Empires and Barbarians

by Peter Heather

  Shortly afterwards, mixed groups of raiders exploited a second line of attack, crossing the Black Sea to Asia Minor by ship in three successive years, 255–7.16 The first expedition, unsuccessful, was directed at Pityus on the south-eastern shore of the Black Sea. The second successfully sacked both the previous year’s target and the city of Trapezus (modern Trabzon). These initial raids were undertaken by what our main source calls ‘Boranoi’, a name that perhaps just means ‘northerners’. The third, seemingly much more substantial, expedition of 257, this one explicitly including Goths, caused widespread devastation in Bithynia and the Propontis, inflicting damage on the cities of Chalcedon, Nicomedia, Nicaea, Apamea and Prusa. There is then a gap in our sources – which, for all their problems, again probably reflects a cessation or lessening in the intensity of the attacks – until 268, when an enormous maritime expedition left the northern shores of the Black Sea. It was composed again partly of Goths but also of some other Germani, notably Heruli. The new expedition did not sail straight across the Black Sea but moved along its northern and western coasts, keeping within sight of land and raiding some coastal cities, such as Anchialus, as it went. Other assaults on Tomi, Marcianople, Cyzicus and Byzantium were beaten off. The raiders then forced the Dardanelles, and spilled out into the Aegean. For the first time, northern sea raiders had broken into Rome’s Mediterranean lake. There the expedition divided into three main groups. These attacked, respectively, the northern Balkans around Thessalonica, Attica, and the coastal hinterland of Asia Minor. The Emperor Gallienus began the counterattack in the Balkans, but it was his successor Claudius who inflicted a massive defeat on the Balkan groups in 269, winning the sobriquet ‘Gothicus’ – ‘victor over the Goths’ – for his efforts. The struggle against the Heruli around Athens was led amongst others by the historian Dexippus, while the third group, led by the chieftains Respa, Veduc and Thuruar, was eventually driven back into the Black Sea in 269, but not before it had wreaked havoc. The islands of Rhodes and Cyprus were devastated, as were the cities of Side and Ilium on the mainland. The raid’s most dramatic casualty was perhaps the legendary temple of Diana at Ephesus.17

  The Roman response was fierce. Not only was each of the individual groups defeated, but no major raid ever again broke through the Dardanelles. As with the Danube after the defeat of Decius, one can only presume that effective countermeasures were put in place to seal off the line of attack. Not that this was the end of the Gothic problem. A further attack across the Danube occurred in 270, when Anchialus and Nicopolis were sacked, but the new Emperor Aurelian then led his forces north of the river in 271 and thoroughly defeated a Gothic leader called Cannabaudes, who had presumably been responsible for the latest outrages. Aurelian’s counterattack nipped the new threat in the bud. The mid-270s saw some further sea-borne raids, which plundered the Pontus in particular, but no further assaults over the Danube into the Roman Balkans. Not only had the Emperor’s defeat of the Goths brought some relief, but he had also organized, more or less simultaneously, a planned evacuation of Transylvanian Dacia.18

  As with the parallel withdrawal from the Agri Decumates in the west, our information about the abandonment of Dacia is limited. But both narrative evidence and the coin hoards indicate that most of the attacks of the third century had skirted Dacia’s frontiers and entered the Balkans proper, or crossed the Black Sea into Asia Minor, rather than directly affecting the province itself. This withdrawal would again appear to have been more by strategic design, therefore, than a headlong retreat from direct military disaster. On one level, Aurelian probably had it in mind to shorten his frontier lines. Dacia was a projecting salient north of the Danube, which needed defending on three sides. By evacuating it, the Roman frontier in south-eastern Europe could be reduced by something like eight hundred kilometres. It also gave the troublesome outsiders a new prize to squabble over, diverting them from making further attacks on Roman territory. Writing in the fourth century, Eutropius notes that Dacia was ‘now’ (369) divided between the Taifali, Victohali and Tervingi. Aurelian’s combination of military success and strategic withdrawal took much of the steam out of the cross-border attacks, but it was to be another generation before order was fully restored on Rome’s Danube frontier.19 As on the Rhine, further campaigns by the Tetrarchs and Constantine were required fully to force the Goths and others into the semi-client status in which we encountered them in the last chapter.

  But who exactly were the Goths who feature so strongly in the third-century action, and what underlay these two or three generations of large-scale disturbance on the east European frontiers of the Roman Empire?

  There is no doubt at all that the emergence of Gothic domination represented a complete revolution in the nature of the threat facing the Roman Empire across its Lower Danube frontier. In the first and second centuries, Rome had mostly faced a mixture of nomadic Iranian-speaking Sarmatians and settled Dacian-speakers in this theatre of operations. By the fourth century, groups labelled ‘Goth’ had become the main focus of Roman campaigning and diplomacy in the region. The Gothic Tervingi, as we saw in Chapter 2, became the Empire’s main client beyond the Lower Danube, and as the events we have just summarized demonstrate so clearly, the intervening century had seen a huge increase in the military threat posed to the Empire across both its land and its water frontiers. There had been no attacks via Dacia, across the Black Sea or through the Dardanelles on anything like a similar scale in the first and second centuries.

  The traditional response to these observations has always been to suppose that Germanic migration was a key ingredient of this strategic revolution. ‘Goths’ were nowhere a presence north of the Black Sea in the first and second centuries AD, when Sarmatians and Dacians are the only two groups to be mentioned in the region. The only Goths we hear about at this time were established in northern Poland. So, game, set and match, you might think, to migration? Well, not exactly. It has recently been argued by Michael Kulikowski that the traditional view of the developing situation north of the Black Sea is a ‘text-hindered’ fantasy. This is a term borrowed from the jargon of archaeologists (although Kulikowski is himself not one), and is used to describe a situation where the interpretation of archaeological evidence has been bent out of shape by a determination to make it conform to the available historical evidence. In this instance, among the range of far from wonderful historical materials available to us for the third century is a sixth-century Gothic history, written by a man called Jordanes, which records the migration of Goths to the Black Sea under a certain King Filimer. This account, Kulikowski argues, not only has little credibility in itself, but has also unduly influenced how historians and archaeologists have looked at the other evidence. Without it, in his view, the other archaeological and historical evidence would not make anyone think in terms of migration. What really underlay the troubles of the third century, and the emergence of Gothic domination in the fourth, was not migration at all but sociopolitical reorganization among the region’s existing population: in fact, of broadly the same kind that produced the new Germanic confederations of the late Roman west.20 Is he correct?

  Two elements of the argument are convincing. First, there’s not the slightest doubt that socioeconomic and political reorganization – ‘development’ – were an important dimension of the story. The Gothic Tervingi of the fourth century had a complex, confederative political structure, developed social hierarchies, and an economic profile both in production and exchange that went far beyond the norms of first-century Germania. Their political structures were based on hereditary power, and robust enough both to survive major defeats and to develop coherent strategies for overturning their worst consequences. Second, Kulikowski is right enough that little reliance can be placed on Jordanes. Jordanes was writing three hundred years after the event, and can be shown to have produced a completely anachronistic view of the Gothic world of the fourth century, on which more in a moment.21 If he can be so wrong about fourth-century Gothic history, th
is must call his account of the third century into question, even if we don’t have enough contemporary sources to be able to check it systematically. Even conceding these points, however, there is still more than enough good-quality evidence to establish that Germanic migration from the north was a major factor in the strategic revolution of the third century.

  It does need to be emphasized, first of all, that the change in the nature of the forces Rome was facing across its Lower Danube frontier was much more profound than a mere change in labels. In the first two centuries AD, the eastern foothills of the Carpathian range – modern Moldavia and Wallachia – were occupied by a number of Dacian groups who had not been brought under direct Roman rule at the time of Trajan’s conquest of Transylvania. In the course of the third century, they generated a new degree of political unity among themselves and came to be known collectively as the Carpi. The main Sarmatian group immediately north of the Black Sea was the Roxolani, who, together with the Iazyges, had dismantled the dominance of the Germanic-speaking Bastarnae in the region at the start of the first century AD. Where the Iazyges had subsequently moved on to the Great Hungarian Plain, west of the Carpathians, the Roxolani stayed east, exercising hegemony over the ancient Greek cities of the Pontus, which retained some independence into the third century. Both Sarmatians and Dacians became at least semi-subdued Roman clients after Trajan’s conquest of Transylvanian Dacia, even though they were not formally incorporated into the Empire. The sudden dominance of Goths and other Germanic-speakers in the region represented, therefore, a major cultural shift. And there is no doubt that the new Gothic masters of the landscape were Germanic-speakers. The Gothic Bible translation was produced for some of them by Ulfila, the descendant of Roman prisoners captured by the Goths from Asia Minor, and its Germanic credentials are irrefutable. The appearance of the Goths thus represents a massive change in the complexion and identity of the forces lined up on Rome’s north-eastern frontier.22

  This, of course, was not the first time that Germanic-speakers had provided the dominant population stratum in the region. The Bastarnae, subdued by the Sarmatians around the beginning of the first millennium, had also been Germanic. So in theory it might be possible to explain the rise of Gothic domination north of the Black Sea in the third century as the re-emergence of those Germanic groups who had been subordinated here in the first. However, a pretty extensive range of evidence suggests, on the contrary, that the immigration of new Germanic-speakers played a critical role in the action.

  In the period of Dacian and Sarmatian dominance, groups known as Goths – or perhaps ‘Gothones’ or ‘Guthones’ – inhabited lands far to the north-west, beside the Baltic. Tacitus placed them there at the end of the first century AD, and Ptolemy did likewise in the middle of the second, the latter explicitly among a number of groups said to inhabit the mouth of the River Vistula. Philologists have no doubt, despite the varying transliterations into Greek and Latin, that it is the same group name that suddenly shifted its epicentre from northern Poland to the Black Sea in the third century. Nor was it the only group name to do so at this time. Goths get pride of place in our sources and in scholarly discussion, but other Germanic groups participated in the action too. We have already encountered the Heruli, and late third-and early fourth-century sources record the presence in and around the Carpathians, in addition, of Germanic-speaking Gepids, Vandals, Taifali and Rugi. The Rugi, like the Goths, had occupied part of the Baltic littoral in the time of Tacitus, and the likeliest location for Vandals in the same period is north-central Poland, to the south of the Goths and Rugi. The presence of Vandals and Rugi in the Carpathian region, alongside Goths, represents a major relocation of some kind on their part, and all were moving south and east from Poland towards the Pontus. The Heruli are not mentioned by Tacitus, but in the fourth and fifth centuries a second, non-Danubian, group of Heruli again lived far to the north-west, suggesting, again, that our Danubian Heruli may have got there via some kind of migration. The Gepids and Taifali, like the Heruli, are first encountered at the end of the third century, and we will return to the significance of these ‘new’ Germanic-speaking groups later in the chapter.

  There is, of course, more that we would like to know, but despite obvious deficiencies the historical evidence in its entirety strongly indicates that a wave of Germanic expansion – moving broadly northwest to south-east – underlay the strategic shift that led Aurelian to abandon upland Transylvania. This has to be deduced. There is no explicit description of Germanic migration in contemporary Roman sources, which confine themselves to accounts of its effects – attacks by these new groups across the Roman frontier. If ‘Goth’ was the only Germanic group name from north-central Europe to shift its location in these years, you might get away with the argument that it’s a case of accidental resemblance, but, as we have just seen, it isn’t only ‘Goth’. This being so, there is no reason not to accept what the historical evidence is prima facie telling us. In a reversal of the effects of the arrival of the Sarmatian nomads in the first century AD, the hegemony of Germanic-speakers east of the Carpathians, lost in the overthrow of the Bastarnae and their allies, was restored by the migration of Goths, Rugi, Heruli and other Germanic-speaking groups in the third century.23

  This interim conclusion is only strengthened by two broader aspects of the historical evidence. First, the rise of Gothic power north of the Black Sea eventually led some indigenous groups to evacuate the region entirely. As we shall soon see in more detail, large numbers of Dacian-speaking Carpi (but not all of them) from the Carpathian foothills were admitted into the Roman Empire in the twenty-five years or so after 290 AD. An increased level of competition between groups already indigenous to the region might conceivably have generated such an exodus, but it is much more consistent with the after-effects of substantial Germanic immigration. Second, the new Gothic populations of the region remained highly mobile, even after moving into the plains south and east of the Carpathians following the exodus of the Carpi. In the 330s, the Gothic Tervingi contemplated moving lock, stock and barrel to the Middle Danubian region, and from the 370s, as we will explore in the next chapter, relocated along with their fellow Gothic Greuthungi to new homes in the Roman Empire. This later mobility is relevant, because, as we have seen, comparative studies have consistently shown that migration is a cultural habit that builds up in particular population groups. Finding Gothic populations mobile in the fourth century provides a further reason for accepting the evidence that they – or their ancestors – had been so in the third. Neither of these points would be conclusive by itself, but both reinforce the historical evidence that Gothic migration played a large role in recasting the strategic situation north of the Black Sea in the third century.24

  The archaeological legacy of the Cold War, moreover, again allows us to expand the discussion beyond the limits of the historical sources. Between c.150 and c.220/230 AD, there occurred a further large-scale south-eastern expansion of the Wielbark cultural system into Polesie and Podlachia first of all, and then on into Volhynia and northern Ukraine. This entirely dwarfed in its geographical scale the earlier Wielbark expansion from around the time of the Marcomannic War. At the same time, Wielbark sites and cemeteries in western Pomerania were falling out of use, so the shift in the Wielbark centre of gravity was huge (Map 6). Given that certainly the Goths and probably at least the Rugi, too, among the newly dominant Germanic groups of the Black Sea region had their origins within the Wielbark system in the first and second centuries, these finds are highly suggestive, in fact, of the route followed by some of the Germanic-speakers who ended up by the Black Sea. A ribbon of Wielbark cemeteries of more or less the right date has been traced south along the upper reaches of the River Vistula, and then on to the Upper Dniester (Map 6). These certainly tie in chronologically with the sudden appearance of Gothic attackers outside the walls of the city of Histria in 238.25

  The really striking development in the north Pontic archaeology of this period, howeve
r, was not the further spread of the Wielbark system per se, but the generation of a series of new cultural systems incorporating some Wielbark features. The most important of these was the Cernjachov, which by the middle of the fourth century had spread over a huge area between the Danube and the Don (Map 6). This is another case where the date and identity of the system used to be much fought over, but its basic characteristics are now well established. Over five thousand settlements have been identified, and many large bi-ritual cemeteries excavated, the remains showing beyond doubt that the system flourished from the second half of the third century down to the year 400, or just a bit later. Chronologically, as well as geographically, its remains coincide with Gothic dominion in the late Roman period as described in trustworthy contemporary sources, and it is now universally accepted that the system can be taken to reflect the world created by the Goths – and probably our other Germanic-speakers too – north of the Black Sea.

  Some elements of the new system strongly recall, or are identical to, their counterparts in the Wielbark system to the north-west, but it is important to recognize that the latter did continue on in its own right; there was nothing like a total evacuation of northern Poland. Some of the pottery is identical, with handmade bowl-shaped Wielbark ceramics being particularly prevalent in early Cernjachov levels. Otherwise, many of the fibula brooch types, and the style of female costume (brooches worn as a pair on each shoulder), are identical with those found in Wielbark areas. Some house-types, particularly the longhouses shared by both humans and animals (German Wohnstallhäuser), are likewise common to certain areas, at least, of both systems. It is also striking, although at present a full comparative study is lacking, that the two customs which distinguish Wielbark cemeteries from those found in surrounding areas of north-central Europe are also found in Cernjachov territories. In cemeteries of both systems, two types of burial ritual coexisted: inhumation and cremation. Likewise, the population of Wielbark areas did not bury weapons (or any other iron objects) with their male dead, and the absence of this habit was also a feature of Cernjachov burial ritual.