Featured post

Gender in the Merovingian World

Friday, 24 August 2012

Two worlds become one: A ‘Counter-Intuitive’ View of the Roman Empire and ‘Germanic’ Migration

[This is a draft of a piece I am writing for a special issue of German History on migration and German history.  It needs another thorough edit and it lacks critical apparatus.  Essentially the argument is that the Roman Empire and barbaricum were inextricably linked throughout the Roman Iron Age.  By late antiquity Germanic-speaking trans-Rhenan areas were inundated with imperial influence.  Migration was two-way and in various forms, all of which, including large-scale 'folk movement' were normal, part and parcel of the imperial frontier's dynamics.  Those of you who know my work will not be surprised by any of that.  The new aspect, although it it is really a development of what I've been writing recently on migration, the North Sea Cultural Zone, etc., is the 'counter-intuitive' conclusion that the relationship between the existence of a formal frontier and significant migration is quite the opposite of the one we have grown used to thinking of.  The collapse of the frontier took with it the mechanisms for migration.  Therefore I have to modify my 2007 epigram that 'the end of the Roman Empire produced the Barbarian Invasions and not vice versa'.  The end of the Roman Empire put an end to the barbarian migrations.  This conclusion helps us contribute more responsibly to modern debate on migration.  It also contributes to a discussion of the formation of Germany.  The end of migration changed the political dynamics of the regions between Rhine and Baltic.  The latter became more inward-facing and from these, eventually, emerged 'Germany'.]


The place of the so-called ‘Barbarian Migrations’ in German history is a topic riddled with irony. The Völkerwanderung are central to any survey of migration, the nation state, of nationalism and German history. Yet, the events of the fourth, fifth and sixth centuries themselves have precious little to do with nations, with states or even with Germany, at least insofar as any of these terms is currently understood. Nor indeed, contrary to its usual label, can this period be distinguished from others as the migration period. Even in more traditional visions of the period, the linkage between the formation of a German identity and the migration of peoples contains its share of irony; in most national myths the primordial migration brings the founders of the nation to the land which bears their name. In the Völkerwanderungszeit, by contrast, the migrants who became so important to later German politics and identity left the region subsequently known as Germany. A region riven with conflict between independent, effectively (if not, technically, legally) sovereign polities was unified partly by appeal to a shared national identity, though the historical attractiveness of such an identity lay largely in the fact that this ‘nation’ was believed to have left its homeland and conquered much – indeed most – of the rest of Europe. The migrating ‘Germans’ played at least as important a role in the political foundation legends (and debates about them) of Spain, France and Italy. So, for all that the conquests of the Germanic barbarians became a source of German national pride, the German nation itself was descended from those of the Germani who had stayed at home! Further irony consists in the fact that the idea that all speakers of a Germanic language could be treated together as a unified ‘Germanic’ people stemmed ultimately from the fact that outsiders, principally Romans, had dealt with all such barbarians interchangeably for their own chauvinistic reasons.

The historiography of the ‘Barbarian Migrations’ has been well studied; a brief résumé will suffice. As is fairly widely known, the term Völkerwanderung was coined in 1557 in Vienna by Wolfgang Laz (Lazius; 1514-65). Cut off from the Roman civilisation then allegedly being revived in more southerly regions, antiquaries across the north of Europe sought ways of including their territories within classical history and of creating a distinct ‘Germanic antiquity’. Had not, indeed, the Germans conquered the Roman Empire and established their kingdoms in its wreckage? The entire Middle Ages could thus be seen as a Germanic creation. In the nineteenth century, the great project editing the texts of the Middle Ages labelled its volumes ‘the Historic Monuments of Germany’ (Monumenta Germaniae Historica). These ideas were powerful tools in German unification, in overcoming centuries of armed hostility between, say, Bavarians and Prussians. Some were used to justify the annexation of Alsace and the Moselle in 1871. The binary opposition between Romania and Germania remained crucial in structuring early medieval studies throughout Europe.

The notion that the wandering Germans had a foundational role in western civilisation, predating that of Rome, found its apogee in the work of Gustav Kossinna (1858-1931). A mix of linguistics, folklore and archaeology led Kossinna to promote the idea that Indo-European civilisation was created by migrating Aryan proto-Germans. The Nazis made much use of these and similar ideas, justifying, beyond general ideas of racial superiority, the annexation of northern France and the proposed renaming of Sevastopol as Theoderichshafen (because the Germanic-speaking Goths had come from that region). The Nazis employed other strands of history relating to the Migrations, particularly the idea that Germanic ‘peoples’ were formed by a war-leader and his followers. This contradicted a rival view of the peoples of Germania as, essentially, proto-democratic communes of freemen – in opposition to Roman hierarchy and tyranny. Archaeology was pressed into service to bolster all these claims.

When the horrific results of this abuse of history were apparent, in the Second World War’s aftermath, a certain back-tracking took place. Rather than being viewed as pure, ‘Germanic’ peoples, the gentes who migrated into the western Roman Empire were interpreted as poly-ethnic. In this (as subsequently dubbed) ‘ethnogenesis’ interpretation, warriors of all origins could be incorporated into a group by subscription to a body of foundation legends, customs and law. This was less of a break with the history of the Nazi period than was once thought; the role of the leader and his Gefolgschaft remained paramount. Archaeological methods avowedly eschewed Kossinna and his fellows’ Mischargumentation (the admixture of various forms of information) for a ‘purer’ form of archaeological reasoning. In fact, this method was predicated on a series of historical assumptions about the Germanic peoples. Nineteenth-century preconceptions about the Germans and their migrations remain fundamental to much writing about the post-imperial period, even if they are rarely directly acknowledged.

Many of these ideas have come under close scrutiny as researchers have re-examined the written and especially the archaeological evidence. The concept of ethnic identity has been thoroughly reassessed, starting from anthropological work on the mutability of ethnicity. The amount of change produced by the migrations and the extent to which post-imperial and medieval western culture and social structures were derived from ‘Germanic’ influences have been radically questioned. In extreme cases, even the reality of migration has been denied.

However, something of an academic counter-revolution against the advances in the study of the ‘Migration Period’ has taken place. Oxford-trained historians have led the way, publishing books repeating the same argument: the barbarian migrations involved real ‘peoples on the move’, and they brought down the Roman Empire. This has stimulated traditionalist archaeologists into a backlash against more nuanced interpretations of the material record. Whatever their authors’ politics, there can be no doubt that these works have been written sufficiently carelessly as to provide succour to far-right extremists. The barbarian migrations have become a popular metaphor amongst racists – a Google search for ‘nouveaux barbares’ will yield many pages of ‘Front National’ diatribe against Muslim and eastern European immigrants. The Norwegian mass-murderer Anders Behring Breivik dwelt more on the Crusades as his preferred historical model but it is nevertheless significant that he described the killings as ‘a small barbarian act to prevent a larger barbarian act’, the latter being the supposed take-over of Europe by Muslim immigrants. In a speech in Rome, the far-right Dutch politician Geert Wilders spoke at length about the Barbarian Invasions and the Fall of Rome as a lesson from history about immigration’s dangers. The metaphor is not limited to extremists; it has worked its way into more mainstream conservative comment, in direct relationship to modern scholarship on Barbarian Migrations. A German newspaper reviewed a museum exhibition set out according to traditional paradigms, arguing that it showed how little comfort we, with ‘a new Völkerwanderung into our Imperium’ could draw from the history of the Empire, with hardly any sign of ‘multiculturalism’, where the migration of peoples meant, besides ‘acculturation’, above all ‘plunder, burning and death on a massive scale’.

The use of scientific (or pseudo-scientific) methodologies to examine migration adds an alarming extra dimension. The use of DNA, whether ‘ancient’ (from excavated material) or ‘modern’ samples (from living populations), is being used to track migration. The danger, barely addressed (at best dismissed as a purely political, ‘ideological’ objection), is of reducing ethnicity to biology and thus to something close to the nineteenth-century idea of race, at the basis of the ‘nation state’. Yet a DNA chain no more has an ethnic identity than does a bronze buckle. What is at stake in fifth- to seventh-century western European history is not the reality of migration; migration is a constant of human existence. It is why migration and ethnic change became so powerfully linked and such an important feature of socio-political change. Science can tell us nothing about any of that.

A reconsideration of the migrations and their place in history is therefore timely. Here I will set out a new framework for thinking about Roman-Barbarian relationships and the way in which these shaped migratory patterns and mechanisms in the late and post-imperial periods. My intention is to provide a basis for a more politically and ethically-responsible intervention by historians in modern political debate.


As intimated above, fundamental to traditional interpretations is the view of the migrating barbarians as ‘Germanic’ – even as ‘Germans’. Partly this stems from Roman ethno-geography, most notably Tacitus’ Germania, which bracketed together all those ‘fair-haired races’ who could not be included under the heading of ‘Gauls’ or ‘Celts’. That definition proved tricky even then; Graeco-Roman writers readily admitted that Gauls and Germani were closely related. Linguistically, we can justify a grouping on the basis that all these peoples spoke a related form of Indo-European language, whether East, West or North Germanic. Such a modern definition, however, does not equate with the classical idea of the Germani. At least the ruling stratum of the Goths, who have in recent decades become something of a paradigm for ‘Germanic migrations’, spoke a Germanic language but they were not considered Germani by Graeco-Roman authors, who usually saw them as ‘Scythians’ or as descendants of other peoples recorded in the same region like the Getae.

No sense of a shared Germanic identity existed amongst trans-Rhenan barbarians, except possibly within the Empire and under the aegis of classical ethnography, when the Romans, on the basis of their world-view, grouped such barbarians together into military units. Usually the different political groups east of the Rhine and north of the upper Danube were perfectly happy to fight each other, especially for Roman pay. Walter Goffart has argued that awareness of a common identity and past, based upon linguistic kinship, is first attested during Frankish campaigns against the Muslims in Spain and southern France. The similarity between Gothic and Frankish personal names was noted. The comprehensive rejection of the idea of a unifying Germanic ethos and identity amongst pre-migratory Germani removes the classic basis for the nineteenth-century view of the German people as rooted in distant history. It is now well established that there is no ‘German history’ before the ninth-century creation of the Eastern Frankish kingdom and the fostering of an identity separate from that of the West Franks. Unlike French and German, which distinguish between ancient and modern ‘Germans’ (germains and allemands; Germanen and Deutsch), English no longer has such a linguistic resource. Henceforth, therefore, I will use the Latin Germani to describe the groups bracketed under that heading by the Romans, and enclose the term Germanic in inverted commas except where discussing linguistics.

Nonetheless, ‘Germanic’ and even ‘German’ remain in use in discussions of the Migration-period, masking a swathe of uncritical assumptions. This is especially, though far from universally, true amongst archaeologists. Certain cultural features, principally furnished inhumation (burial with grave-goods) and types of sunken hut (Grubenhäuser) are habitually described as ‘Germanic’. Their presence or appearance in a region is thus taken to denote the appearance of ‘Germanic’ immigrants. The inheritance of the nineteenth-century (and earlier) notion of pan-Germanic culture is the ludicrous idea that all Germani had some sort of access to a common range of cultural traits, upon which they could draw at will. The appearance of, for example, inhumation with weapons in an area known historically to have been settled by Germani can be described as an indication of that settlement, even if the people in question never used such a rite in their homeland. The description is justified by the fact that somewhere in the huge Germanic-speaking region between the North Sea and Ukraine, another ‘Germanic’ group did bury its dead in this way. When shorn of, however unconscious, Germanist assumptions, the archaeological data rarely provide prima facie support for the interpretation. The attempt to change this intellectually careless state of affairs is making only slow process.

Imperial-barbarian relationships

The persistence of the assumption of a unified ‘Germanic’ culture is one area where the notion of a binary opposition between Roman and barbarian worlds is manifested. Its roots lie, again, in classical ethnography. At least some Roman authors viewed the world in these terms, with the imperium Romanum surrounded by hostile gentes. The idea that natural antagonism dominated Roman-Barbarian relationships pervades popular and academic views of late antiquity. A common interpretation of the late imperial frontiers sees them as a straining dam. It envisages the barbarians of Germania Magna piling up against them until, in the early fifth century, they could no longer hold back the tide. When that happened – maintaining the aquatic metaphors beloved of studies of the migrations – the barbarians ‘flooded’ in, in ‘waves’, ‘swamping’ the Roman provinces.

This vision has been redressed in academic work; sadly, that correction has rarely been taken beyond the halls of academe. When it has, it has frequently been dismissed as mere ‘liberal’ political correctness – sometimes, sadly, by academic historians writing in public fora. There are a number of elements to the revisionist argument. The first concerns the nature and even the existence of an imperial Roman frontier ‘policy’, or ‘Grand Strategy’, to use the term employed by Edward Luttwak in a notorious volume published in 1976. Luttwak, a US defence analyst, viewed the Roman situation very much through the prism of the then current situation in Western Europe, where NATO forces confronted those of the Soviet Bloc across a long, fortified frontier. Whilst the contemporary resonance of Luttwak’s ideas was transparent – even down to the notion of ‘defence in depth’, Luttwak’s preferred NATO ‘Grand Strategy’, which (funnily enough) he saw as adopted in the later Roman period – the idea that the Roman and Barbarian worlds confronted each other as separate, opposed entities, rather like the Western powers and their Soviet antagonists, remains deeply entrenched. Just as western propaganda presented the ‘Russians’ as ready to invade the West the very instant the latter let its guard down, the idea implicit in much writing about the Late Roman Empire is that the barbarians were similarly, perpetually watching and waiting for their chance to overrun imperial territories.

Why this should have been the case is (as with the theories of automatic Soviet aggression) rarely explored in any depth. Barbarians were just like that. Few people have expressed the preconception as clearly as did Henri Pirenne in the 1920s: the barbarians were irresistibly drawn towards the Mediterranean, ‘happy regions where the mildness of the climate and the fertility of the soil were matched by the charms and the wealth of civilization’. The reality of such a proto-Germanic ‘Drang nach Suden’ need not detain us. Suffice it to say that this putative intention is largely derived, ex post facto, from the contingent fact that some Germanic-speaking barbarian groups finished their wanderings on the shores of the Middle Sea. The mechanisms of migration will receive further discussion below. For now all that needs to be said is that the traditional assumption of a natural barbarian desire to conquer the Roman Empire lacks empirical foundation.

Luttwak’s concept of a Grand Strategy has received detailed critique. At a basic operational level, it has been argued that the Romans had no notion of a Grand Strategy. Frontier activity tended to be local, ad hoc and contingent. The idea that the Later Empire saw the development of a strategy of ‘defence in depth’ has been dismantled. The deeper frontier belts of the later Empire resulted more from the demands of supply in a period when the armed forces were largely paid in kind and received their supplies from the state rather than buying them in local markets. Spreading the troops more evenly through a largely militarised and state-controlled frontier region, such as Northern Gaul, made more sense in this context. While the danger of barbarian raiding cannot and should not be discounted, the defence of interior strong-points also stemmed from the need to protect the late imperial tax and supply network’s nodal points. Banditry and rebellion (often fused in Roman ideology) were as much a danger as barbarian attack – perhaps more so. The old idea that the short walled circuits of late Roman cities resulted from haste and emergency, during the barbarian invasions of the 270s, has been rejected on numerous grounds. That the Romans moved from ‘preclusive security’ to ‘defence in depth’ is belied by the heavy expenditure on frontier fortifications during the fourth century, as late as Valentinian I’s reign (364-75). Finally, the separation of Roman forces into border forces (limitanei and ripenses) and field armies (comitatenses) is unlikely to have reflected the postulated strategy of holding up barbarians in a deep fortified belt until mobile field armies could move up and destroy them. Apart from the fact that the distinction between the two types of troops was more fluid than might be supposed, the speed of communications and movement in Late Antiquity undermine Luttwak and his followers’ attempt to transpose modern strategy onto the Roman Empire. Mobile field armies composed of the Empire’s best troops and located close to the emperor himself developed during the turbulent third century and certainly resulted mostly from the desire to keep a large force of the best troops close to the ruler and thus away from any challengers for power.

If there were general strategic principles in operation, governing to some extent the otherwise ad hoc and piecemeal activities of individual emperors, these were more likely concerned with internal imperial politics than with any actual ‘Grand Strategy’ confronting ‘the barbarian threat’. In practical military terms, although barbarians could cause widespread damage and demoralisation during their raids, and this should not be understated, the military balance of power lay overwhelmingly in the Romans’ favour. With a total military manpower in excess, it is estimated, of 400,000 men, the Empire had, by antique standards, bottomless reserves of troops to draw upon. Even on the Rhine, the number of troops available far outweighed anything that even a large barbarian confederacy could assemble. Roman sources habitually discuss the size of barbarian armies in the tens of thousands – occasionally hundreds of thousands. It is astonishing how many professional historians continue to accept these estimates. The largest nucleated settlements known in barbaricum east of the Rhine housed about two hundred people. There were no urban sites. In much of the trans-Rhenan zone, for most of the Roman Iron Age, agricultural surplus and its control were insufficient to support wealthy, established élites or much craft specialisation and organised industry. That, in this context, small or even middling groups could have mobilised armies of twenty- to thirty-thousand men and move them even as far as the frontier without causing catastrophic famine in their homelands, or maintain them without starving once within Roman territory, defies belief. Julian was able to starve out a force of only 600 Franks. While the Late Roman Empire had 400,000 men or more under arms, its largest field armies seem only to have numbered about 20,000 troops; expeditionary forces of a couple of thousand were frequently deemed adequate to deal even with serious regional trouble. Well-organised, taxing states with more complex economies, banking loans, urban settlements and more developed agricultures did not habitually raise armies of 30,000 men or more before the sixteenth or seventeenth century. The confederate Alamannic army that fought the Roman army at Strasbourg in 357 cannot possibly, therefore, have approached the 35,000 men claimed by Ammianus Marcellinus. If it matched the 13,000 on the Roman side, that will have been remarkable. Add to that the undoubted superiority that the Romans held in the fields of armament, fortifications, siege techniques and logistics and one arrives at a better idea of how serious the military threat posed by the Germanic-speaking barbarians really was.

The proof of this argument lies in the relative importance attached by Roman commanders to Roman and barbarian enemies. Invariably they left invasions by trans-Rhenan barbarians to be cleared up afterwards, no matter how damaging they were (as with the Alamannic attacks during the civil wars of the 350s). Primacy had to be given to the subjection of the larger, better-equipped forces of Roman enemies and this remained true through the fifth century. Indeed the Rhine frontier could be effectively denuded of defending units, as perhaps in the 390s and the first decade of the fifth century. Claudian’s panegyric for Stilicho admits that he left it defended mainly by the fear of his name and by treaties with barbarian frontier kings. Only the prevalence of the modern ‘straining dam’ theory of Roman history makes this fact seem surprising. For most of the past two centuries, France and Germany, or Canada and the USA, have not permanently had to line up huge field armies along their shared borders in a form of ‘dynamic tension’, lest their removal led to automatic invasion by the other side. There was, likewise, nothing natural or automatic about the hostile movement of Germanic-speaking barbarians across the imperial borders.

So, why did the Romans focus so heavily upon the frontier and the barbarian threat? One of the most important historiographical developments of recent decades has been the increase in awareness of the ideological role of the frontier and the barbarian ‘bogey man’. The barbarian threat – especially the ‘Germanic’ barbarian threat – was largely, in John Drinkwater’s words, a Roman artefact. The third century’s political instability had made it clear that the unusual cultural and economic features that had held the Empire together were no longer effective. Political fission and multiple Empires had briefly seemed realistic possibilities. The revived late third- and fourth-century Empire based itself upon a large bureaucracy. This 20-30,000-strong civil service operated as a huge interlocking pyramid of patronage networks with the emperor at the pinnacle. The attractions of promotion, in terms of privilege and local precedence, and the short terms for which many offices were held made this system an enormous resource for the emperors. Local ambitions were best served by imperially-moderated competition for the benefits brought by involvement in government and thus the Empire was bound together as a political unit. The principal role of this crucial bureaucracy was to collect taxation; taxation was primarily required to pay the army. In turn the justification for the army was defence from barbarian attack. No emperor could admit that he needed his army to cow his subjects and potential rivals! Central to any emperor’s claims to good rulership – and thus to the legitimacy and attraction of association with him – was a role as the domitor gentium (pacifier of the nations). Victory over barbarians was especially important and expenditure on frontier fortifications a valuable sign of imperial good management. Consequently, throughout the period up to 388, the emperors remained close to their frontiers in the ‘inside-out’ late Empire, close to their élite field armies and also, importantly, easily accessible to the local aristocracies of strategically-important frontier provinces, like Pannonia and Gaul. Patrick Geary famously said that ‘the Germanic world was perhaps the greatest and most enduring creation of Roman political and military genius’; Alexander Callander Murray has responded that the late Roman world was in many ways created around the barbarian threat. In constructing the barbarian world, the Roman Empire defined itself.

Much of the military basis for seeing the Roman and barbarian worlds as mutually-antagonistic entities can thus largely be discounted. Violence and warfare occurred but are unquantifiable. The whole frontier was rarely in turmoil at once; by far the majority of such barbarian raids as occurred were likely small-scale rustlings or robberies necessitating ‘police actions’ in response, rather than large incursions requiring fully-fledged campaigning. The default setting for Roman-Barbarian relations across the Rhine-Danube frontiers was most probably quiet coexistence.

Plentiful evidence supports this notion. Fourth-century human movement is far more commonly revealed, archaeologically, going from the Empire to Germanic-speaking barbaricum. For all the weight of received wisdom, movement in the opposite direction is mostly implicit. Such evidence takes several forms. Most interestingly, items of Roman military uniform are found in the cremation (and, later, inhumation) cemeteries of modern Lower Saxony and the northern Netherlands, testifying to circular ‘career-migration’, attested in the written sources too. Saxons left home, served in the Roman army and returned to their native land. Such was the status acquired by having served Rome that when these men died their families displayed that imperial connection in their funeral. Only differential funerary custom makes Lower Saxony and neighbouring areas stand out on archaeological distribution maps. Had grave-goods been incorporated in the burial rites of the rest of Germania Magna we would doubtless see such objects more evenly scattered across the region. It is well-known that the late Empire recruited heavily amongst the Germanic-speaking barbarians. This underlines yet further the military imbalance between Rome and her neighbours. However small barbarian manpower was, it was reduced further by the Empire’s recruitment of so many young warriors, among whom the attractions of Roman service doubtless far outweighed those of membership of a chieftain’s warband. Such troops served Rome loyally, not least in attacks on their co-linguists: further nailing shut the coffin of the idea of a unified proto-German ‘people’.

Movement is also revealed by the distribution of other types of Roman artefact. Settlements close to the limites are saturated with Roman imports of all sorts, bearing witness to a lively system of exchange. Imperial products are, of course, less frequent the further away from the Rhine one moves, but they reached a long way. A type of bronze bowl manufactured in the Meuse valley is known as the Vestland Type because of the frequency with which it occurs in that region of Norway, for example. Roman imports are found around the Baltic. On Fyn it is clear from their distribution that the regional élite controlled access to prestigious Roman artefacts. The site at Gudme-Lundeborg, through which such imports entered the region, is emblematic of a particular type of small, high-status settlement involved in trading with the Empire. Another example is Dankirke on the west coast of Jutland. The commerce that reached Jutland and Fyn was presumably seaborne but other trade is further, demonstrated through the movement of goods up the ‘amber routes’ between the Baltic and the Empire. The growth and prosperity of some settlements in barbaricum near the frontier, such as Wijster, has been linked to trade with the limes. This might also have lain behind the organised iron-working revealed at Heeten in the Netherlands. The exchanges attested by this evidence can reasonably be interpreted as commercial.

Other Roman-barbarian interactions are attested by the Roman silver and gold objects in prestigious inhumation burials, especially in central Germania. Although these might have originated as loot, it is as plausible to view them as diplomatic payments paid by the Empire to its friends in barbaricum. These relationships are also revealed by huge gold medallions bearing the same imperial iconography as is found on the later Empire’s normal gold coinage, the solidi, and usually of a weight that corresponds to a multiple of the latter. It would seem erroneous to view these as coins rather than as enormous, prestigious units of bullion struck for diplomatic purposes.

The deployment of Roman objects in public ritual, seen in the Saxon cremations and the lavish central German inhumations, underlines very clearly the importance with which links with the Empire were viewed in late third and fourth-century barbaricum. Rome’s late antique cultural dominance is also revealed by the fact that just across the Upper Rhine and Danube frontiers, in the territory of the Alamanni, local leaders made imitations of just the sorts of official brooches that Saxon soldiers took back home with them. We can set this evidence against that of the written sources to suggest that after three hundred years of close proximity to the imperial superpower, society, politics and culture amongst the Germani were so soaked in Roman influence that legitimate power was difficult to express except via imperial idioms. According to Ammianus, an Alamannic ruler even named his son Serapio, without the latter seeming to mind being saddled with the strange moniker. In more general ways it might not even have been that unusual. Until perhaps the last quarter of the fourth century it was normal for Germanic-speaking and other barbarian recruits into the army to adopt Roman names.

Furthermore, the new third- and fourth-century confederacies – Saxons, Franks, Alamans and Goths (like, perhaps, the Picts) – appeared directly on the imperial frontiers. The role of the imperial frontier in these units’ formation has recently been much discussed. The Alamanni might have come together in some way under Roman auspices, during the imperial withdrawal from the Agri Decumates. It was, however, probably more usual for the Empire to play a part in group-formation through the de facto proximity of the limes and via possibly unintended results of imperial ‘foreign policy’. The cross-border trade discussed above probably allowed greater wealth to be concentrated in the hands of local rulers, but Roman diplomacy was doubtless even more important.

The Romans had always played their barbarian neighbours off against each other. During the third-century troubles, diplomatic payments to barbarians across the Rhine probably became larger and more common as rivals for power purchased either quiet borders, while they removed troops to fight elsewhere, or allies to disturb the tranquillity of their rivals’ frontiers. These policies continued through the fourth century and into the fifth. The Romans could intervene directly in barbarian politics close to the limes and this usually enabled them to prevent the barbarian confederacies from coalescing. However, when the Romans were distracted by civil war, the confederate identities that had emerged in the third century permitted larger barbarian units to form. The stakes in barbarian politics had been raised.

It is important not to get carried away with a cosy view of the Roman-barbarian frontier. Murderous and very damaging raids on the Empire occurred. When they were launched, imperial punitive expeditions into barbaricum were probably even more lethal. The barbarian could be, and was, presented as the very antithesis of the civilised Roman, with much in common with wild animals. Roman troops were unleashed across the frontier with orders to kill every living thing they found; captured barbarians could be thrown to wild beasts in public spectacle, or forced to kill each other in the arena. When activated, the Roman ideology of the barbarian put the latter very squarely in a state of exception, of ‘bare life’, in the phrases coined by Giorgio Agamben. That said, although the binary civilised:barbarian opposition could be mapped geographically onto the imperial frontiers, it was considerably more flexible and less well policed. Nor should it be fused with the Romans’ ‘taxonomic’ ethnography of the different peoples of the world, in which the limes played a less defined role. Barbarians, like animals and women, were praised for their closeness to the ideal, central pole of civilisation represented by the Roman male, as well as damned for distance from it. Unlike women and animals, indeed, they could be brought so close to that ideal as to be incorporated in it, with no notable trace of their different origins. Thus the very troops slaughtering men, women and children in Alamannic villages or dragging captives off to the arenas might themselves be Germanic-speaking barbarians – possibly even Alamans. Although always present as a resource to be called upon, the ideological distinction between the civilised Roman and the ferocious barbarian was more often left hovering – however chillingly and menacingly – in the background of cross-border relations.

The Roman Empire and Migration

The preceding discussion has demonstrated the inextricable connection between the ‘Germanic’ barbarian and the Roman worlds, indeed the dependence of the former upon the latter. Rather than two hostile, opposed blocs, we should envisage a single world with a Roman core and northern barbarian periphery. This provides a very different framework into which to fit the documented movements of barbarians into Roman territory. We can look at the dynamics from the perspective of individuals and of larger groups.

Let us take individuals first. In recent decades, what is known as ‘migration theory’ has made an appearance in the studies of late antique population movement. This has largely been pressed into service by traditionalists wanting to maintain the idea of large-scale folk-wanderings, especially into Roman Britain. Its employment has rarely gone beyond demonstrating that migrations occur but in fact this ‘theory’ (in reality a body of general, comparative observations about the nature of migration) has much more to offer. Ironically, though, it tends to tell against the arguments of those who make most use of it, in that it provides little support for reading material cultural change in simple terms of the movement of people from barbaricum into the Roman Empire. Particularly important for our purposes are migration theory’s insights that processes of population movement rarely if ever constitute one-way traffic; that migrants follow established routes rather than ‘flooding’ over the borders on a wide front (something that surely applies a fortiori to antique population movement); that they are drawn to pre-existing immigrant communities; and (underpinning all the previous comments) that the flow of information is crucial to migration. A typology of migration has also been set out, which helps to make the discussion more precise.

What is interesting about this is that it suggests that barbarian immigration was in many respects considerably easier when the Roman Empire was functioning and its borders still effectively maintained. There was a steady stream of information flowing from the Empire to barbaricum (manifest in the archaeological data discussed earlier). For those, like the Saxons mentioned above, who undertook circular ‘career migration’, this was vital. It was essential to know if, and whom, the imperial army was hiring, where to go to be enrolled, and how to get there. For families moving for economic reasons, again, information about the feasibility of settlement, routes and the safety of travel, and existing communities was important; in addition to controlling access, those manning the imperial borders seem to have been able to organise the settlement of those allowed into the Empire. Like the movement of traders in the opposite direction (or those soldiers’ movements back home after their service expired), these migrations required a peaceful and well-ordered frontier. Population movement that was (in terms of the social units involved) small-scale thus flourished under the Empire.

Large-scale population movement was also integral to the imperial world. As early as Julius Caesar’s day, the Romans’ management of their borders and their promotion of allied leaders or groups had led to such migration. Those who lost out in barbarian politics had long sought shelter within the Republic’s or, after it, the Empire’s frontiers, especially if they had at some point been Rome’s friends. Unsurprisingly, they took their families with them, as well as being joined by those of their compatriots who had supported their faction. Large groups were, then, settled within the Empire and it is moot whether such groups were smaller than those that entered in the so-called Migration Period. Forty-thousand Suevi and Sicambri were (allegedly) settled in 8 BC; fifty-thousand ‘Getae’ in AD 5; an inscription records 100,000 barbarians settled in Moesia under Nero; and so on. The same prudence needs to be shown towards these numbers as to those of barbarian armies; what matters is that these were larger social groups than simple families and that the Romans described them in the same sort of terms as they used for the fifth-century movements. As before, the effective management of the frontier was crucial; the groups admitted were moved and their settlement organised and administered by the imperial government. These provided, one assumes, focal communities to which later migrants from the same area were drawn. All told, it is quite likely that significantly more migration took place across the Rhine and Danube before the collapse of the western Empire than afterwards, as long as we keep in mind the nature of such movement.

The greater population movements of the late fourth and fifth centuries in large part followed the same patterns. Although an attempt has recently been made to revive the antique ‘domino theory’ first espoused by Ambrose of Milan, with the Huns pushing the various peoples in front of them, such a view is tendentious at best. A more sophisticated examination suggests that the Gothic crisis 376-82 should be understood in the terms of normal Roman practice. The trans-Danubian region was thoroughly destabilised by Emperor Valens’ campaigns in the 360s. Rather than bursting unexpectedly onto the scene as a deus ex machina, the Huns became another element in an already faction-ridden situation north of the Danube, which took perhaps thirty years to resolve. It is clear that the pro-Roman Tervingian Gothic faction led by Fritigern and Alaviv, like the Greuthungian group that formed around the child-king Videric, lost out in this civil strife and took the well-worn route of moving to the imperial frontier and demanding asylum. It was the mismanagement of the Greuthungian receptio (in no small measure attributable to the emperor’s distance from events) rather than its scale that led to the Gothic revolt and eventual victory at Adrianople (378). In any case, close scrutiny of the more contemporary data, rather than teleological interpretations based upon subsequent events and sources (themselves seeking explanations for later actions) suggests that the Goths were eventually worn down and surrendered, being settled in fairly conventional manner. Similarly, a solid case can be made that the Gothic invasion of Italy under Radagaisus and the so-called Great Invasion of c.406 followed a fairly typical pattern. The final emergence of the groups based around the Huns as the dominant faction north of the Danube led to the losers in such politics, like the Tervingians before them, seeking refuge within the Roman frontiers. The collapse of Roman frontier management in the decades either side of 400 also played a part, removing the usual checks and balance that maintained a rough balance of power. As had happened before, the end of effective, active Roman ‘foreign policy’ produced crisis and tension in the barbarian world and the emergence of a larger and more dangerous unit there.

It is interesting, in this connection, that the barbarian groups who moved furthest in the fifth and sixth centuries came not from the large border confederacies but from a sort of ‘middle band’ of territories in the interior of Germania Magna: Vandals, Sueves, Burgundians, Lombards. It was in this zone that Rome’s role in maintaining political stability might have been greatest and thus where the end of effective frontier management might have been most keenly felt.

The crucial role of the Roman frontier and its dynamics in governing migration to and from Germania is underlined by the fact that significant movements by large groups into the former imperial territories more or less end with the political and military disintegration of the frontier during the first half of the fifth century. Thereafter (and indeed during the disintegration), such movement as occurred across the Rhine was short-range ‘drift’ by Franks and Alamans, largely small-scale in nature, if cumulatively significant. The Saxon movement across the North Sea was probably similar in nature. This probably ought not to surprise us. With the collapse of the Empire the distances over which information travelled probably reduced considerably, shortening the range of human movement commensurately. The economic collapse attendant upon the fifth-century imperial crisis only underlined this by shortening the distances over which exchanges took place. A final support for the thesis proposed here, relating migration to the core-periphery relations between the Empire and its neighbours, might be found in the way that the last of the classic ‘Germanic’ Völkerwanderungen, that of the Lombards, also moved from trans-Danubian barbaricum into (at that point) imperial territory.

The ways in which Roman and barbarian worlds were intertwined is made very clear by the ways in which the crisis of the Roman Empire around 400 produced crisis in barbaricum. The migrations of Goths, of Vandals, Alans and Sueves, of Burgundians and, later, Lombards have been mentioned. Saxon migration was related to the political, social and economic instability produced by the crises in the north-western provinces around 400. As in Britain and northern Gaul, archaeology there reveals similar symptoms of such crisis: settlements and cemeteries were abandoned, new inhumation rites involving more significant grave-goods deposition appear. There are indications that the Saxon confederacy fractured. In these regards Saxon migration has some points of contact with those of the Vandals and others. The divergent trajectories of fifth-century development in different areas of barbaricum can be related to the varying relationships between those areas and the Empire before 400.

Another point worth stressing is that, as before, movement across the former frontier was two-way. The similarity between the archaeologies of fifth-century eastern Britain (indeed as far north as Scotland), northern Gaul and North Sea Coastal Germany can only be explained by postulating human movement across and around the North Sea in all directions and the creation of a North Sea cultural zone. Written sources also testify to movement back into barbaricum in the fifth and sixth centuries. If anything, movement from the formerly imperial provinces into what had been barbaricum might have become proportionately even more important in the post-imperial centuries. Politically, this was certainly the case, as the late fifth and sixth centuries saw the establishment of a Frankish hegemony east of the Rhine, which ultimately became the eastern portions of the Carolingian Empire.


In 2007 I wrote that we should reverse the usual formula: ‘The ‘barbarian migrations’ were … the product of the ‘end of the Roman Empire’, and not vice versa.’ I would now correct this conclusion. The imperial crisis around 400 can be said to have produced the last great ‘folk movements’ of the old style, but closer examination suggests that it would be truer to say that ‘the Fall of Rome’ put an end to ‘the barbarian migrations’. The analysis that produces such a conclusion should enable students of the processes of the so-called Völkerwanderung to contribute in a more responsible fashion to modern debate. The counter-intuitive aspect of this study is that it posits a relationship between the existence of a formal frontier and large-scale migration that is the opposite of the one we are accustomed to envisage. For centuries, population movements of varying sizes and characters were a normal part of the dynamic of the Roman frontier. When that frontier collapsed, the mechanisms for such movement went with it. A perhaps more important contemporary resonance is that we have to view these types of migration within a broad system of intimate core-periphery Roman-barbarian relationships, rather than seeing population movement as a product of two hostile, opposed entities. This, it seems to me, is a more responsible and ethical basis upon which to move from the late imperial past to migration in the twenty-first century world system of ‘the west and the rest’.

For the purposes of German history, this rethinking is important. If the migrations played a part in the emergence of Germany it was in a quite different way from that which we are used to supposing. The collapse of the Empire changed the nature and possibilities of population movement and their relationship to politics within the trans-Rhenan regions. As a result, new political dynamics – which we might provisionally describe as more inward-looking – came into being in these areas. These would create a more integrated zone of political interaction, quite different from the Germania Magna of the Roman era. It was that which later became Germany.


  1. I like this lots but not sure I understood most of it! Good work Guy!
    Adie Williams

  2. I recently read Peter S. Well's "The Barbarians Speak", which explores similar ideas with the intriguing illumination of other encounters between large empires from later times - the Spanish with the Inca, for instance - to expand our understanding of the Roman Empire's encounter with Germanic groups.
    My own interest is in Merovingian food, which has lead me to a number of such nuanced views of these interactions. Yet one continues to find the term "barbarian hordes" alive and kicking in numerous sources, even though Clovis' father held rank in the Roman army and wrote (apparently bad) Latin verse and these "hordes" had been comfortably established in places like Picardy for some time before Clovis' takeover. And paid in Roman coin.

  3. The army that Clovis' father commanded was operational much further south than Picardie, even. It was Clovis' grandson Chilperic, rather than his father, Childeric, who wrote the bad poetry, by the way - though its easy to get those names confused!

  4. Damn. And here I was faulting Toussaint-Samat for confusing Radegund (saint) with Fredegund (poisoner).
    Thanks for the correction.
    Otherwise, I knew Childeric had fought far beyond Picardy (it seems that with a few exceptions the Franks had been stalwart allies - or servants? - of the Romans until Clovis). But my impression was that actual settlement only went about that far south (before the - if it was one - conquest).


Note: only a member of this blog may post a comment.