Peace in our Time?
The idea, in the light of such observations, that the transition of the Roman west from unitary Empire to multiple successor states was a largely peaceful process, for one thing, will not stand comparison with the evidence. Its initial premise that the outside groups who eventually founded the successor states had originally been invited across the frontier rests on the flimsiest of bases. There is no shred of evidence that any Roman official invited in Radagaisus’ Goths, the Rhine invaders (Vandals, Alans and Sueves), the Burgundians, or Uldin the Hun. In other words, every intrusive outsider involved in the crisis of c.405–8 was an uninvited guest, and all were resisted with might and main. The same is true of all the smaller components of the earlier frontier crisis of c.375–80: the Taifali, Farnobius’ Goths, the Sarmatians, and the Huns and Alans who allied with the rebellious Goths in the autumn of 377. And similarly with one of the two major Gothic groups who crossed the Danube in late summer or early autumn 376, the Greuthungi of Alatheus and Saphrax. They were originally excluded by force, but took advantage of an opportunity presented by increasing tension between the Roman state and the Gothic Tervingi to get across the river.
The only outsiders in this entire saga that actually crossed into the Empire with imperial permission were the Gothic Tervingi, and even here it is very likely that Valens had no real choice. The Emperor was fully committed to war with Persia in the summer of 376 when the Goths arrived on the Danube and asked for asylum. You’d have to think him a complete idiot to suppose that he would have been happy to see one major frontier go up in flames when he was already at war – with most of his army – on the other. As one source reports, the decision to admit the Tervingi was taken only after rancorous debate, and it looks, in the circumstances, like damage limitation. The emperor did not have enough troops to hope to exclude both the Tervingi and the Greuthungi, and was looking to divide and rule by letting one in and excluding the other. The point is confirmed by the various contingency plans that were put in place to neutralize any military threat the Tervingi might pose, especially the strategic control of food supplies and orders to attack the Tervingi’s leadership in case of trouble. It is true that in the fourth century (and before) emperors had periodically used contingents from their Gothic and other client kingdoms in their wars, even civil ones, but this provided no motive for admitting large groups of armed men permanently on to Roman soil – a much more dangerous proposition than recruiting armed bands from across the frontier and sending them home again once a campaign had finished.4
If it is demonstrably not true that the barbarians who made their way across the imperial frontier in the late fourth and fifth centuries did so at Roman invitation, a more sophisticated version of the idea has been proposed for the crisis of 405–8. This argues that the Empire issued a kind of implicit invitation by loosening its hold on the relevant frontier regions. It’s a bit like the argument heard at the time of the Falklands War in the early 1980s, in which the British decision to scrap the minesweeper Endeavour on financial grounds was read by the Galtieri junta as a sign that Britain would not resist an Argentine takeover. The ship had previously spent its time flying the flag in the South Atlantic. Applying this kind of analogy to 405–8 makes for a much more possible and interesting argument, but not in the end a persuasive one. In particular, the precise triggers of barbarian invasion are argued to have been the withdrawal of Roman military forces from the frontier region of northern Gaul, and the end of, or a substantial reduction in, subsidies paid to its frontier clients. The problem with this, however, is that the invasions of 405–8 were not for the most part mounted by those living on the immediate frontier, the prime recipients of such subsidies, but by other entities from outside the frontier region – sometimes, like the Alans, from far beyond it. There were also enough Roman forces in Britain and northern Gaul to propel the usurper Constantine III to within a cat’s whisker of controlling the entire western Empire in the autumn/winter of 409/410, and, in any case, the first attack (that of Radagaisus) didn’t target the supposedly semi-evacuated area. In short, there is no reason to think that the unprecedented pulses of barbarian intrusion had anything to do with a Roman invitation, explicit or implicit. The outsiders moved on to Roman soil in acts of violent self-assertion.5
What followed on from the original invasions was nothing very different. The hundred years separating the arrival of the Goths on the Danube in 376 from the deposition of Romulus Augustulus in 476 witnessed the working-out of many different political processes, within which there was certainly no underlying aim of bringing down the Empire. But all of these processes involved periodic violent confrontation, often substantial, between the intruders and the Roman state. Looking at it from the immigrants’ perspective, the politics of this century had two main stages. The first consisted of putting up a good enough fight to prevent the Roman authorities from destroying your group’s independence at the initial point of contact. The Tervingi and Greuthungi managed this between 376 and 382. Military might and the capacity to run away to North Africa were central to the ability of the groups who crossed the Rhine in 406 to survive their initial clashes with Roman (and Visigothic) forces in Spain. The Burgundians, on the other hand, were resettled further into Roman territory in the 430s by consent, it seems, but only after they had been devastated by the Huns, and the Roman general Aetius seems to have had a role in sponsoring those attacks.
If these groups managed to survive their first encounters with Roman might, many others did not. A number of Gothic subgroups were destroyed piecemeal between 376 and 382, Radagaisus’ force was bodily dismantled in 405 and many of its members sold into slavery, though some of the survivors later rejoined Alaric. The Rhine invaders suffered such heavy casualties between 416 and 418, likewise, that, as we saw earlier, three previously separate groups – Hasding Vandals, Siling Vandals and Alans – combined into one. However you look at it, surviving an initial encounter with the Roman state was no cakewalk. On my count, between 376 and their eventual establishment in Gaul in 418, the Goths that combined to create the Visigoths had between them fought eleven major campaigns and a host of smaller ones.6
This overall level of violence was central to two specific features of the migratory activity characteristic of this first stage. On the one hand, it helps explain why immigrant groups tended to engage in repeat migration. Continued movement was part of a survival strategy as they either manoeuvred for a compromise with the Roman state (the moves of Alaric’s Goths, for instance, from the Balkans via Italy to Gaul), or looked for safer and more prosperous venues from which to continue to defy it (the Vandal coalition’s move to North Africa). Second, it is impossible to explain why, without continued large-scale conflict, so many different immigrant groups came to operate together in a smaller number of larger confederations. The new political units formed on Roman soil – the Visigoths, the Vandal–Alan coalition, the Ostrogoths – all had in common the fact that they represented larger units that were better able to confront the military power of the Roman state, both to ensure their members’ survival and to extract from it more advantageous terms.7
More violence was central to stage two of migrant political activity: maximizing their position once they had ensured initial survival. These two stages tended to elide into one another, since not even the first Gothic immigrants of 376 entered the Empire without a range of ambitions that went beyond mere survival, but the second is distinct enough to be worth identifying. It can be characterized as the emergence of a framework of Roman–barbarian diplomatic relations, which had moved beyond any possibility of the particular immigrant group’s destruction. In the case of Alaric’s Visigoths, this stage was reached somewhere between 395 and 418, and is highly visible in the nature of the group’s subsequent diplomatic contacts with the Roman state. From 418, diplomacy focused only on how much territory the Visigoths were going to dominate, and on what terms: not, any longer, on whether their existence was going to be tolerated or not.
Even so, stage two was still marked by repeated military conflict: first in southern Gaul – where the regional capital at Arles provided an attractive target for the Goths in the 420s and 430s – and then more widely between the Loire and Gibraltar in the late 460s and 470s, when the Goths under Euric (467–84) founded a huge and independent kingdom. The Vandal–Alan coalition, by contrast, only began to reach the second stage from the mid-440s, when the west Roman state was forced to acknowledge its North African conquests, and, in fact, never enjoyed it as securely as the Visigoths. The death throes of the western Empire involved two serious attempts to reconquer the Vandal kingdom in 461 and 468. The Franks and Anglo-Saxons, on the other hand, never had to confront the Roman state head-on, and so in a sense moved direct to stage two. Nonetheless, they still pursued their ambitions by a violence-fuelled mixture of conquest and expropriation.8
Switching the perspective now to that of the Roman state, the connection between immigrant violence and the collapse of the western Empire could not be more direct. In simple terms, Rome taxed a comparatively developed agricultural economy in order to fund its armies and other structures. There were other sectors to the economy, but no one thinks that agriculture constituted anything less than 80 per cent of Gross Imperial Product, and many scholars would put that figure higher. In this context, the activities of the immigrants had direct effects on imperial tax revenues, and in so doing materially diminished the state’s capacity to survive. Any loss of territory to an immigrant group, such as the Spanish provinces to the Rhine invaders in the 410s, meant that the area concerned was now no longer contributing to central imperial coffers. Additionally, provinces caught up in any fighting, even if not conquered outright, were also much less able to pay their taxes. Nearly a decade after they had been occupied for only two years by Alaric’s Goths, the provinces around the city of Rome were still being assessed at only one-seventh of the normal tax rate. A similar rebate was also granted to two North African provinces which were not part of the Vandal–Alan kingdom of the 440s, but had been occupied by them for three years in the mid-430s. A six-sevenths reduction was perhaps generally granted, therefore, to provinces that had been heavily fought over.9
Once you start adding up the tally of lost and damaged provinces with the western Empire’s landed tax base in mind, the extent of the problem posed by immigrants quickly comes into focus. As early as 420, Britain had been definitively lost to central Roman control, along with the Garonne valley granted to the Visigoths. In addition, most of Spain had been taken or fought over by the Rhine invaders, and much of central and southern Italy damaged in the course of the Visigoths’ stay there between 408 and 410. The reduction in tax revenues caused by all these losses shows up beautifully in a late Roman resiger of military and civilian officals called the Notitia Dignitatum, which includes a listing of the western Empire’s armies dating to the early 420s. By this point, about half of the field army regiments, as constituted in 395, had been destroyed in the intervening quarter-century. But over half of the replacement units now incorporated into that army – 62 out of 97 – were simply old garrison troops upgraded on paper to field army status. Not only had field army losses not been replaced with troops of top quality, but neither is there any sign that the upgraded garrison forces had been replaced at all. Quality and quantity had both declined drastically as a direct effect of the erosion of the Empire’s tax base.10
Worse was to follow. By 445, the western Empire’s richest provinces – Numidia, Byzacena, and Proconsularis in North Africa – had succumbed to the Vandals, and part of Pannonia (modern Hungary) to the Huns, while Burgundians and some other Alans had been granted smaller areas in Gaul in the mid-430s. At this point, something close to 50 per cent of the western Empire’s tax base had been eroded, and the money was running out. Not surprisingly, this is the era in which western legislation both complained about the unwillingness of landowners to pay their taxes and attempted to claw back existing tax breaks. Any disinclination on the part of landowners to pay up is obviously an important phenomenon, especially since there is good reason to suppose that normal tax rates were having to increase at this time. Furthermore, new taxes were being invented. But to argue from this that the unwillingness of the rich to pay their taxes was a central cause of western imperial collapse, as has sometimes been done, is to put the cart before the horse. Tax privileges for the rich and well connected had always been part of imperial politics: enriching your friends was one reason why they backed you to win power. The phenomenon assumed an unwonted importance in the 440s only because so many provinces had already been lost to immigrants, or so damaged by warfare, that the western Empire’s revenues had shrunk to dangerously low levels.11
This aggregate loss in the state’s military and political effectiveness in turn contributed to a new strategic situation that allowed the immigrants further to expand the areas under their control, and dramatically so from the mid-460s. By that date the western army, bled dry by declining tax revenues, was a shadow of its former self and unable to confront with any chance of success the Visigoths, Vandals and others, particularly the Franks, who had just started or were finishing carving out power bases on former western territory. Looked at in terms of its effects upon tax revenues, and hence upon the Empire’s military establishment, there is no mistaking the direct line of cause and effect that runs from the immigration of armed outsiders to the collapse of the Roman west.
Against this backdrop, the increasing tendency of local Roman aristocrats to do deals with the various immigrants as the fifth century progressed can only be considered, like aristocratic unwillingness to pay high taxes, a very secondary phenomenon in the story of Roman collapse. Again, it is important to put these deals in context. The local aristocrats involved in them were all essentially landowners, whose estates, the fundamental source of their wealth, were for the most part situated in one locality. These physical assets could not be moved. So if that locality started to fall within the expanding sphere of influence of one of the immigrant groups, the relevant landowners had little choice. They had either to come to some accommodation with the immigrants’ leadership, if they could, or risk losing the land that was the source of all their wealth and status. Such accommodations were not automatic. In lowland Britain, as we have seen, the old Roman landowning class completely failed to survive the Anglo-Saxon takeover.12
Attempts to make the end of the western Empire into a largely peaceful process, carried forward by the withdrawal of the local elite from continued participation in central state structures, are unconvincing. On the contrary, all the various political processes of the fifth century were implemented by violence. Those elites were caught in the middle, with little choice but to make accommodations with the new powers in their lands, whether they wanted to or not, and if they could. A key distinction sometimes missed here is that between the central Roman state and local Roman landowner. Looking at just the latter, it is possible to document many stories of accommodation. These occurred, however, only after, and because, immigrant groups had fought their way across the frontier, and so stripped the western Empire of its tax base that it no longer had sufficient revenues to keep worthwhile armies in the field, leaving provincial landowners completely exposed.
Know Your Barbarians
When it comes to the immigrants of the late fourth and fifth centuries, there is again real substance to some of the revisionist arguments. Most of these groups were new political entities, not ‘peoples’. Ostrogoths and Visigoths, the Franks of Clovis, the Vandal and Alan alliance, and the Sueves of Spain: all were new entities forged on the march. A new political order was created among Anglo-Saxons, likewise, during their takeover of Britain. Of all the kingdom-forming groups who established successor states to the western Roman Empire, it is only the Burgundians for whom we lack explicit evidence of a major sociopolitical reconfiguration on the move, and even this may be due to a lack of information rather than any smooth continuity in their fifth-century history, whi
ch was pretty chequered.13
But if the immigrant groups weren’t ‘peoples’, neither does it fit the evidence to take an equally simple, if opposite, point of view and write them off as small-scale will-o’-the-wisp entities of little historical significance. Many were substantial. The few plausible figures we have, confirmed by their capacity to stand up to major Roman field armies, all suggest that the largest groups were able to put into the field forces numbering over ten thousand fighting men and sometimes over twenty thousand, especially after the amalgamation processes of the fifth century had run their course. The group identities operating within these large assemblages were not as straightforward as old nationalist orthodoxies imagined. Not even all the fighting men enjoyed the same status. At least in the larger alliances, there were two distinct status groups among warriors, and quite probably a third, of non-militarized slaves, besides. How many slaves there might have been is impossible to know, but we can’t just assume that they were few in number. Some of the kingdom-forming groups even crossed major cultural boundaries, the long-time alliance of Germanic Vandals and originally Iranian-speaking nomadic Alans being the classic case in point. What exactly happened when Vandal met Alan in the Middle Danube in the run-up to 31 December 406 is extraordinary to contemplate.14
But to conclude from these undoubted truths that the new group identities meant little is mistaken. Full participation was not allowed to all group members, as the existence of lower-status warriors and slaves within the new group identities makes clear. Neither of these lesser-status groups had as much invested in the identity of the group as its higher-status warriors. But neither was full participation the preserve of just a few individuals. Royal families came and went much too easily for us to characterize group identity as short-term loyalty to a particular dynasty. Even after deposing their last Amal ruler, the Ostrogoths retained their identity. I would argue that the prime carriers of, and beneficiaries from, the group identities being negotiated and renegotiated during this period were precisely the higher-status warrior groups. Some indications suggest that these may have comprised something like a fifth to a third of all armed males. And even if subject to periodic renegotiation – essentially political, perhaps, rather than cultural – nothing suggests that the kinds of group identities these men constructed were easy to destroy. Among the larger groups, the Ostrogoths did not, as has recently been argued, easily fade away into the Italian landscape after 493; while, amongst the smaller, Heruli and Rugi both showed considerable capacity to survive, in their different ways, even major defeats. Though not ‘peoples’ in the classic sense of the word, then, the immigrant groups were substantial not only in size but in structural resilience. Here again, the degree of violence characteristic of the era had an important role to play.15
Empires and Barbarians Page 45