But the political context, too, played a crucial role in shaping the British outcome. On the continent, migrants were forced to operate in large concentrations because they needed to put substantial military forces into the field, whether to survive in the face of Roman imperial power or to escape from Hunnic control. No such constraints operated in Britain. Local government in the Roman provinces, not least the raising of taxes, operated on the basis of the city territory – the civitas – and it was thus that sub-Roman Britain seems to have continued. It was the tax revenues of these city units, no longer handed over to the Roman state, which provided, it seems, the Anglo-Saxon mercenaries’ pay. The boundaries of some of the easternmost, therefore presumably the earliest, Anglo-Saxon kingdoms – East Anglia, Essex, Lindsey and Kent (Map 11) – roughly correspond to likely civitas boundaries, suggesting that they might have been founded by migrants taking over the relevant civitates (cities) as still-functioning entities. But these civitas territories were not very large, and could never have supported large armed forces. This state of affairs must have evolved in the time of Aurelius Ambrosius. From this point, perhaps, you did have to organize a larger migration force if you wanted to carve out a nice piece of lowland Britain from the now more organized Romano-British. Even so, no sixth-century Romano-British king could field anything like the armies available to the Roman state or Attila the Hun, so the difference in requisite magnitude was still enormous, and any British military-cum-political unity proved only temporary. One of Gildas’ main declared motives for writing was precisely the fact that the political leaders of his own generation (and he names five kings) were squandering the Ambrosian inheritance by petty internal squabbling.38
As far as we can reconstruct it, therefore, the Anglo-Saxon migration flow was no Völkerwanderung and little resembled the old culture-historical model of mass movement combined with ethnic cleansing. It was a long-drawn-out phenomenon, not an event as Bede’s single date for the Adventus Saxonum might lead you to think. Many of the groups involved were perhaps small, especially at the beginning, but probably increased in scale as obstacles were encountered. Women and children also participated. The motivation was a desire to profit from the wealth of lowland Britain’s developed agricultural economy, but secondary factors affected at least the speed of the flow at different moments, and there was also a strong political dimension, since the wealth of Britain could only be accessed fully by seizing control of the land. For all the shortage of information, therefore, the Anglo-Saxon migration into lowland Britain clearly took the form of a predatory population flow. And, as the comparative literature on migration would suggest, it is also fairly easy to work out the profound effect of factors such as the availability of information, logistics, and the evolving political-cum-strategic context in which it was operating.
But what about the question we have so far avoided? Was the Anglo-Saxon migration flow a case of elite transfer, or need we think in terms of a different migration model?
The Limits of Emulation
It’s worth considering what is at stake. In its classic form, the elite-transfer/cultural-emulation model would suggest that incoming Germanic-speakers represented only a small percentage of the total population that either partly or wholly replaced the indigenous Romano-British landowning elite. The bulk of the Romano-British population remained in place and massively outnumbered the immigrants, but, over time, absorbed the latter’s material and non-material culture until immigrant and native became indistinguishable. What you essentially see is an overwhelming majority of Romano-British voluntarily exchanging their group identities to become Anglo-Saxons. The underlying purpose of this model is specifically to show that massive material and non-material cultural transformation in a Germanic direction can reasonably be explained by only a limited number of Anglo-Saxon immigrants having come across the North Sea. More generally, it is part of a broader argument downplaying the importance of Anglo-Saxon migration as an agent of major change. For in most of the variants put forward, what happens in Roman Britain before the Anglo-Saxons arrive (for example, the collapse of Roman structures), and the reaction of the natives to their arrival (their voluntary decisions to become Anglo-Saxons), are at least as important as the migration flow itself. As such, the model was originally, and in its variant forms has continued to be, framed in reaction to past overuse of the invasion hypothesis.39
The other side of the argument has become harder to define, since no one now believes that the incoming Anglo-Saxons wiped out or expelled virtually all the native population. Anglo-Saxon ‘mass’ migration just isn’t what it used to be. In a very real sense, it is now defined negatively, against the elite-transfer model. Fundamentally, it amounts to the proposition that there were too many Anglo-Saxon migrants for them to be categorized as just an aristocratic elite, and, more generally, that they – rather than the indigenous population by their own free choice – were responsible for the cultural and other transformations that unfolded in lowland Britain. So Anglo-Saxon numbers are at the heart of the argument, along with the overall nature of their relations with the indigenous population. Were the Romano-British free to respond to the arrival of the Anglo-Saxons as they chose, or were the immigrants aggressive and politically dominant? Of these two issues, the question of numbers is at first sight the more problematic, since it is precisely on this issue that none of our sources provides straightforward information. We have only very broad estimates of the native Romano-British population in c.400 AD, and little of substance on the scale of the subsequent Anglo-Saxon migration flow. But if we don’t obsess about precise numbers, a more productive way forward begins to emerge.
The place to start is with the revolution in the organization of the countryside which unfolded in the fifth and sixth centuries. Late Roman Britain was divided into a number of large and medium-sized estates, many run from villas: substantial country houses-cum-estate-centres. Land, as in much of the Roman world, was unequally distributed, with large quantities in the hands of a relatively small landowning class. By c.600, this property distribution had been replaced with another, worked out on a very different basis. Not only had all the villa buildings fallen out of use, but the estate boundaries, too, had failed to survive. In only one or two cases has it even been suggested that old Roman estate boundaries were still in use in the Anglo-Saxon period, and none of these has been convincingly proven. In effect, the economic map of the countryside had been totally redrawn. By 600 AD, Anglo-Saxon kings had delineated larger zones for the purposes of taxation, but much of the farming was based on much smaller units than the old Roman villas, and it was not until the ninth century that large centrally organized estates began to re-emerge in the English countryside. These were the first of the manors, which went on to become a dominant feature of the countryside by the time of Doomsday Book.40
This shows us immediately that the Anglo-Saxon takeover was not a simple case of elite transfer after the classic pattern of the Norman Conquest, half a millennium later. By the time the information recorded in Doomsday Book was gathered, twenty years after Hastings, the indigenous Anglo-Saxon aristocracy had been swept away, its lands transferred to William the Conqueror’s chief henchmen: the so-called tenants-in-chief. The process reached down to more local elites, the gentry, because the tenants-in-chief in turn enriched their more important followers by granting them economic rights over large amounts of the land that they themselves had received. This secondary process was as much a political necessity as the Conqueror’s original gifts to his tenants-in-chief, because it was the loyal service of all these men that had made the conquest possible, and they were consequently expecting a share in the profits of their joint military enterprise. As a result, the Anglo-Saxon gentry as well as the aristocracy lost ownership of their land, even if some still survived as tenants on property they had previously owned.
But a central feature of the process was that none of this property-swapping disturbed existing estate boundaries, or changed the workin
g pattern of the manorial economy. As functioning farming units, the manors continued to operate. There were some alterations in detail, and a convincing argument can be made that elements of the Anglo-Saxon peasantry suffered a considerable demotion. Fundamentally, however, manorial estate boundaries and the general working of the rural economy were left undisturbed by the massive transfer of property rights which resulted from the Normans’ great victory. This was the best possible economic outcome for the Norman conquerors. The main activity of a manor was arable agriculture – growing grain – using a centrally directed labour force, but it still needed its pasture-lands and woods to provide for the animals and people without whom the estate could not function. Any disturbance of these arrangements would have lowered agricultural outputs and the new estate owners’ incomes.41
The Anglo-Saxon takeover of the fifth and sixth centuries, by contrast, did not see anything like such a straightforward transfer of property. The new Anglo-Saxon landowning elites did not simply take possession of existing villa estates, even though this would, in economic terms, have been much the best option. Like the manors of the eleventh century, Roman villa estates were integrated farming units whose output had provided the wealth for a very prosperous class of rural landowners. Changing villa boundaries meant disturbing the actual functioning of the rural economy, and there is good evidence that this process, as you would expect, led to substantial declines in rural output. While the overall area under cultivation did not change markedly (this is shown by pollen analysis), even if some marginal areas were abandoned, more complicated estate structures fell out of use. Some of the Roman-era drainage mechanisms around the Thames at Dorchester, for instance, were not maintained into the Saxon period, and a less sophisticated ‘scratch plough’ generally replaced the heavy Roman ploughs previously in use. The latter were expensive items of capital equipment because of the costs incurred in generating enough spare fodder to overwinter the draught animals required to pull them. Presumably the smaller farming units of the early Anglo-Saxon era could not afford to maintain them, even if they had wanted to.42 This also goes a long way towards explaining why the old towns of Roman Britain lost what was left of their urban character. They had not been centres of industrial production, anyway, but, ‘agro-towns’, existing to serve certain functions in a relatively developed agricultural economy, in return for which some food surpluses found their way out of the countryside to feed the urban population. If you disturbed the organization of the rural economy, particularly by simplifying its functioning and reducing total output, you would be cutting away at the roots of this kind of urbanism, so that it is not surprising that the towns as such disappeared in the post-Roman period, even if they sometimes retained an administrative importance, with new Anglo-Saxon royal palaces being established within their boundaries.43 The next question, therefore, is an obvious one. Why did the Anglo-Saxon takeover break up existing Roman estate structures, even though this brought with it substantial economic costs?
One line of approach has located the answer to this conundrum in developments internal to Britain before the Anglo-Saxons arrived. Some have seen the British rebellion of 409 reported by Zosimus as a kind of peasants’ revolt which not only threw off central Roman control but overturned the established social domination of the villa landowning class. The villas, of course, would be a natural casualty of such an uprising. More recently, Guy Halsall has argued that the villa estate structure of lowland Britain crumbled as a direct consequence of its separation from the imperial system under whose umbrella it had emerged, but he identifies a different sequence of development. In his view, the position of the villa owners at the top of the social heap depended on the relationships they had forged with the Empire, and when those ties were broken after 410 AD, they had to work much harder to maintain their elite status. The profits from their estates, which they had previously used to build and decorate their elaborate villas and for other forms of conspicuous consumption, or to trade for valuable items of Roman manufacture (Mediterranean foodstuffs, fine pottery and so on), now had to be distributed locally instead, as gifts to establish networks of supporters. These networks replaced the Empire in structural terms, allowing the landowners to maintain their position in the new conditions, but were relatively expensive, leaving little or no surplus for them to invest in the old forms of conspicuous expenditure. As a result, villas and trading patterns quickly disappeared, and a competitive burial ritual – ‘furnished inhumation’ – grew up among them in lowland Britain as these men staged lavish funerals, depositing much finery with the dead as part of the struggle to maintain their social position.44
The plain peasants’ revolt version of the argument is not convincing. Despite the prevailing chaos, the contemporary Roman west throws up almost no evidence of this kind of activity, but plenty of local elites taking power into their own hands, and precisely in contexts where central imperial authority had failed to answer local needs. In 409 or thereabouts, Constantine III had long since abandoned his British base and his attention was focused firmly on Italy and Spain, where he was trying simultaneously to supplant the Emperor Honorius and to deal with the Rhine invaders who were now established south of the Pyrenees. To my mind, and here I am in full agreement with Halsall, the British rebellion is much more likely to have been of this standard kind, a response to Constantine’s neglect rather than a social revolution that was in any sense anti-Roman. Equally important, the Life of St Germanus of Auxerre portrays a distinctly Roman-looking elite of lowland Britain seeking help from a still Roman continent against both invasion and heresy in the 420s and 430s. Romance (simplified Latin) remained the spoken language of lowland British political life well into the fifth century, and I am also inclined to believe that Gildas’ famous reference to the British seeking help from the Roman generalissimo Aetius, ‘three times consul’, rests on something concrete. All this would mean that a Romanized British landed class was still looking in a Roman direction, and had held on to some of its Roman structures as late as the 440s. This makes the class-warfare argument unappealing.45
Halsall’s version of internal systems collapse is a much more possible explanation for two of the central phenomena in the fifth-century transformation of the British lowlands: that the villas disappeared and that inhumation furnished with graveoods came into fashion. In assessing the argument, however, it is only reasonable to point out why it exists. Halsall is the scholar we met in Chapter 1 who has argued that to sidestep migration in explanations of archaeological change ‘is simply to dispose of an always simplistic and usually groundless supposition in order to enable its replacement with a more subtle interpretation of the period’. His explanation of developments in lowland Britain is entirely in line with this world view, since it is fundamentally internalist. The villas disappeared because of a crisis within lowland British society, which prompted hugely expensive competitive funerary display, and Anglo-Saxon immigrants are denuded of any major role in the action. But while the invasion hypothesis was certainly overused in the past, there can be problems with an a priori determination to deny migration any role of importance. The danger is that any argument will command assent among like believers just because it moves the historical spotlight away from migrants, whatever its intellectual and other qualities.46 And in this case, I would argue that there is a much more straightforward explanation for the disappearance of the villas available if you are not worried about being considered a simple-minded migrationist, and the virtues of economy and Occam’s razor should not be forgotten. Equally important, the alternative explanation also does a better job of accounting for all the available data.
For one thing, it’s not clear that there was scope in the fifth century circumstances for Halsall’s vision – of the villas literally crumbling away through a purely British, internal political process – actually to have worked its way through in the decades after 409. According to the Gallic Chronicle of 452, Saxon attacks had already begun around the year 410, and the vil
las – large, isolated country houses of the wealthy – were both highly vulnerable to attack and obvious targets. We have encountered in Chapter 2 the looted contents of one of their number which Alamannic raiders failed – for once – to get back across the Rhine. More generally, whenever Roman frontier security broke down in any region, villas were the first to suffer.47 There is every reason to expect, therefore, that any increase in outside attack would have affected the villa network immediately. To my mind, this makes it unlikely that a fairly lengthy process of internal erosion would have had time and space after 409 to unfold untouched by outside attack.
Equally important, the collapse of the villa network and the rise of furnished inhumation are not the only phenomena that need to be explained. What Halsall’s argument (or any version of internal-systems collapse) doesn’t easily explain is the degree of cultural change that accompanied the socioeconomic revolution of the fifth and sixth centuries. Not only did the villa estates of lowland Britain disappear, but by 600 AD the region’s Latin-speaking Christian elite had been replaced by Germanic-speaking non-Christians. Halsall of course recognizes this, and accepts that there has to have been a significant degree of Anglo-Saxon migration to account for these shifts even if he provides no real mechanism to explain them, and is trying, overall, to decouple migration from what he sees as the more fundamental process of socioeconomic transformation. The profound nature of these cultural changes, however, needs to be taken fully into account.
For one thing, most of the items buried with the dead in the furnished graves were Germanic, but this is only part of the story of Germanization. What’s really striking about the written Anglo-Saxon language, which survives in a variety of texts dating from just after the year 600 all the way down to the Norman Conquest, is how little it was actually influenced by indigenous British Celtic. Loanwords are few and far between, and there is almost no Celtic influence on its grammatical structures. This is telling us something important: that the spoken language in the various local dialects of the new landowning elite of lowland Britain as it had emerged by 600 AD, the tongue that provided the basis for the extant written form of the language, was not only thoroughly Germanic, but also firmly insulated from contact with the native Celtic tongues of Britain. And in this era, language was transmitted within families, especially through mothers on whom most childcare devolved – one reason, as we have seen, why the Anglo-Saxon flow must have included large numbers of women. Incidentally, this also explains why, in later medieval examples where migration did generate large-scale language change, this occurred only when a peasant population (even an elite one of free, if fairly small-scale, landowners) was involved in the action, and was never generated by very small-scale aristocratic elite transfer along Norman Conquest lines alone.48
Empires and Barbarians Page 39