Empires and Barbarians

by Peter Heather

  None of this evidence fundamentally contradicts the broader picture, then, that the social structure of northern Gaul would not be dominated by a small aristocratic elite until after 600 AD, unlike the regions south of the Loire where descendants of the old Roman aristocrats remained firmly in place. Culturally too, discontinuity was manifest. In wide areas of the north-east, episcopal succession was not continuous from the late Roman period into early medieval times (Map 12). So a period of positive paganism, or at least Christian interruption, must be envisaged in these lands. At the same time, the language line was also on the move. Germanic dialects became prevalent to the west of the old Roman frontier on the Rhine.68

  Material culture, too, north of the Loire was significantly different from southern areas of the kingdom. In the late fifth and the sixth century, inhumation that was furnished – sometimes spectacularly – became the vogue, replacing Roman-style burial. Men were buried not only with some personal items, but also with weapons: normally a long sword (spatha), javelin (angon), axe (francisca) and shield (of which only the conical boss usually survives). Women were buried fully clothed with their jewellery, their clothing fastened with a brooch at each shoulder. The brooches themselves were often framed in cloisonné work, with semi-precious stones mounted in individual settings. This was a Roman form of decoration in origin, but had become widely popular in barbarian Europe as a characteristic element of the ‘Danubian style’ that evolved in the Hunnic Empire. Even burial sites changed. In the sixth century, many of these new furnished burials were within new cemeteries well away from old habitation sites, the graves arranged in ordered lines (hence their technical German term Reihengräber, ‘row-grave cemeteries’).69 These centralized cemeteries presumably reflect some sense of community among an otherwise more dispersed rural population, rather like the large cremation cemeteries of East Anglia. All this firmly indicates that a new, non-Roman social order had come into being, and there is no doubting the degree of discontinuity it represents. What were its causes?

  Parts of north-eastern Gaul had suffered heavily in the raids of the later third century, and, unlike most of the Roman west, rural prosperity in some of its zones seems not to have recovered. But this was true only of a relatively restricted area to the west of the Lower Rhine. Throughout the fourth century, by contrast, Trier and the entire Moselle valley remained a hub of prosperous romanitas: in town, country and culture. The city itself was an imperial capital for many years. Further to the north-west in Picardy, likewise, an active villa culture seems to have survived the disasters of the third century, while the frontier continued to be heavily and actively defended by extensive fortifications and large numbers of troops. While the third-century crisis had caused some lasting disruption, the region between the Rhine and the Loire as a whole was by no means abandoned by the Empire, and active Roman life continued across much of it.70 Substantial structures of Roman life remained to be toppled here, therefore, before any new order could emerge.

  One body of evidence has sometimes been thought to show that a bout of pre-Merovingian Frankish immigration played a direct role in the process of imperial dissolution in these territories. Excavations of some late Roman cemeteries in the region have thrown up furnished inhumation burials ranging in date from c.350 to 450 AD. Unlike their Merovingian counterparts, these earlier inhumations are relatively few, just small clusters in cemeteries where the mass of burials entirely lacks any gravegoods. Males – whose graves predominate in these clusters – were buried with weapons and Roman military belt sets; a smaller number of females were buried alongside some of the men, with jewellery and personal items such as glass and pottery. These burials were first identified as a group by Hans-Joachim Werner, who argued that they were the graves of Franks known from historical sources to have been forcibly resettled on Roman territory in the 290s, as so-called laeti. He also saw the continued distinctiveness of these men and their descendants as an important contributory factor to the later Frankish conquest of the region in the time of Clovis: a sign that a first phase of Frankish settlement had disrupted the normal patterns of Roman life. As was pointed out by H.-W. Böhme, however, the graves date from a generation or two after the settlements of laeti mentioned in written sources, and, more importantly, are the remains of individuals of reasonably high status, whereas laeti were not even fully free. Böhme suggested, therefore, that the graves belong to the category of higher-status barbarian immigrant called foederati, linking them to a succession of Frankish officers known to have risen to high Roman rank in the fourth century.71 The graves, he argued, belonged to their slightly less distinguished peers. Nonetheless, Böhme’s argument kept the overall connection between the burials and an important group of immigrant Franks.

  Recently, however, Guy Halsall has challenged the whole idea that these graves belonged to immigrants at all, on the reasonable grounds that furnished inhumation was not the funerary rite practised by Franks beyond the frontier in the late Roman period. In fact, Frankish burials of the period c.350 to 450 (and, indeed, earlier) are undetectable in the confederation’s heartlands between the Rhine and the Weser. Enough work has been done in these areas to suggest that this is not just a gap in the evidence. Almost certainly, Franks in the wild disposed of their dead in an archaeologically invisible manner, quite likely cremation followed by a scattering of the ashes. It is also the case that the belt sets and weaponry contained in the furnished male burials on Roman soil were all of Roman manufacture. The idea that burial with weapons was a Germanic habit, Halsall argues, is an anachronistic back-projection from later Merovingian practice, when furnished inhumation did spread through the Frankish world. Rather than indicating that these graves were occupied by non-Romans, this fourth- and fifth-century group of furnished inhumations shows that a new, competitive burial practice was spreading among the would-be social leaders of the region. As imperial structures offered increasingly less support to those whom they had previously benefited, competition for social superiority began, and a new furnished burial ritual was one element in the process.72 This is obviously a variant of the same argument used in the case of fifth-century Britain, but in Gaul the new burial rite certainly came in long before there was any large-scale Frankish immigration.

  On balance, neither of these interpretations seems entirely convincing. The disparate nature of the find spots of these graves – cemeteries attached to military installations, rural settings, and even some in urban graveyards – indicates that, in life, the people being buried in this fashion were not a unified group, but individuals operating in a variety of contexts. It is certainly very difficult to see them, therefore, as any kind of Frankish fifth column. Both the chronology of the burials and the nature of the male gravegoods also suggest, more generally, that they were individuals working within and not against Roman imperial structures. But nor is the social-stress argument completely convincing either. For one thing, the furnished burials start too early (c.350 AD) to be associated with any major decline in Roman imperial power in the region, which even Halsall would not date before the late 380s, and I and others would actually date to the aftermath of the crisis of 405–8.

  The burials are also relatively few. If they were the products of a process of social competition, it was a very low-key one. And that the individuals concerned might just possibly have been Germanic immigrants of some kind (though not necessarily Franks) is suggested by the female graves that accompany some of the males. Not every male has a female counterpart, but in Picardy as many as half do, which is a strikingly healthy proportion. And while the goods buried with the men are undoubtedly Roman-made, their accompanying womenfolk were buried with ‘tutulus’-style brooches, which are otherwise found only in a group of rich Germanic burials from the Lower Elbe, far beyond the Rhine in Saxon country. The Elbe brooches are mostly of slightly different types from those found in northern Gaul, and the latter may be earlier in date. In that case, the brooches would not provide any reason for thinking these burials Germ
anic, since Roman fashions were often adopted by Germanic elites beyond the frontier. But, for the moment at least, the jury appears still to be out on these more technical matters, and if this brooch type does prove to be basically non-Roman we may still be looking at the burials of migrants who did well in the Roman system. Either way, Halsall is entirely convincing both that the burials have nothing obvious to do with later Merovingian-era burial habits and that, even if Germanic, they would provide no compelling evidence for a large-scale late Roman settlement of Franks between the Rhine and the Loire that facilitated Clovis’ later triumphs.73

  If the north–south divide within the sixth-century Merovingian kingdom cannot be traced back to a preceding Frankish settlement north of the Loire in the late Roman period, part of the explanation lies in the fifth-century political history of the region, in which Franks played a part. Arguments for a large-scale withdrawal of Roman power before the crisis of 405–8 are no more convincing here than they are in the case of Britain.74 But as the potent mixture of invasion and usurpation began to work havoc with the power base of the western Empire, then parts at least of northern Gaul, like Britain another fringe region of the Empire, began to feel a similar loss of protection. Thus Armorica – north-western Gaul (now Brittany) – revolted at the same time as Britain in 409/10, likewise perhaps from the control of the usurping Emperor Constantine III. But whereas Britain was cut permanently adrift from the Empire at this point, serious efforts were made in the 410s to bring northern Gaul back under the imperial umbrella, when the worst of the initial crisis had been weathered. And throughout the first half of the fifth century, periodic efforts were made to maintain imperial control north of the Loire: a mixture of direct interventions against breakaway groups, maintaining some regular Roman forces in the region, and occasionally implanting irregular ones.75

  In the longer term, however, these ongoing attempts to project imperial authority in northern Gaul were steadily undermined by the knock-on effects of the crisis of 405–8. As we shall see in the next chapter, the imperial centre progressively lost control of its key revenue-producing districts, and with them its capacity to maintain significant military forces and control its regional commanders. As a direct result, it could no longer protect the key structures of Roman civilian life. This all came to a head in the mid-450s in the additional chaos generated by the collapse of Attila’s Empire (Chapter 5) – the context in which Childeric rose to prominence in the 460s. Northern Gaul thus navigated its way from Roman past to Frankish future via a political process that was both long-drawn-out and highly contested. It began with the Rhine crossing of 31 December 406, and didn’t really come to an end until Clovis consolidated his power in the decades either side of the year 500. In the meantime, the region had seen many contestants for power: Roman central authorities, local self-help groups (often labelled Bagaudae after third-century bandit groups), barbarian invaders and settlers, and, eventually, Frankish military forces. The process was also essentially a violent one. So we should not wonder that the region’s Roman landowning elite suffered huge disruption. Their villas were rich and vulnerable, and here, as everywhere else where the capacity of Rome’s armies to provide security withered away, the villa network failed to survive the process of imperial collapse.76

  The role of the Franks themselves in all this would appear to have been substantial, but not primary. As we have seen, Frankish forces become prominent only towards the end of the process, in the 460s. This pattern stands in marked contrast to the history of lowland Britain, where the disappearance of any imperial protection was quickly followed by the arrival of Anglo-Saxon raiders, mercenaries and migrants who displaced sitting Roman landowners in a pretty direct fashion. Unlike those Anglo-Saxons, therefore, the Franks cannot be fingered in any simple way for the destruction of Roman life north of the Loire, which began long before the Franks emerged as a major military power. Indeed, given that Roman interventions in Frankish politics had probably aimed, as with the Alamanni, at preventing the emergence of larger and more dangerous coalitions among them (Chapter 2), the unified Franks should themselves be thought of as basically a post-Roman phenomenon, in the sense that Clovis’ career would not have been possible had the Empire maintained its full military and political capacities.77 But if there is good reason to separate the erosion of Roman life in northern Gaul from the effects there of rising Frankish power, what role did Frankish immigration play there in the Merovingian sixth century?

  Skulls and Sarcophagi

  In trying to answer this question, we are pretty much dependent upon archaeological evidence. Gregory of Tours does not deal with the issue of Frankish settlement, and the archaeology poses the same basic methodological problem that we have just encountered in lowland Britain. The new burial habit in northern Gaul coincides chronologically with the rise of Frankish power, but was everyone buried with gravegoods a Frankish immigrant? If so, we would be looking at something like a Völkerwanderung, since the new Reihengräber, replete with furnished burials, became widespread across northern and eastern Gaul (Map 12).78

  There have been many attempts to resolve the problem. Nineteenth- and early twentieth-century investigators were convinced that skull shapes could do the trick. Indigenous Celts, they argued, were brachycephalic (round-headed), whereas immigrant Germans were dolichocephalic (long-headed). Other scholars looked at details of burial rite. The use of sarcophagi was thought a uniquely Roman habit, and the same label was attached to burials without gravegoods, of which there are some in most Reihengräber. Sadly, none of these older methods works. There are no simple ethnic differences in skull shape, and explicitly documented Franks have been found in sarcophagi. Nor are the burials without gravegoods any more useful. These show a marked tendency to cluster on the edges of the Reihengräber, and modern excavation methods have shown that use of these cemeteries started in the middle and worked outwards. The real explanation for the absence of gravegoods is chronological. From the seventh century onwards (as was also the case in longer-lived Anglo-Saxon inhumation cemeteries – if cemeteries can have a long life), there was a marked decline in the use of gravegoods, probably under the influence of Christianity, until funerary rites returned to the unfurnished burial typical of the late Roman period.79

  If no simple method has yet emerged for telling Romans from Franks, there is some entirely convincing evidence that some of the men buried with weapons in the sixth century were of indigenous Gallo-Roman stock. An excellent case study is the large, carefully excavated cemetery at Krefeld-Gellep on the Lower Rhine in northern Germany. This is one of few burial sites that remained in continuous use from the late Roman into the Merovingian period. In c.500 AD, close to an existing late Roman cemetery, a second cemetery was inaugurated with the interment of one richly furnished burial and subsequently continued in use with a pretty standard collection of Merovingian furnished burials. Opening the second cemetery, however, did not lead the first to close, and what happened there is striking. Furnished inhumation – with standard Merovingian collections of weapons and jewellery for males and females, respectively – quickly became the normal mode there too. The original excavator and secondary commentators have all concluded, surely correctly, that in the first cemetery an existing late- or now post-Roman population adapted itself to the new cultural norms inaugurated by the rich burial in cemetery two: a beautiful case of elite emulation at work. And what happened so demonstrably at Krefeld-Gellep is likely to have been happening right across northern Gaul, in other areas where existing late Roman cemeteries were entirely replaced by Reihengräber, and where, consequently, it is not possible to demonstrate a similar cultural evolution. Many of these new Reihengräber surely must include populations of Gallo-Roman descent who adapted themselves to the new Merovingian-era norms.80

  A further body of evidence which seems, in part at least, to reflect this process consists of those cemeteries whose use began with one particularly rich burial, like that second cemetery at Krefeld-Gellep. Thank
s to a half-century of careful work on the chronology of Merovingian material culture, this pattern has been documented at a whole string of sites: Mézières in the Ardennes, Lavoye on the Meuse, Pry, Gutlingen, Chaouilley in Lorraine, Rübenach, Héruvillette and Bale Berning. It might be tempting to think of many of the burials in such cemeteries as those of Frankish immigrants, but you can’t just assume so. This has been demonstrated at Frénouville in Calvados. Here, a freak genetic marker in the form of a cranial suture shows that substantially the same population remained in place before and after the rise there of furnished inhumation, even though the new rite appears in an entirely new Merovingian cemetery. The furnished-burial habit was imported into Frénouville, it would seem, by just the one elite family responsible for the rich ‘founding’ burial in the new cemetery, but by the mid-sixth century it had spread right across an indigenous population that now also used the new cemetery.81 The evidence is unimpeachable, therefore, that, in part, the new habit of furnished inhumation spread so widely over northern Gaul because the indigenous Gallo-Roman population adopted it with enthusiasm.

  But what was the origin of this rite, and why was it adopted? One strand within modern archaeological theory tends, as we have seen, to interpret clusters of relatively rich burials as evidence of social insecurity and competition. Because status was not clear, families aimed to put on a competitive show for their neighbours. Such an interpretation is clearly possible in this instance. The rise of the Merovingian kingdom enforced, as we have seen, rules on social status that were new compared with its Roman predecessor, and you can see precisely why this might have sparked off social competition.82 But this is not the only possible explanation, nor in this case even the most persuasive one. For one thing, the same remaking of the social order happened south of the Loire, but this did not spark off burial competition. You could reasonably counter this with the fact that there the old elites survived with many more of the building blocks of their social distinction intact, so that there was no need for them to compete; but there are other more significant objections to the social-stress theory as well. Above all, the neatly laid-out Reihengräber look much more like highly organized communal spaces than arenas of intense social competition. Their carefully ordered lines strongly suggest that burial was being policed in some way. As indeed do the goods put in the burials: particular categories of dead – adult but not elderly males and young women at the height of childbearing – were consistently buried with more goods than others. Again, it has been suggested that this is because these deaths were more stressful, but the law codes tell us that this is when wergild was at its height, and that could just as well be the key point.