of such artefact studies, in conjunction with documentary and representational data, in elucidating the nature of life and experience in military communities, particularly if used in a more sophisticated framework of social theory. For example, some of my own work explores specific dress details and bodily practices which did not simply reflect and express Roman soldierly identity, but were fundamental in constructing and articulating it ( James 1999). Finally, a direct effort to look at the activities of the military in its wider geographical and social context is to be seen in Bradford University’s Newstead project. This attempted to look in detail at the social make-up, use of space and economy of a particular military base and its satellite ‘civil’ community, in the context of the indigenous society and settlement landscape, which have also been subjected to detailed work. The full results are keenly awaited, in the light of a series of important papers already published (Clarke 1994; 1997; 1999; Clarke and Jones 1996). CURRENT REALITY: THE ROMAN MILITARY IN SCHOLARSHIP IN THE EARLY TWENTY-FIRST CENTURY The varied approaches outlined above hold out the possibility of a much fuller understanding of the Roman military in relation to the wider contemporary world, through the creation of multiple histories. However, persistent structural problems within academia threaten their full exploitation. ‘Multivocalism’ is surely desirable in the writing of histories, especially of colonial contexts, but in scholarship at present there seems to be less a constructive polyphony than a cacophony exacerbated by a strong degree of mutual deafness. Exploiting the potential benefits of multiple perspectives demands the routine crossing of theoretical and disciplinary boundaries, but this is not happening on anything like an adequate scale. Rather than crossfertilization, we see over-specialization of individuals, and fragmentation of the archaeological community in particular into numerous small theoretical factions, often displaying lack of openness to alternative approaches ( Johnson 1999, fig. 12.2). The problems of compartmentalization, and the urgent need for greater interdisciplinarity, are all too evident, for example between Roman and iron age archaeologists in Britain (a problem highlighted in the case of Scotland by Hanson and Breeze 1991, 72). Among later prehistorians dislike of Rome can be intense and visceral; a few years ago one prominent iron age specialist said to me, only half jokingly, ‘I hate the Romans.’ Such attitudes inhibit joint investigations of the Roman part-conquest of Britain. But all is still far from well within Roman studies themselves. The community of scholars working on the Roman world, military or civil, whether describing themselves as classicists, epigraphers, ancient historians, or archaeologists, is today divided by mutual scepticism, often based on misunderstanding and incomprehension. For some, it remains ‘business as usual’ within the traditional 34 writing the legions framework. While few scholars would articulate it openly, there are still some who, at least implicitly, identify with imperial powers such as Rome. In this kind of approach (still also a widespread popular view), Roman civilization, and the armed forces in particular as its literal and metaphorical cutting edge, are seen as admirable. To parody somewhat, the first truly professional army brought the wonders of civilization and central heating to the poor benighted primitives around the periphery of the classical world, wisely stopped them killing each other, and defended them from the incorrigible and the uninitiated of the outer darkness. Speidel, commenting on the ‘achievement’ of the Roman army, has written in this congratulatory mode: ‘Not only did the army uphold for centuries a consistent order from Scotland to Nubia, and from the Caucasus to Morocco, but that order is reflected and preserved in rich works of literature and law, in abundant archaeological sites and finds, and above all [sic!], in countless inscriptions, reliefs and papyri’ (Preface to Speidel 1984, unpag.). In this view, the Romans were essentially like us, fundamentally civilized and peace-loving except when obliged manfully to defend themselves (an assumption that lies behind Luttwak’s Grand Strategy, in which the Roman empire was used as an analogy of NATO, and vice-versa: Luttwak 1976). Within its own terms, the Roman military clearly was spectacularly successful, but the nature and basis of that achievement must be examined critically; such identification solely with the Roman view is no longer tenable. Since we are aware that the British empire used the Roman empire as a model (Hingley 2000), during the rise to ascendancy of post-colonial world-views many have consequently rejected such Romanocentric orientations, the civil counterpart of which has been the notion of acculturative ‘Romanization’. This has been seen as a ‘natural’ post-conquest process of acculturation by a ‘superior’ culture, which ‘natives’ adopted unquestioningly, if with varied degrees of success or competence. Such notions have been subjected to sustained attack from post-colonial perspectives (for useful collections of recent perspectives, see Webster and Cooper 1996; Mattingly 1997). While post-colonial thinking in general seeks to move beyond a simple reversal resulting in identification with the conquered, in practice there is a serious danger that, viscerally if not overtly, some have switched to the opposite extreme, to seeing the Romans — especially soldiers — as a fundamentally nasty Them, from whom we wish to distance ourselves as much as possible, as we distance ourselves from the enormities of our own empire. However, as Woolf has observed, ‘In recognising and rejecting any sense of identification we may have with the Roman empire, it is not necessary to side with their opponents. Rather, we should step outside the debate and assess its structure and terms’ (Woolf 1993, 181). In practice, most people are not so polarized in their views on the subject: many, perhaps the majority, still simply find the whole subject inherently unpleasant and uncomfortable, and lacking in interest or relevance. Even among those studying the Roman military, mutual incomprehension reigns. A good illustration is the historian Goldsworthy’s assertion of the primacy of texts; his statements that the literary record ‘needs to be properly studied before we can hope to understand the archaeological evidence of camp-sites’ (Goldsworthy 1996, 77) and that archaeological information may be dismissed ‘as too clumsy to reveal what the writing the legions 35 soldiers in the fort were doing on a day-to-day basis’ (Goldsworthy 1996, 11) are simplistic, to say the least. We have seen how archaeology can make major contributions to exactly these issues, and indeed tell us of ‘people without history’ in and around the forts who would remain invisible no matter how hard we analyse Vegetius or Pseudo-Hyginus. Even an experienced archaeological fieldworker like David Kennedy, in his survey of evidence for the eastern Roman army, still discusses the record entirely in terms of installations and texts; artefacts such as military clothing, fittings, horse-harness, arms and armour, let alone ceramics or other ‘domestic’ and environmental material, are not even mentioned as potential sources of information (Kennedy 1996a, 12–15), despite some major relevant discoveries (e.g. Dura-Europos, Syria: Rostovtzeff et al. 1936, 439–66; James 1986; 1997; forthcoming b; Hebron, Palestine: Weinberg 1979; Nawa, Syria: Abdul-Hak 1955). It is easy to understand how such mutual incomprehension between disciplines and traditions has arisen. In the mid-twentieth century it was possible for scholars of the Roman world, whether they concentrated on texts or material evidence, to share common bodies of methodology, (largely unarticulated) theory and basic data (knowledge of the major classical texts, the essentials of epigraphy, classical art, numismatics, and pottery, especially Samian). Today this is no longer feasible. It is impossible — it has not really been possible for a generation — for any one scholar to master the theory, methodology, exponentially expanding literatures and data-sets of all the disciplines which are now essential to the study of a field like the Roman military. Within British archaeology (to say nothing of wider fields), the languages spoken by the various theoretical schools are now very divergent, and too often are inaccessible to those working in other traditions; when encountering the work of other groups, ignorance becomes compounded with frustration and too easily breeds contempt. In a spirited exchange I had in 1998 with a senior colleague on why Roman military traditionalists should make a concerted effort to explore the insights of the current generation of overtly theoretically minded archaeologists, the impenetrability of their jargon was cited as a reason for ignoring their work. Certainly, too many writers employ unnecessarily complex language, as a mark of initiation into a particular grouping, and, more or less knowingly, as a weapon to exclude others ( Jenkins 1992, 157–72). Yet, of course, much ‘jargon’ is appropriate technical language which, as in all scholarship, necessarily assumes that the reader has some knowledge of that particular field or tradition. Indeed, exactly the same charge may be levelled in both directions. There is no difference in kind between modern use and misuse of philosophical or sociological terminology and, for example, the quotation of Latin without translation: both cases assume prior knowledge in the reader; both further relate to the sociological (and political) structure of the discipline, and to the claiming, and the wielding, of academic power (expectation of knowledge of Latin presupposes classical education, for centuries almost a caste-mark of the elite). It seems to me that all scholars need to make an effort, to avoid unnecessary incomprehensibility in writing, and to learn something of each others’ approaches. Another serious problem is the neglect of non-English academic literatures. This is due to a general decline in foreign language attainment, exacerbated by the sheer 36 writing the legions volume of publication in English; we are becoming victims of our own productivity, itself partly driven by financial and political, not academic, imperatives. It is easy to neglect French, German, Italian and other literatures on grounds of lack of time alone. Aside from these general obstructions to mutual comprehension between contemporary scholarly communities, there remain the special factors surrounding the study of warfare and military institutions. Many continue to avoid military aspects largely because of their unpleasant connotations. It is essential to address the brutality inherent in empire building, and not just to focus on the internal practices, and the apparent efficiency and organizational skills of the Roman military, or on the ostensibly nonviolent, integrational side of Roman political and social history ( James 2001b, 198; Alcock 2001, 228). Indeed, avoidance of dealing with unpleasant realities of human life through history, not least violence and war, is as psychologically suspect as an excessive interest in them; denial can be as damaging as unhealthy obsessions ( James 1993). TOWARDS REFORMING THE STUDY OF THE MILITARY ASPECT OF THE ROMAN WORLD: RECONCEPTUALIZING AND THEORIZING THE FIELD OF STUDY At the start of the twenty-first century, then, we find ourselves with established traditions of military studies which badly need overhauling; new schools of Roman archaeology which have not yet got to grips with the martial aspect of the Roman world; a host of new insights available from related fields; and a continuing stream of rich data on which to work. Surely it is time to take a fresh look at the whole subject of the military, drawing on as much of this potential as possible. Any real advance will require profound re-examination of all our concepts and attitudes, from the basics such as the nature of human violence, and evidence for levels of violence from the household upwards, to what we mean by states of war and peace. We need to be far more aware of the wider socio-political context of the military. We must deal more with the question of the nature of ancient warfare, and the need to look at both sides of conflicts, contra the past tendency to over-concentrate on the rich internal detail of Roman military epigraphy and archaeology, at the expense of understanding the local civilian context, the nature and capabilities of Rome’s opponents and their interactions both on and off the battlefield. It seems to me that progress in this will best be made by: 1. Increased emphasis on comparative perspectives Limesforschung has always had a strong, but tightly bounded, comparative aspect. Care must also be taken to maintain the time-honoured links with German scholarship, and develop those others — not least with eastern Europe and the Middle East — which the Limeskongressen have helped to foster (see below). However, this must be broadened from the traditional focus on other, primarily European, Roman frontiers, to include much more on other regions of the empire, neighbouring cultures, and different, contrasting or potentially analogous periods and cultural contexts. The latter emphasizes the importance of the continued writing the legions 37 involvement of (ex-)soldiers, as well as military historians of other periods, for the perspectives they can provide. 2. Maintaining openness to ideas from elsewhere More generally, care should be taken to continue to look for inspiration and new perspectives from outside the immediately relevant disciplines, to ensure the study of the Roman military does not again become insular. SOME SUGGESTED STARTING-POINTS: RECONCEPTUALIZING ‘THE ROMAN ARMY’ Currently, in many quarters, the most basic concepts underpinning our understandings of antiquity are being critically re-evaluated, including, not least, the very meanings of the term ‘Roman’ itself, and the nature and utility of the concept of Romanization. For me, this process of fundamental overhaul has highlighted the urgent need for Roman military studies to examine its own fundamental concepts, a debate already long underway for the nature of ‘frontiers’ (e.g. Whittaker 1994; Freeman 1996). We need to ask questions as basic as: what was the Roman military really for?; how was it actually constituted, what did it consist of and what ‘glue’ really held it together?; and even, did ‘the Roman army’ exist at all? The following (which for clarity concentrates on the early imperial period, where discourse on the subject is overwhelmingly concentrated) offers up for debate suggestions for an alternative framework for thinking about the military institutions of the Roman world. It draws on reconsideration of classical texts, in the light of recent work on the archaeology of Roman soldiers and military communities, overt consideration of comparative cases, and aspects of social theory. WHAT WAS THE ROMAN MILITARY FOR? As we have seen, the question of the real purpose of the Roman military is already being critically examined, in, for example, Isaac’s attack on unspoken Western assumptions about the imperial army being a modern-style national defence force. Goldsworthy has argued strongly that we should call a spade a spade: ‘the waging of war. . . is the ultimate function of all armies’; ‘The [Roman] army existed to wage war’; ‘The primary purpose of any army is to defeat the enemy in battle’; ‘Warfare was the ultimate purpose of the Roman army.’ (Goldsworthy 1996, 2, 11, 116, 286). While Goldsworthy’s work has been a vital antidote to the traditional tendency to avoid confronting the realities of ancient warfare, this is surely to overstate the case in the opposite direction. His thesis fits quite well with the history of the middle republic, but is much less valid for the imperial period. As Isaac (1992) in particular has advocated, we are obliged to accept that, in much of the empire, the military’s main actual employment was as an army of occupation, or perhaps more accurately for much of the empire (including Italy), as an instrument of internal security. It is clear that, from the third century a.d. and probably earlier, it was taking an increasingly important role in internal surveillance and policing work on a scale seriously underestimated in most accounts (MacMullen 1963, 52–72; James 2001a, 82). 38 writing the legions We can resolve these divergent views quite simply, by suggesting that the ultimate function of the army, especially in its own eyes, ideology and rhetoric, was to wage war and defeat ‘the enemy’ in battle; but that de facto its primary purpose or ‘mission statement’, more or less explicit in the soldier’s oath, was to protect the emperor, to promote his interests and to enforce his authority, whether against foreign enemies, unruly provincials, criminals, ‘brigands’, rioters or against usurpers and other highplaced Roman ‘rebels’. These aims were achieved through the threat, as much as the application, of lethal or sub-lethal force at all scales. Full-scale battle was a relatively rare expedient, most routine objectives being achieved by less drastic and expensive methods (although not necessarily less drastic or expensive to its opponents or victims). It is suggested that this was how the Roman military, from the emperor to the ranks, understood itself and its task; while it focused in its language, material culture, and display behaviour on the perceived ‘gloriousness’ of battle against the outsider, the reality of its activities, perhaps played down for reasons of self-respect and ‘public relations’, was much more often internal repression and internecine conflict. HOW WAS THE MILITARY REALLY CONSTITUTED, OR, DID ‘THE ROMAN ARMY’ EXIST? Today we commonly write of ‘The Roman Army’, at least implicitly capitalized and, as we have seen, often implicitly understood as a single vast monolithic organization, analogized with modern armies. This concept, equally embedded in German and French literatures, is closely tied to the idea of the army as a ‘war machine’ (Goldsworthy 1996, 8, 285), which evokes a single, perfectly co-ordinated entity, in which Roman soldiers themselves are mere cogs, brutal, brutalized, unthinking and obedient automata. Yet the idea of ‘the Roman army’, singular, is an entirely modern one; the Romans themselves had no such term for their armed forces as a whole and, so far as I can see, no such concept existed during the republic or the early empire at least ( James 1999; 2001a). The closest equivalent Latin term to the English word ‘army’ is exercitus. When they wrote of ‘the/an army’ (exercitus), Romans meant an individual provincial armygroup (e.g. the exercitus Syriaticae: Whittaker 1994, 68), or a particular campaign army, probably a composite force from several provinces; each such exercitus was an autonomous corps. When they generalized about the military, Romans such as Tacitus and indeed Augustus (who, after all, created the early imperial armed forces) spoke in plurals; they wrote of ‘the armies’ (exercitūs, Tacitus Annals, 1,3; 4,5; Augustus Res gestae, 26, 29). Otherwise, with varying degrees of technical precision and poetic license, writers employed other plurals such as legiones (‘the legions’), cohortes (‘the cohorts’) and numeri (‘the regiments’ or ‘units’). Very often they also spoke of milites, ‘the soldiers’ as a mass, group or class (e.g. Augustus Res gestae, 16, 17, 28). This most generalizing term for the military focuses, I believe significantly, on a category of people, not an organizational structure. Romans had no term for a unitary military organization, institution, or entity equivalent to our term ‘the Roman Army’ — because it did not exist. It seems to me that this is no mere semantic point, but highlights a fundamental misconception running through almost all modern discourse on the Roman military. writing the legions 39 To speak of ‘the Army’, singular, is a serious misrepresentation of the nature of Roman thinking and institutions, indeed of military organization itself, through its backprojection of models of armies from much later, very different, nation-state societies. The idea of ‘the Roman Army’ is an anachronistic reification, which materially obscures the routine reality of multiple armies in the Roman world for most of its history; Speidel has noted the significance of multiple provincial armies, and called for their study (1992b, 16–17), but little emphasis has yet been given to the subject. During the imperial period the armies were, for long periods, to a large degree autonomous organizations with their own histories and, probably, traditions, not directly linked by any central military hierarchy per se, but interlinked more informally at the level of personnel, by movements of men, not least the mobile centurionate. The reality of ‘the army’, insofar as it existed at all, lay in ‘the soldiers’ themselves, as a social grouping and a real, self-aware force in Roman society (and not an unthinking instrument of state power: James 1999). Simply to adopt terminology closer to that of the Romans themselves, to speak routinely of ‘Roman armies’, plural, and of ‘Roman soldiers’, would itself be a step of no small importance in improving our thinking on what the Roman military was ‘really like’. These two conceptual shifts have profound implications for helping us rethink much wider processes, not least the nature and mechanisms of ‘Romanization’ in the frontier regions (see below). WAS THE MILITARY A MACHINE OR A SOCIAL ORGANIZATION? As a return to the texts challenges the assumption of a monolithic army, so it also discredits the commonly held ‘machine analogy’ for understanding its operation. That Roman soldiers certainly did not behave like unquestioning automata, on or off the battlefield, is evident, and has been well documented by, for example, Grant (1974), Campbell (1984) and Goldsworthy (1996, e.g. 283). Even with a draconian disciplinary code, and the vast social gulf between generals and rankers, soldiers from long before Caesar’s time to the late empire were frequently unruly if not mutinous, and not afraid to speak their minds. Often proud of their profession, arrogant and boisterous in their power, they were difficult to control, and not infrequently were themselves a menace to public order. Besides denoting a military force, exercitus can also refer to physical exercise or exertion; additionally, it can mean ‘disciplined’, ‘troubled’, and, significantly, ‘troublesome’, neatly encapsulating the images which Roman soldiers undoubtedly harboured of themselves, and those, less positive, which many civilians associated with them (e.g. Campbell 1984, 246–54; see also Alston 1998). If the machine analogy is discredited, what kind of conception might be put in its place? Surely the obvious alternative is to approach the military as a social organization or community, which it clearly was (Goldsworthy and Haynes 1999; James 1999); and to focus on the soldiers who formed it not as automata, cogs in the machine, but as self-aware social agents — which also fits far better with the glimpses we have of them in the historical record. Further, this military social organization was of course fully a part of the social system of the Roman world, if often uncomfortably so; each is best understood if both are considered together. This may be illustrated by asking what, then, was the force which kept the military system together and functioning, if not 40 writing the legions solely the ‘electronic impulses’ of orders from the top instantly obeyed, and the iron framework of discipline and punishment enforcing unquestioning subordination and obedience? Once the question is reformulated in social terms, the answer seems to me quite simple. In Roman imperial society as a whole, social control of subordinates, whether soldiers or peasants, slaves or freedmen, depended in part on discipline, indoctrination and physical brutality, to be sure. But in the absence of the tools of surveillance and coercion available to the modern nation state, authority was often uncertain, weak, and unable to enforce its will (as is revealed by surviving imperial laws, repeatedly seeking to ban various transgressions, their very repetition revealing the limits of power). Yet Roman society did work, successfully if with varying degrees of internal conflict and tension, for centuries. How? It is well known that, while the asymmetries of power were extreme, the Roman world worked largely by negotiation (e.g. Terrenato 2001), and was articulated not least by patronage. Subordinates could and did resist, but also tended voluntarily to follow their superiors because they accepted that it was in their mutual interest to do so; they did not have to like or enjoy such subordination, but it allowed society to function successfully enough. As part of Roman society, it is argued that the military functioned on exactly the same basis. From our own world of examination-based aspirations to meritocracy, it is widely assumed that patronage systems are inherently corrupt, and that they could not produce efficient or just internal operation of any social organization. It may be doubted, viscerally, that so consistently effective an organization as the Roman military could possibly have operated on such an apparently ramshackle basis: emphasis is always placed on high levels of training and severe discipline, both of which were indeed available in the system. But the documentary evidence we have is consistent with the view that negotiation and the promise of patronage were extremely important, perhaps actually the most important source of both unit cohesion and individual motivation, alongside, and related to, the material rewards of pay and donatives. Commanders and centurions disposed of considerable powers of patronage, from the granting of leave to access to lucrative posts and promotions. In the Vindolanda tablets, we can see the links of military patronage not only operating at the level of equestrian commanders and senatorial legates ( Tab. Vindol. II, 225), but also involving centurions and apparently ordinary provincials ( Tab. Vindol. II, 250). Traditions of mutual social obligation were naturally carried into Roman armies during the republic, when they were still close to their origin as the propertied citizenry under arms, led by their elected magistrates. Notwithstanding the capital powers conferred on generals by their possession of imperium, Roman commanders continued to employ sometimes quite theatrical tactics to persuade their men to follow them, especially in circumstances where loyalty was under stress, when civil conflict was in the air (Campbell 1984). The mutual obligation between commander-in-chief and soldiers, and the effective hold they had over each other, was seen most glaringly during the civil wars of the first century b.c., and again during the military anarchy of the third century a.d. It is seen throughout the early imperial period in the common image of the emperor as commilito, ‘fellow-soldier’, one of the most important of the ways in which he sought to keep the loyalty and trust of the milites (Campbell 1984, writing the legions 41 32–51). It remained as essential for Roman military leaders — from the emperor himself to governors and unit commanders — to control the troops by winning them over through expressions of solidarity and mutual interest, as by bribery, training or punishment. One of the most famous illustrations of this was the mutiny of Claudius’s invasion force on the eve of the assault on Britain, when the imperial legate failed to persuade the men to embark, and they only did so, in boisterous mirth, when a direct appeal from the emperor was delivered by an imperial freedman (Dio Cassius 60.19.1–3; the curious story of Caligula’s treatment of the troops on the beach at Boulogne is probably a garbled account of an earlier mutiny at the prospect of a Channel crossing: Dio Cassius 59.25.1–3; Suetonius Gaius 46.1). All officers, then, were inevitably placed in the position of patrons of their men, and were expected to look after their interests and to see that they were rewarded materially for their loyalty and zeal; their authority, and sometimes even their lives, depended on their ability to fulfil this role, as much as on discipline or any other factor. It is difficult to imagine something further removed from the commonly held view of the Roman military, so much like our perception of the ‘lean, mean, fighting machines’ of modern times, instantly responsive to command (at least in theory). Yet there is a thoroughly documented comparative example of just such a persuasion- and patronage-based society producing a military organization which in its day dominated the world, producing a continuous flow of brilliant commanders, remarkably efficient military units and highly motivated men which came routinely to defeat any enemy, against odds; and this on a scale and for a duration which stands direct comparison with the Roman military. That force was the Royal Navy of the eighteenth and nineteenth centuries. In his brilliant analysis of the structure and operation of the Royal Navy, which focuses on the Seven Years’ War (1756–63), N. A. M. Rodger has shown in detail the workings, strengths and acute shortcomings of a military system which had, on paper, a ferocious disciplinary code, yet which in practice was astonishingly weak in authority and ability to enforce its will over any of its members from admiral to rating, even through direct orders (Rodger 1986). The meticulously documented tendency of all levels in the naval hierarchy to loud insubordination, even riot and mutiny, usually without serious disciplinary consequences, is amazing to anyone who has formed their conception of military systems form modern European armies; yet that Navy achieved levels of operational efficiency which rival those of modern forces, and its string of spectacular successes in battle were instrumental in the establishment of the British empire. The key to the paradox, as Rodger shows, was that, in the Navy as in contemporary British society as a whole, individuals and groups vociferously attached to their rights and liberties were often willing to acquiesce in their allotted places in the hierarchy, because class-consciousness was weakly developed, and everyone believed that they stood to gain by agreeing to work together. Advancement and material rewards could be expected from working within the system of obligation and patronage; officers were able to reward and advance loyal and talented subordinates, whose success enhanced their patron’s prestige in return. Here is an excellent example of the ethos of a military force made comprehensible by placing it in its contemporary social context. It also bears striking similarities to 42 writing the legions what we know of Roman armies, really of all periods; a vociferous and unruly soldiery, usually drawn from men of some social standing who knew their rights and social value, living in a society based on patronage, and held together by thoroughly understood and agreed bonds of mutual obligation running both up and down the hierarchy. The Georgian Royal Navy seems to me to provide a far better analogy for the Roman armies than the rigid ranks of brutalized, brainwashed peasant soldiers of Frederick the Great. In the final analysis, we must of course seek to discover what was unique about the Roman system, in its own terms, and in its own context; but exploration of varied analogies like this can alert us to potentially useful alternative models. ‘The soldiers’ and multiple armies under the early empire, then, not ‘The Army’; soldiers who were active social agents, not robots; men with their own values, aspirations, families, and social networks within and beyond their regiments. I would further suggest that such a reconceptualization is equally applicable to much of the history of the republic, and the later empire as well as the early imperial period. A MONOLITHIC INSTITUTION, OR MULTIPLE MILITARY COMMUNITIES, MULTIPLE ‘ROMANIZATIONS’? Making the soldiers-as-people central, rather than the military-as-institution, immediately changes the perspective both on the military itself, and on the nature of its interaction with other groups. Refocusing on the men and their social world brings dependants and neighbouring communities into the frame as well, resulting in a very different, ‘bottom-up’ view of the Roman military. On enlistment, men became part of the peculiar society of milites, an increasingly self-aware ‘imagined community’ of commilitones (fellow-soldiers) with its own special ideology, symbols and value-system ( James 1999; for ‘imagined communities’ see Anderson 1991, 5–7). It is clear that ‘the soldiers’ were seen as a distinct group by Roman society as a whole, as well as by the milites themselves. Soldierly identity was distinct from the ethos and structure of the imperial regime, armies and regiments, but was of course influenced by it; its complexities are discussed in more detail elsewhere ( James 1999; see also Goldsworthy and Haynes 1999). There is great potential for studying the nature of soldierly identity, not least its embodiment, from the rich body of archaeological, representational and documentary evidence available. The empire-wide ‘imagined community’ of milites was of course subdivided in various ways, the most important of which was by unit. Units might be brigaded together in winter quarters, or they might be broken up into vexillations for greater or lesser periods, but nonetheless that tangible military world for most soldiers was the community of his own particular regiment, in which, assuming he survived, he would normally expect to spend most of his adult life. For many soldiers, long years in a relatively stable regimental community were probably the reality. Such regimental communities were also, I would argue, much larger, more diverse, and perhaps less sharply bounded and distinguished from the ‘civil matrix’ than is usually assumed. The Vindolanda tablets have demonstrated the presence of commanders’ families, including wives, children, and slaves in the fort (officers’ wives: Tab. Vindol. II, 291–94; other women writing: 257, 324; children: 291; slaves: 301, writing the legions 43 302, 347). There are also women’s and children’s shoes from the Period III praetorium, probably representing the familia of the prefect, Flavius Cerealis, his wife Sulpicia Lepidina and perhaps other related womenfolk, children, including household slaves: Van Driel-Murray 1995, 9). Centurions were allowed to marry (Hoffmann 1995, 110); and the discovery of female and/or children’s shoes in barrack rooms at Vindolanda has, as we have seen, highlighted the possibility that, despite the legal ban, in practice soldiers, too, had families in the forts. Indeed, apart from the men actually on the regimental rolls and people working in the ‘service industries’ routinely assumed to be present beyond the gates, military bases may have been thronged with far greater numbers of non-soldiers than have usually been allowed for. There is clear evidence that soldiers of all ranks could retain servants, presumably mostly of slave status, and some of these, notably the grooms which perhaps all cavalrymen maintained, may have had a paramilitary function themselves (Speidel 1992c; Maxfield 1995, 9–11). Many soldiers will have had other locally-resident dependants, including personal slaves, freedwomen or semi-official wives (often closely related categories), their own children (Carrié 1993, 114; Maxfield 1995, 16–20; Varon 1994) and perhaps other blood-relatives (e.g. elderly parents, minor and unmarried siblings). There also appear to have been slaves owned by the unit as a whole (Maxfield 1995, 8). All of these groups centred their lives on the regiment as much as the soldiers themselves, and, it is suggested, will have taken a major part of their sense of identity from this association, where the main breadwinners were the soldiers. In this sense, all of these people formed a ‘military community’, a social entity with complex internal structure. Such complex social groupings are, apparently, exactly what we are seeing archaeologically in the detailed work at Newstead (Clarke 1994, 81). Given the diverse ethnic and national origins of Roman regiments, especially the auxilia but also at least some of the legions (e.g. legio XXII Deiotariana, raised from Galatian royal troops: Webster 1985, 105), the imperial military was always ethnically very mixed, and it is likely that the culture of each of these regimental communities was a unique interplay of Roman or Italian and other cultural traditions, even before we begin to consider how they may have interacted with ‘host’ communities in the frontier zones. Archaeologists have become increasingly aware of the complexity and diversity of the societies within and beyond the empire, below the scale of provinces and states; iron age Britain, in particular, is seen to consist of a group of quite heterogeneous polities. The evidence suggests that the military garrison-communities which were established among provincial societies were also more diverse than we have previously assumed (Haynes 1999). It is therefore clear that the nature of interaction of ‘civil’ and ‘military’, of ‘Roman’ and ‘native’, will be unique in each area, not just down to provincial level, but down to much smaller regions, even individual forts and the indigenous communities of their immediate surroundings. We must expect the processes of ‘Romanization’ (or, more neutrally, cultural interaction and change) in military zones to be much more complex and varied than was thought, and it is possible that the trajectories of each zone will prove to be so different that they defy useful generalization. To investigate these military : civil patternings, we need more regional and comparative studies, such as those conducted for the recruiting areas of 44 writing the legions north-western Iberia by Millett (2001), and for the Batavi and Ituraeans by Haynes (2001). We will probably have to reckon, then, with highly complex interplays of factors when we consider the nature and role of military communities in the life of the Roman world: there were the centripetal pressures from the imperial regime to follow official ideologies and values (e.g. through propaganda and ritual); peer-pressures among the soldiers themselves to conform to an evolving but quite strongly normalizing military subculture, the empire-wide imagined community of milites; in at least some cases there was continuity of social and cultural links between ‘ethnic’ units and sometimes distant home-provinces (e.g. the Syrians at Intercisa: Fitz 1972); and by no means least, the enormously varied patterns of interaction around garrison sites, with local communities probably both sides of the frontiers. The detailed studies which are required to examine these issues largely remain to be undertaken, but it is already clear that the patterns emerging from Newstead in the second century a.d. — a relatively isolated community with little apparent evidence of much cultural contact at all with the indigenous population of its hinterland (Clarke 1999) — could hardly contrast more strongly with that for Dura-Europos in the third century; a garrison of troops largely raised from a closely-related neighbouring community, implanted into a large, pre-existing Graeco-Syrian city ( James forthcoming b; contra Pollard 1996). The shift of focus from army-as-institution to soldiers-as-people is, I think, fundamental for understanding all aspects of the Roman military, from the purpose and operation of forts (as instruments for surveillance and control of the soldiers inside as much as of provincials and enemies outside) to our idea of what motivated Roman soldiers to behave as they did, whether as actors in Roman society and politics (e.g. Campbell 1984), on the battlefield (Goldsworthy 1996) or in sacking cities (Ziolkowski 1993). It is a key part of widening the spectrum of research towards a more truly holistic study of the military aspect of Roman civilization. DISCIPLINARY CONTRIBUTIONS TO AN INTERDISCIPLINARY PROJECT What should be the aims of a renovated field of Roman military studies, and how might the various relevant disciplines contribute? Clearly we need to continue and develop the study of the nature and operation of Roman armies as instruments of conquest and imperial control. However, the theme of this paper is that such a study can only flourish and advance as an integral part of a wider effort: we also need to understand that the military consisted of huge numbers of people and innumerable communities, great and small, living in constant contact with others, contacts which ranged from the murderous and oppressive to the commercial, the amicable and the matrimonial. This broader perspective would better equip us for the task of studying the lives of soldiers and their dependants, and trying to understand their social relations with their ‘host’ communities (in either the hospitable or parasitized sense). The several relevant existing disciplinary traditions all have major potential contributions to make to such an enterprise. Not in any order of priority or precedence: writing the legions 45 1. Ancient history and palaeography Ancient historians and palaeographers have, in a sense, already claimed their place in such an enterprise. As we have seen, they have led innovation and progress in recent years, in so many aspects from the role of Roman ideology and cosmology in underpinning Roman martial behaviour to the minutiae of soldierly life seen in Egyptian and Syrian papyri and the Vindolanda tablets (e.g. Fink 1971; Alston 1995; Bowman 1994; Tab. Vindol. II). In the last, we are fortunate to have an entirely new kind of evidence for the northern frontier armies. It is symbolically appropriate that the Vindolanda tablets should have turned up almost literally in Eric Birley’s backyard — the Birley family live at the site of Chesterholm/Vindolanda, and the tablets were found in excavations conducted by his son, Robin — and that this new chapter opened within months of his retirement. The new texts provide major insights into how the soldiers of the army of Britain thought about themselves, challenging accepted interpretations, e.g. forts were not forts in the early empire, but winter quarters ( Tab. Vindol. II, 225). The fact that such tablets have also now turned up at other sites (Caerleon: Hassall and Tomlin 1986, 450–51; and Carlisle: Tomlin 1998) holds out the real possibility that, now we know how to spot them, further significant finds may be made in the near future, and Britain will continue to be, as it presently is, the most important new source of original military texts in the empire. 2. Epigraphy/prosopography and Limesforschung These must remain important subdisciplines, but should neither dominate nor be marginalized. Future work on the Roman military will, necessarily, in large measure build on the detailed structural analyses produced by past scholarship in these areas, and they continue to have much of interest to contribute. A good example is Margaret Roxan’s recent work using epigraphic data to test widespread assumptions about retirement patterns of auxiliaries (Roxan 1997), crucial to understanding social interaction of soldiers and provincials, and ‘Romanization’. However, while these disciplines remain fundamental, with the development of a sophisticated and autonomous body of archaeological theory and methodology, they must increasingly share the stage with the broader study of material culture, and become more integrated at a theoretical level. 3. Historiography Another important aspect of this putative reformed Roman military studies would be historiographical research. As I hope is clear from the foregoing provisional survey, in military scholarship, as currently in the wider study of the Roman world, there is a need for in-depth research into the development of the discipline. Major tasks which immediately present themselves are the writing of an up-to-date treatment of Wall Zone scholarship as a whole, and a good critical biography of Eric Birley and/or a history of the Durham School in particular, if political sensitivities would yet allow it. Particularly with regard to Birley himself, and his intellectual legacy, none of the common attitudes I have encountered serves posterity well — a 46 writing the legions tendency to hagiography, vilification or an indifference which would consign all to oblivion. 4. Archaeology Due to prolonged theoretical stagnation, the study of the military through material remains has both the greatest potential, and the biggest task, for it has yet to establish in a major way that it can make significant new contributions to broader contemporary debates about the Roman past. If archaeology cannot produce dramatic results with the rich data sets at its disposal, it will thoroughly deserve the scepticism with which many have regarded it in recent decades. We need to apply to the archaeological evidence the conceptual shift advocated here, from seeing the military as a machine, over-concentrating on its institutional and systemic aspect, to considering it as a complex social organization deeply enmeshed within the wider socio-political matrices of the Roman provinces and neighbouring regions. The following offers some thoughts on what an archaeological agenda might include, for wider debate. (I have already discussed the particular case of the military of Roman Britain elsewhere: James 2001a.) The archaeology of warfare itself remains in many ways a surprisingly underdeveloped aspect; in material terms, we know vastly more about Roman armies in camp than on campaign or in battle. To be sure, there is much more work to be done on the material culture of war, on technologies and capabilities, e.g. the important recent work on saddles which show Roman-period cavalry were potentially far more effective than had been assumed (Connolly and Van Driel-Murray 1991). And even more remains to be said about martial material culture as a field of interaction across frontiers; it is perhaps primarily through studying dress and arms, using frameworks that go beyond the merely typological and functionalist, that we may approach ancient soldiers’ understandings of, and attitudes towards, themselves and each other. There also remains much to be said archaeologically about the experience faced by ordinary soldiers of both sides in ancient campaigning and battle. While marching camps may make early imperial campaigning archaeologically visible to a degree rare in other periods, their transience limits their information potential. However, while many battlefields and siege sites across the empire were explored long ago, important new discoveries are being made (notably the location of the Varian disaster of a.d. 9: Schlüter and Wiegels 1999), while the archaeology of such sites is a field currently undergoing considerable theoretical and methodological development (Freeman and Pollard 2001; see especially Coulston 2001 on Roman ‘conflict landscapes’). However, most military-related archaeology tells us about soldiers not on active campaign, but in winter quarters or permanent bases, or engaged in ‘peacetime’ activities within provinces, which in the imperial period generally did constitute most of the routine reality of life for soldiers in the ranks, their dependants, and those groups (‘civilians’, ‘provincials’, ‘barbarians’, etc.) with whom they interacted. Under the new perspective advocated here, the archaeology of military installations can tell us much more than has hitherto been widely assumed. Of course, it will continue to illuminate aspects of the military as an instrument of the state: the layout writing the legions 47 and construction of military sites are largely to be seen in terms of the creation of theatres of power and institutional control. However, the lives of the communities they housed were largely lived out through the creation, manipulation and disposal of portable material culture and materials, in and around such built environments. Therefore, at military sites, different types of archaeological context, and categories or patterns of material evidence, may be related to different aspects of the institutional life of the military, and the collective and personal worlds of soldiers and their dependants ( James 2001a, 85). To understand the internal dynamics of military communities in more detail, and equally their relations with the local or regional ‘civilian’, ‘provincial’ or ‘barbarian’ populations with whom they directly interacted, we would like to be able to define such groups archaeologically. This is not possible on simple spatial grounds; we cannot just call everything inside the walls of bases ‘military’ and everything outside ‘civilian’ since it is clear that soldiers spent much time outside, and non-soldiers were routinely inside, such boundaries. To investigate such matters, we have to look in more sophisticated ways at material culture and its patterning in, around and beyond military sites. Can we identify military communities archaeologically, especially to find evidence of soldiers operating in wider society, away from morphologically or epigraphically military sites? Can we identify ‘military’ assemblages per se, i.e. characteristic material repertoires attesting the presence of military contingents (Allason-Jones 2001; James 2001a)? Growing awareness that artefacts do not ‘contain’ fixed cultural meanings, and that identities like ‘soldierliness’ are not expressed in fixed ways through material culture but rather are lived out as evolving sets of practices, has led Gardner to challenge the notion of the existence of simply-definable military assemblages per se (Gardner 2001). However, he does not rule out the possibility of ‘military’ practices and actions leading to characteristic depositional patternings across multiple categories of artefacts, which may require sophisticated spatial and statistical analysis to detect and define. He, and others, are actively at work on such studies (Gardner 2001; Cool et al. 1995). It may be that we will get furthest by using material culture data closely with other categories of information, documentary and representational, as I have sought to do in discussing the embodiment of soldierly identity ( James 1999). Such work, often on very rich assemblages, constitutes case studies in the ‘hot’ archaeological topic of the nature and material articulation of identities in past human societies, providing an example of how Roman military studies can develop in ways of interest and relevance to wider archaeology. We need to investigate the internal structuring of military communities and dynamics of local interactions, across wide expanses of space and time, through comparative studies of similarities and differences. To what extent was their commonality across the empire in military life and material culture, and what did it signify? Some similarities are remarkable, in the planning of headquarters buildings, or the fine details of many helmet and boot designs from Scotland to the Euphrates. Conversely, what does regional and local variation, increasingly apparent in military as well as civil aspects of Roman provincial archaeology, tell us about the diversity of lives of Roman soldiers and their families? What were the dynamics between the centripetal forces of state institutions, the normative pressures of aspiration to soldierly 48 writing the legions solidarity, and the centrifugal tendencies of diverse regional experiences and ethnic origins? To conduct such regional and comparative studies, between different provinces and across the frontiers, will involve digging more bases and their environs, certainly, but in different ways, with different aims, as well as specifically contextualizing the bestexplored by looking at military sites in their landscapes. For example, Hanson and Breeze have highlighted the shortcomings of our knowledge of the contemporary indigenous archaeology in Scotland, and have advocated work on rural settlements and blocks of landscape containing Roman military sites (1991, 72–73). It is to be hoped that the Newstead project will fulfil this in part, supplemented by work elsewhere at places like, perhaps, Inveresk (Hanson and Breeze 1991, 73). For comparative purposes, we should seek to develop more overseas co-operative field projects. British scholars have been seizing newly available opportunities to work in central and eastern Europe, e.g. Hanson, Haynes, Wilmott and others in Romania. Further afield, Ian Haynes ( pers. comm.) identifies Cappadocia as near-virgin territory to explore, if politics permit. Another example with great potential for overt comparison and contrast across the length of the frontiers is Dura-Europos on the Euphrates. Here, much of the military equipment is identical to that seen in Britain, but the cultural background of the Roman garrison, its environmental and political context, and the nature of its cantonment inside a large, bustling Greco-Mesopotamian city could hardly differ more from places like Newstead. CONCLUSION AND PROSPECT Today scholars tend to be wholly partisans of the traditional approach to Roman military studies, fundamentally hostile to it or, worse, simply dismiss it as uninteresting anachronism. Instead we need constructive critique. What now appear to be errors or inadequacies in the assumptions and writings of our predecessors often inspire irritation, scorn, amusement or horror. But it is all too easy to be tempted into smugness at our own enlightened superiority, portraying scholars of earlier generations as either heroes or villains on the anachronistic basis of how their views compare with later ideas which we regard as better (Gould 1978). The proper purpose of academic criticism of the work of our predecessors is surely both to recognize its limitations but also to build on its achievements, in order to develop new understandings. In particular, it should encourage us to be self-critical, to ask whether we, too, wittingly or unwittingly, are similarly selecting misconceived perspectives which coming generations will regard as egregious errors and unbelievable myopia. The critique of traditional Roman military studies presented here is not intended to disparage it, or to attempt to replace it with another, single, supposedly superior paradigm. It is intended to evaluate it in its historical context, from a very different standpoint, and to learn from it. During the course of the twentieth century, the study in Britain of the Roman military, exemplified by the Durham School, made great strides, yet many potential avenues of study were virtually excluded from discourse, either devalued or rendered effectively invisible, as of no significant account. The field became overspecialized on writing the legions 49 prosopography and frontier installations, and lost sight of the context of the armies, of soldiers in society, and even of warfare. It seems to me that one of the main lessons of the Durham School’s dominance of Roman military studies in Britain is that strong schools with single approaches, and very powerful individuals, can be a mixed blessing for the advancement of any field. The influence of an individual can become an obstacle to the general progress of scholarship, and Eric Birley’s shadow long lay across the study of the Roman armies, in a manner reminiscent of Newton’s domination of the Royal Society in the early eighteenth century, giving energetic leadership and direction, but also constraining scholarship, and generating conflict (Hall 1992, 337–38; Westfall 1993, 264–65, 272). I have argued here for a much broader approach, drawing on as many disciplines as may be of relevance, to produce new conceptions of the military aspect of Roman antiquity for new generations with post-imperial outlooks, and interests and priorities different from those of the past. This will be best achieved through the establishment of an inclusive, broadly defined field of Roman military studies, drawing on the strengths of all relevant disciplines and sources of data. The realities of specialization bring their own difficulties of co-ordination, but at least there is less chance of potentially fruitful avenues being stifled. We all know that, even collectively, we struggle to explore, let alone to comprehend, the temporal and spatial vastness and complexity of the Roman world. There is room, and a need, for many approaches to the Roman military which was a key component of that enormous entity. As Speidel has observed, ‘Indeed, considering the all-important role the Roman army played during the first five centuries of our era, Roman army researchers are far too few’ (Speidel 1992b, 18–19). The prospects for substantial advance are excellent, and hold out the likelihood that the study of the military aspect of Roman antiquity will be able to make significant contributions to, and so become properly reintegrated with, wider Roman archaeology. It will have much to contribute to the study of archaeologies of identity, and to the explanation of cultural change in and around the Roman empire: the enormous numbers of milites and their dependants played a major role in the forging of Roman civilization and identity in many parts of the empire, and their study will contribute to the continuing debate about cultural change, and whether it should be termed Romanization (Keay and Terrenato 2001; James 2001b) or something else (e.g. ‘creolization’: Webster 2001). It is probably asking too much of human beings to demand that they should deal openly and dispassionately with a subject so emotive, and so full of implications of violence and suffering; to some degree, most will always tend to shy away from these aspects, some to avoid the military completely, others veering into the military’s ‘safer’ aspects, where swords were sheathed. Yet, whichever answer we are tempted to make to the instinctive if essentially meaningless question which Sellar and Yeatman (1930) might have asked, ‘Was the Roman military a Good Thing or a Bad Thing?’, we can at least aspire to detachment, and agree that the Roman military was a Very Big Thing indeed, and thoroughly worthy of academic study. ACKNOWLEDGEMENTS This article is based on a paper originally given as a seminar for the Centre for Roman Provincial Archaeology, Durham University Department of Archaeology, in December 1996. I should like to 50 writing the legions thank the Leverhulme Trust and the University of Durham for funding the Special Research Fellowship during which this work was conducted, as a spin-off of my research into the Roman military presence at Dura-Europos, Syria. Thanks also to Martin Millett for information and encouragement, and Lindy Brewster for access to the Durham University Museum of Archaeology’s exhibition material on Eric Birley. I am also grateful to a number of others who provided invaluable critical comment on an early draft, including Mark Hassall, Matthew Johnson, Margaret Roxan, Richard Reece and John Wilkes. The following also kindly provided essential information: David Thomas, Margaret Roxan, Jeremy Taylor and Vivien Swann. Many thanks also to Ian Haynes and Andrew Gardner for stimulating discussions and feedback on a later draft. Finally, I am particularly grateful to David Breeze for his generous advice, support and encouragement. Of course, none of the above necessarily agree with the views expressed here. BIBLIOGRAPHY ABBREVIATIONS Tab. Vindol. II Bowman, A. K. and Thomas, J. D. 1994: The Vindolanda Writing-Tablets ( Tabulae Vindolandenses II), London: British Museum Press CLASSICAL SOURCES Augustus Res gestae Dio Cassius Historia romana Vegetius Epitoma rei militaris Tacitus Annals Horace Odes Suetonius, The Twelve Caesars: Gaius MODERN WORKS Abdul-Hak, S. 1955. Rapport préliminaire sur des objets provenant de la nécropole romaine situeé a proximité de Nawa (Hauran), Annales Archéologiques de Syrie, 4/5, 163–88 Adams, C. 2001. Feeding the wolf : logistics and the Roman army, J. Roman Archaeol., 14, 465–72 Alcock, S. 2001. Vulgar Romanization and the dominance of elites, in S. J. Keay and N. Terrenato (eds) Italy and the West: Comparative Issues in Romanization, 227–30, Oxford: Oxbow Alföldy, G. 1995. Eric Birley, 12. Januar 1906–20. Oktober 1995, Saalburg Jahrbuch, 48, 140–44 Alföldy, G. Dobson, B. and Eck, W. (eds), 2000. Kaiser, Heer und Gesellschaft in der Römischen Kaiserzeit: Gedenkschrift für Eric Birley, Stuttgart: Steiner. Allason-Jones, L. 1988. Small finds from turrets on Hadrian’s Wall, in J. C. Coulston (ed.) Military Equipment and the Identity of Roman Soldiers; Proceedings of the Fourth Roman Military Equipment Conference, Brit. Archaeol. Rep. international series, 394, 197–233, Oxford: British Archaeological Reports Allason-Jones, L. 1999. Women and the Roman army in Britain, in A. K. Goldsworthy and I. Haynes (eds) The Roman Army as a Community, J. Roman Archaeol. supplementary series, 34, 41–51, Portsmouth, Rhode Island: Journal of Roman Archaeology Allason-Jones, L. 2001. Material culture and identity, in S. T. James, and M. J. Millett (eds) Britons and Romans: Advancing an Archaeological Agenda, Counc. Brit. Archaeol. res. rep., 125, 19–26, York: Council For British Archaeology Alston, R. 1995. Soldier and Society in Roman Egypt: a Social History, London and New York: Routledge Alston, R. 1998. Arms and the man: soldiers, masculinity and power in Republican and Imperial Rome, in L. Foxhall and J. Salmon (eds) When Men were Men: Masculinity, Power and Identity in Classical Antiquity, 205–23, London and New York: Routledge Anderson, B. 1991. Imagined Communities; Reflections on the Origins and Spread of Nationalism, London and New York: Verso (revised edition). writing the legions 51 Anderson, M. S. 1998. War and Society in Europe of the Old Regime 1618–1789, Stroud: Sutton (rev. edn) Anon. 1989. Fouilles d’Alise-Sainte-Reine 1861–1865, Mémoires de l’Académie des Inscriptions et Belles-lettres, nouvelle série, tome, IX, Paris: Institut de France Ball, W. 2000. Rome in the East: Transformation of an Empire, London and New York: Routledge Beevor, A. 1991. Inside the British Army, London: Corgi Best, G. 1998a. Editor’s preface, in M. S. Anderson War and Society in Europe of the Old Regime 1618–1789, Stroud: Sutton (rev. edn), 5–10 Best, G. 1998b. War and Society in Revolutionary Europe 1770–1870, Stroud: Sutton (rev. edn). Birley, E. 1935. Fourth report on excavations at Housesteads, Archaeol. Aeliana fourth ser., 12, 204–59 (reprinted in Birley 1953a, 69–86) Birley, E. (ed.) 1952. The Congress of Roman Frontier Studies, 1949, Durham: Durham University Birley, E. 1953a. Roman Britain and the Roman Army: Collected Papers, Kendal: T. Wilson Birley, E. 1953b. Britain under Nero: the significance of Q. Veranius, in E. Birley Roman Britain and the Roman Army: Collected Papers, Kendal: T. Wilson, 1–9 (originally published in Durham University Journal, June 1952, 88–92) Birley, E. 1953c. The origins of equestrian officers: prosopographical method, in E. Birley Roman Britain and the Roman Army: Collected Papers, Kendal: T. Wilson, 154–171 (originally published in Durham University Journal, June 1951, 86–95) Birley, E. 1958. Archaeology in the North of England: Inaugural Lecture of the Professor of Romano-British History and Archaeology delivered in the Appleby Lecture Theatre on 6 May, 1958, Durham: University of Durham Birley, E. 1961. Research on Hadrian’s Wall, Kendal: T. Wilson Birley, E. 1984. Foreword in M. P. Speidel, Roman Army Studies I, unpaginated, Amsterdam: Gieben Birley, E. 1988a. The Roman Army: Papers 1929–1986 (ed. M. P. Speidel), Amsterdam: Gieben Birley, E. 1988b. The epigraphy of the Roman army, in E. Birley The Roman Army: Papers 1929–1986 (ed. M. P. Speidel), 3–11, Amsterdam: Gieben (reset, repaginated republication of a paper first published in 1952 Actes du Deuxième Congrès d’Épigraphie Grecque et Latine, 226–38, Paris) Bishop, M. C. and Coulston, J. C. N. 1993. Roman Military Equipment, London: Batsford Le Bohec, Y. 1989. L’Armée Romaine sous le Haut-Empire Paris: Picard Le Bohec, Y. 1994. The Imperial Roman Army, London: Routledge Bond, B. 1998. War and Society in Europe 1870–1970, Stroud: Sutton (rev. edn) Bourke, J. 1999. An Intimate History of Killing: Face-to-face Killing in Twentieth-Century Warfare, London: Granta Bowman, A. K. 1994. Life and Letters on the Roman Frontier, London: British Museum Press Breeze, D. J. 1985. Roman forces and native populations, Proc. Soc. Antiq. Scot., 115, 223–8 Breeze, D. J. 1993a. Why did the Romans fail to conquer Scotland?, in D. J. Breeze, and B. Dobson, (eds) Roman OYcers and Frontiers, Mavors, X, 365–84, Stuttgart: Franz Steiner Verlag, (first published in Proc. Soc. Antiq. Scot., 118, 1988, 3–22) Breeze, D. J. 1993b. The logistics of Agricola’s final campaign, in D .J. Breeze, and B. Dobson, (eds) Roman OYcers and Frontiers, Mavors, X, 574–95, Stuttgart: Franz Steiner Verlag (first published in Talanta, 18/19, 1987/8, 7–28) Breeze, D. J. 1996. Eric Birley (1906–1995) Britannia, 27, xi-xiv Breeze, D. J. and Dobson, B. (eds) 1993. Roman OYcers and Frontiers, Mavors, X, Stuttgart: Franz Steiner Verlag Bruce, J. C. 1875. Lapidarium Septentrionale; or, a Description of the Monuments of Roman Rule in the North of England, Newcastle-upon-Tyne: Society of Antiquaries of Newcastle upon Tyne Campbell, D. B. 1984. The Emperor and the Roman Army, 31 BC — AD 235, Oxford: Clarendon Press Campbell, D. B. 1997. The Roman art of war (review article), Britannia, 28, 479–483 Campbell, D. B. 2002. War and Society in Imperial Rome 31 BC to AD 284, London and New York: Routledge Carman, J. (ed.) 1997. Material Harm: Archaeological Studies of War and Violence, Glasgow: Cruithne Press Carman, J. and Harding, A. (eds) 1999. Ancient Warfare, Stroud: Sutton 52 writing the legions Carrié, J.-M. 1993. The soldier, in A. Giardina (ed.) The Romans, 100–137, Chicago and London: University of Chicago Press Clarke, S. 1994. A quantitative analysis of the finds from the Roman fort of Newstead — some preliminary findings, in S. Cottam, D. Dungworth, S. Scott and J. Taylor (eds) TRAC 94 Proceedings of the Fourth Annual Theoretical Roman Archaeology Conference, Durham 1994, 72–82, Oxford: Oxbow Clarke, S. 1997. Abandonment, rubbish disposal and ‘special deposits’ at Newstead, in K. Meadows, C. Lemke and J. Heron, J. (eds) TRAC 96: Proceedings of the Sixth Annual Theoretical Roman Archaeology Conference, SheYeld 1996, 73–81, Oxford: Oxbow Clarke, S. 1999. Contact, architectural symbolism and the negotiation of cultural identity in the military zone, in C. Forcey and R. Witcher (eds) Proceedings of the Eighth Theoretical Roman Archaeology Conference, Leicester, 1998, 37–45, Oxford: Oxbow Clarke, S. and Jones, R. 1996. The Newstead pits, J. Roman Military Equipment Stud., 5, 109–24. von Clausewitz, C. 1832. Vom Kriege, Berlin Connolly, P. and Van Driel-Murray, C. 1991. The Roman cavalry saddle, Britannia, 22, 33–50 Cool, H. E. M., Lloyd-Morgan, G. and Hooley, A. D. 1995. Finds from the Fortress, The Archaeology of York, 17/10, York: CBA for York Archaeological Trust Coulston, J. C. N. 2001. The archaeology of Roman conflict, in P. W. M. Freeman, and A. Pollard (eds) Fields of Conflict: Progress and Prospect in Battlefield Archaeology, Brit. Archaeol. Rep. international series, 958, 23–49, Oxford: Archaeopress Creasy, E. 1894. The Fifteen Decisive Battles of the World: From Marathon to Waterloo, London: Richard Bentley (fortieth edition) Curle, J. 1911. A Roman Frontier Post and its People: The Fort of Newstead in the Parish of Melrose, Glasgow: Maclehose Davies, R. W. 1989. Daily life of the Roman soldier, in R. W. Davies, Service in the Roman Army, 33–68, Edinburgh: Edinburgh University Press with the Publications Board of the University of Durham Delbrück, H. 1975. History of the Art of War within the Framework of Political History: Vol.1, Antiquity (trans. W. J. Renfroe, Jr.), London: Greenwood Press (originally published 1900) Dobney, K. 2001. A place at the table: the role of vertebrate zooarchaeology with a Roman research agenda for Britain, in S. T. James, and M. J. Millett (eds) Britons and Romans: Advancing an Archaeological Agenda, 36–45, Counc. Brit. Archaeol. res. rep., 125, York: Council for British Archaeology Dobson, B. 1978. Die Primipilares: Entwicklung und Bedeutung, Lauf bahnen und Persönlichkeiten eines römischen OYziersranges, Köln: Rheinland-Verlag Dobson, B. 1995. Eric Birley, Durham Archaeol. J., 11, 97–98 Dobson, B. 1998. Eric Barff Birley, Proc. Brit. Academy 1997, 214–32 Domaszewski, A. von 1967. Die Rangordnung des römischen Heeres, 2. Durchgesehene Auflage. Einführung, Berichtungen und Nachträge von B. Dobson, Köln und Graz: Böhlau (originally published 1908) Drinkwater, J. F. 1996. ‘The Germanic threat on the Rhine frontier’: a Romano-Gallic artefact?, in R. W. Mathisen, and H. S. Sivan (eds) Shifting Frontiers in Late Antiquity, 20–30, Aldershot: Variorum Dudley, D. R. and Webster, G. 1965. The Roman Conquest of Britain, A.D.43–57, London: Batsford Ehrenreich, B. 1997. Blood Rites: Origins and History of the Passions of War, New York: Henry Holt; London: Virago Ellis, J. 1990. The Sharp End: The Fighting Man in World War II, Pimlico, London (rev. edn) Elton, H. 1996. Frontiers of the Roman Empire, London: Batsford Erdkamp, P. 1998. Hunger and the Sword: Warfare and Food Supply in Roman Republican Wars (264–30 B.C.) , Amsterdam: J. C. Gieben Ferguson, R. B. 1984. Warfare, Culture, and Environment, New York and London: Academic Press Ferguson, R. B. and Whitehead, N. L. (eds), 1992a. War in the Tribal Zone: Expanding States and Indigenous Warfare, Santa Fe, New Mexico: School of Advanced Research. writing the legions 53 Ferguson, R. B. and Whitehead, N. L. 1992b. The violent edge of empire, in R. B. Ferguson, and N. L. Whitehead, (eds) War in the Tribal Zone: Expanding States and Indigenous Warfare, 1–30, Santa Fe, New Mexico: School of Advanced Research Fink, R. O. 1971. Roman Military Records on Papyrus, Cleveland: Case Western Reserve University Press Fitz , J. 1972. Les Syriens à Intercisa, Brussels: Latomus Foucault, M. 1979. Discipline and Punish: The Birth of the Prison, Harmondsworth: Penguin. Freeman, P. W. M. 1996. Roman frontier studies: what’s new? (review article), Britannia, 27, 465–70 Freeman, P. W. M. 1997. Mommsen to Haverfield: the origins of studies of Romanization in late 19th-c. Britain, in D. J. Mattingly (ed.) Dialogues in Roman Imperialism: Power, Discourse and Discrepant Experience in the Roman Empire, J. Roman Archaeol. supplementary ser., 23, 27–50, Portsmouth, Rhode Island: Journal of Roman Archaeology Freeman, P. W. M. and Pollard, A. (eds) 2001. Fields of Conflict: Progress and Prospect in Battlefield Archaeology, Brit. Archaeol. Rep. international series, 958, Oxford: Archaeopress Frere, S. S. 1987. Britannia, London and New York: Routledge and Kegan Paul (3rd edition) Freud, S. 1968. Why war?, in L. Bramson and G. W. Goethals, (eds) War: Studies from Psychology, Sociology, Anthropology, 71–80, New York and London: Basic Fried, M. H., Harris, M. and Murphy, R. (eds) 1968. War: The Anthropology of Armed Conflict and Aggression, Garden City, New York: Natural History Press Gardner, A. 2001. ‘‘Military’ and ‘Civilian’ in Late Roman Britain: an Archaeology of Social Identity’ (unpublished doctoral thesis, University of London) Giddens, A. 1985. The Nation-State and Violence: Volume Two of a Contemporary Critique of Historical Materialism, Cambridge: Polity Press Goldsworthy, A. K. 1996. The Roman Army at War 100BC — AD200, Oxford: Clarendon Press Goldsworthy, A. K. and Haynes, I. (eds), 1999. The Roman Army as a Community, J. Roman Archaeol. supplementary series, 34, Portsmouth, Rhode Island: Journal of Roman Archaeology Gould, S. J. 1978. On heroes and fools in science, in S. J. Gould, Ever Since Darwin, London: Burnett. Grafton, A. 1997. The Footnote: a Curious History, London: Faber Grant, M. 1974. The Army of the Caesars, London: Weidenfeld and Nicholson Groenman-van Waateringe, W., van Beek, B. L., Willems, W. J. H. and Wynia, S. L. (eds) 1997. Roman Frontier Studies 1995 Proceedings of the XVIth International Congress of Roman Frontier Studies, Oxbow monograph, 91, Oxford: Oxbow Guilaine, J. and Zammit, J. 2001. Le Sentier de la Guerre: Visages de la Violence Préhistorique, Paris: Seuil. Hall, A. R. 1992. Isaac Newton, Adventurer in Thought, Cambridge: Blackwell. Hanson, V. D. 1989. The Western Way of War: Infantry Battle in Classical Greece, Oxford and New York: Oxford University Press Hanson, W. S. 1997. Across the frontiers: addressing the ambiguities, in W. Groenman-van Waateringe, B. L. van Beek, W. J. H. Willems and S. L. Wynia (eds) Roman Frontier Studies 1995: Proceedings of the XVIth International Congress of Roman Frontier Studies, 373–8, Oxbow monograph, 91, Oxford: Oxbow Hanson, W. S. and Breeze, D. J. 1991. The future of Roman Scotland, in W. S. Hanson and E. A. Slater (eds) 1991, Scottish Archaeology: New Perceptions, 57–80, Aberdeen: Aberdeen University Press Härke, H. 1995. ‘The Hun is a methodical chap’: reflections on the German tradition of pre- and proto-history, in P. J. Ucko (ed.) Theory in Archaeology: A World Perspective, 47–60, London: Routledge Härke, H. (ed.) 2000. Archaeology, Ideology, and Society: The German Experience, Frankfurt am Main: Peter Lang Harris, W. V. 1979. War and Imperialism in Republican Rome, 327–70 B.C., Oxford: Clarendon Press Hassall, M. 1999. Homes for heroes: married quarters for soldiers and veterans, in A. K. Goldsworthy and I. Haynes (eds) The Roman Army as a Community, J. Roman Archaeol. supplementary series, 34, 35–40, Portsmouth, Rhode Island: Journal of Roman Archaeology Hassall, M. W. C. and Tomlin, R. S .O. 1986. Roman Britain in 1985: II, Inscriptions, Britannia 17, 428–54 54 writing the legions Haynes, I. 1999. Introduction: the Roman army as a community, in A. K. Goldsworthy and I. Haynes, (eds), The Roman Army as a Community, J. Roman Archaeol. supplementary series, 34, 7–14, Portsmouth, Rhode Island: Journal of Roman Archaeology Haynes, I. P. 2001. The impact of auxiliary recruitment on provincial societies from Augustus to Caracalla, in L. de Blois (ed) Administration, Prosopography and Appointment Policies in the Roman Empire, Proceedings of the International Network Impact of Empire, 62–83, Amsterdam: J. C. Gieben Hibberd, D. 1992. Wilfred Owen, the Last Year 1917–1918, London: Constable Hingley, R. 2000. Roman OYcers and English Gentlemen, London and New York: Routledge Hoffmann, B. 1995. The quarters of the legionary centurions, Britannia, 26, 107–51 Holden, W. 1998. Shell Shock: The Psychological Impact of War, London: Channel 4 Books Hopkins, K 1978. Conquerors and Slaves, Cambridge: Cambridge University Press Howard, M. 1991. The Lessons of History, Oxford: Clarenden Press Huntley, J. P. 1999. Environmental evidence for Hadrian’s Wall, in P. Bidwell (ed.) Hadrian’s Wall 1989–1999, 48–64, Newcastle: Cumberland and Westmoreland Antiquarian and Archaeological Society, Society of Antiquaries of Newcastle Upon Tyne Huntley, J. P. and Stallibrass, S. 1995. Plant and Vertebrate Remains from Archaeological Sites in Northern England: Data Reviews and Future Directions, Architect. Archaeol. Soc. Durham Northumberland res. rep., 4, Durham: Architectural and Archaeological Society of Durham and Northumberland Hynes, S. 1998. The Soldier’s Tale: Bearing Witness to Modern War, London: Pimlico Isaac, B. 1992. The Limits of Empire: The Roman Army in the East (revised edn), Oxford: Clarendon Press James, S. T. 1986. Evidence from Dura-Europos for the origins of late Roman helmets, Syria, 63, 107–134 James, S. T. 1993. How was it for you? Personal psychology and the perception of the past, Archaeol. Rev. Cambridge, 12, 87–100 James, S. T. 1997. Military equipment from the Yale/French Academy excavations at Dura-Europos, 1928–36, Doura-Europos Études, IV, 1991–3, 223–228, Beirut: IFAPO James, S. T. 1999. The community of the soldiers: a major identity and centre of power in the Roman empire, in C. Forcey, and R. Witcher (eds) Proceedings of the Eighth Theoretical Roman Archaeology Conference, Leicester, 1998, 14–25, Oxford: Oxbow James, S. T. 2001a. Soldiers and civilians: identity and interaction in Roman Britain, in S. T. James, and M. J. Millett (eds) Britons and Romans: Advancing an Archaeological Agenda, Counc. Brit. Archaeol. res. rep., 125, 77–89, York: Council for British Archaeology James S. T. 2001b. ‘Romanization’ and the peoples of Britain, in S. J. Keay, and N. Terrenato (eds) 2001 Italy and the West. Comparative Issues in Romanization, 187–209, Oxford: Oxbow James, S. T. forthcoming a. ‘A bloodless past: the pacification of Iron Age Britain’, in C. C. Haselgrove, and R. E. Pope (eds) The Earlier Iron Age in Britain and the Near Continent, Oxford: Oxbow James, S. T. forthcoming b. Excavations at Dura-Europos, Final Report Volume VII, The Arms and Armour, and other Military Equipment, London: British Museum Press James, S. T. and Millett, M. J. (eds) 2001. Britons and Romans: Advancing an Archaeological Agenda, Counc. Brit. Archaeol. res. rep., 125, 77–89, York: Council for British Archaeology Jenkins, R. 1992. Pierre Bourdieu, London: Routledge Johnson, M. 1999 Archaeological Theory: An Introduction, Oxford: Blackwell Keay, S. J. and Terrenato, N. (eds) 2001. Italy and the West. Comparative Issues in Romanization, Oxford: Oxbow Keegan, J. 1976. The Face of Battle: A Study of Agincourt, Waterloo and the Somme, London: Jonathan Cape (second edn 1988), London: Barrie and Jenkins Keegan, J. 1993. A History of Warfare, London: Hutchinson Keeley, L. 1996. War before Civilization: the Myth of the Peaceful Savage, Oxford and New York: Oxford University Press Kennedy, D. 1996a. The Roman army in the East, in D. Kennedy (ed.) The Roman Army in the East, J. Roman Archaeol. supplementary ser., 18, 9–24, Ann Arbor: Journal of Roman Archaeology writing the legions 55 Kennedy, D. 1996b. Parthia and Rome: eastern perspectives, in in D. Kennedy (ed.) The Roman Army in the East, J. Roman Archaeol. supplementary ser., 18, 67–90, Ann Arbor: Journal of Roman Archaeology Kennedy, D. and Riley, D. 1990. Rome’s Desert Frontier from the Air, London: Batsford Keppie, L. 1984. The Making of the Roman Army: From Republic to Empire, London: Batsford King, A. 2001. The Romanization of diet in the Western empire: comparative archaeological studies, in S. J. Keay, and N. Terrenato, (eds) 2001 Italy and the West. Comparative Issues in Romanization, 210–23, Oxford: Oxbow Kipling, R. 1906. Puck of Pook Hill London: Macmillan Lee, A. D. 1996. Morale and the Roman experience of battle, in A. B. Lloyd (ed.) Battle in Antiquity, London: Duckworth, 192–217 Lesser, A. 1968. War and the state, in M. H. Fried, M. Harris, and R. Murphy (eds) 1968 War: The Anthropology of Armed Conflict and Aggression, 92–96, Garden City, New York: Natural History Press Liversidge, M. and Edwards, C. (eds) 1996. Imagining Rome: British Artists and Rome in the Nineteenth Century, London: Merrell Holbertson Luttwak, E. 1976. The Grand Strategy of the Roman Empire, Baltimore and London: Johns Hopkins University Press Luttwak, E. N. 1993. Logistics and the aristocratic idea of war, in J. A. Lynn (ed.), Feeding Mars: Logistics in Western Warfare from the Middle Ages to the Present, 3–7, Boulder, San Francisco and Oxford: Westview MacMullen, R. 1963. Soldier and Civilian in the Later Roman Empire, Cambridge, MA: Harvard University Press MacMullen, R. 1984. The legion as a society, Historia, 33, 440–56 Mann, J. 1979. Power, force and the frontiers of the empire, J. Roman Stud., 69, 175–83 Mann, J. C. 1996. Britain and the Roman Empire, Aldershot: Variorum Marchand, S. 1996. Down from Olympus: Archaeology and Philhellenism in Germany, 1750–1970, Princeton: Princeton University Press Marwick, A. 1974. War and Social Change in the Twentieth Century, London: Macmillan Mattingly, D. J. 1992. War and peace in Roman north Africa: observations and models of state-tribe interaction, R. B. Ferguson and N. L. Whitehead (eds) War in the Tribal Zone: Expanding States and Indigenous Warfare, 31–60, Santa Fe, New Mexico: School of American Research Press Mattingly, D. J. (ed.) 1997. Dialogues in Roman Imperialism: Power, Discourse and Discrepant Experience in the Roman Empire, J. Roman Archaeol. supplementary ser., 23, Portsmouth, Rhode Island: Journal of Roman Archaeology Maxfield, V. A. 1995. Soldier and Civilian: Life Beyond the Ramparts: The Eighth Annual Caerleon Lecture, Cardiff : National Museum of Wales du Mesnil du Buisson, R. 1936. The Persian mines, in M. Rostovtzeff, A. Bellinger, C. Hopkins, and C. Welles (eds) The Excavations at Dura-Europos, Preliminary Report on the Sixth Season, 1932–1933, 188–205, New Haven: Yale University Press du Mesnil du Buisson, R. 1944. Les ouvrages du siège à Doura-Europos, Memoires de la Société Nationale des Antiquares de France, ninth ser., 1, 5–60 Miket, R. and Burgess, C. (eds) 1984. Between and Beyond the Walls: Essays on the Prehistory and History of North Britain in Honour of George Jobey, Edinburgh: John Donald Millar, F. 1977. The Emperor in the Roman World, London: Duckworth Millar, F. 1993. The Roman Near East, 31BC — AD337, Cambridge, MA and London: Harvard University Press Millett, M. J. 1990. The Romanization of Britain, Cambridge: Cambridge University Press Millett, M. J. 2001. Roman interaction in North-Western Iberia, Oxford J. Archaeol., 20, 157–70 Milner, N. P. 1993. Vegetius: Epitome of Military Science, Liverpool: Liverpool University Press Moreland, J. 2001. Archaeology and Text, London: Duckworth Moyes, W. A. 1996. Hatfield 1846–1996, Durham: Hatfield College Trust Musée du Louvre, 1989. Jacques-Louis David 1748–1825, Paris: Editions de la Réunion des Musées Nationaux 56 writing the legions Niegl, M. A. 1980. Die Archäologische Erforschung der Römerzeit in Österreich, Denkschriften der Österreichische Akademie der Wissenschaften, 141, Vienna: Österreichische Akademie der Wissenschaften Osgood, R. and Monks, S. with Toms, J. 2000. Bronze Age Warfare, Alan Sutton, Stroud Parker Pearson, M. forthcoming. Warfare, Violence and Slavery in Prehistory, Brit. Archaeol. Rep. international ser., Oxford: British Archaeological Reports Peddie, J. 1994. The Roman War Machine, Stroud: Alan Sutton Pollard, N. 1996. The Roman army as ‘total institution’ in the Near East? Dura-Europos as a case study, in D. Kennedy (ed.) The Roman Army in the East, J. Roman Archaeol. supplementary ser., 18, 211–27, Ann Arbor: Journal of Roman Archaeology Phillips, T. R. (ed.) 1985. Roots of Strategy, Harrisburg, PA: Stackpole Reece, R. 1988. My Roman Britain, Cirencester: Cotswold Studies Reece, R. 1997. The Future of Roman Military Archaeology: The Tenth Annual Caerleon Lecture, Cardiff : National Museum of Wales Ribera i Lacomba, A. and Calvo Galvez, M. 1995. La primera evidencia arqueológica de la desrucción de Valentia por Pompeyo, J. Roman Archaeol., 8, 19–40 Rich, J. and Shipley, G. (eds) 1993. War and Society in the Roman World, London and New York: Routledge Riches, D. 1986. The phenomenon of violence, in D. Riches (ed.) The Anthropology of Violence, 1–27, Oxford: Blackwell Richmond, I. A. (ed.) 1958. Roman and Native in North Britain, Edinburgh: Nelson Richmond, I. A. 1962. The Roman siege-works of Masada, Israel, J. Roman Stud., 53, 43–54 Robinson, H. R. 1975. The Armour of Imperial Rome, London: Arms and Armour Press Rodger, N. A. M. 1986. The Wooden World: An Anatomy of the Georgian Navy, London: Collins Rostovtzeff, M., Bellinger, A., Hopkins, C., and Welles, C. (eds) 1936. The Excavations at DuraEuropos, Preliminary Report on the Sixth Season, 1932–1933, New Haven: Yale University Press Roth, J. P. 1999. The Logistics of the Roman Army at War (264BC to AD235), Leiden: Brill Roxan, M. M. 1997. Settlement of veterans of the auxilia — a preliminary study, in W. Groenmanvan Waateringe, B. L. van Beek., W. J. H. Willems and S. L. Wynia (eds) Roman Frontier Studies 1995 Proceedings of the XVIth International Congress of Roman Frontier Studies, Oxbow monograph, 91, 483–91, Oxford: Oxbow Books Roy, W. 1793. The Military Antiquities of the Romans in Britain, London: Bulme Sabin, P. 2000. The face of Roman battle, J. Roman Stud., 90, 1–17 Salway, P. 1967. The People of Roman Britain, Cambridge: Cambridge University Press Sartre, M. 2001. D’Alexandre à Zénobie: Histoire du Levant Antique IVe Siècle av. J.C. — IIIe Siècle ap. J.C., Paris: Fayard Sauer, E. 2000. Alchester, a Claudian ‘vexillation fortress’ near the western boundary of the Catuvellauni: new light on the Roman invasion of Britain, Archaeol. J. 157, 1–78 Sauer, E. 2001. Alchester, Current Archaeology, 173, 189–91 Saunders, D. A. 1981. Introduction to Edward Gibbon, The Decline and Fall of the Roman Empire (abridged edn), Harmondsworth: Penguin de Saxe, M. 1757. My Reveries upon the Art of War (trans. T. R. Phillips), in T. R. Phillips, (ed.) 1985 Roots of Strategy, 177–300, Harrisburg, PA: Stackpole Schallmayer, E. (ed.) 1997. Hundert Jahre Saalburg : vom Römischen Grenzposten zum Europäischen Museum, Mainz: Von Zabern Schama, S. 1989. Citizens: a Chronicle of the French Revolution, New York: Knopf. Schlüter, W. and Wiegels R. (eds) 1999. Rom, Germanien und die Ausgrabungen von Kalkriese, Osnabruck: Universitätsverlag Rasch Schmidt, B. E. and Schröder, I. W. (eds) 2001. Anthropology of War and Conflict, London and New York: Routledge Sellar, W. A. and Yeatman, R. J. 1930. 1066 and All That, London: Methuen Service, E. R. 1968. War and our contemporary ancestors, in M. H. Fried, M. Harris and R. Murphy (eds) War: The Anthropology of Armed Conflict and Aggression, 160–67, Garden City, New York: Natural History Press writing the legions 57 Shulman, M. 1995. The B-team that mapped the war, The Standard [London evening newspaper], 24 November 1995. Slusser, R. M. 1993. An American at Bletchley Park, in F. H. Hinsley and A. Stripp (eds), Codebreakers: The Inside Story of Bletchley Park, 74–76, Oxford: Oxford University Press Sommer, C. S. 1984. The Military Vici in Roman Britain: Aspects of their Origins, their Location and Layout, Administration, Function and End, Brit. Archaeol. Rep., British ser., 129, Oxford: British Archaeological Reports Sommer, C. S. 1993. 100 Jahre Reichs-Limeskommission: Zur Intensität der Archäologischen Untersuchungen an der Römischen Grenzen im Südlichen Obergermanien, Fundberichte aus Baden-Württemberg, 18, 235–39 Speidel, M. P. (ed.) 1984. Roman Army Studies I, Amsterdam: Gieben Speidel, M. P. (ed.) 1992a. Roman Army Studies II, Mavors, VIII, Stuttgart: Steiner Speidel, M. P. 1992b. Work to be done on the organization of the Roman army, in M. P. Speidel (ed.) Roman Army Studies II, Mavors, VIII, 13–20 Stuttgart: Steiner [First published in Bull. Inst. Arch. London, 26 (1989), 99–106] Speidel, M. P. 1992c. The soldiers’ servants, in M. P. Speidel (ed.) Roman Army Studies II, Mavors, VIII, 342–52, Stuttgart: Steiner [originally published in Ancient Society, 20 (1989), 239–48] Stallworthy, J. (ed.) 1984. The Oxford Book of War Poetry, Oxford: Oxford University Press Stanfield, J. A. and Simpson, M. G. 1958. Central Gaulish Potters, London: Oxford University Press Terrenato, N. 1998a. The Romanization of Italy: global acculturation or cultural bricolage? in C. Forcey, J. Hawthorne and R. Witcher (eds) TRAC 97 Proceedings of the Seventh Annual Theoretical Archaeology Conference, Nottingham 1997, 20–27, Oxford: Oxbow Terrenato, N. 1998b. Tam firmum municipium: the Romanization of Volaterrae and its cultural implications, J. Roman Stud., 88, 94–114 Terrenato, N. 2001. Introduction, in S. J. Keay and N. Terrenato (eds) Italy and the West. Comparative Issues in Romanization, 1–6, Oxford: Oxbow Tomlin, R. S. O. 1998. Roman manuscripts from Carlisle: the ink-written tablets, Britannia 29, 31–84. Turner, J. F. 1992. Rupert Brooke, the Splendour and the Pain, London: Breese. Van Driel-Murray, C. 1995. Gender in question, in Rush, P. (ed.) Theoretical Roman Archaeology, Second Conference Proceedings, Aldershot and Brookfield, USA: Avebury, 3–21. Van der Veen, M. 1998. A life of luxury in the desert? The food and fodder supply to Mons Claudianus, J. Roman Archaeol., 11, 101–116. Varon, P. 1994. Emptio ancillae/mulieris by Roman army soldiers, in E. Dabrowa (ed.) The Roman and Byzantine Army in the East, 189–95, Krakow: Uniwersytet Jagiellonski, Instytut Historii War Office, 1944. The German Order of Battle, 1944, London: War Office (reprinted 1994, London: Greenhill Books) Watson, G. R. 1969. The Roman Soldier, London and New York: Thames and Hudson Webster, G. 1969. The Roman Imperial Army of the First and Second Centuries A.D., London: Black (third edition 1985, London: Black) Webster, G. 1980. The Roman Invasion of Britain, London: Batsford Webster, J. 2001. Creolizing the Roman provinces, American J. Archaeol., 105, 209–25 Webster J. and Cooper, N. (eds) 1996. Roman Imperialism: Post-Colonial Perspectives, Leicester: School of Archaeological Studies Weinberg, S. 1979. A hoard of Roman armour, Antike Kunst, 22, 82–86 Wells, P. S. 1999. The Barbarians Speak, Princeton University Press, Princeton Westfall, R. S. 1993. The life of Isaac Newton, Cambridge: Cambridge University Press Wheeler, R. E. M. 1943. Maiden Castle, Dorset, Oxford: University Press Whittaker, C. R. 1994. Frontiers of the Roman Empire, a Social and Economic History, Baltimore and London: Johns Hopkins University Press Wilkes, J. J. 1975. Eric Birley: a bibliography 1928–74, Britannia, 6, xi-xxviii Wood, N. 1965. Introduction to Niccolò Machiavelli, The Art of War, Indianapolis: Da Capo, ixlxxix 58 writing the legions Woolf, G. 1993. Roman peace, in J. Rich and G. Shipley (eds) War and Society in the Roman world, 171–94, London and New York: Routledge Ziolkowski, A. 1993. Urbs direpta, or how the Romans sacked cities, in J. Rich, and G. Shipley (eds) War and Society in the Roman world, 69–91, London and New York: Routledge