Empires and Barbarians

by Peter Heather

  One caveat, though, needs to be added. As generally conceived, wave of advance is a model of random movement, whereby the buildup of population at one point leads subgroups from that settlement to move on in the next generation to the nearest available piece of suitable land. One application of this model to the spread of Europe’s first farming populations suggested that the mathematics of such a process dictate that a population spreading over a landscape by this means might make an aggregate advance of around a kilometre a year. But Korchak Slavs went from the fringes of the Carpathians to the Elbe–Saale region, a distance of around 900 kilometres, in only a hundred and fifty years. This is a sufficiently faster rate to suggest that some of the assumptions normally inherent in the wave-of-advance model did not apply in this case. One possible explanation is that Slavic movement – like the spread of Frankish settlers in northern Gaul – was not random. A Byzantine military treatise called The Strategicon of Maurice reports that some Slavs preferred to inhabit wooded uplands rather than more open plains, and the ribbon of Korchak sites through upland central Europe might be taken as some confirmation of this statement. If true, the choice of destination for each new Korchak generation was limited to similar and particular environments, and in fact this does all make a kind of sense. Most of the open plains of Europe were dominated by larger political powers (whether Byzantines, Franks or Avars), so if you wanted to live independently in the kind of small community characteristic of the Korchak world, lowlands were not an option. For Korchak Slavs, migration was a means to carry on traditional lifestyles, including a very small scale of social organization, quite probably at a time of population expansion.43

  Seventh-century Slavic settlement in the Balkans, by contrast, was undertaken by considerably larger units: tribes, for want of a better word. Around Thessalonica, a series of larger named Slavic groups were already settled in the valley of the River Strymon in the middle years of the seventh century. Our source here, the Miracles of St Demetrius, also tells us that another Slavic group mentioned earlier, the Belegezitae, held land somewhere further south. Further south in the Peloponnese, likewise, named Slavic groups existed in the early ninth century – the Milingas and the Ezeritae. The same pattern is also found in ninth-century Slavic Bohemia, and the wider regions covered by the Anonymous Bavarian Geographer. In these central European cases, and possibly also the Peloponnese, the larger named groups probably did not migrate as whole units into the regions where the literary sources later find them, but were later evolutionary creations within that landscape from a much more fragmented, wave-of-advance type of migration. Bohemia, at least, was originally settled by Slavic migrants generating a ‘pure’ Korchak outcome, so that its more organized structure in the ninth century was apparently a later development. It does not seem possible, however, to explain the size and organization of the seventh-century Belegezitae or the other groups settled around Thessalonica as the products of a post-migration process. These areas have produced no evidence of an initial Korchak outcome, and the historical evidence for the tribes’ existence dates from shortly after the initial migration. The text of the Miracles is contemporary and local, recording an incident of c.670, while the settlement itself, as we’ve just seen, cannot have happened before the 610s. In this case, the time lag, barely two generations, seems insufficient for a whole new sociopolitical order to have emerged from a flow of extended family units.44

  So how should we envisage these larger groups? Historical sources consistently describe the early Slavs as living in small sociopolitical units, but how small was small? Some were very small. The Korchak system was probably being transported about the European landscape in the sixth and seventh centuries by social units less than a hundred strong. As an important recent study has rightly stressed, however, other parts of the Slavic-speaking world underwent a major sociopolitical revolution in the sixth century. In the many pages of his histories devoted to a wide range of Slavic activities in the period c.530–60, our east Roman historian Procopius names no individual Slavic leaders at all. In the last quarter of the sixth century, however, the pattern suddenly changes. In a number of different east Roman sources various Slavic leaders appear, with enough circumstantial detail to show that we are dealing with substantial political figures. The territory ruled by a certain Musocius, for instance, took three days to cross, suggesting that it covered somewhere between 100 and 150 kilometres. The rule of another leader, Ardagastus, was solid enough, likewise, to survive for the best part of a decade between 585 and 593. Perigastes had enough forces under his command to kill a thousand east Roman soldiers, while another named figure, Dabritas, was confident enough of his military strength to kill the diplomatic envoys of the Avar Khagan, boasting with the suave masculine charm typical of the period: ‘What man has been born, what man is warmed by the rays of the sun who shall make our might his subject? Others do not conquer our land, we conquer theirs.’

  Territories extending over a hundred kilometres, even with relatively low population densities, indicate social units of several thousand individuals, and this is confirmed by the one plausible overall figure to survive. After east Roman assault destroyed the domain of Ardagastus, a quarrel broke out between the Romans and the Avars over who should have control of the prisoners. It was eventually decided in the Avars’ favour, and the Romans duly handed over five thousand individuals. This figure is consistent with the new Slavic kings of the late sixth century commanding populations of something like ten thousand, but not several tens of thousands. And while not huge, we are clearly talking here of an entirely different order of magnitude from the kind of social units involved in spreading Korchak culture further north. And if we can estimate from the Ardagastus incident a rough figure for the larger Slavic groups that had coalesced on the fringes of east Roman territory by c.600, this would suggest that the four groups settled in the region of Thessalonica comprised between them several tens of thousands of Slavic immigrants. For what it’s worth, this also fits with Byzantine reports that a later pacification of the area, in the 680s, involved transferring thirty thousand Slavs to Asia Minor.45

  Serbs and Croats might represent yet a third type of migrant group caught up in the Slavic diaspora of the sixth and seventh centuries. There is obviously a huge margin for error built into the tenth-century traditions retold by Constantine Porphyryogenitus, but if there is any truth to them at all, the Serbs and Croats were breakaways from the Avar Empire which had previously used them in a military capacity. Avar campaigns between 570 and 620 were many and varied, and this would provide a plausible context for a further bout of sociopolitical evolution among those Slavs caught up in this latest nomad war machine to establish itself in central Europe, sufficient to produce this third type of Slavic migrant group that was either large enough or militarily specialized enough to throw off Avar domination. It might be the same kind of force as the five thousand militarily ‘elite’ Slavs who launched a surprise attack on Thessalonica. If so, Serb and Croat migration might have taken the form of a kind of elite transfer, with a militarily effective force breaking out of the Avar Empire and establishing its own niche in the Balkans.46 This is speculative, but well within the bounds of possibility, and we do have independent contemporary evidence that the evolution of Slavic society was throwing up military specialists at this time. At the very least, it underlines exactly how many and varied were the migratory processes that get lumped together retrospectively to account for the ‘Slavicization of Europe’.

  A comparison of the historical and archaeological evidence thus sets up a seeming paradox. Those regions of Europe that saw the complete transfer of a Korchak-type material-cultural system also witnessed a Slavic migration process involving only very small social units. On the other hand, the historical evidence for much larger Slavic social units on the move (whether ‘tribal’ or military specialists – if such, indeed, were the Serbs and Croats) relates to those areas where archaeologists have not uncovered any large-scale transfer of ‘com
plete’ Korchak-type socioeconomic systems. This is at first sight surprising. The larger the migration unit, you might suppose, the greater its capacity to import and maintain its own distinctive way of life. When you think about it, though, the larger Slavic social units were actually very recent creations, generated by processes of rapid sociopolitical and economic development that were unfolding among those Slavs closest to the Roman frontier or who became caught up in the Avar Empire. We will return to these processes in Chapter 10, but there is every reason to suppose that much of the momentum behind them was generated by a dynamic interaction between the groups involved and the opportunities and dangers that came their way from an unprecedented proximity to the bigger and materially rich east Roman and Avar Empires. In other words, it was precisely the larger Slavic groups rather than the small-scale farmers of the Korchak world who would have been more open to the kinds of influences and processes that would have caused their patterns of material culture to evolve away from older Korchak-type norms.

  Because of the lack of information, there is no point in spending much time on the issue, but it is worth reflecting on what all this suggests about the kind of Slavic migration units which were operating in those contexts that are entirely undocumented in the surviving historical sources: north-central Europe in the seventh and eighth centuries, and European Russia in the eighth and ninth. In central Europe, between the Vistula and the Elbe, archaeologists have revealed a third kind of outcome. The Sukow-Dziedzice culture certainly saw the absorption by incoming Slavs of some existing patterns of indigenous material culture, notably its repertoire of pottery. But the Mogilany culture, which started the process of Slavicization, is really a Korchak variant, and even the Sukow-Dziedzice culture, in its earliest levels, did not depart far from these norms. In its original ‘islands’ (Map 18), settlement originally came in the form of small open villages, similar in size to the Korchak norm, but the buildings were usually above ground rather than sunken huts. Although they absorbed more of native culture than elsewhere, the original Slavic migration units probably differed little in size from those that created Korchak Europe. Why they should have departed from Korchak norms as far as they did, is a question we will return to in a moment.

  For European Russia, the only evidence we have for the migration process is again archaeological and hence, by its nature, indirect. But some of the settlement patterns, like those from Korchak Europe, are again highly suggestive of the type of social unit engaging in the expansionary process, and hence shed at least some light on that process itself. Take, for instance, the Borshevo-era hilltop site at Novotroitskoe in the Psiol valley in northern Ukraine. Here, the steep sides of the hill form natural defences, and the excavators found clustered together about fifty sunken-floored huts dating from the eighth and ninth centuries. This indicates that the total population of the settlement must have been just a few hundred people. The choice and nature of the site itself are enough to suggest that prevailing conditions were far from peaceful, as does the end of this initial period of its occupation, when it was apparently destroyed by raiders. Novotroitskoe is not an isolated example. Romny-Borshevo settlements were customarily situated in highly defensible situations on hilltops or in swamps, often walled, and they generally hosted a similarly dense clustering of population.

  All this suggests two things. Most obviously, the progress of Slavic-speakers over this landscape was far from uncontested. You go to the trouble of building this type of settlement only if you need to, and its eventual fate does suggest that it was necessary. Second, and this follows from the first observation, Slavic expansion in this region was probably being conducted by groups big enough to build and maintain settlements of this type. If expansion was contested, small groups could not just pitch up in a new area. They had to come with sufficient strength to construct a settlement in which they could protect themselves and their families.

  Despite the lack of historical description, therefore, it seems that the migration units operating from the eighth century in north-western Russia were considerably larger than those that had earlier spread Korchak culture and its variants across the central European uplands and east of the Carpathians into southern Russia and Ukraine. Their defended settlements stand in marked contrast to the small undefended ones of the Korchak, Penkovka and even Kolochin systems of the sixth and seventh centuries, emphasizing the degree to which the later centuries marked a new era in the nature of Slavic expansion. We are still looking here at an expansion that began with something akin to a wave of advance rather than the sudden occupation of an entire landscape, but it evolved over time, until larger social units were eventually moving into contested areas. Overall, Slavic expansion into European Russia may well have taken a form we’ve seen in other contexts, ancient and modern, where a flow of small-scale social units builds up momentum and is forced to reorganize itself into larger groups when it eventually encounters serious opposition, as the Goths and others did in the third century, the Vikings in the ninth and the Boers in the nineteenth.47

  The range of evidence available for the nature and scale of Slavic migration flows bears not the remotest resemblance to anything you might consider an ideal data set. But this is all part of the fun of early medieval history, and in any case, it is still sufficient to show that Slavic expansion took a variety of forms, as we would anyway have to suppose given the wide variety of contexts it encompassed. At one end of the spectrum we have the transfer of replica Korchak-type settlements from the foothills of the Carpathians across wide tracts of central and eastern Europe from the Elbe to Ukraine. In the Romny-Borshevo era further to the north and east, by contrast, more substantial settlements were the norm, constructed by Slavic population units several hundred strong. Different again was the movement of entire ‘tribes’ into the former Roman Balkans in the seventh century, where the groups may have been up to ten thousand strong, if it is correct to think of them as the kind of unit run by an Ardagastes or a Perigastes taking to the road. With so few sources, the details of all this could be argued over endlessly, and there is a large margin for error. But the Slavicization of Europe clearly involved a wide range of migratory activity, with unit sizes extending from family groupings at one extreme to social units in the thousands at the other.

  Immigrant and Native

  Composed around the year 600 AD, the east Roman military treatise often attributed to the Emperor Maurice (582–602) includes a fascinating comment on the approach of some early Slavic groups to prisoners taken on their raids:

  They do not keep those who are in captivity among them in perpetual slavery, as do other nations. But they set a definite period of time for them and then give them the choice either, if they so desire, to return to their own homes with a small recompense or to remain there as free men and friends.48

  This immediately raises the basic intellectual problem involved in trying to understand the astonishing rise to prominence of Slavic-speakers all the way from the Elbe to the Volga in the early Middle Ages. On the one hand, there is no one who supposes that this could have happened without an element of migration: actual Slavic-speakers on the move across the landscape. On the other, old culture-historical, quasi-nationalist visions of the Slavs as a ‘people’, a single population group that started from one geographical point and then went forth to multiply over vast tracts of the European landscape, are not credible. In similar vein, although they did add up in aggregate to substantial population movements, the earlier, largely Germanic migrations of the fourth to the sixth century were certainly not large enough to create entirely empty landscapes in the vast tracts of Europe that were affected by Germanic culture collapse. Most of these areas reappear in Carolingian sources under new, Slavic, management, but the original Slavic migrants were mostly interacting with an indigenous population. The two key issues we need to explore, therefore, are, first, just how big a demographic event was Slavic migration itself; and, second, what kinds of relationship did incoming Slavs form with the ind
igenous populations they found at their various points of destination?

  Comprehensive information is not available, but there is good reason to suppose that incoming Slavs did encounter a substantially reduced population in areas affected by Germanic culture collapse, and even, in a few localities, some entirely abandoned landscapes. For just a few areas, general settlement surveys are available. In Bohemia, for instance, there seems to have been a substantial decline in population in the late Roman period. Twenty-four major find-spots (mostly cemeteries) are known from the early Roman period, but this declines to just fourteen in late antiquity. Slavic immigrants into Bohemia encountered not an empty landscape, therefore, but certainly a smaller population than the region had previously carried. Elsewhere, pollen diagrams provide further insights. Pollen is carried in the wind and, on landing, will sink to the bottom of standing water. A core can then be extracted from the bottom, particularly of lakes, allowing changes in the varieties of pollen deposited over time to be charted. Continuous activity from an indigenous farming population shows up as an undisturbed pollen sequence, with no great rise in tree or grass pollen, and no major change in the range of cereals being produced. Pollen diagrams are unavailable for much of eastern Europe, but they do indicate that in some places a substantial indigenous pre-Slavic population was not displaced. Samples from the Baltic island of Rügen and from Saaleland show more or less total continuity from the Roman into the Slavic periods, even though both passed into Slavic control at some point before 800 AD. But in other areas, a different picture has emerged. In large parts of Mecklenburg in the former DDR, the pollen diagrams indicate great disruption to established farming patterns in the same period. Here, at least, it would seem, Slavic-era immigrants more or less started farming the landscape again from scratch. Similar evidence for disruption and forest regeneration has also emerged from Biskupin in modern Poland.49