Empires and Barbarians

by Peter Heather

  When social scientists started thinking about group identity, they generally assumed that human population groups became politically and culturally distinct from one another as a result of physical separation. One major advance since the Second World War is the insight that active group identities are regularly generated out of precisely the opposite scenario: intense contact in the form of competition.

  Developing a group identity is often about becoming part of an entity strong enough to protect a particular set of interests. And you don’t have to look very far in the events of the later fourth and the fifth century to find violence – the epitome of competitive contact – brokering the renegotiations of identity that produced the new kingdom-forming barbarian alliances. Some of the new identities (particularly those of the Visigoths and the Vandal–Alan coalition) were formed by migrants who needed to operate on Roman soil in larger groupings so as to preserve their independence against traditional imperial policies designed to dismantle threatening concentrations of outsiders. Others were born out of the collapse of Attila’s Empire which again generated fierce competition, this time among the many armed groups gathered by the Huns on the Middle Danubian plain. And yet a third was generated among groups looking to take over the landed assets of the collapsing western Empire. The Visigoths and Vandal–Alan coalition were well placed to play this more profitable game, having united originally so as to survive, but new groups formed to participate in the same land-grabbing exercise – particularly the Ostrogoths, Franks and Lombards. On a smaller scale, the Anglo-Saxons moving into lowland Britain also fall into this category.

  All of these new group identities were born in violence, and even if recently renegotiated they were reasonably durable, at least among the higher-status warriors who were the prime beneficiaries of the ambitions they had been formed to pursue. That does not mean, of course, that every member of the group, even those higher-status warriors, was equally committed to the new identities, or that they were indestructible. Neither of these is true of any modern group identity, either. But those forged in the later fourth and the fifth century were real political phenomena, not mere ideologies or dynastic fantasy.16

  The migrations undertaken by these groups were correspondingly substantial. As we have seen, some of the historical evidence for large, mixed population groups taking to the road with massive wagon trains is much too weighty to be dismissed. Ammianus on the Goths of 376, in particular, is too well informed and demonstrably capable of describing a variety of barbarian activities for us to dismiss his account as migration topos. As the revised notions of group identity would suggest, the large kingdom-forming concentrations of population did not move from point A to point B untouched by the process. They recruited extra manpower as they went, which they slotted in, as appropriate, to the various positions available within the group: in Germanic-dominated groups, it would seem, either as higher-status free warrior, freed lower-status warrior or non-militarized slave. But accepting this does not license us to dismiss their migrations as relatively small-scale phenomena. While certainly different from migration units you find today, large mixed population groups do make sense in their own context, given the general level of development of contemporary non-Roman society and the kinds of enterprise being undertaken.

  Three principal types of migration emerge from the narrative. The first comprises those mixed groups of outsiders who moved across the imperial frontier because of the direct or indirect threat posed to their existing territories by the build-up of Hunnic power. The Tervingi and Greuthungi of 376 fall into this category, as do, in my view, the Goths of Radagaisus who invaded Italy in 405/6; and the Vandals, Alans and Sueves who moved over the Rhine shortly afterwards. The many different strands within these two pulses of migration eventually reorganized themselves into two large confederations: the Visigoths and the Vandal–Alan coalition, as noted earlier. These could each field something in the region of ten to twenty thousand warriors, and both had women and children along besides, not to mention slaves. The motivation for all these groups was essentially political and negative – fear of the Huns – but they were also busy calculating, increasingly from direct experience, what it would take for them to carve out a profitable niche within Roman territory. The constituent groups behind the Visigothic alliance, moving from Ukraine to southern France via the Balkans and Italy, the others from central Europe (or from much further east in the case of the Alans) to North Africa via Spain, also provide the most spectacular examples of long-distance movement. Their treks took the form of discrete jumps, with considerable pauses in between, rather than one continuous movement, because migration was part of a developing survival strategy. The chronological gaps also reflect the distances involved in these treks, since information had to be acquired at each jumping-off point about the new options that were opening up. The Vandals – from modern Hungary or thereabouts – can have had little sense of how to get to North Africa from Spain when they first set out, or even perhaps that you could.17

  The second category consists of those groups, many again involving women and children, who moved out of the Middle Danubian heartlands of the Hunnic Empire in the chaos following Attila’s death. Again, some of these were substantial. The Amal-led Goths from Pannonia comprised over ten thousand fighting men, besides women and children. The Sueves, Heruli and Rugi who made their way into the army of Italy or joined the Amal bandwagon certainly mustered at least a few thousand warriors each, and of these the Heruli and Rugi, at least, were again moving with women and children.18 These groups’ motivation was, again, partly political and negative: fear of the other parties involved in the competition unleashed by Hunnic collapse. At the same time, there was a powerful element of opportunism. The Amal-led Goths took calculated decisions first to throw themselves into east Roman territory, and second, having united with the Thracian Goths, to move on to Italy. In both cases, their decisions were based not only on the limitations and difficulties of their current situation but also at least as much on the greater degree of prosperity potentially available at the projected destinations. This category is distinguished from the first not only by a greater element of opportunism, but also by the distances involved. The long march of Theoderic’s Goths from Hungary to Thessalonica, to Constantinople, to Albania and then on to Italy is impressive, but it was not quite of the same order as the epic trek of the Vandals, or the trials and tribulations of the Visigoths.

  Frankish and Anglo-Saxon migration into north-eastern Gaul and lowland Britain, respectively, took a third and different form, although there were some significant variations between the two. Distances were shorter; and the characteristic migration unit within the flows, smaller. The archaeological evidence suggests that the most intensive Frankish settlement happened in areas of Roman Gaul within a radius of only a hundred kilometres or so from the limits of their previous holdings. Anglo-Saxon groups obviously had to cross the Channel and/or the North Sea, but this too was a relatively short hop. The range of motivations in play was also different. The North Sea may at this time have been eroding some continental coastlands, making some long-cultivated areas unusable.

  For the most part, however, the motivations behind Frankish and Anglo-Saxon settlement were positive and predatory. Both followed the elimination of effective Roman state power at their respective destinations; for both, previously, the Empire’s armies, fleets and fortifications had made it impossible for them to do anything more than raid. Both flows were filling power vacuums in fairly adjacent landscapes, attracted by the relative prosperity of the target area’s more developed economy and the fact that landed wealth was easier to access there. Franks and Anglo-Saxons had no need to operate in migration units on anything like the scale of those in the other categories, therefore, although it does seem likely, since conquest and settlement were simultaneous in Britain, that the Anglo-Saxon migration groups were larger than their Frankish counterparts. Both migration flows certainly also included women and children as well as wa
rriors. These positively motivated expansions, carried mostly by smaller units over shorter distances, stand in marked contrast to the more spectacular, longer-distance moves made by larger concentrations of population whose motivations were much more mixed.19

  Brave New Worlds

  All of these migrants represented minorities in the new kingdoms they created. In those generated by longer-distance migration, the minorities were tiny. The Ostrogoths weighed in at some tens of thousands, but no more than a probably generous maximum of one hundred thousand, though this might be an underestimate if slave numbers were substantial.20 The population of late Roman Italy is normally reckoned at a few million. If we estimate it for the sake of argument at five million, then the immigrant Ostrogoths would have amounted to no more than 2 per cent of the total. You could fiddle endlessly with these figures, but the basic ratio won’t change very much. Ostrogothic immigration generated only a fractional increase in the total population of post-Roman Italy. The same was true of the Vandal–Alan coalition and the Burgundians, both of whom (the latter certainly) seem to have been second-rank powers compared to the Ostrogothic kingdom. From the perspective of the Italian, North African and Gallic populations, these migrations would have been experienced not even as elite replacements but as partial elite replacements. Within the kingdoms the newcomers created, many of the indigenous landowners of Roman descent remained in place, along with much of their Roman culture and even some Roman governmental institutions. The Visigothic kingdom of Gaul and Spain also falls into this category, although its origins were different. On being settled initially in just the Garonne valley in 418, the Goths must have represented a greater addition to the local population, even if still a minority. Within the larger kingdom created by Euric’s campaigns, which stretched from the Loire to Gibraltar, Gothic immigrants probably represented an even smaller minority than the Amal-led Goths in Italy.

  The Franks in northern, especially north-eastern, Gaul, and especially the Anglo-Saxons in what became England, represented a bigger influx of population in percentage terms, though they were still very much a minority. Even so, it is hard to get their numbers much above 10 per cent of the total population in the areas affected, and, in some places it may have been considerably less. Here, the new landowning elites that had emerged by the second half of the sixth century, and especially in Gaul, may again have included some genetic descendants of the old indigenous Gallo-Roman and Romano-British populations. But this must not obscure the point that north-eastern Gaul and lowland Britain represent an entirely different case from Italy, North Africa, Spain and the rest of Gaul. In the old Roman north-west, the elite class and its cultural norms were completely remade between 400 and 600 AD, and the landed assets upon which elite status was based completely reallocated, as the old villa estates were broken up into different-sized parcels. If the migrations involving Goths, Vandals and Burgundians were experienced at the receiving end as no more than comparatively unrevolutionary partial elite replacements, those of the Franks in north-eastern Gaul and the Anglo-Saxons in lowland Britain represented mass migration with profound social, political and cultural consequences for the areas concerned. If we take these different types of case together with the complete elite replacement represented by the Norman Conquest, there are actually three types of situation that need to be differentiated: partial elite replacement, elite replacement that doesn’t disturb major socioeconomic structures, and the mass migration of Franks and Anglo-Saxons.

  But even if discussion were to be confined just to the cases of partial elite replacement, there remain levels on which the population movements of the late fourth to the early sixth century would still have to be considered, both individually and in aggregate, as mass migrations. To start with, there is the fact that, between them, the migrants destroyed the long-standing Roman imperial edifice, or at least the western half of it. The Empire was always hampered by its economic, political and administrative limitations, but there is not the slightest evidence that it would have ceased to exist in the fifth century without the new centrifugal forces generated by the arrival within its borders of large, armed immigrant groups. Collectively the immigrants must be ‘credited’, if that is the correct word, with administering a huge political shock to the Roman world, or at least to the central state, even if some Roman landowners and even some more local and regional Roman institutions were left intact after the state’s collapse. This state was a powerful organism that had shaped life within its borders for five hundred years on every level from culture and religion to law and landownership. The point is easily overlooked, because the Roman Empire was large and ramshackle, operating with frankly pathetic communications and bureaucratic technologies, which meant that it could not run its localities with anything approaching close day-to-day control. Nevertheless, its structures had long set the macro conditions within which particular social, economic and cultural patterns evolved. A comprehensive listing would require another book, but the existence of the Pax Romana and state-generated transport systems dictated particular patterns of economic interaction, its legal structures defined property ownership and hence social status, its career structures – demanding a sophisticated elite literacy – underpinned an entire education system, and so on. Even religious institutions were largely dictated by the state. As Christianity matured into a mass religion in the fourth and fifth centuries, its authority structures became closely intertwined with those of the Empire. Given all this, the consequences of imperial extinction could not but be profound, setting local society and culture in western Europe off in quite new directions, some of which we will explore briefly in the second half of this chapter.21 In the qualitative sense used in migration studies, the overall impact of Hunnic-era migration qualifies in all respects as mass migration.

  The impact of the Franks in northern Gaul and of the Anglo-Saxons in lowland Britain also counts as mass migration, in the same qualitative terms but at a more local level. Reallocating the landed resources of these regions from scratch among intrusive elites caused many more cultural, economic and political transformations, which again qualify for the ‘mass’ label. Within the other successor kingdoms which saw only partial elite replacement, the local shock experienced was much less, but there was still some transfer of economic assets. Scholars of earlier generations were pretty much unanimous that this took the form – as in Britain and northern Gaul – of landed property passing from its previous Roman owners to at least some of the immigrants. This is another dimension of the barbarian issue that has come in for substantial revision in the last scholarly generation, again as part of the general tendency to downplay the significance of the fall of the Roman west. In this case, the revisionist view argues that, initially at least, incoming barbarians were rewarded not with landed estates taken from their former Roman owners, but with a share of the tax revenues raised from these estates – a type of exchange that would have generated much less friction than actual land seizures.22 There is no space here for a full discussion of the technical evidence relating to this issue, but in my view the revisionist case remains substantially unproven. There were important local variations, and adjusting the tax regime was probably one measure that sometimes came into play. Nonetheless, I am confident that a transfer of landed assets – as the most recent restatement of the revisionist position partly concedes – was at the heart of paying off the newly dominant immigrant element in all the major successor kingdoms.

  That is not to say that every member of each immigrant group received the same degree of remuneration, or even, necessarily anything at all. Doomsday Book, to draw again on the better-documented Norman analogy, shows that William’s more important supporters received much bigger payouts than their lesser peers, and that there were many rank-and-file Norman soldiers who received no land at all. As this parallel emphasizes, in general terms immigrants will surely have been rewarded in proportion to their pre-existing importance within their groups. But even if the big rewards were given at
first only to the most aristocratic among them, they will in turn have had their own chief supporters to reward (just as the tenants-in-chief did after 1066). Among the immigrant groups of the late fourth and the fifth century direct landed rewards from the king may well not have gone further down the social scale than leading members of the higher-grade (free?) warrior class, though its lesser members and even some or all of the lower-status warriors are likely, on the Norman Conquest model, to have received something from the higher-status warriors to whom they were attached.23 The detailed evidence also suggests that land transfers within the kingdoms created by partial elite transfer were restricted to specific localities.

  While the Vandal–Alan takeover of the richest provinces of Roman North Africa (Proconsularis, Byzacena and Numidia) represented a substantial shock for the political systems of the entire region, any deeper socioeconomic shock was geographically much more limited. After the conquest, Geiseric expelled some of the larger Roman landowners of the region and appropriated their estates to reward his important followers. This disruption at the local level, however, was confined to the province of Proconsularis. In Byzacena and Numidia, indigenous landowners seem to have been left in place; and any shock there was, limited to the operation of the new kingdom’s central political systems. Much of the land in Proconsularis was owned by absentee Italo-Roman senatorial families, so this was probably a good place to find land at minimum cost in terms of political hostility provoked. Strategically, too, Proconsularis faced towards Sicily and Italy, from where any significant Roman counterattack was likely to come.24