in Late Antiquity

Although there has been a tendency in modern scholarship on the Roman Empire in late antiquity (early third to early seventh century BCE) to view the period through the lens of transformation rather than violent upheaval, warfare undoubtedly became more frequent, at least compared with the first two centuries BCE, and impacted on regions of the empire long insulated from significant military conflict. The empire of late antiquity faced more significant external challenges, as well as more regular bouts of civil war. Increased use of archery, with its potential to inflict mass casualties, was a distinctive feature of battle in this period; siege warfare became more common, so that civilian populations experienced the violence of war more directly; and expansion in the size of the army placed increased pressures on recruitment and logistical support – pressures which resulted in greater use of force by the state to maintain the military establishment. Changes in the structure of the army also meant that troops were more frequently billeted on the civilian population, who thereby became more exposed to casual violence at the hands of their own troops. In these different ways, late antiquity can be considered a period of Roman history when military violence became more prevalent.

its own right. 1 Brown's positive vision, however, was one which emphasised social and cultural transformation, and privileged developments in religious life and artistic expression over more traditional narratives focused on political and military events. At the same time, the assumption that barbarian invaders were intent on destroying Roman civilisation was being questioned, alongside minimising of their numbers, with the cumulative consequence that the violence of the barbarian invasions was increasingly downplayed. The most prominent reflection of this approach was Walter Goffart's 1980 study of barbarian settlement in the western empire, with its telling subtitle 'The techniques of accommodation', its opening description of that settlement as a 'peaceful and smooth process', and its closing suggestion that the fifth century was 'less memorable for invasions than for the incorporation of barbarian protectors into the fabric of the West.' 2 While this approach can be seen as an understandable attempt to provide a more balanced view of the impact of the barbarian invasions in place of traditional stereotypes of marauding hordes flooding into the empire and wreaking death and destruction, it in turn provoked a reaction from those who thought it was important not to lose sight of the violence that accompanied the invasions -'the horrors of war', to quote the title of an early chapter in one of the most sustained ripostes to the minimising school of thought: Bryan Ward-Perkins' provocatively titled study The Fall of Rome and the End of Civilization. This appeared in 2005, but the beginnings of a reaction can be seen in the early 1990s, with a renewed emphasis on the violent dimension of interaction between barbarian groups and the Roman empire. 3 Since this debate has concerned the impact of the barbarian invasions of the fourth and fifth centuries, it has primarily focused on the western half of the empire. However, a more holistic perspective, encompassing both the east and the full chronological range of late antiquity, from third to seventh century, can leave no doubt as to the importance of warfare and violence in late antiquity. Above all there is the Persian dimension. It was the overthrow of the Parthian Arsacid regime by the Sasanian Persians in the 220s which marked a significant reconfiguration of the strategic position of the Roman empireand for the worse.
Remarkably quickly, the new regime in Persia began pursuing a more aggressive strategy towards the empire, inflicting a succession of major defeats by the middle of the third century and contributing significantly to imperial instability. At the other end of late antiquity, the early part of the seventh century, prior to the Islamic invasions, witnessed more than two decades of war between the Roman empire and Persia during which Persian forces occupied all the eastern provinces of the empire and even laid siege to Constantinople itself in 626. The emperor Heraclius managed to hold out against Persian forces and then surprisingly quickly turned the tables on their overstretched resources to secure an unlikely victory in 628, but it is hard to avoid the conclusion that this protracted war left both powers in a weakened state which facilitated the unexpected Arab successes of the 630s. In the intervening centuries the tempo of warfare between the Roman empire and Persia varied significantly, with, for 4 example, the fifth century largely peaceful, but there can be no doubt that the interactions of these two powers generated significant periods of hard-fought conflict. 4 Nor were barbarian invasions from the north limited to the fourth and fifth centuries. The mid-third century had also seen major inroads along the Rhine and Danube, and across the Black Sea, which had further contributed to the empire's instability, even if order was eventually restored by the end of the century with minimal territorial losses. Similarly in the latter part of the sixth and early seventh century, the empire faced significant challenges from nomadic Avars and Slavs in the lower Danube basin. Furthermore, although expansionist warfare of the type familiar from the Roman Republic was rare on the part of the empire during late antiquity, there was a significant exception during the mid-sixth century when the Emperor Justinian launched campaigns to regain control of the western provinces lost to barbarian invaders during the fifth century, and these campaigns resulted in significant warfare and violence in north Africa and Italy. Finally, in addition to conflict with external enemies, there were also major instances of civil war in nearly every century of late Late antiquity was, then, a phase of Roman history when warfare was particularly prevalent.
This has been implicit in the outline provided in the preceding paragraphs, but it can be reinforced by comparing late antiquity with the period which preceded itthe Principate (late first century BCE to early third century CE). This was the period traditionally associated with the phrase pax Romana. Although the idea of a 'Roman peace' was an ideological construct fostered in the interests of emperors and imperial rule, and by no means meant an end to military-related violence in the Roman world, 6 it was the case that the Roman Empire of the Principate waged expansionist wars far less frequently than had been the case during the Republic and experienced only limited enemy incursions, while the period also marked an end to the chronic civil wars which had devastated the final century of the Republic, even if there were brief, albeit bloody, recurrences in 68-9 and 192-3. There were also some significant provincial revolts, notably in Judaea and Britain, but these were exceptional.
Why then was there an intensification of warfare during late antiquity? As far as external conflict is concerned, the emergence of Sasanian Persia in the east was undoubtedly of fundamental importance. For the first time since the middle republican period (third and second centuries BCE), the Roman empire confronted an enemy commanding comparable material and military resources which enabled it to pose a very serious challenge to Roman power. It was, moreover, a particularly aggressive opponent, especially during the first century after its emergence, and while that can in part be attributed to the determination of successive Persian rulers to consolidate the position of their new regime by the tried and 6 G. Woolf, 'Roman peace' in J. Rich and G. Shipley (eds), War and Society in the Roman World (London: Routledge, 1993), pp.171-94;A. Goldsworthy, Pax Romana: War, Peace and Conquest in the Roman World (Yale University Press, 2016). tested method of military success, 7 it was also a response to repeated Roman aggression against the Parthian Arsacid regime in the second century, during which Roman armies had three times invaded deep into Parthian territory, thereby increasingly discrediting the Arsacids and helping to lay the ground for their overthrow by the Sasanians. So while the impact of Sasanian Persia might seem to be an exogenous factor outside of Roman control, the empire had inadvertently played a significant role in its creation.
The same can be said about the parallel impact of barbarian groups from the north in the third century. The economically underdeveloped and politically fragmented communities north of the empire in the first century gradually enhanced their agricultural productivity and population size, so that, with increasing social stratification and political centralisation, they were able to pose more serious military threats to the empire by the mid-third century. Those developments owed much to the effects of commercial and diplomatic interchange with the empire, which facilitated improvements in agriculture and channelled imperial subsidies to clearly identifiable leaders. The empire understandably preferred to deal with individuals whose authority held substantial sway because this made the conclusion and maintenance of agreements easier -but it also had the unintended consequence of encouraging the emergence of more powerful barbarian groups. And because the empire's approach to relations with these barbarian groups was fundamentally exploitative (e.g., a significant element of commercial interchange was the channelling of barbarian slaves into the empire) and also involved periodic bouts of imperial military aggression, these more powerful barbarian groups were by no means favourably disposed towards the empire. Indeed as in Persia, leadership was validated through successful warfare. 8 The external military challenges of the mid-third century in turn contributed significantly to increased incidence of civil war within the empire. Emperors had traditionally been drawn from the senatorial elite, among whom military competence was valued, but was less important as a qualification for imperial office than dynastic ties. The multiple problems which the empire faced on its frontiers during this period very quickly changed this, so that military experience became a desideratumto the extent that it trumped all other requirements, including senatorial status. In this way, attaining imperial office became a realistic goal for senior military officers of ability and ambition, even if from a lower social background. Greater competition for imperial office ensued, and since the competitors were military men with troops under their command, periodic bouts of civil war were an almost inevitable consequence. This was particularly the case during the mid-third century, but it remained a reasonably regular occurrence during the fourth and fifth centuries, despite the efforts of emperors to re-establish the principle of dynastic continuity. It was less of a problem during the sixth century, largely, it seems, because emperors adopted a strategy of placing individuals connected to them by ties of blood or marriage in positions of senior military command, but there was a recrudescence in the early seventh century against a background of renewed military crisis, on a scale comparable to that of the mid-third century. 9 2. What was distinctive about warfare in late antiquity?
While it is evident that the frequency of warfare intensified in late antiquity, there remains the question of the extent to which its character was distinctive from warfare in earlier periods of Roman history, especially in relation to the theme of violence. The overall picture is in fact one of substantial continuities, but it is possible to identify certain shifts in emphasis. Starting from the perspective of the empire's most important enemies, the forces of Sasanian Persia relied above all on their strengths in cavalry and archery, and in these respects were no different from their Parthian predecessors; the one area of warfare where they quickly proved to be much more adept than the Parthiansand this was an important changewas in the conduct of sieges, about which more will be said below. Northern barbarian groups continued to rely broadly on the same weaponry and tactics as their forebears during the Principate, while the empire had also previously confronted the horsemanship and archery skills of steppe nomads before the advent of the Huns and the Avars, in the form of such groups as Sarmatians and Alans in the first and second centuries. As far as the Roman army itself is concerned, heavy infantry deployed in close order remained of central importance throughout late antiquity. However, these troops were increasingly expected to acquire greater versatility in their weapons skills, especially missiles and archeryareas of expertise traditionally associated with light infantrywhile there was also an increase in the number of specialist units of archers. It is also apparent that Roman cavalry numbers increased in late antiquity, 9 Lee, War in Late antiquity, pp.66-73.
with growing emphasis particularly on mounted archers, no doubt in response to the impact of the Huns. 10 Of the two most important shifts in emphasis relevant to pitched battlesthe greater roles played by cavalry and archeryit is the latter which is the more significant in terms of violence in battle, since it was almost impossible to make horses charge into a solid infantry formation. The principal roles of cavalry on the battlefield in late antiquity have been identified as: pre-battle skirmishing with and countering of enemy cavalry; placing psychological pressure on enemy infantry formations with a view to hastening their loss of cohesion; and either harrying a defeated enemy in retreat, or alternatively providing cover for the retreat of Roman forces. 11 While these various activities could undoubtedly include significant elements of violence, especially where horse archers were involved, it was the greater use of archery, whether on foot or horseback, which had the potential to have a more fatal consequences, 12 but more significant is the bigger picture of how concentrated firepower from Roman foot archers deployed in a 'crescent' formation inflicted heavy casualties on Gothic forces in Italy at the crucial battle of Taginae in 552, and then again two years later on a Frankish and Alamannic army at Casilinum. 13 While archery had undoubtedly been an element of Roman warfare before late antiquity, it does seem to have been used with increasing effectiveness in late antiquity as a deliberate tactical decision against enemies such as the Goths and Vandals who were unable to match Roman firepower. This is not to say that barbarian archery in this period was always ineffective: Gothic archery had played a part in the catastrophic Roman defeat at Adrianople in 378, albeit when the progress of the battle left Roman infantry tightly bunched so as to provide an easier target for Gothic arrows. 14 12 Details and discussion in B. Shaw, 'War and violence' in G. Bowersock, P. Brown and O. does not provide a continuous account, it is nonetheless very detailed and includes features which convey a convincing sense of the highs and lows of this type of warfare from the perspective of defenders, and the traumas it could entail. 18 From among numerous vignettes, the following give some idea of the vividness of his narrative. The impact of the second day of the Persian assault is presented in gruesome detail: Everyone tended his own wounds as best he could or as medical help was available; some of the seriously wounded gave up the ghost from loss of blood after a long struggle; others, mangled by sword-thrusts, were treated without success, and when they at last expired their dead bodies were thrown aside; in some cases of extensive injury the surgeons forbade any attempt at treatment, which would only inflict further useless pain; a number who faced the hazard of pulling out the arrows endured torments worse than death.
Towards the end of the siege, another perspective on this type of warfare is presented as he describes the approach of formidable bodies of Persians supported by lines of elephants, whose noise and size make them the most frightful objects the human mind can conceive. But while we were beset on all sides by the combined pressure of armed men, siege-works and wild beasts, round debris. The elephants, too, were forcibly repulsed. As soon as the firebrands thrown at them touched their bodies, they bolted and their mahouts lost control. But though we subsequently burned their siege-works the enemy gave us no rest. 19 Ammianus' account of the siege of Amida is significant not only as a particularly good example of a 'face of battle' narrative, but also because, as previously noted, siege warfare was more broadly a notable feature of warfare in late antiquity, comprising, on one estimate, more than half the military engagements of the period. More specifically, it was a distinctive feature of Roman conflict with Sasanian Persia because, unlike the Parthians whose ability to conduct sieges appears to have been limited, the Persians very quickly demonstrated Constantina because Persian plundering twelve months earlier had left the surrounding countryside so devastated that it was unable to provide supplies to support his forces, while during the Roman counter-siege of Amida following its capture by the Persians in 503, some of the surviving inhabitants were reduced to such a level of hunger that they resorted to cannibalismnot just of those who had died in the fighting or of malnutrition, but actively killing the elderly or young for food. The author also notes how unburied corpses from battles in the countryside encouraged scavenging by wild animals, who then became emboldened to attack solitary travellers and venture into villages to snatch children. 21 As these episodes imply, late antique siege warfare had a serious impact not only on the troops involved, but also on the civilian population of besieged cities and their hinterland.
The successful capture of a city would almost inevitably be followed by plundering and, especially if the siege had lasted for a significant period (as with Amida in 359 [73 days] and

502-3 [3 months]), by the random slaughter of inhabitants by soldiers venting their anger.
Surviving soldiers and civilians were usually enslaved, with all that that implied in terms of deracination from family and community. Specific features of the Persian treatment of some captured Roman cities were less commonfor example, the crucifixion of the Roman commander of Amida and his senior officers in 359and the sources provide particularly vivid insights into the harsh treatment of captivesfor example, elderly and infirm prisoners who had difficulty keeping up on the journey to Persia had their calf muscles or hamstrings severed and were left to die. Nonetheless, it must be acknowledged that pillaging and the killing and enslavement of inhabitants were standard features of the denouement of sieges throughout antiquity. At the same time, a case can be made for the distinctiveness of Persian practices when it came to enslavement. Whereas the normal expectation in antiquity was that war captives would make their way onto the slave market where they would mostly be purchased by private individuals, the Persians frequently relocated the entire populations of captured Roman cities to Persia, where they were put to work on state projects, above all the construction of irrigation schemes and of cities. This seems to have been part of a deliberate policy to supplement the workforce available to Persian kings, both numerically and in terms of skills. 22 Northern barbarians were generally seen as less of a threat with regard to siege warfare, at least during the earlier centuries of late antiquity, partly because of lack of technological knowledge and partly because of their inability to organise the logistical infrastructure necessary for sustaining a successful siege. This did not, however, prevent them from trying their hand when circumstances seemed favourable, as when the Goths, after their victory at Adrianople in 378, tried unsuccessfully to capture Constantinople. If anything distinguished late antiquity in this respect, it was the vulnerability of imperial centres to attackand sometimes capturewhich would have been unthinkable during the Principate. So Rome was famously blockaded three times by the Goths in 409-10, and eventually sacked on the third occasion, albeit only, it seems, through treachery, and in a similar manner the Vandals pillaged the city in 455. Rome was also subject to a number of sieges during the conflict between eastern Roman and Gothic forces in the 530s and 540s, while in 626 Constantinople was besieged (ultimately unsuccessfully) by combined Avar and Persian forces. Whether these attempts were successful or not, they gave the inhabitants of these large metropolises, to varying degrees, direct experience of war and its consequences in ways which had not been the case for many centuries. 23 Sieges and pitched battles understandably attract the most attention in the narratives of late antique historians, so it is important in concluding this section to note that much of the warfare and associated violence in this period consisted of low-level raiding which was not likely to have an impact at a strategic level, but could help to keep the enemy on the back foot and could certainly have just as serious an impact on non-combatants as other forms of warfare. So, for example, Roman forces based near the fourth-century Rhine periodically made incursions into barbaricum to destroy villages and crops, just as barbarians undertook raids into imperial territory in search of bootyas graphically illustrated by the discovery, in the river gravel at Neupotz in the early 1980s, of barbarian wagons which seem to have been returning home at some point in the mid-third century laden with substantial quantities of Roman plunder, as well as prisoners. This type of small-scale warfare was certainly not unique to late antiquity, but there are some episodes from this period which show it in a particularly grisly light. While in charge of Gaul during the later 350s, the junior emperor Julian is reported to have incentivised troops undertaking raids across the Rhine by offering them a financial reward for every barbarian head they brought back; and during the conflict males as young as twelve years old. 24 However, these actions seem to have been atypical. In terms of what was distinctive about warfare and violence in late antiquity, the focus is better placed on such aspects as the increased use of archery and the increased incidence of sieges, with all that followed from these modes of warfare.

'Institutionalised' military violence in late antiquity
The focus so far has primarily been on warfare and its violent elements, above all in battles and sieges. However, military violence was not restricted to these obvious contexts. It is hardly surprising that the use of compulsion and violence should have played an important role in the maintenance of the army, an institution whose raison d'être was the use of force.
The most obvious area where this was evident was the conscription of manpower into the army, but there were also other areas relating to the maintenance of the army which will be considered below.
Since armies in many periods of history have relied on conscription, the use of conscription in late antiquity might not seem especially significant. However, this is to lose sight of the more immediate historical context of the period, as well as some of the specific features of conscription in late antiquity. While conscription may sometimes have been used during the Principate more than traditionally assumed, it was not a constant feature of the period, especially in the latter stages, during the Severan period of the late second and early third 24 Roman incursions: Amm. 27.10.7,30.3

century -perhaps because significant increases in soldiers' pay and other improvements in
conditions of service under the Severans encouraged more volunteers to enlist. 25 Evidence for renewed use of conscription emerges under the emperor Diocletian at the end of the third century, but drafting of manpower had presumably become the norm again during the military turmoil of the (poorly documented) mid-third century. As in other areas of government, Diocletian formalised arrangements and integrated them into his new fiscal regime, part of which required landowners to provide recruits according to size of their estates. The renewed use of conscription reflected the need for a larger army, which may have increased in size by between twenty-five and fifty percentnecessary to meet the multiple threats which the empire now faced. One way in which reliance on conscription could require the use of force can be seen in circumstantial detail from a Coptic biography of the Egyptian monk Pachomius who, as a young man in the early fourth century, was conscripted into the army; along with others, he was transported down the Nile to Alexandria, and en route the group stopped overnight at various points where they were kept in prison, clearly to prevent them from deserting. Another strategy for discouraging desertion during the fourth century was the introduction of the tattooing of recruits on their hand or arm. In addition to the physical violation involved in tattooing, there was also the psychological damage arising from the use of a permanent method of marking previously reserved for slaves. 26 Recruiting pressures only increased as the fourth century progressed, with major losses of manpower in the civil war battle of Mursa (351) and as a result of the defeats of the emperor Julian's Persian expedition (363) and of Adrianople (378). It was in this context that a bishop in Asia Minor referred in passing to the habitual assaults and injuries inflicted by soldiers on local peasants during the process of conscripting recruits, and that the emperor Valens sanctioned the use of cudgels against monks who resisted military service (375). 27 The desire to avoid military service was apparently such that it became commonplace for individuals to respond with self-inflicted violence, in the form of digital amputation which rendered them unable to grasp a weapona practice which the imperial authorities tried to discourage by various strategies. Conscription continued to play an important role in maintaining the late Roman army throughout the remainder of late antiquity, even if specific evidence relating to its forcible imposition is more limited (there is a report from the 580s of clergy being coerced into military service and of recruiters dragging children from their parents). 28 antiquity ', JRS 77 (1987), 139-55, at 143-4, 149; C. Zuckerman, 'The hapless recruit Psois and the mighty anchorite, Apa John, BASP 32 (1998), 183-94, at 184-6. 27 Greg. Nys., Hom. in XL Mart. (PG 46.784c); N. Lenski, 'Valens and the monks: Cudgelling and conscription as a means of social control ', DOP 58 (2004), 93-117. 28 Amputation: Amm. Marc. 15.12.3;Cod. Theod. 7.13 (with one emperor decreeing that recruits who turned up with missing fingers were to be burned alive); it is therefore surprising that a general in the 370s punished deserters by having their right hands cut off (Amm. Marc. 29.5.49); later centuries: M. Whitby, 'Recruitment in Roman armies from Justinian to Byzantine and Early Islamic Near East,vol.3, (forcible recruitment at 81).
Unsurprisingly, the use of compulsion extended to other activities which supported the army.
So, for example, the men who worked in the state arsenals producing weapons and armour (the fabricenses)another distinctive late antique developmentwere also tattooed. There is also evidence that soldiers were sometimes used to assist in the collection of taxes; 29 since they were presumably deployed in this way so that the threat of force could be used against reluctant taxpayers, and since the army was the government's main item of expenditure, this can be seen as another example of compulsion helping to maintain the armed forces. Finally, the munera sordida ('dishonourable duties') which civilians of lower status could be obliged to undertake by imperial officials now included the grinding of grain and baking of bread for troopsa task sometimes required on a significant scale, as the inhabitants of Edessa experienced in the early sixth century. 30 The exaction of these tasks did not necessarily involve violence, but compulsion can easily metamorphose into force, and force into violence, depending on circumstances. A fundamental change in the organisation of the Roman army in late antiquity also had significant ramifications in this area. By the early fourth century, army units had been recategorised as belonging either to the mobile, field army forces (comitatenses) or to troops based in frontier provinces (ripenses, later limitanei). The latter were stationed in permanent camps and forts, as in earlier centuries, but the former, when not on campaign (and sometimes also when campaigning), were billeted in cities and towns. House owners were obliged to allow soldiers to occupy one-third of their dwellingan arrangement which could be guaranteed to give rise to tensions between 'host' and 'guest' because of the inherent difficulty of demarcating a proportion of a property. However, on top of this, it is apparent that (unsurprisingly) soldiers also tried to use their status and the latent threat of force to extract supplies from home ownersmost commonly, firewood, mattresses and olive oil. It is clear from a succession of imperial laws aimed at controlling soldiers' behaviour that this was a losing battle and that the soldiers often resorted to violence. The contemporary chronicle attributed to Joshua the Stylite provides the most detailed and graphic illustration of how far matters could get out of control. Troops billeted in Edessa in 505 are reported to have ejected house owners from their properties, and resorted to stealing food, clothing and cattle, and even raping local women: 'those who came to our assistance ostensibly as saviours…looted us in a manner little short of enemies.' 31 Since violent behaviour by Roman soldiers towards civilians was hardly a late antique novelty, it is worth emphasising that late antiquity will have witnessed a notable intensification of this phenomenon because of the fundamental changes in military 31 Billetting in cities: Lee, War in Late Antiquity, pp.163-75; imperial laws: Cod.Theod. 7.9.1-4 (esp. 1, which refers to 'seizing of items by violence'); Edessa: Ps.-Joshua, Chronicle, pp.103-4. organisation noted abovespecifically, the creation of field armies and the basing of their units in urban centres. Similarly, the need for a much enlarged army and more centralised organisation of its support structures meant greater use of compulsion in areas such as conscription and logistics. So whatever continuities from earlier centuries of Roman history may be detected, there was nonetheless a significant step-change in 'institutionalised' military violence in late antiquity.

Christianity as a mitigating influence?
A final question which warrants brief consideration is whether another distinctive feature of late antiquitythe increasing impact of Christianityextended to its influencing attitudes to, and the use of, violence in warfare. With the brief exception of the pagan Julian (361-3), emperors from Constantine in the early fourth century onwards gave their official support to the Christian church, and demonstrated that support in various ways, including material resources for charitable activities and church construction. However, that support did not extend to embracing the pacifist elements evident in the New Testament (Old Testament warrior heroes like King David were more appealing role models), nor did church leaders expect emperors to relinquish their military role. Besides, in this phase of Roman history the emphasis was much more on defence than on imperialist expansion, which made it possible to view warfare as a justified activity defending a Christian empire against pagan enemies; but even when there were opportunities for expansion, as with Justinian's sixth-century expeditions to regain the west, religion could be co-opted in justification, with Justinian presenting his campaigns as quasi-crusades against barbarian incomers who had meanwhile misguidedly embraced Arianism, a heterodox version of Christianity. Nor did Constantine's adoption of Christianity initiate a thorough-going Christianisation of the military establishment. Soldiers' religious loyalties during the fourth century seem to have been conservative, and emperors were more concerned to retain their political loyalties than to impose a new religion on them. The religious complexion of the army gradually changed after the fourth century, but obviously not in such a way as to take the violence out of warfare or out of soldiers' behaviour towards civilians. 32 There are two areas where it could be argued that Christianity did play a positive role in the context of warfare, although the first of these is perhaps ultimately less convincing. This is the role of Christianity in bolstering morale, whether it be a general encouraging troops about to face the enemy in battle by parading a holy icon, or a bishop reassuring defenders resisting a siege by leading them in prayers. 33 However, this is not really so different from the use of pagan religious rituals to strengthen morale in the Roman army of the Republic or the non-combatant refugees dislocated by the impact of war. 34 It is in these areas above all that one can see the sorts of developments highlighted by the approach to late antiquity popularised by Peter Brownthe increasing role of bishops in society at large and the charitable activities of the churchintersecting with the consequences of the undoubted violence which arose from the much greater incidence of warfare in late antiquity.

Conclusion
Military violence was endemic throughout Roman history, and so it would be unwise to try to draw too sharp a distinction between late antiquity and earlier periods. Nonetheless, a good case can be made for the greater frequency of warfare in late antiquity, certainly compared with the Principate of the early centuries CE. Moreover, because late antiquity was a period when the empire often found itself in the unfamiliar position of being on the back foot militarily, that warfare impacted on regions of the empire which were removed from the frontiers and which had largely been insulated from military conflict during the first two centuries CE. Siege warfare was a particularly common form of combat during late antiquity, with the result that civilian populations were much more exposed to the direct experience of the violence of war, while the expansion in the size of the army created pressures on recruitment and logistics which resulted in increased use of state force to maintain the army.
Finally, the creation of field armies in the early fourth century, which were then often billeted on civilians, exposed a greater number of communities to the casual violence of their own soldiers. In these different ways, then, late antiquity can be viewed as a period of Roman