Empires and Barbarians

by Peter Heather

  Recitach’s assassination thus completed an astonishing process of amalgamation. Theoderic’s uncle Valamer had probably been the first member of the family to achieve an unusual pre-eminence by killing, subduing and forcing out rival Gothic warband leaders in order to unite the Amal-led Goths: manoeuvrings that occurred either in Ukraine before the Goths’ move to Pannonia, or in the Middle Danube after Attila’s death (if these Goths were already established there). None of these warbands can have numbered much more than a thousand fighting men, and perhaps even only several hundreds. Within two lifetimes, therefore, uncle and nephew had taken the Amal line from one among a set of warband leaders to pre-eminent Gothic kings commanding in excess of twenty thousand warriors. It was this much larger force, complete with women, children and wagon train and amounting to between fifty and a hundred thousand souls, that took the road for Italy in the autumn of 488.

  There’s more you’d like to know, of course, but for the mid-first millennium this is pretty decent evidence. It also gives us some parameters for considering the other forces that came and went from the Middle Danube as the Hunnic Empire rose and fell, and it’s instantly clear that none of the other population groups on the move in this period was quite as big as this truly monstrous force. No source gives us figures for any of those smaller groups of former Hunnic manpower that entered the Roman Empire in the 460s – the forces of Hormidac, Bigelis, and of the two surviving sons of Attila. But none could establish the kind of independent position enjoyed by Theoderic’s Goths, and many ended up scattered in small clusters along the Danube frontier. It is hard to envisage that any could have fielded more than a thousand or two fighting men, and most perhaps mustered only a few hundreds.65

  Somewhat larger, though still nowhere near as big as Theoderic’s force, were the population clusters set in motion by the defeats of the Heruli and Rugi. The one other plausible-looking figure we have from the events that followed the death of Attila is from as late as 549. When the Herule allies of the Gepids and of Byzantium faced each other in battle that year, the two contingents numbered, respectively, fifteen hundred and three thousand men. This postdated a second split among the group, the first of which, you will recall, had sent an unspecified number of Heruli spinning off to Scandinavia. It also seems unlikely that either of the remaining concentrations of Heruli left in the Danube area would have been willing to commit its entire military manpower to war on someone else’s behalf. Before the splits occurred, and before their heavy defeat at the hands of the Lombards, therefore, the Heruli may have been able to field somewhere between five and ten thousand warriors, making them just a touch less powerful, perhaps, than the Amal-led Goths before Theoderic recruited the Thracian Goths into his following. We have no figures at all for the Rugi, but the fact that they could be defeated so thoroughly by Odovacar indicates that they amounted to no more than a medium-rank power in the Danubian scheme of things, so again perhaps a force of a similar size to or slightly smaller than the Heruli.66

  The most difficult to envisage of all the comings and goings during this era are those of the Lombards. That Lombard power eventually became dominant in the Middle Danube is clear enough, but the historical process behind this development is opaque. Late Lombard sources report that the seizure of Rugiland and the subsequent occupation of Pannonia, not to mention the earlier moves that had brought them that far from the mouth of the Elbe, were all invasions led by individual kings – the invasion hypothesis trundled out once more. On the other hand, all the contemporary evidence suggests that Lombard royal authority was not a very powerful phenomenon. After the move to Italy, second-rank leaders murdered the king and operated without central royal authority for a decade. It is quite possible, therefore, that independent initiatives on the part of intermediate leaders played an important part in the action, particularly in its earlier stages. Like Jordanes’ account of third-century Gothic expansion, later Lombard accounts have surely become infected with a migration topos that recasts the action in the form of one king, one people, one move.67

  On the other hand, migrant Lombards were never moving into a complete power vacuum as they came south down the Elbe, and by the time they got to the lands of the Heruli they were taking on a not inconsiderable power in head-to-head confrontation. Lombard expansion into the Middle Danubian region may well have been analogous, therefore, to third-century Germanic expansion towards the Black Sea (Chapter 3). While some of the action was carried forward by separate groups, some or many of which may have been small, especially at the beginning, the migration flow also had the capacity to generate larger groups at crunch moments to fight major battles. It looks like another example, in other words, of the classic pattern widely observable in groups from third-century Goths to ninth-century Scandinavians to nineteenth-century Boers, where the successes of initial intruders into a landscape encourage others, and eventually higher-status leaders enter the fray with larger followings. The lack of historical sources means that we have no indication of the overall numbers involved in these moves, or even whether they were primarily all-male warbands or groups encompassing women and children as well. By the time of the move to Italy in the 560s, whole families were certainly involved, and, since at least from the defeat of the Heruli in 508 large military forces were being assembled, the presumption must again be that militarized manpower beyond the scale of that available in specialist warrior retinues was required. If so, mixed social groups will have played a substantial part in the action in all but the very earliest phases of Lombard expansion.

  The archaeological evidence relevant to Lombard migration is not much more informative. The characteristic funerary ritual in Bohemia by the late Roman period was inhumation. In the late fourth and earlier fifth centuries, however, some cremation cemeteries started to appear there which bear strong similarities to those found further north where the Lombards originated (the northern Elbe, northern Harz, Altmark and Mecklenburg regions). These intrusive funerary rites could be the result of some indigenous Bohemians deciding to cremate their dead, but given that Lombards had certainly made their way to the Middle Danube in some numbers by the end of the fifth century, the cemeteries probably provide us with an indication of their route68 – hardly overwhelming evidence, but, as we have seen, archaeological finds will almost never provide entirely unambiguous evidence of migration. The material cultures of the populations of the northern Elbe were too similar to one another for shorter-distance population flows within the region to show up with any clarity, so that it is not possible to say where, precisely, the first northern intruders into Bohemia came from. And, in any case, the migrating groups may well have recruited from right across the region.

  The archaeological evidence from the Middle Danube after the Lombards took power there, likewise, is in one sense clear enough: in the course of the sixth century, a coherent set of well-dated remains centred on old Roman Pannonia spread over those territories where historical sources report Lombard domination. These without doubt reflect the Lombard kingdom. On the other hand, there is nothing very distinctive about them compared with other Middle Danubian remains, especially those stemming from areas which, the historical sources tell us, were dominated by Gepids. This does not mean that the differences between the Lombard kingdom and its Gepid rival were insignificant. What the resemblance really shows is that sixth-century Lombard material culture followed a similar trajectory to that of the Huns in the fifth. Over time, it lost its original distinctiveness and firmly adapted itself to Middle Danubian norms, which reinforces the idea, perhaps, that the Huns of Attila’s time are archaeologically invisible because they too had adopted new material cultural norms. In the case of the Lombards, their original cremation rite was replaced with a new habit of burying unburnt bodies in cemeteries laid out approximately in rows, oriented broadly east–west (German: Reihengräber). Lombard women wore their clothes – at least those they were buried in – in the same Danubian fashion as everyone else, with a pair of brooches one on e
ach shoulder. Handmade ceramics with idiosyncratic designs of the kind marking out different northern Elbe groups in the early Roman period made way for wheel-made pottery of a fairly uniform Middle Danubian design. The most that can be argued, and this is in line with modern ethnographic parallels, is that particular fibula designs became symbolic of Lombard and Gepid allegiance, since two entirely different designs are found, with their distribution patterns confined to each half of the Middle Danubian plain.69

  A range of migratory phenomena can be seen intertwined in the rise and fall of the Hunnic Empire. Some of the moves were made by large, concentrated groups, notably those of the Amal-led Goths. In 473 several tens of thousands of people left Hungary for the Balkans, possibly the same group that had moved to Hungary from Ukraine about twenty years before; and in 488 an even larger group, close to a hundred thousand souls if you add in the Thracian Goths and the refugee Rugi, set off from the Balkans for Italy. Other moves were made by smaller population groups, refugees from the military defeats that had dismantled old hegemonies, notably the Huns and Sciri in the 460s, the Rugi in the 480s and the various groups of Heruli after 508. And to complete the picture, the period also saw one predatory flow of migration of the kind we have met before, in the form of the Lombards.

  Even though the historical sources give us few decent figures, many of these movements of armed immigrants into and out of the Middle Danubian region represented mass migration at least in the qualitative terms used in comparative migration analysis. The overall ‘shock’ of Attila’s tribal gathering in the first half of the fifth century is visible archaeologically in the so-called Danubian style, and, in narrative terms, in the attacks the Huns launched into the Mediterranean using their unprecedented concentration of military manpower. New political and social relations were generated in the region under Hunnic domination, representing a further level of shock. The whole creaky structure relied on a flow of Roman gold, extracted by war and intimidation, to lubricate its operations. War and its profits kept the mass of the Huns’ armed subject groups in line via a potent mixture of intimidation and reward, and intense political and indeed cultural dislocation are visible in all of this.

  Much of the undocumented, or insufficiently documented, population displacement of the era of Hunnic collapse, likewise, amounted to mass migration in qualitative terms. Odovacar’s intervention came as a huge political shock for the Rugi, since it destroyed their kingdom and set survivors off on two forced treks, each of several hundred kilometres, first to join Theoderic in the Roman Balkans and then on, in his train, to Italy. The intervention of the Amal-led Goths had earlier had similar effects upon the Sciri. That all the Sciri and Rugi left the Middle Danube region following these defeats is unlikely, but their independence was extinguished and enough Sciri left for the army of Italy to contribute to a changing balance of forces there. Hence, in due course, Odovacar became the effective ruler of the first post-Roman successor state on Italian soil. The Lombards’ arrival in the Middle Danube, likewise, was a shock for the Heruli, who also saw their independence and their unity destroyed, and many of them felt forced to move on. In pretty much every case, then, though there are few figures worth a damn, we are dealing with groups possessing substantial military power whose migratory responses to the rise and fall of the Hunnic Empire generated substantial restructuring of the political systems operating not only in the Middle Danubian region itself, but also in adjacent and not so adjacent areas of the Balkans, the northern shores of the Black Sea, and even within Italy itself. The detailed narrative evidence available to us thus broadly confirms the picture that emerged from the analysis of the operation of group identities in the Hunnic Empire. The group labels we encounter in our sources belonged to functioning concentrations of human beings, some of them tens of thousands strong, whose lives were wrenched out of shape by the tumultuous events of the rise and fall of Hunnic domination in central Europe, and who often took to the road as a result.

  Several different types of migration can be observed, from concentrated mass pulses to more extended flows, but many clearly went far beyond the bounds circumscribed by wave-of-advance or elite-transfer models. Though not all are covered in the same detail as the Amal-led Gothic diaspora, it is clear that many of these moves were hugely traumatic, whether measured in terms of distance, violence or loss of political independence. Viewed from the migrants’ perspective, much of the action was ‘mass’ in a more absolute sense as well. For many of the migrant groups, as we have seen, there is either good (Amal-led Goths, Rugi), or reasonable (Heruli, Huns, Lombards), evidence that they comprised men, women and children. In some cases, such as the Amal-led Goths, these groups numbered several tens of thousands of people, and in many cases, as in 376 and 405/6, they moved in compact masses.70 None of the participating groups was a ‘people’ in the old sense of the word, and there is much evidence that the process of migration, as any reading of the comparative literature would lead us to expect, caused splits among the migrants, who were faced with enormously difficult decisions. Some of the Amal-led Goths refused to move south into the Roman Balkans in 473, for instance, preferring the leadership of Thiudimer’s younger brother Vidimer. They moved west instead, where they were eventually absorbed into the Visigothic kingdom. Not all the Goths in the Balkans, likewise, were ready to move with Theoderic to Italy in 489. Some preferred a Byzantine allegiance. And the repeated splitting of the Heruli is eloquent testimony to just how difficult these decisions to move actually were, leading some to Scandinavia and others to subordination to the Gepids or east Rome, depending on the outcome of wars and the conditions offered by potential hosts.71 But despite all the problems with the evidence, the only reasonable conclusion to derive from the rise and fall of the Hunnic Empire is that the migratory phenomena outside the Roman Empire were just as substantial as those that characterized the crises of 376–80 and 405–8.

  Ways and Means

  The reasons why some of the migratory processes should have taken this form, so different from any encountered in the modern world, are similar to those that explain its appearance in other first-millennium contexts, and don’t need extensive discussion again. Take, for example, the two moves of the Amal-led Goths, first into the east Roman Balkans in 473, then on into Italy in 488/9. Both were underpinned by a substantially economic and hence voluntary motivation. The first was undertaken with the aim of supplanting the Thracian Goths as Constantinople’s favoured allies, in order to lay hands on the benefits they enjoyed. Amongst other perks, the Thracian Goths received subsidies measured in thousands of kilos of gold per annum, whereas those of the Amal-led Goths out in Pannonia amounted to just a few hundreds. In the move to Italy, likewise, Theoderic had it in mind to enrich himself and his followers at the expense of Odovacar and such Roman fiscal structures as remained in operation. Theoderic’s extant building works at Ravenna, and his many other known monuments besides, bear eloquent testimony to just how much disposable income continued to be delivered to Italy’s early-sixth-century ruler. He also recycled some of the tax income to invent salaried posts for his more important Gothic followers: a device surely designed to ensure their political support. Both of these strategies for economic advance were entirely dependent, however, on having sufficient military muscle to transform an existing political situation – to persuade the Emperor Leo to choose a new set of Gothic allies in the first instance, to defeat Odovacar’s army in the second. And certainly in the second case there was an extra political dimension, since relations between Theoderic the Amal and the Emperor Zeno had reached deadlock. Neither trusted the other, and a series of confrontations had shown that neither could easily rid himself of the other.72 In these migrations, economic and political motivations cannot easily be separated, and, to have any chance of success, Theoderic had to field a substantial army. As we have seen before, the number of specialist warriors that could be supported by the non-Roman European economy in this era was not sufficient for large-scale campaigning. Free
men and their families thus became integral to the enterprise.

  The play of motivations behind Lombard expansion looks very similar. As far as we can tell, their move into the Middle Danube was not made in response to any kind of threat, but inspired by the region’s attractions. The Middle Danube had long formed part of the inner periphery around the Roman Empire, and over the first four centuries AD had steadily built up levels of wealth and development far beyond anything to be found at the mouth of the Elbe. The apogee of Attila hugely accentuated this imbalance. The amount of gold stashed away in Middle Danubian burials of the Hunnic period is without precedent in the Germanic world. And this can only be a fraction of the total amount, much of which was stored, presumably, in the treasuries of the kings who now ruled in the region. Even if we lack explicit evidence, it’s much more than a guess that Lombard migration had in mind a share of this booty, still being reinforced by the smaller diplomatic subsidies that continued to be paid by Constantinople after Attila’s death. But acquiring any part of this wealth required, as usual, the application of main force to alter existing political configurations – in other words, the Heruli needed to be defeated. While Lombard expansion may have started with warband-size groupings seeping south, both the Lombards and other immigrants caught up in the flow had to reform themselves into a more cohesive group, at the latest by the time they left Rugiland, whence they proceeded to destroy the kingdom of the Heruli.73 Even where largely economic, and hence voluntary, these kinds of migration always had a political dimension. Did the migrants pack sufficient military punch to succeed in the enterprise they were about to undertake or did they not?