4

Crisis and Continuity, 400-550

On 25 February 484, Huneric, king of the Vandals and Alans, and ruler of the former Roman provinces of North Africa, issued a decree against the ‘Homousian’ (we would say Catholic) heresy of the Roman population of his kingdom. The Vandals were Arian Christians, and they regarded the beliefs of the Roman majority as sufficiently incorrect that they needed to be expunged. Huneric, accordingly, adapted the emperor Honorius’ law of 412 against the Donatists of Africa, which had been a major Catholic weapon in the days of Augustine, and used it against the Catholics themselves. Huneric was explicit about this:

It is well known that the casting back of evil counsels against those who give them is a feature of triumphant majesty and royal strength . . . It is necessary and very just to twist around against them what is shown to be contained in those very laws which happen to have been promulgated by the emperors of various times who, with them, had been led into error.

Huneric’s mode in this decree, and in the persecution it began (which seems to have quietened down after his death in December of the same year), was consistently playful: you did this yourselves; it is therefore right that it should be done back to you. Indeed, his whole preparation for it was a deliberate echoing of the 410s. Honorius in 410 had called for a conlatio, a formal disputation, between Donatist and Catholic bishops, which took place in Carthage in June 411; its acts largely survive, and they show a striking mixture of ceremonial power-plays, insult and argument, followed by a judgement against the Donatists - and then repression a year later. The Donatists must have known that they were probably being set up; and when in May 483 Huneric called the Catholic bishops to a similar debate in Carthage for the February of the following year, the latter certainly knew what was coming. Both the Donatists in 411 and the Catholics in 484 tried to pre-empt discussion by presenting a manifesto ; but Huneric, if we believe the account of his fervent opponent Victor of Vita, had already prepared his decree, thus cutting short the debate. If this is true, it was Huneric’s only deviation from his replay of the Honorian drama. Huneric was enjoying being a Roman emperor in persecuting mode, act by act; and the Catholics knew well what he was doing.

The Vandals in Africa represent a paradox, which is epitomized by this account. The modern use of their name shows the bad reputation they already had, expressed above all in Victor’s polemical account of their cruelty and oppression. Most contemporary accounts of the Vandals were indeed negative, from Possidius’ eyewitness account of their violent arrival in Africa in 429 to the eastern Roman historian Prokopios’ criticisms of their luxurious lifestyle at the moment of the Roman reconquest in 533-4. Under their most successful king, Huneric’s father Geiseric (428-77), who brought them from Spain to Numidia and then in 439 to Carthage and the African grain heartland, their ships (ex-grain ships, no doubt) raided Sicily, conquered Sardinia and sacked Rome in 455. Huneric was not the only king to persecute Catholics; Thrasamund (496-523) did the same in the 510s. Conversely, however, there is evidence to show that the Vandals thought they were being very Roman. Those we know about all spoke Latin. Huneric married Honorius’ great-niece, and had spent time in Italy. The Vandal administration seems to have been close to identical to the Roman provincial administration of Africa, and to have been staffed by Africans (at most they may have adopted a Vandal dress code); the currency was a creative adaptation of Roman models; the kings taxed as the Romans had; the Vandal élites accumulated great wealth as a result, which they spent in Roman ways, on luxurious town houses and churches, as both literary sources and archaeology tell us. Archaeology, indeed, implies little change in most aspects of African material culture across the Vandal century. And, of course, their religious persecution was entirely Roman. Other conquering Germanic peoples were also Arian, notably the Goths, as we have seen, but they saw their religion for the most part as marking out their own identity vis-à-vis their new Roman subjects, who could stay Catholic. Only the Vandals assumed that their version of Christianity should be the universal one, and that others should be uprooted, as the Romans themselves did: hence also the negative tone of contemporary accounts, which are all written by Catholics.

It is thus possible to turn the Vandals into a version of the Romans themselves. They could be seen as in effect a rogue army that seized power in a Roman province and ran it in a Roman way; although the Vandals had themselves never been imperial federate troops, they were very like them, and one would be hard put to it to identify any element in their political or social practice that had non-Roman roots. But we would be mistaken if we thought nothing changed when Geiseric marched into Carthage. There were two major differences. First, the Vandals ruled Africa as a military landowning aristocracy, who continued to see themselves as ethnically distinct. Roman armies which seized power before the fifth century were content to create their own emperor and retire to barracks with rich gifts; but the Vandals became a political élite, replacing and expropriating the largely absentee senatorial aristocracy (and some Roman landowners who lived in Africa too, though most of these survived). Secondly, the Vandals broke the Mediterranean infrastructure of the late empire; they took over the major grain and oil export province of the West, the source of most of the city of Rome’s food. The food had largely been supplied free, in tax; the Vandals were autonomous, however, and kept African produce for themselves - although they were prepared to sell it. The Carthage-Rome tax spine ended. The population of the city of Rome began to lessen precipitously after the mid-fifth century; in the next century it probably dropped more than 80 per cent. And a gaping hole appeared in the carefully balanced fiscal system of the western empire; the Romans faced a fiscal crisis, just when they needed to spend as much on troops as they possibly could. Not to foresee that Geiseric would take Carthage, notwithstanding a treaty agreed in 435, is arguably the main strategic error of the imperial government in the fifth century: the moment when the political break-up of the western empire first became a serious possibility. Hence the belated but intense efforts made to recapture Africa in 441, 460 and especially the large mobilization of 468, which failed disastrously, even though Vandal military strength was not, as far as can be seen, unusually great. Reconquest in 533-4 was easy in the end, but the western empire was gone by then. However Romanized the Vandals were, they were agents of major changes.

This is the key feature of the events of the fifth century, at least in the western empire. Over and over again, ‘barbarian’ armies occupied Roman provinces, which they ran in Roman ways; so nothing changed; but everything changed. In 400 the western and eastern Roman empires were twins, run by brothers (Honorius and Arcadius, the two sons of Theodosius I, ruling 395-423 and 395-408 respectively), with little structural difference between them, and, as we saw in Chapter 2, no fundamental internal weaknesses. In 500 the East was hardly changed (indeed, it was experiencing an economic boom), but the West was divided into half a dozen major sections, Vandal Africa, Visigothic Spain and south-west Gaul, Burgundian south-east Gaul, Frankish northern Gaul, Ostrogothic Italy (including the Alpine region), and a host of smaller autonomous units in Britain and in more marginal areas elsewhere. The larger western polities were all ruled in a Roman tradition, but they were more militarized, their fiscal structures were weaker, they had fewer economic interrelationships, and their internal economies were often simpler. A major change had taken place, without anyone particularly intending it. The purpose of this chapter is to investigate how - but not with hindsight. The events of the fifth century were not inevitable, and they were not perceived as such by the people who lived through them. No one saw the western empire as ‘falling’ in this period; the first writer specifically to date its end (to 476) was a Constantinople-based chronicler, Marcellinus comes, writing around 518. We shall look at those events in four chronological tranches, up to 425, up to 455, up to 500, and up to 550, so as to try to pin down what were the principal changes, but also stabilities, at each stage. We shall then deal with the issue of what these changes meant.

Neither Honorius nor Arcadius was any sort of political protagonist, nor in fact were their successors as emperors, and it was not until the 470s that effective rulers occupied supreme political positions again. Others ruled through them. In the West, the strong-man at the start of the fifth century was Stilicho, military commander (magister militum praesentalis) of the western armies since 394: a powerful dealer, which he needed to be. For the whole of his ascendancy he faced Alaric, king of the Goths (c. 391-410), in the latter’s attempts to establish a stable location for his people. Gothic groups had first come into the empire in 376, as we saw in Chapter 2; after their victory at Adrianople in 378, they were left alone in the 380s in Illyricum and Thrace, the modern Balkans. Alaric was the first Gothic leader to serve with his own followers in a Roman army, for Theodosius in 394. This military arrangement came unstuck by 396, however, and Alaric’s Goths (we call them the Visigoths, to avoid confusion with other Gothic groups, though they did not call themselves this) spent two decades trying to regain, by force, a recognized position in the empire. They attacked Greece, then moved north, and entered northern Italy in 401. Stilicho defeated them and drove them back into Illyricum in 402, but they returned in 408. Nor were they the only ‘barbarians’ in the empire by now; other groups, probably persuaded to take their chances across the border by the development of Hunnic power, came in during the same decade. In 405 an army led by Radagaisus, again largely Gothic, crossed the Alps into Italy from the north; Stilicho defeated and destroyed them near Florence in 406. Stilicho needed a larger army for all this than Italy possessed, especially as he himself also wanted to make Illyricum part of the western, not the eastern empire, and he pulled troops from the Rhine frontier to meet this need. This was probably a mistake, for it was followed by an invasion of central European tribes led by the Vandals, over the Rhine on New Year’s Eve 406, an irruption into western Gaul and then (in 409) into Spain which was almost unresisted; and also in 407 another invasion of Gaul, this time by a usurper, Constantine III (406-11), at the head of the army of Roman Britain. Faced with these multiple crises, whispering campaigns against Stilicho began, and after a mutiny he was executed in 408.

Stilicho was brought down by problems that were not entirely of his own making; the western leadership immediately after his death only made errors. Stilicho was half-Vandal in origin, and was regarded by some as too favourably disposed to ‘barbarians’; those who were in his Italian army were either massacred or fled to Alaric. Alaric was dominant in Italy in 408-10, but the Romans would not consistently make peace with him, even though he blockaded Rome three times. In the end he sacked Rome in 410, an event which shocked the Roman world much as 11 September 2001 shocked the United States, a huge, upsetting, symbolic blow to its self-confidence; but it was without other repercussions, and was only one step in the long Visigothic road to settlement. The Goths tried to go south to Africa, then went north into Gaul instead, under their new leader Athaulf (410-15); there they found, and contributed to, a still greater confusion, with as many as four rival emperors in 411, most of them the protégés of different ‘barbarian’ groups. Slowly, the legitimist Roman armies regrouped under a new magister militum, Constantius (411-21), who picked off the usurpers one by one and forced the ‘barbarian’ groups to come to terms. Athaulf’s Visigoths were, as Roman armies were, dependent on Mediterranean grain, and the Romans blockaded them into submission in 414-17; they ended up fighting on behalf of the Romans against the Vandals in Spain, who were partially destroyed in 417-18, until, in 418, they were finally settled around Toulouse. Constantius married Honorius’ sister Galla Placidia, who had previously been married to Athaulf, and he became co-emperor shortly before his death in 421. Military rivalries continued, but the crisis was quietening down. By 425, after a disputed succession, Honorius’ nephew, Constantius and Placidia’s young son Valentinian III, was western emperor (425-55), with his mother as regent.

The East faced less trauma in this period. The Balkans was a military district, and was always the most invaded part of the eastern empire; there were also Hunnic attacks on it, both before and after the Goths left. But Constantinople, on the edge of the Balkans, was well defended, and the wealth of the East was in the Levant and Egypt, a long way from the northern frontier. Above all, Sassanian Persia, Rome’s traditional enemy to the east, was at peace with the empire for almost the entire fifth century, probably because it faced its own threats elsewhere, which allowed the eastern empire a greater strategic security. Eastern politics were often fraught, sometimes violently so, as with the anti ‘barbarian’ hysteria in the capital which in 400 destroyed the magister militumGainas, and, soon, his rival Fravitta as well, a foretaste of Stilicho’s fate later in the decade. But from then onwards most of the political leaders of the East were not soldiers but civilians, ruling for Arcadius and his equally inactive son Theodosius II (408-50), and indeed empresses were particularly prominent in Constantinople, in this period Arcadius’ wife Eudoxia in 400-404 and Theodosius’ sister Pulcheria in the 410s-420s. Each of these, among other acts, brought down ambitious and uncompromising patriarchs of Constantinople, respectively John Chrysostom in 404 and Nestorios in 431. This in itself shows that the eastern empire was developing a different political style from the West: the patriarch of Constantinople, only established in 381, was already a protagonist in secular politics in a way that the pope in Rome would not be for another century. The fact that the western empire was run from Ravenna, not Rome, meant that Roman city politics were less central to it; the importance of church councils and doctrinal debate as a focus for unity and dissension was also greater in the East, giving bishops in general more of a political voice than they as yet had in the West. The church-state relationship would remain much more intimate in the East in the future, too, except, much later, during the Carolingian period in the West, as we shall see in Chapter 17.

In 425 the East was stable and had begun the long economic revival that would continue into the late sixth century or early seventh. But the West had achieved, after a decade of turmoil, a substantial stability as well. Most of the frontier was still manned by Roman troops. There were ‘barbarian’ groups settled in the empire, it is true, separate from the Roman military hierarchy, the Visigoths between Bordeaux and Toulouse and the remnants of the Vandal confederacy in western Spain, Suevi in the north and Hasding Vandals in the south; but all these had been defeated, and the Visigoths at least were in formal federate alliance with Rome. Only in the northern provinces of the West, north of the Loire, was the situation still unstable. The far northern border of Gaul was increasingly settled by Franks from just over the Rhine; in the north-west there were intermittent peasant revolts, of groups called Bagaudae, which began in the confusion of the 410s and continued into the 440s, presumably an exasperated reaction against continued taxation at times of military failure; and Britain had been abandoned by the Roman administration after 410. These areas were even more marginal for the West than the Balkans were for the East, however. Orosius, a Christian apologist writing in Spain in 417, could already use the cliché that ‘the barbarians, detesting their swords, turned to their ploughs and now cherish the Romans as comrades and friends’, and this did not seem a false vision in the next decade. In that same period, 413-25 to be exact, Augustine wrote his monumental City of God, initially in reaction to the Sack of Rome; it was neither a triumphant tract about Christian Roman victory (as was Orosius’ text) nor a polemic about the dangers facing Roman ill-doing. Augustine was, indeed, careful not to ascribe too much importance or longevity to the great Roman imperial experiment, for the heavenly city is separate from earthly political forms. But his book nonetheless presumes a considerable confidence in the imperial future. The world itself might end, of course, and, Augustine assumed, would indeed do so soon enough; but there is no hint here that an end to the empire was expected or feared by anybody.

Things shifted in the next generation, up to 455. In the East, politics stayed quiet, except for regular Hunnic attacks in the Balkans. This period was marked by the ambitious compilation of the current laws of the empire, the Theodosian Code, completed in 438; these were both western and eastern laws (many of them seem to have been collected in Africa), but they were compiled in Constantinople, and bore the eastern emperor’s name. It was also marked by two defining church councils, at Ephesos in 431 and at Chalcedon in 451, as we saw in Chapter 3, although their definitions were achieved at the expense of alienating large sectors of the Christian community of the Levant and Egypt, who found themselves stigmatized as Monophysite heretics. Pulcheria was a prominent operator behind the scenes in each of these councils. She had a relatively small role at court between them, especially in the 440s, but at Theodosius II’s death she created his successor Marcian (450-57), by marrying him, and she was again influential until her death in 453. Chalcedon, in particular, was a divisive moment; but the fact that the politics of the East hinged on these great ecclesiastical aggregations, rather than on war, is telling in itself.

The West saw more trouble. Military leaders fought over the young Valentinian, with Aetius, based in Gaul, winning out by 433. Aetius ruled the West as magister militum until 454, but his interests remained in Gaul. The responsibility for letting the Vandals move into Carthage essentially lies with him; he reacted, but ineffectively and too late. Aetius’ main concern was the Visigoths, whom he at least temporarily pacified in 439. Other ‘barbarian’ groups in Gaul were also persuaded to accept Roman military hegemony, including the Alans and the Burgundians, whom Aetius himself settled in, respectively, the lower Loire valley and the upper Rhone in 442-3. Gaul remained stable under Roman hegemony as a result of Aetius’ attentions, although it is undeniable that there were more autonomous groups settled there by Aetius’ death than earlier. Italy, too, the core of the West, was actually less menaced by invasion than in the early years of the century. But Africa had been lost, and Spain, too, after the Vandals left in 429, came largely under Suevic control in the 440s; Spain, though, as we have seen, was much less essential to the imperial infrastructure. It is in the 440s that we get our first indications in western legislation that standard taxation was insufficient to pay imperial troops, which heralded tax rises. The Bagaudae reappeared in northern Gaul, and now in north-east Spain as well, the part of the Iberian peninsula still under Roman control. Salvian of Marseille wrote a long hell-fire sermon called On the Governance of God in the 440s which ascribed Roman failures against the (obviously inferior) ‘barbarians’ to their own sins: notably, unjust and excessive taxation, public entertainment and sexual licence. This is the sort of thing extreme Christian preachers always said (and still say), and its detail cannot be taken too seriously; we could not conclude from this, for example, that the western provinces really were being destroyed by overtaxation, and it would be best to see Salvian’s writing as a proof of the continuing effectiveness of the fiscal system. But it is undoubtedly true that Salvian’s vision of the West now included the ‘barbarians’ as stable political players, alternatives to Roman rule, and the same was true of the Bagaudae (though the latter were in reality less stable, and disappear from our sources by 450; Aetius and his ‘barbarian’ allies had defeated them). Salvian thought that Romans often chose to be ruled by ‘barbarians’ in order to escape Roman state injustice. This was probably not common in the 440s, but the concept was possible to invoke; the historian Priskos in the East, when discussing the Huns, did so in the same period as well.

Aetius, in his campaigns against the Visigoths and others, relied quite substantially on the military support of the Huns. The latter had, by the 420s at the latest, largely settled just outside the empire in the middle Danube plain, what is now eastern Hungary, a good strategic point for attack both into the Balkans and the West. But they were not a full-scale danger until Attila (c. 435-53) and his brother Bleda both unified them and reinforced their military hegemony over other ‘barbarian’ groups, notably the Gepids and that section of the Goths we call Ostrogoths, around 440. The 440s marked serious Hunnic attacks in all directions, culminating in major invasions of Gaul in 451 and Italy in 452. The Huns were defeated in Gaul, however (Aetius used the Visigoths against them, as he had previously used the Huns against the Goths), and retreated from Italy, for less clear reasons; in 453 Attila died unexpectedly, and in 454/5 conflict among his sons and his subject peoples led to the rapid break-up of the Hunnic hegemony. The Huns were a terrifying because unfamiliar people, but as a direct military threat to the Romans they were a flash in the pan. The same is true for Attila’s construction of an alternative political focus to the capitals of the empire, which looked impressive at the time, but did not last much more than fifteen years. It could equally be argued that the Huns helped the Romans, not only by fighting for Aetius but also as a force for stability (and thus fewer population movements) beyond the frontier. But this did not outlast 454 either.

The Hunnic empire collapsed, but Aetius was already dead, assassinated by Valentinian III personally in 454, the latter himself killed as a direct result a year later. Aetius was seen by many later as (to quote Marcellinus comes) ‘the main salvation of the western empire’, largely because he was its last commander to convey an impression of military energy over a long period. His errors, especially in Africa, could be regarded as equally fatal. But the 450s still saw a certain level of stability in the West. It now contained half a dozen ‘barbarian’ polities, with all of which any Roman leader would have to deal, though still from a position of strength: all those polities operated by Roman rules, and cared enough about the empire to seek to influence its choice of rulers. This was shown in the crisis after Valentinian’s death, when Geiseric sacked Rome; Theoderic II of the Visigoths (453-66) persuaded Eparchius Avitus, a senator from the Auvergne in central Gaul and one of Aetius’ former generals, at that moment on an embassy to him, to claim the imperial office in 455. Avitus was no cipher, all the same. He did not last long, but there would still have been space for an energetic ruler in the West to maintain at least Aetius’ hegemony, and maybe even to regain that of Constantius, if he could get eastern logistical support (sometimes available), and if he was very lucky.

Imperial luck did not hold, however. The next two decades, into the next generation, are the period when the West finally broke into pieces. Avitus, clearly a Gaulish imperial candidate, had been defeated by the Italian army under Majorian and Ricimer, and the former became emperor (457-61). Majorian took the trouble to get both eastern recognition and the support of Avitus’ Gaulish clientele; he issued legislation which shows reforming aspirations, too. But, if he was energetic, he was certainly not lucky either, for Ricimer, his magister militum, organized a coup against him and had him killed. Ricimer then ruled until his death in 472, through a succession of mostly puppet emperors, although Anthemius (467-72), a military figure from the East, had a certain presence and autonomy until Ricimer fell out with him. It was Anthemius who organized, together with the eastern general Basiliscus (the eastern emperor Leo I’s brother-in-law), the great attack on the Vandals of 468, which was not only a failure but an extremely expensive one. After that, Ricimer concentrated on Italy, which he defended effectively, and left the rest of the empire largely to its own devices, although he maintained links with south-eastern Gaul through his son-in-law the Burgundian prince Gundobad, who succeeded Ricimer briefly as the imperial strong- man before leaving Italy to become Burgundian king (474-516). Ricimer is hard to assess through sources that are both hostile and sketchy, but there is no sign that he had political interests or ambitions which extended beyond Italy; he is a clear sign that imperial horizons were shrinking. After two more short-lived coups, Odovacer, the next effective military supremo in Italy (476-93), did not bother to appoint any emperor of the West, but instead got the Roman senate to petition the eastern emperor Zeno that only one emperor was by now needed; Odovacer then governed Italy in Zeno’s name, as patricius, patrician, a title used by both Aetius and Ricimer, although inside Italy Odovacer called himself rex, king.

The year 476 is the traditional date for the end of the western empire, at the overthrow in Italy of the last emperor, Romulus Augustulus, although 480 is an alternative, for Romulus’ predecessor Julius Nepos held out in Dalmatia until then. But Italy is actually the region of the western empire which lived through least change in the 470s, for Odovacer ruled much as Ricimer had, at the head of a regular army. Italy did not experience an invasion and conquest until 489-93, with the arrival of Theoderic the Amal and his Ostrogoths, and Theoderic (489-526) ruled in as Roman a way as possible, too. The end of the empire was experienced most directly in Gaul. The Visigothic king Euric (466-84) was the first major ruler of a ‘barbarian’ polity in Gaul - the second in the empire after Geiseric - to have a fully autonomous political practice, uninfluenced by any residual Roman loyalties. Between 471 and 476 he expanded his power east to the Rhone (and beyond, into Provence), north to the Loire, and south into Spain. The Goths had already been fighting in Spain since the late 450s (initially on behalf of the emperor Avitus), but Euric organized a fully fledged conquest there, which is ill-documented, but seems to have been complete (except for a Suevic enclave in the north-west) by the time of his death. By far the best documented of Euric’s conquests, though not the most important, was the Auvergne in 471-5, because the bishop of its central city, Clermont, was the Roman senator Sidonius Apollinaris. Sidonius, who was Avitus’ son-in-law, and had been a leading lay official for both Majorian and Anthemius, ended his political career besieged inside his home city, and we can see all the political changes of the 450s-470s through his eyes. A supporter of alliance with the Visigoths in the 450s, by the late 460s Sidonius had become increasingly aware of the dangers involved, and hostile to Roman officials who still dealt with them; then in the 470s we see him despairing of any further help for Clermont, and contemptuous of the Italian envoys who sacrificed the Auvergne so as to keep Provence under Roman control. By around 480, as he put it, ‘now that the old degrees of official rank are swept away . . . the only token of nobility will henceforth be a knowledge of letters’; the official hierarchy had gone, only traditional Roman culture survived.

As an epitaph for the western empire, this is somewhat muted. It is far from clear that Sidonius saw Rome as having definitively ended; and his claim that the traditional hierarchies had gone was certainly exaggerated. But much was changing in Gaul, for all that. Euric’s conquests were soon matched by the Burgundians under Gundobad in the Rhône valley, with Provence a battleground between these two peoples and the Ostrogoths in the decades after 490. In the North, there were still armies which looked to Rome, under Aegidius around Soissons, Arbogast around Trier, and Riothamus, a British warlord, on the Loire; but Aegidius had recognized no emperor since Majorian, and these can be regarded as effectively independent polities, probably using rather fewer Roman traditions than the Goths and Burgundians did. The Frankish kings in the North allied and competed with them, and the most successful of these, Clovis of Tournai (481-511), began to take over rival Frankish kingdoms and the lands of Roman warlords alike.

The north of Gaul had long been the most militarized part of the region, where the army structured exchange, social display and landowning patterns, and this accentuated across the fifth century. Villa culture had ended here by 450, for example, as also in rapidly de-Romanizing Britain, but unlike anywhere else in the West, where the richest rural residences continued until well into the sixth century; this marks the early end of one of the classic markers of civilian élite culture. Sidonius, who knew all the great civilian aristocrats of Gaul, hardly ever wrote to people north of the Loire (one was Arbogast of Trier, whom he praises for maintaining Roman cultural traditions - Sidonius clearly thought that this was hard in the north). The rest of what we know of the north points at very ad-hoc political procedures, as with the saintly Genovefa’s travels to find food for Paris in, perhaps, the 470s, or the bishops who dealt directly with Clovis in the 480s. The south of Gaul was much better organized; Visigothic and Burgundian kings legislated, taxed, shipped grain around, used Roman civilian officials, and created integrated Roman and ‘barbarian’ armies, including Roman generals. But, everywhere in Gaul, the last two decades of the fifth century were definitively post-imperial, in the sense that half a dozen rulers faced each other with no mediation, no distant Rome/Ravenna-based hegemony to look to. Gaul is the best-documented part of the West in the late fifth century, so we can see this most clearly there, even if it was also arguably the region where change was greatest: more than in Italy, certainly, but more even than in Africa, where Vandal rule, popular or not, was solid and relatively traditional. All of these regions were nonetheless post-Roman too; imperial unity and identity was by 500 the property of the East alone.

It must also be recognized when discussing these post-Roman kingdoms that the shift away from Roman government was often rather less organized, or quick, than narratives of conquest imply. Eugippius’ Life of Severinus gives us an instance of this. Severinus (d. 482) was a holy man in Noricum (modern Austria) in the 470s, at a time when the Danube frontier was breaking down, but the main ‘barbarian’ group nearby, the Rugi, had remained firmly beyond the river and restricted themselves to raiding and taking tribute - and also to trading with the Romans. Severinus won the respect of King Feletheus and was able to mediate between Romans and Rugi on several occasions. Life in Noricum was clearly miserable, as well as cold (the imagery of winter is stressed constantly by Eugippius, who was a younger contemporary of Severinus but had left for Italy, and who was writing thirty years later much further south, in Naples). It was a province in which the Romans were concentrated in towns and fortifications, and various ‘barbarians’ roamed the countryside. The Roman army was still in existence, but there was no political leadership, at least in Eugippius’ vision of the province, except for Severinus’ mediating role. This sort of no man’s land may have characterized other areas, too: parts of northern Gaul, parts of central Spain, much of Britain. The social breakdown involved in these regions would have been much greater than that in any area of quick conquest, no matter how violent. But most of the West was nonetheless under the control of more stable (and more Roman) polities, whether Gothic, Burgundian or Vandal.

The East in the late fifth century was a less tranquil place than under Theodosius and Pulcheria. For a start, it had by now rulers who were much more militarized: Aspar, magister militum in 457-71, strong-man for his protégé, Emperor Leo I (457-74), until Leo had him killed, and his successor Zeno, who became emperor in his own right (474-91). Secondly, Zeno had constant trouble with rivals. The main eastern army base had remained the Balkans, but this military region was itself more unstable after the end of Hunnic power, and ‘barbarian’ groups, mostly Goths, were beginning to enter the empire again: two of their leaders, Theoderic Strabo and Theoderic the Amal, each of them with Roman military experience, tried under Leo and Zeno both to gain power in Constantinople and to settle their respective peoples in a favoured part of the Balkans. Zeno was himself from Isauria, a remote mountain region in what is now southern Turkey, and a traditional source of soldiers (and also bandits) which could be seen to an extent as in competition with the Balkans; Zeno had rivals in Isauria, too; tensions with the army thus increased when he succeeded to the throne. Indeed, for a year (475-6) he was out of office, expelled by the general Basiliscus, and he faced several revolts even after that. It was only in the late 480s, shortly before his death, that he managed to quell rivals, and to persuade the main warlord who survived, Theoderic the Amal, to leave with his Gothic army and occupy Italy in 489. These problems meant that Zeno had no hope of intervening in the West himself, even had the fingers of the East not been burnt by the costly failure of the Vandal war in 468. A substantial stability was, however, restored by Anastasius I (491- 518), an elderly but able career bureaucrat who lived to the age of eighty-eight and had time both to quell Isaurian revolts and to put imperial finances firmly in the black. The fact that Anastasius could do this, and without a military base either, must indicate that the eastern political system was essentially solid.

We are now in 500, and the East, despite some trouble under Zeno, was still in a stable state. The West had greatly changed, as we have seen, but there were elements of stability there too. Theoderic ruled Italy from Ravenna, the western Roman capital, with a traditional Roman administration, a mixture of senatorial leaders from the city of Rome and career bureaucrats; he was (as Odovacer had also been) respectful of the Roman senate, and he made a ceremonial visit to the city in 500, with formal visits to St Peter’s, to the senate building, and then to the imperial palace on the Palatine, where he presided over games, like any emperor. Theoderic’s whole modus operandi was largely imperial, and many commentators saw him as a restorer of imperial traditions. This was certainly the view of Cassiodorus Senator (lived c. 485-580), who was an administrator for him after 507 and who wrote an extensive collection of official letters for Theoderic and his immediate successors, which he called the Variae; Cassiodorus deliberately wrote up Theoderic as an upholder of Roman values, but it was easy for him to do so. The administrative and fiscal system had changed little; the same traditional landowners dominated politics, beside a new (but partly Romanizing) Gothic or Ostrogothic military élite.

Theoderic looked beyond Italy, too. He ruled Dalmatia and the Danube frontier; and he was well aware of his cultural connections to the second Romano-Germanic power in the West, the Visigothic kingdom of Alaric II (484-507) in southern Gaul and Spain. Orosius had claimed that Athaulf the Visigoth said in 414 that he had considered replacing Romania with Gothia, but had decided against it, because the Goths were too barbaric, and could not obey laws. Were this story true (which is unlikely), it was reversed by the end of the century. Theoderic in Italy, Euric and Alaric in Gaul all legislated for their subjects, Goth and Roman. The Goths were military figures, it is true, unlike the senatorial stratum (or most of them), and were Arian, not Catholic, Christians, but in other respects they were picking up Roman values fast. In this they were followed by the Vandals and Burgundians, who were both very influenced by the larger Gothic kingdoms by 500 or so. In a sense, Gothia really had replaced Romania, but had done so in large part by imitating the Romans. In the western Mediterranean, in effect everywhere in the West south of the Loire and the Alps, a common political culture survived.

But the world was changing. The end of political unity was not a trivial shift; the whole structure of politics had to change as a result. The ruling classes of the provinces were all still (mostly) Roman, but they were diverging fast. The East was moving away from the West, too. It was becoming much more Greek in its official culture, for a start. Leo I was the first emperor to legislate in Greek; under a century later, Justinian (527-65) may have been the last emperor to speak Latin as a first language. But it is above all in the West that we find a growing provincialization in the late fifth century, both a consequence and a cause of the breakdown of central government. Augustine thought in terms of the whole empire; Salvian took his moral images at least from the whole of the West (though he only really knew Gaul). But Sidonius was definitely a Gaul. Gaulish élites rarely travelled to Italy by now; although Sidonius was urban prefect in Rome in 468, he was the first Gaul to hold the office since perhaps 414, and also the last. His colleagues were even more clearly focused on Gaulish politics, like his friend Arvandus, praetorian prefect of Gaul in 464-8, and his enemy Seronatus, an administrator in central Gaul in and after 469, both of whom threw in their lot with Euric’s political ambitions and were cashiered for it; Euric’s Roman generals Victorius and Vincentius were presumably more successful variants of the same type, provincials who saw advancement in the Visigothic court as simply more relevant than the traditional career hierarchy centred on distant Ravenna. These were political shifts which made a lot of sense to local actors, but they were fatal to what remained of the empire. Sidonius himself left the imperial hierarchy when he became a bishop in 469/70, and the growing tendency for aristocrats in Gaul to look to the episcopate for a career (above, Chapter 3) expresses this local focus very clearly. In the next generation, horizons narrowed again; Ruricius of Limoges (d. 510) and Avitus of Vienne (d. 518), bishops in the Visigothic and Burgundian kingdoms respectively, both left collections of letters, written very largely to recipients inside their respective kingdoms (with the main exception of Sidonius’ son Apollinaris in Clermont, to whom they were both related).

This provincialization was not restricted to Gaul, either. Hydatius of Chaves (d. c. 470) wrote a chronicle which tells almost entirely of Spain, especially the north-west, where he was based. Victor of Vita in Huneric’s Africa saw the Vandals exclusively from the perspective of the Africani; the Roman empire never appears in his text, and even Romani are only referred to when he is being very generic. A common political culture may have survived, but in each former Roman region or province its points of reference were becoming more localized, and its lineaments would soon start to diverge. The easy unity which had taken the biblical scholar Jerome in the late fourth century from Dalmatia to Trier, then Antioch, Constantinople, Rome and finally Palestine, from where he wrote letters to his Mediterranean-wide ascetic clientele for thirty years, had gone. I shall come back to this issue in more general terms later in this chapter.

The high point of the Gothic western Mediterranean was around 500. It was destroyed by two men, Clovis the Frankish king and the eastern emperor Justinian; let us look at them in turn. Clovis reunited northern Gaul, including some non-Roman territories, during his reign; in 507 he attacked the Visigoths, defeating and killing Alaric II at the battle of Vouillé, and virtually drove them out of Gaul (they only kept Languedoc, on the Mediterranean coast). The Burgundians held on for a time, but in the 520s Clovis’s sons attacked them too, and took over their kingdom in 534. Theoderic reacted by occupying Visigothic Spain, nominally ruling for Alaric’s son Amalaric (511-31), but Spain’s political system went into crisis for two generations. It is hard to see that Theoderic’s Spanish extension was more than the temporary reinforcement of the Mediterranean coast against the Frankish threat; already by 511 the hegemony of the Goths in the West had largely gone, except in Italy. Clovis’s Merovingian dynasty would dominate post-Roman politics in the West for the next two centuries. We shall look at its history in the next chapter. For now, it is enough to stress one important geopolitical consequence of Clovis’s success: northern Gaul, long a military borderland, rather marginal to the Roman world except in the mid-fourth century when Trier was the western capital, became a political heartland territory, a focus for great landed wealth and political power. It was initially a focus for Gaul alone, but across later centuries it was one for the whole of western Europe.

Justinian, Anastasius’ second successor, took Anastasius’ large budgetary surplus and devoted most of his forty-year reign to imperial renewal. There is a bounce about his accession in 527 that had not been visible for any emperor since Julian. As we saw in Chapter 2, starting in 528 he had Theodosius II’s law code revised in a year, and by 533 the writings of the Roman jurists were codified as the Digest, still today the master text of Roman law. Furthermore, a string of new laws (Novels) surveyed and revised the administration of the empire in the 530s, and also tightened laws on sexual deviance and heresy, even Jewish heresy, provoking Samaritan revolts and severe repression in northern Palestine in 529 and 555. Justinian was no liberal, and a growing humourlessness and intolerance of religious difference is visible in the East from this time onwards; he was nonetheless an innovator, and the complaints of traditionalists during his reign about the uncultured radicals in his administration indicate that his organizational changes had some effect. Justinian was also a builder, always an important part of political display in the Roman tradition. He is not the only one in this chapter; Zeno, Anastasius and perhaps also Theoderic the Ostrogoth were particularly active; but the scale of Justinian’s building outmatched them all, as with the huge churches he built in Constantinople (such as Hagia Sophia, see below, Chapter 10), Ephesos and Jerusalem. These building campaigns are well documented in a panegyric work, Prokopios’ On Buildings; as a result, archaeologists have been prone to date almost every major late Roman building in the East to the second quarter of the sixth century, and careful redatings have been necessary to uncover other patrons both before and after him. All the same, the money and the commitment were there to do a lot.

Given the self-confidence of these acts, it is not surprising that Justinian was also interested in war. He faced Persian wars, the first serious conflicts for well over a century, in 527-32 and 540-45, and intermittently thereafter up to 562. Persia was always the major front for the eastern empire (the Balkans were also attacked in his reign, but this was hardly new, and was regarded as less crucial). It was expensive both in resources and in post-war reconstruction, and many emperors would have restricted their attention to Persian defence. But Justinian used the period of eastern peace in 532-40 to attack the West as well. His general Belisarios took Vandal Africa quickly, in 533-4, and moved straight into Ostrogothic Italy; he had almost completely conquered it by 540. Theoderic’s last years had shown up tensions with traditionalist figures too, and the aristocratic philosopher Boethius, among others, was executed for treasonous communication with the East in 526; infighting between Theoderic’s heirs in 526-36 led to a more serious alienation of some of the aristocratic élite from the Ostrogothic regime, many of whom ended up in Constantinople. But whereas the conquest of Africa was largely a success, Italy was not. Most of the non-Gothic Italians were at best neutral about Justinian’s armies, and the Goths regrouped after 540 under Totila (541-52), when the renewal of the Persian war pulled Roman troops away from the peninsula. The 540s saw Italy devastated, as Roman and Gothic armies in turn conquered and reconquered sections of the peninsula, and when war largely stopped in 554 Italy, now Roman again, had a fiscal system in ruins, a fragmented economy and a largely scattered aristocracy. This was not handled well, then. But Justinian had nonetheless absorbed the central Mediterranean back into the empire, and when his armies also occupied part of the Spanish coast in 552, almost the whole of the Mediterranean returned to being a Roman lake.

Justinian was and is a controversial figure. He was hated by many, notably those whom he disagreed with on religious matters and persecuted, who became more numerous as his reign went on. This followed his growing hostility to Monophysites, especially after the death of his influential wife Theodora (herself a Monophysite) in 548, and then his equally controversial attempt to take a doctrinal step in the Monophysite direction at the fifth ecumenical council of Constantinople in 553, which alienated much of the West. Less serious (and far too influential on modern scholars) was Prokopios’ set-piece anti-panegyric, the Secret History, which depicts Justinian and Theodora as wicked geniuses, in highly coloured and sexualized terms, with Justinian characterized as a demon. Today, Justinian is above all accused of ruining the empire financially, thanks to his anachronistic wars in the West; the eastern empire after his death in 565 is often seen as weakened, both militarily and economically, a state of affairs that would result in the political disasters of the years after 610. We shall look at the seventh-century crisis in Chapter 11, but it does not seem to me to have much to do with Justinian. The western wars were not anachronistic, for the Roman empire was still a meaningful concept even in the West, nor were they particularly expensive; Africa was won on a shoestring, and remained Roman for more than a century longer, and the Italian war would have been less of a mess if Justinian had put more, not less, money into it. Justinian’s successors, notably Tiberius II (578-82) and Maurice (582- 602), held off the Persians, their main opponents, as effectively as Justinian had done. They also kept out the Avars, the new holders of ‘barbarian’ hegemony in the middle Danube, who from the 560s turned the most recent invaders of the Balkans, mostly Slavic-speaking (but also Turkic- and Germanic-speaking), into the greatest military threat in the area since the Huns. They abandoned most of Italy to a new people, the Lombards, but given Italy’s state this was not necessarily a strategic failure. Furthermore, money was sufficiently loose into the 570s for Tiberius (though not Maurice) to be noted as an extravagant spender. Justinian’s reign does not seem to have been a negative turning point for the empire. But the controversy over it does at least mark respect: Justinian put his stamp on a generation, all over the Mediterranean, and, unlike most rulers, the events of his reign seem to have been the result of his own choices. His protagonism gives the lie to the view that the break-up of the fifth-century West in itself marks the failure of the Roman imperial project.

The foregoing pages give a bare summary of the events of a hundred and fifty years; we must now consider what they mean. I shall concentrate more on the West, because it was there that the greatest changes took place, although the stability and prosperity of the East must act as a permanent reminder to us that the Roman empire was by no means bound to break up. In recent decades this view, already discussed in Chapter 2, has indeed become a dominant one among historians. This means that the invasions and occupations of the western provinces must be at the heart of our explanations of the period. But in recent decades we have also moved away from catastrophist views of the ‘barbarians’, encapsulated in André Piganiol’s famous lines at the close of his book on the late empire, written (significantly) just after the Second World War, ‘Roman civilization did not die a natural death. It was assassinated. ’ Recent work has in fact depicted the new ethnic groups in very Roman terms, a view which I fully accept and shall develop further shortly. This does not lessen the simple point that the Roman empire in the West was replaced by a set of independent kingdoms which did not make claims to imperial legitimacy; but it does force us to ask why each of these kingdoms could not have just reproduced the Roman state in miniature, maintaining structural continuities that could, in principle, have been reunited later, by Justinian, for example. For the fact is that most of them did not do so. One thing that archaeology makes very clear, as we shall see, is the dramatic economic simplification of most of the West: this is visible north of the Loire in the early fifth century, and in the northern Mediterranean lands during the sixth. Building became far less ambitious, artisanal production became less professionalized, exchange became more localized. The fiscal system, the judicial system, the density of Roman administrative activity in general, all began to simplify as well. These are real changes which cannot be talked away by arguments that show, however justifiably, that the ‘barbarians’ merely fitted Roman niches. They are matched by shifts in imagery, values, cultural style, which makes the seventh century in the West noticeably different in feel from the fourth or even the fifth: we are by now out of the late Roman world and into the early Middle Ages. How this could be, given the lack of innovation desired by most of the new ethnic groups, is the issue we need to confront.

To start with, there is an evident continuum between the leadership of the fifth-century western (and indeed eastern) empire and the ‘barbarian’ kings. The fifth-century emperors were mostly ciphers, controlled by military strong-men, Stilicho, Constantius, Aetius, Ricimer, Aspar, Zeno, Gundobad, Orestes (Romulus Augustulus’ father). It is interesting that none of these tried to seize the throne by force, as military figures regularly had in the third century, and only two (Constantius and Zeno) became emperor even by more regular means. One commonly advanced reason for this is that, as ethnic ‘barbarians’, they were not entitled to imperial office; but, quite apart from the fact that not all of them were of non-Roman descent, there is no contemporary basis whatsoever for an exclusion of this kind. Basiliscus, briefly eastern emperor in 475-6, may indeed have been Odovacer’s uncle, and thus a Scirian, from a subject people of Attila’s Huns; Silvanus, a failed usurper in 355, was certainly a Frank. More likely they held off from seizing power because of a trend towards a view that imperial legitimacy was allied to genealogy, a view which can be traced back to Constantine’s family in the mid-fourth century; it would have seemed safer to control an emperor (or a series of emperors, as Ricimer did) than to usurp the throne. And it probably was; these strong-men had much longer periods of authority than most third-century emperors. An important element in late Roman genealogical legitimacy was marriage, so all the strong-men intermarried with the imperial families, hoping to put their sons on the throne; Constantius and Zeno both managed this. (Zeno became sole emperor himself, of course, but only as heir to his own short-lived son.) But this is equally true of the ‘barbarian’ royal families, most of whom had, or soon established, links of marriage to the Romans, often doubtless with the same intent. This genealogical network makes a nonsense of cultural difference, at least at the imperial or royal level. So does the fact that nearly every emperor of the East for more than a century after 450 (with only one exception, Zeno) came from the melting pot of the Balkans, where new identities were being refashioned all the time, as also did a high percentage of the imperial strong-men and the ‘barbarian’ leaders alike. And there were cross-overs in personal terms: both Gundobad the Burgundian and Theoderic the Ostrogoth had careers in and around the imperial court before becoming kings of independent ex-Roman provinces.

The importance of intermarriage as a criterion for succession also put a good deal of stress on imperial women. We have seen that Galla Placidia and particularly Pulcheria were powerful in the early fifth century, and both legitimized their imperial husbands. So did Ariadne, daughter of Leo I and wife successively of Zeno and Anastasius. Verina, Leo’s wife, was Basiliscus’ sister. Theodora, herself a powerful political operator despite her husband Justinian’s dominance, seems to have promoted her kin as well, although she died too long before her husband for any of them still to be in place to succeed him. Sophia, widow of Justin II (565-78), certainly chose his successor, Tiberius II, and perhaps Maurice too. There was a space for female political action here, which was taken up many times. It is thus not surprising that Anicia Juliana (d. 527/8), a rich private citizen in Constantinople but a descendant of Valentinian III and of a whole host of empresses (and also wife of a descendant of Aspar), and bearing the title of patricia by 507, should have had an impact on Justinian: her church of Hagios Polyeuktos, in the centre of Constantinople, built around 525, was the largest church in the city until Justinian built Hagia Sophia a decade later, probably in part as a response. This space for female power, however ambivalent (for it was always that), was more of an eastern than a western feature; the military crises of the West favoured a more male military leadership. Women in the West who could dominate a militarized politics would appear later, with the Lombards after 590 and the Merovingian Franks after 575, but their prominence had different reasons.

To return to the ‘barbarian’ leaders, and to their peoples: what exactly was non-Roman, ‘barbarian’, Germanic, about them at all? There is at present enormous debate about this, with an endless variety of positions even among those who accept that the new ethnic groups sought to accommodate themselves to Roman rules as much as they could: from the belief that there was a substantial kernel of non-Roman values and traditions, associated with the dominant element in any invading or settling group, which could survive for centuries, to the belief that Germanic ethnic markers were only a renaming of the military identity of Roman soldiers, and that there was nothing traditional about them at all. It does at least need to be recognized, with this second position, that most of the new ‘barbarian’ groups in the fifth-century empire had a history of employment in the Roman army; the most successful soldiers among them, such as the Visigoths, were effectively indistinguishable from a Roman military detachment. (‘Barbarian’ armies regularly travelled with their families and dependants, but, although it was theoretically illegal, it would be unwise to presume that Roman armies in practice did not.) We can, however, see a clear distinction in our sources between regular army forces, which, whether of Roman or ‘barbarian’ origin (as we saw in Chapter 2, there was on the frontiers, whence soldiers usually came, little difference between them), were part of a standard military hierarchy and career structure; and the followers of King X or leader Y, who identified with their leader, generally had a distinct ethnic name, and were accepted into the Roman army as a discrete group. This is the distinction between Odovacer and Theoderic, for example, successive rulers of Italy. Odovacer was the candidate of the Roman army of Italy, which merely happened to consist of ethnic Heruls, Sciri and Torcilingi; Odovacer was himself at least half-Scirian, but he had a Roman military background, and is never called leader of the Sciri, or of any other group in Italy. He became a king, formally autonomous, but he recognized Zeno, and could fairly easily have been refigured as part of the Roman empire. Theoderic, by contrast, was a king of the Goths, whose people came with him from the start, no matter how many imperial titles he also had. That people was as mixed as Odovacer’s supporters; it certainly contained Rugi (who maintained an identity through intermarriage for fifty years after Theoderic’s conquest of Italy), Gepids, Huns and doubtless men of Roman descent as well, and, after Theoderic’s conquest, it will have absorbed all or most of Odovacer’s following. But it was attached to a leader, and had a name, ‘Gothic’, Ostrogothic in our terminology; this name characterized the people as a whole, no matter what their origin, and also Theoderic’s kingship. It was peoples like this, heterogeneous but - an essential feature - tied together by a single leader, which took over the western provinces, and indeed renamed them, the regnum Francorum instead of or alongside Gaul, the regnum Vandalorum instead of or alongside Africa. If they stayed in charge of their lands long enough, as the Franks and Visigoths did, though the Vandals and Ostrogoths did not, they tended to forget their disparate origins, and ‘become’ Frankish or Gothic - and also, crucially, not Roman.

It is this process that has been called ‘ethnogenesis’ by Herwig Wolfram and his school: the recognition that ethnic identities were flexible, malleable, ‘situational constructs’; the same ‘barbarian’ in sixth-century Italy could be Rugian, and Ostrogothic, and (though only after the east Roman reconquest) even Roman. Such people would have picked up different identities successively (or contemporaneously), and these would have brought with them different modes of behaviour and loyalties, and even, eventually, different memories. As Walter Pohl has recently put it, the ‘kernel of traditions’ that made someone Ostrogoth or Visigoth was probably a network of contradictory and changeable beliefs; there does not have to have been a stable set of traditions in each group as it moved from beyond the frontier, to discontinuous service in the Roman army, then to settlement in a Roman province. By 650 every ‘barbarian’ kingdom had its own traditions, some of them claiming to go back centuries, and those doubtless were by then core elements in the founding myths of many of their inhabitants; all the same, founding myths not only do not have to be true, but also do not have to be old. Each of the ‘Romano-Germanic’ kingdoms had a bricolage of beliefs and identities with very varying roots, and these, to repeat, could change, and be reconfigured, in each generation to fit new needs. Historians tend to give more attention to the account that Clovis’s grandfather was the son of a sea-monster, a quinotaur, than to the account that the Franks were descended from the Trojans, which seems more ‘literary’, less ‘authentic’; but the first record of each of these traditions appears in the same seventh-century source, and it would be hard to say that one was more widely believed - or older - than the other.

From all of this, one has to conclude that post-Roman identities were a complex mixture, and they had a variety of origins: Roman, ‘barbarian’, biblical; and also both oral and literary. What they had to do was less to locate an ethnic group in the past, than to distinguish it from its contemporary neighbours. This means that to ask what was non-Roman or ‘barbarian’ about the new ethnic groups is in part the wrong question; Arianism, for example, was a very Roman heresy, but by 500, for most people, it had become an ethnic marker, of Goths or Vandals. The Gothic language itself was by 500 in large part a liturgical tradition, associated precisely with that ex-Roman Arianism, rather than with ‘Gothic-ness’ in an ethnic sense; many Goths just spoke Latin, without their Gothic-ness being affected either positively or negatively. Indeed, unlike in the twentieth and twenty-first centuries, language was not, as far as we can see, a strong ethnic marker anywhere in our period. Plenty of Franks in 600, say, still spoke Frankish (a version of what we now call Old High German), but very probably not all did, and many were certainly fully bilingual. Gregory of Tours, the most prolific writer of the sixth century in Gaul, who was a monoglot Latin-speaker, never gives the slightest indication that he had trouble communicating with anyone else in the Frankish kingdoms. Neither he nor anyone else in the Frankish world, until the ninth century in fact, makes anything of communication difficulties between primary speakers of Latin and Frankish; it must have happened, but it was not a problem for Frankishness.

This does not mean that the ‘barbarian’ groups brought nothing of their earlier cultures into the empire, all the same. There is a whole historiography which discusses the German-ness of early medieval social practices, such as large kin-groups, or feud, or personal followings, or meat-eating, or certain concepts of property, or certain types of brooch or belt-buckle. Almost all of this is phoney if seen as a sign of innate identity, as if the Franks of 700 were exactly the same as the Franks of 350. Some of it is inaccurate, too: most early medieval property law had impeccable Roman antecedents, or at least close Roman parallels; similarly, ‘Germanic’ metalwork sometimes has Roman antecedents, and, even if it does not, does not provide us with any guide to the ethnic identities of the people who wore it. But it would be equally unhelpful to cancel all of this by sleight of hand, and to present the new ethnic groups simply as variants of Roman society itself. A stress on aristocratic meat-eating, for example, genuinely does seem to be an innovation of (among others) the Franks; it was not part of Roman cuisine, where status was conveyed by the complexity and the cost of ingredients, but first appears in a treatise about diet written for the Frankish king Theuderic I (511-33) by a doctor of Greek origin called Anthimus, and it continued throughout the Middle Ages.

A particularly important innovation was the public assembly, the formal meeting of the adult male members of a political community, to deliberate and decide on political action and war, and, increasingly, to make law and judge disputes. The Romans had plenty of large-scale public ceremonials, as we saw in Chapter 3, but in the post-Roman kingdoms assemblies had a wider significance, in that they represented the principle that the king had a direct relationship with all free Franks, or Lombards, or Burgundians; these derived from the values of the tribal communities of the imperial period, but continued in the very different post-Roman world. We can thus trace a continuum of political practice which links the Franks and Lombards, not with Rome in this case, but with the less Romanized or un-Romanized peoples of the early medieval North; the Frankish or Lombard placitum assembly, or the Burgundian conventus, has parallels with the Anglo-Saxon gemot, the Scandinavian thing, the Irish óenach. These assemblies were not really of all free men, the traditional kingdom-at-arms of Romantic mythology, but they could be wide gatherings for all that, and they derived their power to legitimate political and judicial acts precisely from the fact that many people were there. From 500 to 1000, and sometimes later, public politics in the West was underpinned by the direct participation of wide sections of free, male, society. This went together with an assumption that wide sections of the free had military obligations, which was largely a product of post-Roman conditions, as we shall see in more detail later. But the link between military commitment and assembly politics must have made sense already to the ethnic armies of the fifth century; the generalization of assembly imagery in every Romano-Germanic kingdom (even the heavily Romanized Visigothic state) itself allows us to presume it.

Notwithstanding these new features, ‘barbarian’ leaders fitted into a Roman world, more and more as the fifth century wore on, and as local Roman élites adjusted to new political situations. It is striking how Roman these élites could make their new rulers in their writings; nearly every new ruling ethnic group had its apologist who was prepared to describe ‘barbarian’ kings in resonantly Roman terms, as with Sidonius’ famous prose panegyric on the Visigothic king Theoderic II, stressing his seriousness, his accessibility to ambassadors and petitioners (and his board-games), and playing down his Arianism. There were not large numbers of ‘barbarian’ invaders in any province; all raw figures are guesswork, but historians generally propose up to 100,000 for major ruling groups like the Ostrogoths or the Vandals, and around 20,000- 25,000 for the adult males who made up their armies, in provinces whose indigenous populations numbered in the millions. Putting together the ethnic flexibility of so many of the actors of the period, the Romanizing images of so many of our texts, and the small demographic impact of the invaders - one in ten? one in twenty? one in fifty? - it is easy to imagine that they had no effect at all on the social practices of each province. But if we argue this line too schematically, we risk ending up with no explanations for change at all. And change, in the fifth century, certainly took place.

This change did not derive mostly from cultural differences, all the same. Regions which experienced the miserable insecurities described earlier for Noricum would have seen substantial social breakdown even if no ‘barbarians’ ever settled. But in conquered provinces, the majority in the West, change derived most of all from the structural position of each ‘barbarian’ group. As noted earlier, the ‘barbarian’ armies that took over provinces had different aims from the Roman armies that seized power for their generals in previous centuries. They wanted to settle back on the land, as their ancestors had done, before the generation or so of intermittent movement and conquest. Their leaders, and probably a good proportion of the middling Goths or Vandals or Franks as well, also wanted to be a ruling class, like the rich Roman aristocrats in each of the provinces they occupied. To fulfil this aim, itself a very Roman one, they needed estates, and, as conquerors, they were in a good position to obtain it. Although the exact details of the land-settlement of each ‘barbarian’ group are obscure and hotly debated (indeed, they must have been very variable), by 500 or so it is clear that Gothic and other ‘barbarian’ aristocrats had extensive properties, and were keen to extend them further; Cassiodorus’ Variaeinclude several instances of Ostrogoths abusing their political and military authority and expropriating the lands of others, for example. Beginning in the fifth century, there was a steady trend away from supporting armies by public taxation and towards supporting them by the rents deriving from private landowning, which was essentially the product of this desire for land of conquering élites. In 476, according to Prokopios, even the Roman army of Italy wanted to be given lands, and got it by supporting Odovacer. Prokopios may well have exaggerated; the Ostrogothic state in Italy certainly still used taxation to pay the army, at least in part, probably more than any other post-Roman polity did by the early sixth century. Overall, however, the shift to land was permanent. After the end of Ostrogothic Italy, there are no references in the West to army pay, except rations for garrisons, until the Arabs reintroduced it into Spain from the mid-eighth century onwards; in the other western kingdoms, only occasional mer cenary detachments were paid until well after the end of the period covered by this book. Some of this land may have been fiscal, that is, public property, and distributed by kings; some may have been part of a regular land-settlement, in which fixed proportions of the property of Roman landowners were ceded to the ‘barbarians’, probably in lieu of tax; some (as in Vandal Africa) may have simply been taken by force. Either way, a move to a landed army, and thus a landed politics, began here; so also did a move to a ‘barbarian’ ethnic identity on the part of landowners, whatever their origins.

The major post-Roman kingdoms still taxed, into the seventh century. But if the army was landed, the major item of expense in the Roman budget had gone. The city of Rome, another important item, was only supplied from Italy after 439, and lost population fast, as we have seen. The central and local administration of the post-Roman states was perhaps paid for longer, but in most of them the administration quickly became smaller and cheaper. Tax still made kings rich, and their generosity increased the attractive power of royal courts. But this was all it was for, by 550 or so. Tax is always unpopular, and takes work to exact; if it is not essential, this work tends to be neglected. It is thus not surprising that there are increasing signs that it was not assiduously collected. In ex-Vandal Africa after 534, the Roman re-conquerors had to reorganize the tax administration to make it effective again, to great local unpopularity; in Frankish Gaul in the 580s, assessment registers were no longer being systematically updated, and tax rates may only have been around a third of those normal under the empire. Tax was, that is to say, no longer the basis of the state. For kings as well as armies, landowning was the major source of wealth from now on.

This was a crucial change. Tax-raising states are much richer than most land-based ones, for property taxes are generally collected from very many more people than pay rent to a ruler from his public land. Probably only the Frankish kings at the high points of their power, the century after 540 and the century after 770, could match in wealth the states of the eastern Mediterranean, the Byzantine empire and the Arab caliphate, which still maintained Roman traditions of taxation. And tax-raising states have a far greater overall control over their territories, partly because of the constant presence of tax-assessors and collectors, partly because state dependants (both officials and soldiers) are salaried. Rulers can stop paying salaries, and have greater control over their personnel as a result. But if armies are based on landowning, they are harder to control. Generals may be disloyal unless they are given more land, which reduces the amount of land the ruler has; and, if they are disloyal, they keep control of their land unless they are expelled by force, often a difficult task. Land-based states risk breaking up, in fact, for their outlying territories are hard to dominate in depth, and may secede altogether. This would not be common until the late ninth century or later in the West. Many things would have to change before then, as we shall see in later chapters. But it did happen in the end, above all in the wide lands ruled by the Franks.

The shift from taxation to landowning as the basis of the state in the West was the clearest sign that the post-Roman kingdoms would not be able to re-create the Roman empire in miniature, however much their rulers would have liked to. Overall, too, these kingdoms did not match the empire in their economic complexity, either. Archaeology shows a steady simplification of economic structure in most of the West by 550 or so. By then, rich urban and rural dwellings (villas) had often been abandoned, or subdivided into smaller houses; artisan production was generally smaller-scale, and sometimes less skilled (this is particularly clear in the case of pottery production, always our best archaeological indicator of artisanal professionalization); goods were exchanged much less between the provinces of the former empire, and inside those provinces, the new kingdoms, the distribution range of artisanal goods was generally much reduced. The pacing of these changes varied greatly from place to place, and not all of them took place everywhere. In northern Gaul, towns decreased in size and villas were abandoned by 450, but production and distribution patterns dipped much less (northern Gaul’s economy had long been separate from that of the Mediterranean), and had stabilized by the sixth century. In Spain, the interior saw a simplification of distribution patterns and a partial abandonment of villas from the later fifth century, whereas the Mediterranean coast saw less change until after 550. In Italy and southern Gaul, the mid-sixth century was the major period of change, but small-scale skilled artisanal production survived, and so did towns. In Africa, the great export region in the late Roman West, little internal change is visible at all until 500 or so, and one can track a survival of the main elements of the Roman economic structure until after 600, even though there is a steady decrease in African exports found in most of the rest of the Mediterranean which begins as early as 450.

These regional differences - which could be multiplied, for our information is getting more detailed all the time, as scientific archaeological excavation becomes commoner in each country - are markers of the different impact the invasions and dislocations of the period 400-550 had on each part of the empire. It was more than one might expect in inland Spain; less than one might expect in Frankish northern Gaul and Vandal Africa. These differences also show that the aristocracies of the newly created kingdoms did not match the wealth of their predecessors or ancestors, partly precisely because it was harder to own far-flung estates now that the empire was divided up (the hyper-rich senatorial élite of Rome ceased to exist, in particular), but this impoverishment was also very variable indeed in regional terms. Seen globally, however, these changes show that the post-Roman kingdoms in the West were unable to match the intensity of circulation and the scale of production of the later Roman empire. The East was very different in this respect; in the early sixth century, towns, industries and the exchange of goods were reaching their height, and continued at that level until the early seventh century. But the empire survived in the East. This correlation is exact: economic complexity depended on imperial unity, in both the eastern and the western empire. The implications that these changes had for local societies in the West will be discussed in Chapter 9.

The existence of ‘barbarian’ élites in each of the post-Roman kingdoms had an impact on Roman élite culture as well: not because the incomers were culturally distinct - as we have just seen, in most respects they were not - but because they were military. The aristocratic strata of the Roman empire had been mostly civilian, as we saw in Chapter 2. This was already less the case in the world of Aetius; Eparchius Avitus, for example, from a major Gaulish senatorial family, had been one of Aetius’ generals before he became emperor, and could be described in very martial terms by his son-in-law Sidonius. But in the post-Roman kingdoms, the secular career structure became steadily more militarized, and more and more ambitious Romans found places in royal armies and entourages alongside the ‘barbarian’ élites themselves, rather than in the steadily simplifying civilian administration. Sidonius himself never did this, but his son Apollinaris fought for the Visigoths at Vouillé, and Apollinaris’ son Arcadius was a supporter of Childebert I of the Franks. The place where civilian aristocratic values survived longest was Rome itself, because the senatorial hierarchy there was partially separate from state service, but even in Italy senators could make the military choice: Boethius’ enemy Cyprian, who had a partly military career, brought up his sons to be soldiers and even to speak Gothic.

These trends persisted; all secular aristocratic hierarchies became military. The only alternative was the church. As we have already noted, aristocrats became bishops in Gaul first, by the mid-fifth century; in Italy this was less common until the Gothic war, but was normal thereafter. This ecclesiastical choice shows the growing wealth of the church, such that it was worthwhile for an élite family to seek to dominate the episcopal office, and thus church land, in a given diocese. It also shows the growing localization of political action, for episcopal power was focused above all inside the diocese, except for the richest and most influential bishops; the church became even more decentralized in the post-imperial West. Being a bishop was sometimes a retirement option (as with both Sidonius and his son Apollinaris in Clermont), but increasingly it became a career choice, with a specifically clerical training: sometimes for younger sons, but sometimes for whole families. The extended family of Gregory of Tours in sixth-century central Gaul included seven bishops in four generations, and only one military figure, the dux Gundulf.

The major result of these trends was that the secular élite culture of the Roman empire lost its role as a marker of status. This is probably why rural villas were abandoned: as a sign of ease and luxury, they were out of date in a more militarized society. Meat-eating came in in this context, too. Élite clothing changed as well; early medieval kings and aristocrats dressed like late Roman generals, not like the older toga-clad senatorial tradition. But above all, to know Virgil and the other secular classics by heart, to be able to write poetry and complex prose, which Sidonius still regarded as essential, ceased to be important; swordsman ship, or the Bible, were far more relevant sources of cultural capital. Our written sources change dramatically as a result, becoming much more focused on Christian themes, hagiography, sermons, liturgy (as they would in Byzantium too). It is not that all forms of literary training ended; even in the West, aristocracies were generally able to read until the end of the ninth century. But we should anyway remain neutral about such changes. As stressed in Chapter 2, it is more important to recognize that a complex education had above all existed in order to mark the Roman elites as special, and, now that that elite identity was changing, it was no longer needed.

These changes usually took place slowly: a hundred and fifty years is a long time, after all. (Only in Italy were the changes really rapid, the result of the catastrophe of the Gothic war, in the 540s above all.) People were not usually aware of them; they adjusted easily to each small shift. It is not at all clear how far the majority of western writers saw the Roman world as having ended in the period up to 550, or indeed later. Writers rarely showed much nostalgia for the past, and, although they were certainly capable of complaining about how dreadful present-day morals were, this is a feature of conservatives of every generation. In any case, as writing became more ecclesiastical, it also became more socially critical, more moralizing; but that was a product of genre, not necessarily of social change, whether perceived or real. Traditional Roman aristocracies, the writers of most of our sources, were after all still in place in most parts of the West; they existed alongside newer families, rising in the church or the army, and of course the new ‘bar barian’ élites, but these latter groups were still copying Roman aristocratic culture. Still, that culture was itself changing. And aristocracies were becoming steadily more localized, drifting apart from each other. In the end - by 650, in every one of the post-Roman kingdoms - they would cease to think of themselves as Roman, but, rather, as Frankish or Visigothic or Lombard. ‘Romans’ were, by then, restricted to the eastern empire, to the non-Lombard portions of Italy (above all Rome itself), and to Aquitaine, the ex-Visigothic part of Gaul, where the Franks settled least. By then Romans were seen as belonging in the past, too; but it took that long for people to recognize that the empire had really gone in the West.

Why the Roman empire vanished in the West and not in the East is a problem that has perplexed centuries of scholars, and will continue to do so. It does not seem to me to reflect social differences between West and East, or the division of the empire. It probably did derive in part from the greater exposure of heartland areas in the West, Italy and especially central and southern Gaul, to frontier invasion; attacks on the Balkans in the East rarely got past Constantinople into the rest of the empire, but attacks on the western military regions, northern Gaul and the Danube provinces, could get further much more easily. Accepting invading groups into the western empire and settling them as federates was a perfectly sensible response to this, as long as those federate areas did not become so unruly that Roman armies had to be held back to fight them, or so large that they threatened the tax base of the empire, and thus the resources for the regular armies themselves. Unfortunately for the West, however, this did happen. The Visigoths in 418 could be a support for the empire, but fifty years later they were inimical to it. As argued earlier, the conquest of the grain heartland of Africa by the Vandals in 439, which the Romans mistakenly did not anticipate and resist, seems to me the turning point, the moment after which these potential supports might turn into dangers. Army resources lessened too much after that; the balance of power changed. By 476 even the Roman army in Italy may have started to think that landowning was desirable. And, not less important, local élites began to deal with the ‘barbarian’ powers rather than with the imperial government, which was by now too distant and decreasingly relevant; the provincialization of politics marked the death knell for the western empire. In the East, the control by the empire of that other huge grain resource, the Nile valley in Egypt, was never under threat in this period, and the logistical structure of the empire remained untouched as a result. When the Persians and then the Arabs took Egypt, and also the Levant, from Roman control after 618, the East would however face a huge and rapid crisis as well. The eastern Roman empire (we shall from that point on call it the Byzantine empire) survived, but it was a close-run thing, and the eastern empire changed considerably as a result.