Romantic nineteenth-century conceptions of early Germanic society, framed at the height of nationalist fervour, propounded the notion of early German Freiheit, ‘freedom’: the idea that Germania before the birth of Christ was a world of free and equal noble savages, with no intermediate nobility but with kings who were directly answerable to assemblies of freemen. This was mistaken. Even in the time of Tacitus, Germanic societies had slaves, though the slaves ran their own farms and handed over part of the produce rather than living under closer domination as unfree labour on someone else’s estate. And although the material remains of the Germanic world in the last few centuries BC show no obvious distinctions of status, this does not mean that there weren’t any. Even in a materially simple culture – and in the third century BC about the greatest sign of social distinction available among the Germani of north-central Europe was to keep your clothes on with a slightly fancier safety-pin – differences of status can still make a huge difference to quality of life. If higher status translated merely into eating more, doing less hard manual labour and having a better chance of passing on your genes successfully, it was nonetheless extremely real, even if it could not be expressed in the possession of much in the way of fancy material goods. I doubt very much, in fact, that the status distinctions we find in Tacitus were new to the Germanic world of the first century AD, even if they can’t be measured easily in archaeologically visible material items over the preceding centuries.22
That said, the evidence is entirely compelling that pre-existing inequalities grew dramatically during the Roman period. We have met some of this already. The new military kings and their retinues, those at least who prospered, were one set of beneficiaries from the new wealth. Archaeologically, their rise is reflected in two ways: burial practice and settlement remains. There is no simple correlation between wealth of gravegoods and status in life. Really rich graves (called Fürstengräber, ‘princely burials’, in the germanophone literature) cluster chronologically with, broadly speaking, one group at the end of the first century and another at the end of the third: the so-called Lübsow and Leuna Hassleben types respectively. It is not credible, though, that a dominant social elite existed only at these limited moments, and it has been suggested that their appearance may mark periods of social stress, when new claims to high status were being made – claims by the individuals running the funeral, of course, rather than the dead persons themselves. Nonetheless, over the long term, changing burial practices certainly reflect the impact of new wealth. Before the last few centuries BC, Germanic funerary rites seem to have been pretty much identical for all, a little handmade pottery and the occasional personal item being all that the cremation burials of the period characteristically contained. In the Roman period, by contrast, not only are there the clusters of extremely rich princely burials, but also a substantial minority of the other burials started to contain increasing numbers of gravegoods, often including weapons with males and jewellery with women. Monumentalizing graves was another strategy for claiming status in some parts of Germanic Europe, particularly Poland, where groups of burials were marked out as special by piling up stones to create barrows, and individual graves by erecting standing stones (stelae). The Wielbark cemetery at Odry, for instance, turned up five hundred flat burials and twenty-nine barrows.23
Settlement archaeology, too, generally reflects the kinds of change under way. At the top end of society, the elite dwellings inhabited by the kings and princes of the Alamanni have been quite extensively investigated. One of the best-known is the Runder Berg at Urach, within the territory of the Alamanni. Here in the late third or early fourth century a hill-top area, with maximum dimensions of 70 metres by 50, was surrounded by a stout timber rampart. Inside were a number of timber buildings, including what looks suspiciously like a substantial hall for feasting retainers and/or fellow kings. The lower slopes housed other buildings, including workshops for craftsmen and possibly dwellings for other servants, and the site as a whole has produced higher concentrations of imported Roman pottery and other elite items than the more run-of-the-mill rural sites. No large dwellings dating to the pre-Roman period have ever been thrown up within the bounds of Germania, but in the early centuries AD they started to become reasonably common. At a lower level of grandeur, at Feddersen Wierde again, one particular house within the village was marked out from all the others in the early second century. It was substantially larger and surrounded by a wooden palisade. The excavators interpreted it as the dwelling of a local headman. Similar examples of particularly large dwellings are known from a number of other sites as well, such as Haldern near Wesel and Kablow, thirty kilometres southeast of Berlin; all date to the Roman period. Within the particularly well-studied territories of the Alamanni, no less than sixty-two elite dwellings of one kind or another, dating to the fourth and fifth centuries, have been identified, of which ten have been excavated; and other similar sites, though less thoroughly studied, have turned up right across Germanic Europe, even as far east as the Gothic-dominated territories north of the Black Sea.24
The general picture, then, is clear enough. Settlements and grave-goods show up an increasing social inequality, and it doesn’t take much thought to see how possession of military might allowed kings and, through them, their retainers to gain privileged access to a more than equal share of the new wealth. By the fourth century, as a direct result, we are faced with a Germanic world that was marked by more social stratification than its first-century counterpart and, in some places at least, greater structural stability in its political organization. It is, in fact, entirely natural that these two phenomena should have gone together. Class definition and state formation have long proved inseparable bedfellows when patterns in the evolution of human social organization have been subjected to comparative study. But how far-reaching had this inequality become by the fourth century, and how should we understand the new political entities that dominated the landscape? Were they ‘states’ in any meaningful sense of the word?
Categorizing human societies and their political systems is a subject with a long and complex history stretching back to Aristotle and beyond. In the modern era, it received a whole new impetus from the significance that Marx and Engels ascribed to the state and its evolution. In classic Marxist analysis, the state is the sum and guarantor of the social, political and legal structures by which the dominant class in any given era perpetuates its control over the prime means of producing wealth at that time: whether we’re talking land in the ancient world, heavy industry in the recent past, or computer software and hardware now. This brute reality is always hidden behind some kind of ideological cloak whereby the elite tells everyone else that the state exists for the benefit of them all, but if you look hard enough, according to the Marxist perspective, it always turns out to be about maintaining the power of the privileged. More recent work has moved well beyond this kind of simple Marxist agenda, with a complex literature devoted to analysing early state forms along a spectrum of size and sophistication marked out by terms such as ‘tribe’, ‘simple chiefdom’, ‘complex chiefdom’ and ‘early state’. Rather than worrying too much about where to place the fourth-century Alamannic and Gothic confederations along this sliding scale, though, we can make better use of this literature in a more general way by identifying four key areas to investigate when seeking to understand the operations of any political system.25
The first, straightforwardly, is scale. What magnitude of human population is being brought together by the political system under discussion? Second, what kind of governmental systems does it employ? Are there any bureaucrats or governmental functionaries, and what kind of powers do they deploy, using what technologies? The third area is the level of economic development and associated social stratification generally at play. Whether you accept the Marxist diagnosis of why this is so or not, it is simply the case that particular types of political system tend to be associated with particular types of economic organization. Large, centralized governmental systems cannot be supported by economies that do not produce an economic surplus of the appropriate size to pay for the existence of the functionaries not engaged in primary agricultural production.26 Fourth and finally, we must look hard at a society’s political relationships. How are rulers chosen and legitimized, and by what mechanisms do they create and sustain their authority? In particular, this area is concerned with the balance between force and consent, and the extent to which rulers need to give something to their subjects – whatever that might be – in return and in justification for the economic and other support that they themselves receive.27
Investigating fourth-century Germania under any of these headings is not straightforward, given the nature of the available evidence. There is generally little of it, and what there is refers primarily to the Alamanni and the Gothic Tervingi, adding the further complication of how far we might legitimately generalize from these cases. But, at the very least, these entities document the limits of the possible among the fourth-century Germani, and there are enough points of conjunction between the two (and with what wider evidence there is) to suggest that it is not unreasonable to draw more general conclusions from their capacities and modes of operation.
Power and the King
On questions of scale, the evidence is far from ideal. But the Alamanni and the Tervingi certainly each had a military capacity – young men of military age – amounting to more than ten thousand individuals. Ammianus tells us that Chnodomarius gathered an army of 35,000 for the battle of Strasbourg. Not all of these were Alamanni, and Roman reporting of barbarian numbers is always questionable, even if, as in this case, not obviously outrageous. But the Roman army numbered 12,000, and that figure – which is more secure – confirms an order of magnitude well over 10,000 for Chnodomarius’ force. The Romans still enjoyed a considerable tactical advantage over the Germani in the fourth century, not least because, as we have seen, the latter did not usually possess defensive armour, so that Chnodomarius would probably not have given battle without at least some superiority in numbers. The figures for the Tervingi are less straightforward, but on at least three occasions the confederation sent contingents of three thousand men to serve in Rome’s wars against Persia, and this is unlikely to have represented anything like one-third of its total military manpower. The Tervingi were also powerful enough to evade the hostile attentions of the Emperor Valens for the three years between 367 and 369, and I would read Ammianus to imply that, even after a split within the confederation, its larger fragment could put at least 10,000 fighting men in the field. All of this suggests that both Alamanni and Tervingi could field well over 10,000 warriors, and perhaps as many as 20,000. Estimates for the size of the overall population of these confederations depend, of course, upon what proportion of the total group you think likely to have borne arms. The minimum multiplier commonly used is something like four or five to one, implying total group sizes in the 50–100,000 range, but I think this is likely, if anything, to underestimate the total population that formed part of these confederations in some capacity or another.28
Nor was any of our Roman sources sufficiently interested to provide a run-down of the governmental structures that made these confederations tick. As will pretty much always be the case throughout this study, therefore, their governmental capacity will have to be deduced largely from the kinds of administrative acts of which the system was capable. In some areas, the Alamanni and Tervingi show an impressive capacity. The least that can be said is that in the face of Roman power, both upheld some concept of their own territorial space. When they were in a position to avoid the most intrusive levels of Roman intervention in their territories, leaders of both the Alamanni and the Tervingi met Roman emperors in summit meetings on boats in the middle of the Rivers Rhine and Danube respectively, meetings which symbolically asserted that the river lines marked clear boundaries between themselves and the Empire. Whether their other boundaries, between themselves and their fellow Germani, were so well defined, in both perception and reality, is less clear but perfectly possible. The River Dniester, for instance, seems to have functioned as a marker between the Tervingi and an adjacent group of Goths, the Greuthungi, and there was enough hostility between the Alamanni and their Burgundian neighbours to suppose that both sides – as Ammianus reports – would have carefully defined their territories. According to him, they used some conveniently placed former Roman boundary markers to define the limits of their territories.29
Within these territorial spaces, at least in response to Roman pressure again, Germanic leaderships were sometimes ambitious enough to impose a degree of cultural uniformity upon their populations. Roman cultural hegemony on the Danube in the fourth century, for instance, occasionally took the form of an interest in spreading Christianity to adjacent lands. On at least two occasions, when they were in a position to act, the leadership of the Tervingi resisted this with determination. In 348, Christian Roman missionaries were expelled and then a second time, after 369, Gothic Christians were actively persecuted to the point of execution, creating in the process a not insignificant number of martyrs. This suggests that the Gothic Tervingi’s sense of their own space, at least, had come to take a fairly active cultural as well as economic and military form.30
The actions of various leaders, moreover, show us that certain institutional powers were in place. Particularly impressive, to my mind, is the evidence for a defined military obligation among the Tervingi. On three occasions, as we have seen, the confederation sent military contingents to Rome’s Persian wars. The individuals who went received some financial compensation from the Roman state, but overall the evidence suggests that this kind of service – on a frontier over fifteen hundred kilometres away, it should be remembered – was a generally resented imposition. Such service was certainly one of the terms of client status which the Goths’ leaders sought to strike out when they were in a position to. Nonetheless, the leadership of the Tervingi was able to make these contingents actually appear, which means that it could both identify individuals liable for military service and force them to show up. The Alamanni, likewise, provided contingents for Roman service on occasion, but we have few details and the distances involved were much smaller. Interestingly, the word generally in use in Germanic languages for ‘doing military service’ is a loan word from Latin, which perhaps suggests that this kind of transferred demand from the Roman state may have been responsible for generating a new kind of compulsory military service among those Germani forced to provide such contingents.31
The leaderships of both Alamanni and Tervingi also had defined rights to basic economic support in the form, presumably, of taxation levied on agricultural production. Rights in this area were necessary to support the kings’ military retinues. By the fourth century, no king with a full-time professional retinue could afford to rely on purely voluntary donations of foodstuffs for their support, as had apparently been the practice in the first century. The extent of Roman imports, not least of wine amphorae found on elite sites in the fourth century, likewise suggests that kings were creaming off a proportion of basic production to exchange for Roman goods for their own consumption. Quite likely, though, Germanic leaders had at least one other major form of economic support. As we have seen, cross-border trade with the Roman Empire had become a substantial phenomenon by the fourth century. For their part, the Roman authorities certainly imposed customs dues on all this economic activity, and it is overwhelmingly likely that Germanic kings did too. We have no explicit evidence to this effect for the Alamanni or the Tervingi, but other Germanic kings of the frontier region were doing this as early as the first century, when the wealth of Vannius king of the Marcomanni was incontrovertibly associated with the presence of Roman merchants at his court, and it is extremely unlikely that their fourth-century counterparts would have failed to do the same. It is hard, otherwise, to explain why trade and its regulation should have figured so prominently in diplomatic negotiations between the leadership of the Tervingi and the eastern Roman Empire; and something made Chnodomarius wealthy enough to buy in mercenary support in addition to the other forces he lined up at Strasbourg.32
Both confederations also had the right to impose labour services on at least parts of their population. Kings of the Alamanni could mobilize labour both for constructing their own defended elite sites, such as the Runder Berg, and when forced to pay off diplomatic obligations by providing labour for Roman state purposes, as in the treaties imposed on them by the Emperor Julian after Strasbourg. Among the Tervingi, likewise, the then judge attempted to fend off Hunnic aggression in the 370s by constructing a substantial set of fortifications – what Ammianus calls the ‘wall’ of Athanaric. This was most likely an attempt to renovate an old Roman fortified line on the River Alutanus, and in the end it came to naught. But the fact that such a project could even be attempted shows that the right to extract labour service was established, as does other physical evidence from the Gothic realms for elite sites similar to the Runder Berg.33 In the Roman world, and later in that of the largely Germanic-dominated successor states to the western Roman Empire, labour service was imposed usually only on the more servile element of the population, meaning that part of it which did not do military service. We have no evidence that this was also the case among the Alamanni and Tervingi, but it seems likely enough.
In certain key areas, then, fourth-century Germanic leaders had well-developed rights. They could define and extract – perhaps from different elements of their populations – military service, labour dues and a percentage of agricultural production. Almost certainly, too, although none of our sources is sufficiently interested to tell us about this, they had rights to be involved in what we would term legal-dispute settlement – in the case of their more important subjects anyway. No leader known in any other context, whose powers can be elaborated in any detail, lacked this kind of authority, so it is probably safe enough to ascribe it to the leadership of the Tervingi and Alamanni as well.34 As to how these various rights were actually administered, neither confederation ran, as far as we can see, to any kind of articulated bureaucracy. No source mentions bureaucrats in the fourth-century Germanic world, though kings certainly had their functionaries, and the rights were possibly exacted with little or no use of any formal literate administration. Writing of various kinds was known to the fourth-century Germani. Runes were in use, some Germani were able to operate successfully in Latin, and, in the mid-fourth century, Gothic was busily being turned into a written language – the first Germanic tongue to be so – for the purposes of Christian missionaries. There is no evidence, however, that any of these literacies was being applied to the exaction and disbursement of revenues in the form of agricultural produce.
But this need not mean, it is worth stressing, that exaction was an essentially random process. How it might have worked on a regular but essentially paperless basis is illustrated by some of the earliest evidence for administration from Anglo-Saxon England. Here the seventh-century agricultural economy was harnessed by dividing the country up into largish revenue-producing districts, each of which had to contribute a given quantity of agricultural produce annually in the form of food renders. The system required an exhaustive surveying process at the beginning, to divide the countryside up; storage space for the goods, and some kind of tallying system to keep track of deliveries; but not that many officials and no great degree, if any, of literacy. It is, in fact, a straightforward mechanism for extracting revenues from a rural economy that is found in various contexts, and there is absolutely no reason to suppose that something of this kind was beyond the capacities of the Tervingi and Alamanni.35 Alamannic territory, as we have seen, was already divided into districts (Gaue, in German), and it is probable that one of their functions was fiscal. In the Alamannic case, of course, we are dealing with multiple kings, many of whom controlled their own cantons. Any revenue collection in this context, presumably, was in the first instance by and for these canton-level kings, although they may then have had to pass on a portion of their take to an overking.
In Anglo-Saxon England and many other early medieval contexts where fiscal systems mainly produced food rather than some more negotiable form of wealth, what is known in the scholarly literature as ‘royal itineration’ was central to their operation. This meant that instead of running a fixed royal court, the king, his leading advisers and his professional retinue moved around the kingdom in a regular cycle, stopping at a series of designated points. These stopping points were also the local collection centres for the food renders, thus greatly reducing the inherent logistic problems of a tax regime based on bulky, heavy food rather than, say, comparatively light and mobile coinage. Instead of the food mountains going to the king, the king went to the mountains. We have no explicit evidence for itineration among fourth-century kings of the Germani, but since the consumption of food renders is so much easier on this basis, it must be a priori likely. It is perhaps a reflection of the intineration process that the Romans could not simply predict where a targeted Alamannic king might be, and an observable correlate of such systems is, obviously enough, the existence of many royal centres, which might also explain why there were quite so many such centres, seemingly, among the Alamanni. There were no more than about twenty-five cantons, implying a maximum of twenty-five kings, but sixty-two elite sites have been identified, and these are all hill forts, while the written sources mention others (so far unidentified) in the lowlands as well.36
State and Society
The consequences of all this economic development for the spread of social power among the Germani are difficult to estimate in their entirety, but two initial observations are straighforward. The overall population of Germanic Europe will have increased markedly over the Roman centuries, as agricultural production grew in intensity and the rest of the economy – at least moderately – diversified, but kings and warbands benefited disproportionately from the extra wealth. The difficulty comes when you try to get a sense of the consequent redistribution of social power. A whole host of evidence suggests, in fact, that the degree of overall change must not be overstated. Both literary and archaeological evidence indicate that other people, apart from kings and their retinues, still mattered in Germanic society of the fourth century.
Some of the relevant evidence consists of narratives of Germanic politics in action. As the famous historian of Rome’s barbarians Edward Thompson observed, Ammianus’ descriptions imply that kings could not simply order warriors about, but had to ‘urge’ and ‘persuade’ them to follow their policies. Also, we have already encountered the Alamannic king who was overthrown by his own followers for not attaching himself to Chnodomarius’ banner. Ammianus explicitly states that this was the result of action by the ‘people’ – plebs, populus – of his canton. This could just about be referring to a restricted political world of royal retinues, although Ammianus’ wording implies not, but Strasbourg involved a military-political community that extended well beyond such limited social circles. The Alamannic army gathered there numbered reportedly thirty-five thousand, as we have seen, and certainly well over ten thousand fighting men. Royal retinues, even of chief kings, numbered just a few hundred. Ammianus refers to sixteen kings and princes assembled for Strasbourg, and even if for the sake of argument we allow each of them a retinue of two hundred (although most will, by definition, have been smaller since Chnodomarius was the most powerful king), that still only amounts to 3,200 fighting men. Military participation was clearly not limited just to kings and small specialist retinues. Nor, it seems, was some kind of elevated social status. Archaeologically, the increase in the quantity of material deposited with the Germanic dead, seen over the Roman period, was not confined to a very small number of rich Fürstengräber. Alongside these highly exceptional burials are found both large numbers of graves with absolutely nothing in them at all, and a fairly numerous category containing a moderate number of personal items: usually pottery and, as mentioned earlier, weapons of some kind for men and jewellery for women. The striking increase in weapons burials in the late Roman period, though not found right across Germania, does lend further weight to the idea that the period saw a substantial increase in the importance of the martial side of male life, consonant with the rise of the retinues, but the total number of such burials indicates that others, apart from kings and retinues, were also treading this path to retained or increased social prominence.37
A large quantity of legal evidence from the sixth and seventh centuries suggests who these others may have been. These texts, or codes, composed in the successor states to the western Roman Empire, provide us with the first full description of the social categories operating in a Germanic-dominated society. Given the date of the texts’ composition, they all reflect Germanic societies that had been through a further stage of interaction with what remained of old Roman imperial economic, governmental and social institutions after the collapse of the western Roman Empire, so there is an obvious difficulty in trying to use them to elucidate the fourth-century Germani. But if anything – and this would be the general consensus, not just my own view – these later interactions will only have increased inequalities of wealth and status in the Germanic world, because the process of taking over former Roman territories led to further unequal acquisitions of wealth on the part of kings and their immediate supporters. That being so, this later legal evidence will tend to underestimate the sociopolitical importance of other social groups not immediately in royal service. It can be used as a guide, therefore, to the maximum level of inequality likely to have been prevalent in the fourth century.
The descriptions of status groups found in these legal materials are strikingly uniform. Kings had a special status, obviously, and being in royal service usually increased status as well. In addition, the codes often referred to a noble class. All of these groups can reasonably be thought of as belonging to worlds analogous to those of the fourth-century kings and retinues. But all the codes (and we do have law codes from a large number of the successor kingdoms) also referred to a class, beneath the nobility, of freemen, who still had considerable rights and responsibilities. These freemen stood above two further classes: permanent freedmen and slaves. Characteristically, freemen did military service (as, in fact, often did freedmen, but not slaves); they could also give trustworthy testimony in cases of legal dispute; and their status was ringed about by safeguards to prevent slaves and freedmen from crossing the boundary without permission.38
The importance of this free class was overemphasized in romanticizing nineteenth-century accounts of Germanic society. Nothing indicates, for instance, that they formed a numerical majority of the male population; and given their obviously privileged position, I would be willing to bet quite a lot of money that they did not. Privileges are enjoyed by minorities, not majorities. Some not very good Ostrogothic and Lombard evidence might suggest that the freemen amounted to something like a quarter or a fifth of weapon-bearing males of these groups in the sixth century (and slaves are excluded from the equation because they did not bear arms). This of course makes freemen a still smaller percentage of the total population. But neither were they a figment of the law-writers’ imaginations. Freemen are encountered in practice right across the post-Roman west as an important group of social actors at the local level in the evidence of legal practice, and also in some of the narrative evidence for warfare between Germanic-dominated groupings and the east Roman state.39 If this was true of the successor states, when a further influx of Roman wealth had increased inequalities again, then it is overwhelmingly likely that freemen were still more important among the fourth-century Germani, before this later process unfolded. We should not imagine, in other words, that increased social stratification in the Roman period had reduced the sociopolitically important stratum of Germanic society to a tiny group of kings and retainers. A broader world of freemen maintained – or had developed – in the changing economic circumstances its range of social and economic privileges. They perhaps show up archaeologically as the owners of the big and prosperous longhouses found in some of the new villages of third- and fourth-century Germania, and as the occupants of the large number of endowed but not massively rich burials.
This fairly complex account of social stratification among the fourth-century Germani has obvious implications for the final key area of analysis: the balance between constraint and consent in Germanic politics.
Evidence for some degree of constraint is straightforward. Kings had warrior retinues. By use of these retinues, they had established a hereditary element to their position. The retinues could also be used more broadly as social enforcers, as we saw among the Tervingi when it came to persecuting Christians. There, in the incident described, the persecution policy went against the general wishes of the village community.40 The leadership of the Tervingi could also, as we have seen, levy military contingents to make the onerous and dangerous trek to fight in Rome’s Persian wars. And what could be a clearer sign that the rise of military kings was not always a consensual process than the weapons find at Ejsbøl Mose?
But just as kings and retinues had not completely eclipsed a broader privileged (freeman?) class, so the political process also had – sometimes, at least – to take account of, and win the broad consent for, their policies from this larger privileged group within the total population. As we have seen, kings could even be overthrown if their policies proved unpopular. The Alamannic king who wouldn’t join Chnodomarius may possibly have been eliminated by his own retinue, but more likely by the broader freeman class of his canton; and similarly, the last member of the old ruling dynasty of the Tervingi, Athanaric, was overthrown in the midst of his fortification work when resistance to his ideas of how to combat the Hunnic menace overflowed into political dissent.41 Both events emphasize that there were marked limits to the powers of the new military kings.
It is not possible to explore the subject in any great detail, but the sources do suggest a few of the mechanisms by which these limits were orchestrated and imposed. To start with, we should probably not draw too distinct a line between freemen and royal retinues. There is considerable evidence that Germanic society operated in age sets for both men and women, with rites of passage marking certain clear stages in an individual’s life, and each stage having its own rights and responsibilities. Older men, even high-status ones, were never buried with weapons, for instance, suggesting that there was an upper age limit to military obligation; and for women, the legal evidence indicates that within each status group child-bearing years were associated with maximum social worth. Pre-pubescent children, likewise, seem rarely to have been buried in cemeteries alongside adults, again suggesting that age and status went hand in hand.42 This is not something that the available source materials will allow us to explore very thoroughly, but it is far from unlikely that at least some males of freeman status customarily served, when younger, in the warrior retinues of kings.
There may also have been other links between the worlds of freemen farmers and royal retinues of which we are not properly informed. Villages certainly provided kings and their retainers with economic support, but kings may well have been expected to hold regular feasts for a broader spectrum of the free class as well as for their immediate retinues. If such feasting remained habitual, then some genuinely reciprocal relations continued between kings and freemen into the fourth century. Again, in places these kinds of behaviour survived into later, still less equal Germanic-dominated successor states, which strengthens the likelihood that they were in evidence in the late Roman period. In early Anglo-Saxon England, itinerating kings were sometimes expected to give the benefit of their presence at more communal feasts, in return for the food supplies they were offered, and these events provided a context for many important social and political exchanges. Looking just at scale, for instance, Alamannic cantons were small enough that their kings can hardly have been isolated figures, cut off from the rest of the population, and I would suspect that feasting and other such interaction would have been unavoidable, and had probably long been a feature of the Germanic world, as they have been found to be in many potentially analogous contexts.43
Assemblies, too, may have played an important limiting role. Germanic political units of the early Roman period customarily worked through councils, at which group policy was debated and decided. Tacitus’ works put a huge emphasis on this institution, and it was clearly much more than a figment of his ever fertile imagination. Particularly striking to my mind is the evidence – several separate occasions being recorded in our highly fragmentary records for the first and second centuries – of the fact that in order to punish a grouping for a revolt, or to prevent one from taking place, assemblies were either prohibited by the Roman authorities, or allowed to proceed only with Roman observers. The fourth-century evidence does not shed much light on the degree to which such assemblies continued, but there certainly seem to have been village gatherings; and the decision of the Gothic Tervingi to seek asylum in the Roman Empire in 376 emerged only after long debate, presumably at a much larger assembly of the socially important. The dispute-settlement procedures envisaged in the successor-state law codes also indicate that regular assemblies were necessary for legal purposes. For all these reasons, I would suppose that an assembly structure continued within the fourth-century confederations, acting as a further brake on the arbitrary powers of kings.44
There is also no evidence that Germanic kings were able to deploy self-justificatory ideologies of sufficient strength to entrench an overarching domination. It has sometimes been suggested, for instance, that they surrounded themselves with a powerful aura of sacrality, which distinguished certain clans as marked out by the special favour of the gods and made resistance to their royal pretensions extremely difficult. But there is actually little evidence of this. None of the three main words used in Germanic languages for ‘king’ carry sacral connotations. They are all, as we have seen, deeply pragmatic: ‘ruler of a people’, ‘ruler of a warband’, ‘ruler of a confederation’. Germanic kings certainly drew on a concept of divine favour – heilag and its various derivatives in the different Germanic language branches – but it was a post de facto kind of concept, which identified itself through practice. If you won battles and therefore power, then you had shown yourself to be heilag, but there is no sign that claiming to be heilagautomatically brought you to power, or prevented anybody else from mounting a challenge to your authority – often to devastating effect, as the narrative evidence again suggests. And if a usurper was successful, then he had proved that he was now heilag.
The one context in which we find a heavy emphasis on the manifest destiny of one particular dynastic line for divinely ordained rule comes in the propaganda produced at the court of Theoderic, the Amal leader of the Ostrogoths in early sixth-century Italy and ruler of one of the first-generation successor states to the Roman west. Such a view of his dynasty is directly written up in the Variae of Cassiodorus and reflected indirectly in the Getica of Jordanes. But when this claim is measured against the actual history of the Amal dynasty, the results are highly instructive. The dynasty had won extensive power in the Gothic world only in the generation or so before Theoderic himself (as we shall see in more detail in Chapter 5), and as soon as it failed to produce suitable male heirs after his death, it was quickly disposed of. Theoderic proved himself to be heilag with a succession of stunning conquests, not least that of Italy itself, but that was not enough to protect the dynasty against incompetent heirs. All the propaganda, produced when Theoderic was trying to secure the succession for his under-age grandson,45 was precisely that – propaganda.
The evidence for age sets, feasting obligations, councils and limited royal ideologies is all very fragmentary, and can only hint at the realities of political life among the Germani. The bottom line, however, is clear enough. While a new elite exploited the economic development of the Roman period to entrench its social prominence, and, in the process, made it possible to build, at least in some areas of Germanic Europe, the larger and more stable political units of the fourth century, we mustn’t overstate its powers. A broader social group outside the nexus of kings and retinues remained important, both socially and economically, and had to be involved in the political process. Not least, it continued to outnumber the royal retinues massively, so that its support remained crucial to larger military enterprises. And in any case, as we have seen, freemen and warrior retinues may well have been interconnected in a variety of ways.
More generally, this broader social group must also have given some kind of consent to the creation of the new and much larger confederations of the late Roman period. Ammianus provides an illustration of this in his account of the attempt of one Alamannic king to distance himself from the confederation before Strasbourg, which led to his own demise. The same is suggested by the fact that not all of the old political associations of the first century were destroyed in creating the new ones of the third and fourth. We have explicit evidence only for the Franks, into which confederation late Roman sources indicate that some of the old units – specifically the Chatti, Batavi, Bructeri and Ampsivarii – had been incorporated. This process, obviously, was never as simple as the old units voting to join a new regional association, since some new units were created as well, the Salii already being mentioned by Ammianus; but nor was there total discontinuity either.46
Looked at against the comparative literature, the fourth-century confederations fall somewhere in the nexus between ‘early states’ and ‘complex chiefdoms’. According to the normal criteria employed, they were too large and too stable, and encompassed too substantial a degree of marked social differentiation, to be categorized as either ‘tribes’ or ‘simple chiefdoms’. And, looked at closely, the differences between early states and complex chiefdoms are essentially ones of degree, where the former have slightly more organization, stability, power and so forth than the latter. The shortage of evidence about the fourth-century confederations makes it extremely difficult to make more precise judgements, and what evidence there is sometimes prompts contrasting conclusions. The extent of their governmental capacities and, especially among the Tervingi, the establishment of dynastic power look quite state-like, for instance, but the lack of specialized royal functionaries and of any evidence for the survival of a relatively broad (freeman?) social elite suggests a complex chiefdom. This is not, however, an issue to become overly fixated on. The important point is that economic and social transformation had generated a new confederative element in Germanic society, or at least in some of those parts of it closest to the Roman frontier, which was capable of combining, for certain functions anyway, many tens of thousands of people. Politically, these new structures built on the past, incorporating sometimes pre-existing social units, but their powers and solidity represented a decisive break with the Germanic past.
One big question, however, remains unaddressed. What kick-started the economic transformations that underlay the confederations, and how precisely did economic development feed through into new political structures?