Featured post

More Posts you might have missed on the other site

Here, in order from oldest to most recent are the not-exactly-numerous posts that have appeared on the other site in the past two and a half...

Monday, 22 July 2019

Reconsidering the Mechanics of Migration: Information and Logistics

[This is the text of a lecture that  I gave at the University of Tübingen last Wednesday.  Many thanks to Mischa Meier, especially, for the invitation and a lovely introduction, and also to Steffen Patzold and Sebastian Schmidt-Hofner.

The paper starts by rehearsing some arguments I have made before about the size of late antique 'barbarian' armies and the restrictions on that imposed by logistics and economics, which it underlines by adducing some comparative material.  It poses the question of how those questions might apply to the migration of larger groups - specifically the sort of large groups who arrived in a specific place at a particular time: the sort of large-scale, short-term migration envisaged in the typical 'Migration of Peoples' (Völkerwanderung) paradigm.

After a brief discussion of the perils, or rather the limitations, of  the sorts of analogy employed in 'Migration Theory', it looks at the types of migration from Barbaricum to the Roman Empire, and their respective logistical feasibility, first from what I call the 'Immediate Zone' close to the frontier and then from the 'Intermediate Zone' in the heart of Germanic-speaking Barbaricum. On this basis it suggests that the movement of 100,000 people (a figure picked simply from its recurrence in studies of the migrations from all sorts of perspective) was, while not impossible, at least over shorter distances, at the limits of possibility, and a decision from which there was probably no going back.

From there it suggests that what would make it less of a gamble would be the ability of the Roman Empire to supply such large groups.  It suggest that the Empire might have had an interest in doing so.  The Empire alone had at least the possibility of doing so, as is further suggested by the fact that the large-scale migrations of the fifth-century overwhelmingly took place within a Roman logistical context.  Finally the argument is underpinned by the fact that once Roman logistical and other structures on the frontier had ceased to exist in the mid-5th century, this type of migration effectively ceased.  The overall argument of my 2014 paper, "Two Worlds Become One: A Counter-Intuitive View of the Roman Epire and 'Germanic Migration'" (German History 32.4 (2014)), that 'the end of the Roman Empire produced the end of the barbarian migrations' is underscored.]

1 Introduction

A. Three well-known scenarios:

I want to remind you of three very well-known scenarios, to which I will return in more detail towards the end of my lecture.

The first is the famous Tetrarchic medallion showing the entry of barbarian families into the Roman Empire, crossing the Rhine at Mainz under the gaze of the emperors.

The second is perhaps the most famous: the image of the Tervingi gathering on the Danube in 376 and petitioning the Empire to grant them entry. I want to remind you all of the extremely well-known accounts of the subsequent starvation and mistreatment of the Tervingian migrants.

The third and final scene is another obvious one: the crossing of the Rhine by the Vandals, Sueves and Alans on the last day of 406 or possibly – as Michael Kulikowski argued and later retracted – 405. [1]

B. Not anti-migration

Before going further, I have been described, in print, as an ‘anti-migrationist’ and, recently, as someone who thinks that migrations only involved a few elite warriors. I find it sad that such writers wish to misrepresent my arguments.

I have argued:
  1. That certain types of archaeological evidence need not and in some cases cannot be read as simple signs of migration
  2. That the fourth-sixth-century migrations of people from Barbaricum were not the only and not necessarily the largest movements of people in the period and may not have been on a larger scale than in the earlier Roman period.
  3. That the migrations of barbarians in the fourth and fifth century did not bring down the western Roman Empire.
But note, first, that none of these arguments logically implies a non-belief in migrations. And second that not one of these arguments implies a particular view of the size of migrations. Nowhere have I argued that migrations didn’t happen or that they were only the movements of small elite groups of warriors. Indeed, I have explicitly argued for larger-scale migrations.

My position has generally been for a variety in the scale and nature of migrations. What I have been sceptical of is the traditional idea of the migrations of entire peoples. I remain sceptical, but that scepticism is something I will explore and question in this lecture.

I want to develop some thoughts I expressed first in Barbarian Migrations and the Roman West and then in a significant modification published in German History in 2014.


2 Numbers and logistics

I want to rehearse some arguments I have made before about numbers, in the military context. I am not the first to express scepticism about the numbers Roman sources give for the size of barbarian forces, yet such numbers are still routinely repeated in modern works. If we leave aside huge figures like 80,000 or 100,000 we can look at the number given by Ammianus for the Alamannic army at the battle of Strasbourg in 357: 35,000 – a number often repeated without the usual scrutiny or reservation. How likely is it that even a confederacy like the Alamans could put together an army that size?

One can think comparatively. Armies of that order of magnitude were not common in better-documented eras until the seventeenth century, although they are certainly attested in the late middle ages. The largest recorded muster of the medieval kingdom of England was just over 32,000 men, in the 1340s. One might mention, in contrast, the capacity of the Zulu kingdom to put armies of over 20,000 men into the field but it must be remembered that these armies assembled quickly, in a different sort of terrain and political structure, and dispersed equally quickly. Could the Alamanni have done something similar? It is worth considering how this could have happened but we need to think hard about those mechanisms and not simply assume them. Such an army would need, like the Zulu, to assemble, fight and disperse quickly. For the Alamanni, a point can be made here, to which I will return: the presence, in however decayed condition, of the Roman road network in the former Agri Decumates, which, perhaps would allow a large army to assemble quickly. Such a force would nonetheless not be sustainable over a long campaign.

The reason for that is logistical. The largest settlements in Iron-Age Germanic-speaking Barbaricum had populations numbering only hundreds. In other words, the economy of Germania Magna was not geared to the provisioning of large numbers of people. An army of 35,000 men would be a population group seventy times the size of the largest settlements in Germania. Historically, armies have sometimes been more numerous than cities; the English muster of 32,000 in 1347 – if they were all gathered at the same place, which is not entirely clear – was two or three times the size of most of the larger English cities, though still be smaller than the estimated population of London at that date. But not seventy times larger…! For that ratio Edward III would need to have assembled something like 700,000 men. The larger late medieval armies were, moreover, raised in urbanised societies with fully monetary economies and complex market, trade and supply networks and systems, taxation, better agricultural techniques with greater yields and so on: things not available to a fourth-century barbarian polity.

Moving of 35,000 men through the late antique landscape – on either side of the Rhine – would consume the grain supplies and livestock of entire communities every day, producing famine, and probably still going hungry. The late Tim Reuter described the effect on the ninth/tenth-century landscape of an army of 20,000 men as equivalent to the down-wind fall-out ellipse of a nuclear explosion; it’s an image worth remembering, not least because the socio-economy context of Carolingian Europe offered more possibilities for supporting a larger army than did that of Iron Age Germania.

A much smaller army, say 10,000 men – or less – could be catastrophic for the population of the regions through which it moved. To see the sort of devastation wrought on Roman provinces that is sometimes attested in the written sources does not require us to envisage enormous invading hordes.

The solution to these problems would be for individual contingents to bring their food with them. How many days’ supplies could be transported by a warrior? With baggage animals, the amount increases but, as the classic studies of military logistics have demonstrated, the demand for fodder and food for the animals’ attendants grows parallel to the number of baggage animals. Charlemagne expected his warriors to bring food for three months’ service in the early ninth century. Again, while that is an interesting comparandum, one needs to take account of broader context. Charlemagne’s was a much more complex and well-organised realm than any in fourth-century Barbaricum: urbanized and monetized to some extent. He also enacted at length about supply carts, package animals, the preservation of fodder for passing troops and the maintenance of bridges and boats. The law about a three months’ supply also applies to the caballarii of a monastery, mounted troops supplied from a large land-owning establishment. Indeed, Charlemagne envisaged that the resources of four farms were necessary to furnish a single warrior. How much of this can be applied to Iron-Age barbarian armies? We would need much better evidence than we currently have before we could assume that much of it was applicable in detail. Charlemagne’s field armies were almost certainly much smaller than 30,000 men. Karl Leyser suggested that such armies were fed as much by cattle on the hoof as by waggon- or pack-horse-loads of grain; this might apply to barbarian forces too. If so, however, issues about fodder and ease of movement across difficult terrain are emphasised.

It’s nonetheless worth remembering that in the 350s a force of 600 Franks caused Julian serious problems.  There’s no logical reason to suppose that Roman writers were more likely to be telling the truth when they mentioned low numbers than when they recorded very high ones but we might accept that this was a ‘small’ force. The Romans defeated it by starving it out. Before that, it can have done enormous damage.  Six hundred warriors would be more than entire Roman communities outside the towns and could move rapidly, consuming settlements’ bread, killing and eating the cattle they were saving for winter, murdering anyone who objected, raping and/or enslaving women and children with impunity. If a couple of Roman field army regiments caught them, the game would be up and swift punishment meted out, but it’s vital remember the damage and trauma even a small raiding force could cause.  Let’s not sanitise this!

This has important implications for barbarian migrations that have rarely been discussed. How feasible was it for a barbarian ‘horde’ of 20,000 warriors plus families and followers – up to the number of 100,000 frequently found in modern discussions (including mine) – to travel any distance or for any length of time?

We must ask serious questions before repeating the numerical estimates that have become almost canonical in the literature. My paper is mostly about asking questions – frequently quite obvious ones.  Many are unanswerable, in some cases perhaps only in the current state of our knowledge but in others they concern things we can never know. Nonetheless the existence and importance of these questions need to hang over future discussions of migration. We can’t simply assume things. We can’t be afraid to challenge age-old consensuses, to point out the difficulties in making assumptions or to admit there are things we can’t know.

Consequently, I am not going to argue for specific answers. Too much writing about the history of the migrations, including some of mine, has been about arguing that one or other interpretation must be correct. We need to stop trying to win arguments and to establish interpretations that can be labelled right or wrong. This attitude has impoverished scholarly and political debate in Britain and produced Brexit. We must recognise that great swathes of the study of what we call the Barbarian Migrations can’t be reduced to totalizing ‘correct answers’ but are areas of discussion where many possibilities must be kept in dialogue. Highlighting answer-less questions reminds us of this.

How feasible is the traditional image of a migration of people?


3. The peril of analogies

Over the past thirty years there has been much use made of so-called Migration Theory – a body of general comparative observations about the phenomenon of migration. This has, to be sure, provided useful information and permitted a more sophisticated understanding of late antique migrations.  There are, however, some limitations, especially as far as large-scale migration is concerned. Many of the most useful insights of migration theory concern what one might call migratory processes: the different forms of migration; the flows of information; the demographic make-up of certain types of migrant groups in particular types of movement; the various push and pull factors involved.  Where thinking about large-scale migration, migration theory tends to discuss a longer-term, cumulative process such as, for example, European migration to North America.  That individual and small-group migration from Barbaricum occurred and, over the whole Roman period, resulted in the movement of possibly hundreds of thousands of people into the Empire seems uncontroversial.  It is quite feasible that areas where extracting a living was hard and, over time, possibly becoming harder were largely depopulated because of migration.  It’s perfectly likely that many of the people departing such areas moved into the Roman Empire, but there’s no reason to suppose that they all did.

It must be said nonetheless that many of the forms of migration studied, and which can be shown to have existed in Iron Age Germania were predicated on the possibility of return. Many analogies for migration occur, furthermore, in very different economic and technological contexts from those of Late Antique Barbaricum. Most importantly, the fact that these types of migration are not merely possible but demonstrable in our period does not, logically, authorise us to declare that the paradigm of The Great Migrations, or Invasions, has been validated, as I have recently seen argued.  This is what philosophers call a ‘category error’: the application of a conclusion drawn from one thing to something else as though they were the same thing, though in fact they are quite different.  The Great Migration, Völkerwanderung paradigm is not about the gradual movement of hundreds of thousands of people over centuries; it is about a specific, short-term, large-scale phenomenon: the arrival in one moment of a single large group – counted in tens of thousands – to which contemporaries applied an ethnic name: the Goths; the Vandals; the Burgundians; the Lombards.  And the argument that the arrival of these groups resulted in large-scale political and cultural change: specifically the End of the Western Roman Empire.  That sort of migratory phenomenon is very difficult to find convincing comparative material for. Though providing much interesting logistical information, things like the westward migration of European Americans or the Boer Great Trek are bad analogies, being the cumulative movement of individually quite small groups rather than of whole peoples.

We must start from Iron Age realities.


4 Different mechanisms of migration and interrelationship in different zones of Germania

It’s crucial to underline the well-recognised fact that the phenomenon of ‘barbarian migration’ was diverse, including a wide range of movements.

I would divide these first by ‘zone’ of Barbaricum. In the past I have adopted a rather crude threefold division of Germania Magna, according to practical nearness to the frontier. This needs refinement and more subtlety but for now it might be useful.

‘Immediate’ zone

I provisionally labelled the first zone the ‘frontier zone’, though I later thought that a better term was needed to avoid perpetuating the idea of two opposed and confronted ‘worlds’. The Immediate Zone might work.

Here, relationships were sometimes fraught. This is the zone from which raids on Roman territory were launched and where, possibly more seriously, Roman military retaliation took place. At the same time, though, this was also a zone where the web of Roman-non-Roman connections was dense. By c.300 Roman material culture had come to dominate this zone. Forceful arguments have been made that Roman imports into Barbaricum before the third century were intermittent and dependent upon particular political events.[2] By the later Roman period, as far as I can see, cross border interactions were much more general.

All kinds of migration are practical in this zone, from the very small to the large-scale. Individuals seeking employment in the Roman army; families seeking a better life than was available on the lands they currently farmed; and so on. We know a lot about the legal relationships between the Empire and various categories of barbarian immigrant and I won’t discuss that; it is well covered in the literature. The key features that I want to discuss later will surely have applied but to a lesser degree.

The one form of larger-scale migration that I want to discuss concerns migration into the Empire of particular groups for political reasons. From the earliest late Republican encounters with the barbarians north of the Alps one of the most important, repeated dynamics was the movement to Roman territory of ousted political factions, particularly those supported by Rome.

The relationships between the British Empire in Natal and the African peoples across the Mzinyathi (Blood) and Thukela rivers before and after the Anglo-Zulu War of 1879 provide very instructive comparisons.[3] In both cases the Empires set up or supported their preferred local leaders and were expected to allow their factions to cross into their territory when defeated. Pro- and anti-imperial factions and attitudes beyond the frontiers swirled and shifted according to circumstance. The military balance of power was overwhelmingly, insurmountably in favour of the imperial powers in any protracted confrontation, whatever short-term, morale-sapping and dramatic victories might occasionally be won by their opponents.

Factions that took refuge in the Roman (or British) imperial orbit could be numerous and politically very significant, but they were not entire peoples. I would like to stress this. The choices before us are not simply between small aristocratic warrior groups (as in the ‘elite transfer’ model) or the migrations of hundreds of thousands of people (as in the traditionalist Völkerwanderung model). In Barbarian Migrations and the Roman West I frequently wanted to stress this intermediate ‘factional’ order of magnitude. It’s a disappointing index of how entrenched, simplistic, polarized and binary the debate on the migrations has become that, if one does not suggest or believe in the movement of whole tribes, one is assumed by default to be arguing for the migration only of small groups of hundreds (or very low thousands) of warriors, or even to deny the existence of migration. What I want to suggest is how significant, complex, difficult, and traumatic even a migration on this intermediate scale could be.

In the famous Tervingian migration of 376, it is crystal clear from the accounts of Ammianus Marcellinus and others that this was not the migration of the whole people. It is not clear that it was even the movement of the majority of the Tervingians. It was clearly a lot of people but that didn’t need to mean the whole or even a large proportion of the trans-Danubian Gothic population.

There are very distinctive things about this migration, nonetheless. One is its economic and ecological situation. Semi-nomadic pastoralism played a large part in the economy north of the lower Danube, as it did in every period we know about between the Bronze Age and the late nineteenth century at least. It was not the only element, as the archaeology of the Černjachov Culture shows, but written and archaeological sources – and the fact that Roman authors, especially those writing in Greek, considered the Goths to be Skythai – underline that a mobile, pastoral existence was a significant part of their way of life.

For at least some elements of the Tervingian population, packing one’s goods into a wagon and moving was not beyond the normal range of experience. Some of those that moved must nonetheless have belonged to more sedentary elements of Gothic society. For all, though, as far as we can reconstruct the circumstances of 376 – and whether you follow a short-term ‘exogenous’ Hunnic explanation or a slightly longer-term ‘trans-Danubian instability plus Hunnic intervention’ interpretation – these circumstances made the movement of Alaviv and Fritigern’s Goths a fairly sudden decision. Organising everything necessary for a sustainable long-term movement was doubtless impossible.  One imagines that they took whatever grain they had and drove their herds and flocks before them, as food.  The distance the Tervingi had to cover to reach the Danube was nevertheless pretty short, across easier terrain and through their own confederacy’s territory.  All this makes the assembly of tens of thousands of Goths on the northern bank of the Danube perfectly plausible.  Even so, it is likely that the Tervingi were going hungry before the Romans ferried them across the Danube.  Once across the river, as we know, things got even worse.

The Tervingian migration was exceptional, but exceptional doesn’t mean unique, and crises like this had happened before on the Danube, not frequently but, at least over the longue durée, regularly – the receptio of the Sarmatians forty years previously had been the most recent instance. The features of Tervingian migration just described facilitated this. The mechanics of migration in the Steppe and neighbouring regions need to be studied separately from those further West.  We need to think more about the differences between these different contexts.


‘Intermediate’ zone

I want however to focus, less on movements like those of the Goths, than on migration from regions further West, behind the ‘Immediate Zone’, in what I call the ‘Intermediate Zone’.  Some of the most significant fifth-century migrations appear to have come from this part of Barbaricum.

It’s clear that contacts with the Roman Empire reached the north of modern Germany.  If less frequent than in areas close to the frontier, they were nonetheless socially and politically important, perhaps more so.  To provide some well-known examples: the official imperial metalwork found in Lower Saxon cremation burials; the trade that flowed up the river networks from the Black Sea and the Danube to Scandinavia; the prestigious Roman silverware found in exceptional burials like those at Haßleben; the importance of Roman cultural influences such as those that led to the adoption of Roman-style inhumation, or the influence of Roman decorative styles; and so on.  Who knows what the inhabitants of the Baltic coast thought the Roman Empire was like – maybe their ideas of the Empire were as mad as the Roman ideas of what the people on the Baltic were like.[4] Nonetheless, the people of central Germania Magna knew about the Empire and knew it was a powerful, culturally-dominant place.

The same reasons for migration as in the areas close to the frontier applied here too: individual or small-group career migration; small-group/familial migration for economic reasons; larger-scale migration of politically-ousted factions. But the mechanics of getting to the Roman Empire were quite different.  Leaving aside seaborne transport down the North Sea/Channel coast there were various options.  From the North German plain, one could move down the Hellweg and similar routes to the Lower Rhine, or one one might travel up the Elbe and its tributaries to Bohemia and thence to the Upper Danube.  Other routes might be across the central German mountains.  Clearly all were used.

They offer different problems and potentials.  The northern German and Elbe route are in some ways easier but they pass through settled regions occupied by other political groups. The central mountain regions were – as far as we know – quite thinly occupied in the Roman Iron Age.  Many provide good archaeological evidence of occupation in the Bronze Age and pre-Roman Iron Age; evidence which vanishes at about the time that the Roman limes were created, though pollen data suggest that there was not a complete absence of settlement.  This might make passing through the region less politically or socially difficult; one could graze animals and hunt. On the other hand, the terrain was more difficult and there were fewer communities upon which to draw for guidance or support in case of disaster.

Many unanswerable questions arise. Were there rights and obligations of hospitality for travellers in Germania Magna? Did approved travellers pass through a region under the protection of the local ruler? Did people hire guides? Did small groups band together for protection, like the caravans moving through Mexico to the US border in the contemporary world? How many migrations in search of a better life ended up in disaster: starving or freezing to death in the mountains; drowning in swollen rivers; murdered, raped, enslaved and sold on by local warriors; how many resulted in an Iron Age equivalent of human trafficking? We don’t think enough about the human dimension of late antique migration; the human realities behind the lines on our maps of contacts.

Three factors exist in dynamic tension: speed; supply; and security.  Individuals or small groups might move quickly and require fewer supplies but might be exposed to attack.  The larger the group, the more secure it might be, but the greater its need for supplies and the slower it would move. The optimum group would possibly be a medium-sized group of warriors, especially if mounted, which could pass quickly through intervening territories, cross difficult terrain more easily, negotiate passage and generally be too threatening to be worth attacking.

But under what circumstances – with what practicality – might a large proportion of, let alone a whole, tribe move? We can reject the idea that any barbarian group realistically thought that it could set off and conquer land in the Roman Empire. It might be that a gradual deterioration in the weather, harvests and so on could provoke a large-scale movement with the decision ultimately to seek admission into the Empire.  Such a decision would need lengthy preliminary discussion, political agreement and planning.  When do you leave?  After lambing/calving in Spring, but with only the grain left after Winter? After the harvest was in but with not long before Winter?  If harvests had been getting worse these decisions would have an extra dimension.  We can rule out movement due to actual famine.  Migration due to that kind of catastrophe tends to produce the tragic imagery of people dying of hunger, begging on the roadside – familiar from the Irish potato famines of the 1840s – not the appearance of mighty, conquering warriors.  In more recent periods we are presented with migrations during famines from the countryside to towns, and thence, sometimes, further afield but these scenarios require monetized, urbanized societies with proper transport networks which simply didn’t exist in late antique Germania Magna.  A catastrophic period of famine might depopulate a region but would be most unlikely to result in a single unidirectional mass-migration.  Weakened by hunger, migrants would easily fall prey, possibly willingly, to groups through whom they passed or among whom they sought sanctuary.

The traditional picture is of migration fuelled by the violent encroachment onto a territory of more powerful neighbours.  Again, it’s difficult to find evidence that this happened to the extent that entire peoples took to the road.  The Hunnensturm did not drive all, and probably not even most, Goths abroad.  Most Goths stayed and by c.400 had accepted or become involved in, Hunnic rule, married into or otherwise joined Hunnic society; all the famous fifth-century Huns have Gothic names.  It seems likely that this was the usual means by which political control changed, though I do not want to sanitise the process: killing, rape, violent displacement, enslavement also occurred.

One must, however, think about practicalities.  How could a group of, say, a hundred thousand people it have been able to move through Germania Magna? I imagine that groups had to be sent out to find routes and to negotiate passage through their lands with the rulers of neighbouring peoples. Elements of this are found in the narratives of migrations in our sources.  Those stories speak of deals done with their threatening aggressors.  All this seems plausible, although it’s also possible that, if I was able to think about these hypothetical problems involved in moving a people through the territories of others, so were the authors of these tales.  What if the neighbours said no?  Could a group like that pass through hostile territory without significant losses from perpetual ambushes, raids and opposed river crossings?

Even without enemy attacks, how long would the waggon train, the column of people, have been?  How many miles could it have covered every day?  I cannot imagine it being more than about 10km a day at most.  Gregory of Tours describes the party that accompanied Princess Rigunth towards Spain for her marriage – numbering perhaps 4000 people – as covering five miles (8 km) in one day. The whole would move at the speed of the slowest waggon.  How big would its campsites have been?  If one compares the probable length of the column (dozens of miles) with the distance it could cover per day (rather less) the column would soon break up into a chain of smaller contingents. A swollen or otherwise temporarily uncrossable river would nevertheless cause them all to pile up again.  How long would it take to cross a river, or climb over a mountain pass?  What was the mortality among the very young and old? Sickness would surely be a major threat.  Dysentery was a major killer in early medieval armies made up of relatively young, healthy men.

There is however an important factor to balance against these difficulties: distance.  We aren’t talking about the Oregon Trail.  Google Maps tells me that the distance from Bremen to Nijmegen is 170 miles or 273 km.  Taking into account the lack of good roads and bridges, and that things might not go so well, I suggest that a generous estimate is that our migration could get from the lands of the Saxons to the Rhine frontier in a couple of tough, unpleasant months.  If they set out in May, when – at least in Carolingian opinion – the fodder was sufficient, they expect to be on the Roman frontier in July.  There would be no going back though.  If the Romans decided not to let them in would the return journey be feasible? Would (in this case) the Franks let these Saxons back through their territory or would they see them as vulnerable?  If they were generous or the Saxons fought their way back, they’d find themselves back home – assuming those homes hadn’t been occupied by enemies or former neighbours – perhaps in October (allowing for a month of negotiations on the frontier), with no harvest to gather in (or which had been reaped by other people), with most of their livestock and grain store consumed.  This would not be an easy Winter.  The decision to put a people on the move was one from which there was no going back, something that understandably one finds in the written stories about migrations.

This hypothetical example follows the shortest and easiest route from the interior of Germania Magna to the limes (and oddly there were as far as I can see no great migrations on this route!).  Every other option involves longer distances and more difficult terrain – principally mountains.  At the opposite end of the spectrum, the route up the Elbe valley from Wolfsburg to Vienna on modern roads is 700 km long, across less easy ground.  That’s longer than the Great Trek which took small groups of Voortrekkers six months to complete.  A single, large-scale late antique migration on that route seems to me to be out of the question.  It would take the best part of a year and probably involve sitting out a Winter.  Unsurprisingly, therefore, we don’t find reliable reports of such huge Völkerwanderungen from one end of the Elbe to the other.  The second half of this route, from Prague to Vienna, would on its own be a similar distance as our first example, but across more difficult terrain.  One would be looking at something more arduous and from which the possibilities of return were slimmer.  That would be at least as true of any other route from the heartlands of Germanic-speaking Barbaricum to the imperial frontier.

My examples assume all went well, that passage was granted, that the locals got out of the way of the migrants as they passed through, and that the migrants didn’t take to rustling cattle, stealing food from or robbing the locals.  It’s easy to see how, if things went wrong, a major political and humanitarian crisis could occur, with locals reduced to potential starvation by the eating up of their food, skirmishing and fighting on the route and a general snowball (rather than domino) effect of displaced people, famine and probably disease in its wake.  To some extent perhaps this was what happened in 376.

What can we conclude from this?  Clearly not that a late antique migration of 100,000 people across Germania Magna was impossible, at least over shorter distances.[5]  I suggest, though, that such a movement could not have been easy or straightforward.  Late antique writers clearly understood that it would be a dramatic and possibly traumatic upheaval and so should we.  Such phenomena need to be placed at the bounds of the possible, in the order of the exceptional, not treated as a general feature of life in late antique Barbaricum.


5 The frontier and immigration

Let’s now reintroduce the Romans into the equation.  In 2014, I argued that we should not see the Imperium and Barbaricum as two opposed and confrontational worlds but as the core and periphery of a single, closely-integrated Far Western Eurasian world.  I also pointed out that the limes were not simply a barrier but also a mechanism of migration.  The frontier was not just a line on a map but part of an information network.  People passing through bring information about the regions linked (not separated) by the Limes: what was happening in the barbarian world; what was happening in the Roman - the opportunities for recruitment; where you needed to go to be accepted; what the best routes were; whether some routes needed to be avoided; what sort of requirements needed to be met.  Migration follows particular routes and entry points, it does not (contrary to popular belief and propaganda) simply wash over a border like a tide.  The tragedy currently being enacted on the southern US border makes this very clear.

The points on the Limes where barbarian migrants entered the Empire were dense nodes on information networks, which reached far into barbarian territory.  After all, Romans could communicate and have diplomatic relationships with the peoples of the interior of Germania, like the Burgundians. It’s interesting to speculate about what roles access to this information played in society in Germania. Did merchants serve as guides? As organisers of ‘caravans’?  As ‘dragomen’? Clearly, these information networks radiating from the frontier were essential to all the kinds of small-scale migration that comprised a big part of the everyday life of the frontier.  How might it have played a part in the larger-scale migration that has been my principal focus?


6 What was the Roman Empire’s involvement?

It is difficult to imagine that the Empire had no role in large-scale migrations, given the information networks just mentioned and the deliberate, planned nature of these movements.  It surely could tell peoples closer to the Empire to refuse passage of such groups if it wished, and provide military assistance if needed.  I imagine that that sort of provision was a part and parcel of the foedera made with barbarian rulers on the frontier and beyond.  If we accept that, which seems uncontroversial, then, by the same token, in other circumstances the Empire surely had the ability to instruct groups to let migrants through – especially if they had a prior relationship with Rome. 

The Romans knew well the advantages of barbarian immigration.  It was a good source of recruits for the army; and of tillers either of land that had been abandoned or of the estates of the rich.  Just as importantly, the admission of barbarian petitioners was an important element of good imperial rule.  Remember the Tetrarchic medallion; or the commonplaces of imperial panegyric.  It was a fine way of showing how the whole world was indeed the Orbis Romanum, with the Empire at its centre.

This is where a more radical proposal emerges.  The best way to mitigate all the logistical problems set out in this paper was surely through Roman support.  The Empire could most easily organize and provide supplies for large groups of migrants. If the Empire wished to admit these migrants, it did not want them starving, emaciated and disease-ridden. It seems plausible, perfectly congruent with what we know of Rome’s frontier policies, that, alongside the usual gifts and sweeteners for the local barbarians, food supplies could be shipped into Barbaricum if needed.

This was most possible in the former Agri Decumates where a vestigial Roman road network existed and projected quite far into Germania – one reason, I am sure, why this is pretty much the only place where the Romans regularly campaigned beyond the frontier.  Tough migration routes from the centre of Germania Magna across the thinly inhabited central mountains could bring you down to the Agri Decumates.[6]


7 The scenarios again

On the basis of these thoughts let’s return to the three scenarios mentioned at the start.

The Tetrarchic Medallion

I am going to assume that we have the depiction of – if not a specific then at least a typical – occasion. The obvious question is: how were the Emperors there on that day? Presumably there must have been holding camps beyond the frontier, either until sufficient numbers of would-be immigrants had built up for a formal ceremony of receptio to be practical or worthwhile or, alternatively, until arrangements had been thrashed out with larger groups and the Emperors had arrived to oversee their entrance into the Empire.  Again, one would be left with the idea that such groups would, at least at this point, be supplied and fed by the Romans.

The Goths on the Danube

The considerations detailed above reinforce my reading of the events of 376, specifically that the crisis was not unprecedented in the magnitude of the group demanding admission but in the scale of Roman corruption, incompetence and mismanagement, all emphasized by the absence of the Emperor and his leading officials, far away in Antioch.  I would now add to that the speed at which events unfolded.  My considerations lead me to suspect that the specific circumstances of the Tervingian migration meant that it could have happened much more quickly than would usually be the case – without the usual advance warning – perhaps over only a matter of weeks.  Again, though, the consequences make abundantly clear how essential was the Roman logistical ability to supply very large numbers of people.

The Great Invasion

My third example was the ‘Great Invasion’ of c.406.  We know little about this other than it took place on the 31 December – the middle of Winter.  That fact is crucial.  Most authorities agree that the migration itself had something to do with the creation of a Hunnic hegemony north of the middle and lower Danube, even if they disagree on the details of that causal relationship.  How did the Vandals, Alans and Sueves (whoever they were) find themselves opposite Worms – far from the Hunnic bases of power – in the middle of Winter?

My hypothesis contains many elements which are not original.  I suggest that the Romans refused some elements of this group – such as the Alans – admission anywhere on the Danube frontier during or in the aftermath of Radegaisus’ invasion and, while feeding and supplying them, nevertheless shunted them westwards along the road north of the Danube frontier.  Other groups – like the Vandals – might instead have crossed the mountains from central Germania and come down into the Agri Decumates on that route.  Whether the ‘Sueves’ were another group from the centre or north of Germania crossing the mountains, or a group from the Danube shunted west, or a disaffected faction of the Alamanni is unknowable, but clearly the Romans ended up gathering diverse groups of displaced barbarians in the former Agri Decumates, on the edges of Alamannic and Frankish territory.  I don’t think we need suppose that all of these had started their treks together even if they ended in one place.  Under normal circumstances this was sound policy.  Grouped in an area where they could be monitored and supplied more easily, the Romans could arrange proper reception and the distribution of different elements to the army or to estates of various sorts in different regions.

These were not normal circumstances.  There was no emperor or effective imperial court at Trier.  Honorius was not going to travel to that part of the world, especially during Radegaisus’ invasion.  Archaeology suggests a profound socio-economic crisis in northern Gaul around 400 and it is likely that much of the Rhine frontier was not properly manned after the catastrophic defeats of the western field army in 388 and 395 and the need to provide troops to defend Italy against first Alaric and then Radegaisus.  If one accepts the traditional date of 31 December 406 for the Rhine-crossing, things had got even worse: the British army had raised a usurper who had invaded Gaul.  In these circumstances it is easy to see how, as in 376, the organization of proper supplies on the Rhine frontier could have broken down.  By mid-winter the assembled groups of barbarians would be in real danger of catastrophe and so forced their way into the Empire.  It’s interesting that the only group recorded as opposing them were the Romans’ allies, the Franks.


8 Migration within the Empire

Before concluding, I want to make a final point that I think underlines my thesis.  In fifth-century history we have numerous well-attested movements of large groups: the Goths of Alaric and his successors between c.395 and 418; the Vandals in the 420s; perhaps the Burgundians to Sapaudia in the 440s; the movements of Theoderic’s Ostrogoths; and so on.  Estimates of the size of these groups are only estimates but the plausible orders of magnitude are in the range of ten to twenty (perhaps thirty) thousand fighting men plus women and children.  The numbers of the latter are impossible to know and, as I have said before, their existence by no means proves that we are looking at a ‘people on the move’, though – as I have also argued – the binary choice of ‘people or army’ is incapable of capturing a complex and evolving situation. Whatever the case, we are looking at the movement of tens of thousands of people.  There can be no debate about that – certainly I have never doubted it even if I might guess fewer tens of thousands than, maybe, Peter Heather would.  Anyone arguing that I think these were the movements only of elite warriors has either not read my work or is misrepresenting it.  My point, however, is that these movements are crucially different from those in Barbaricum.  They are within the context of the Imperium Romanum, its road networks (and the multiplicity of practicable routes), its logistical capacity, its complex monetised, urbanised economy.  Many of these movements were, we know, supervised by the Romans themselves or made while in alliance with the Imperial government.  Not that this made them easy.  Supplying an army of 60,000 men in 362 had taxed the Empire’s logistical capacities to their limits but these were at least movements within Imperial territories.  Again, the safe large-scale movement of people is generally only feasible in a Roman context.


9. Conclusion

In 2014 I argued that not only was migration a normal, central part of the operation of the Roman frontier but that that frontier and the information networks centred on it were crucial to the possibilities of migration from Barbaricum.  Without the organisational structures of the Limes, migration of most of the types we can trace was effectively impossible.

The argument I have presented, which I hope you at least consider worth thinking about, broadens that thesis to encompass the kind of large-scale, short-term movement of people that is central to the Völkerwanderung paradigm – and which I think did occur throughout the Roman period.  Such movements are easiest to envisage as practical, as less of an enormous, existential gamble, either the the support and assistance of the imperial state or within its logistical and economic frameworks.  Accepting this does not require you to discard the idea that the Empire was brought down by, in Peter Heather’s words, the ‘exogenous shock’ of several such movements in a short period of time, if you are that way inclined, though it adds a welcome element of irony to the account.

This further emphasises that Empire and Barbaricum were crucially interconnected.  It densifies the links between core and periphery.  The final argument supporting this is that, once the western limes had collapsed in the middle third of the fifth century the sort of large scale migration that I have been discussing ended in the West. The last Volkswanderung, of the Lombards in 568, was after all a movement within the decayed structures of former imperial provinces towards the newly established imperial Italian frontier.

In 2007 I argued that the the End of the Roman Empire caused the Barbarian migrations rather than vice versa. Throughout Roman history, Roman crises certainly led to many of the kinds of movements that I have been discussing but, as I argued in 2014, this formula was incorrect.  As I hope I have further suggested this evening, it was in fact the collapse of the Western Roman Empire that caused the End of the Barbarian Migrations.


[1] I don’t find the counter-arguments against his 405 date as convincing as he does!

[2] E.g. M. Erdrich, Rom und die Barbaren. Das Verhältnis zwischen dem Imperium Romanum und den germanischen Stammen vor seiner Nordwestgrenze non der späten römischen Republik bis zum Gallischen Sonderreich Römisch-germanisch Forschungen 58 (Philipp von Zabern. Mainz 2001)

[3] On this topic, I cannot recommend highly enough Ian Knight, Zulu Rising. The Epic Story of iSandlwana and Rorke’s Drift (Pan. London, 2011), which is a far wider-ranging and more important history than its subtitle suggests.

[4] It’s an important corrective to the current emphasis on networks to think about what – if anything – the people at the opposite ends of the lines on our maps and diagrams thought or knew about each other.

[5] I think I would rule out a 700km trek up the Elbe, over the mountains into Bohemia across the Czech basin and over more mountains to the Danube.

[6] This was Alamannic territory and might be why the Romans were so keen to keep the Alamans under control in the fourth century – and perhaps also why the Alamanni were so dangerous when this was not possible - and also why they expected their diplomatic gifts to be of a particular quality!

Thursday, 20 June 2019

The Transformations of the Year 600: Book Outline

Nine years ago I started this blog as a way of helping me to focus on the project I was working on, thanks to a grant from the Leverhulme Foundation.  Nine years later - and let's take a moment to ponder the fact that that the free world took less time than that to win the Second World War - I finally have a plan for the book of that project that I am happy with.  It has gone through many, many iterations (including a 2-volume version and even a trilogy!) but in the end, having initially been dead set against writing 'another big brick of a book', I have decided to write a (probably even bigger) brick of a book.  If, you are interested (or indeed care at all), this is what I think it will look like.  I dare say there'll be changes in detail but I think I can run with this.

The Transformations of the Year 600: Far-Western Eurasia c.565-650

Chapter 1: History without ‘events’

                        The closing of the Far West; the ghost of Rome; theory of events; about this book
Chapter 2: Narrative Orientation

Part 1 The Material: The End of the Late Antique State

Chapter 3: Introduction: The Theory of the State
Chapter 4: Government
Kings and things; Officials, Taxation, Warfare and the Army, the law
Chapter 5: The Church
Church structure. Bishops, rural church. Holy men. Monasticism; description of doctrinal differences; non-Christian areas: paganism and Judaism
Chapter 6: The Settlement Pattern (1): The Arena
The fate of the villas; new settlements; caves; beyond the villa zone; hilltop sites; other elite sites; towns; Roman cities; emporia; possibly urban sites beyond the old frontier
Chapter 7: The Settlement Pattern (2): Dynamics
Roman-post-Roman continuity; Settlement patterns and economies; settlement organisation; settlement units; communities; land-ownership; law and inheritance; relationships between towns; between towns and lesser settlements.
Chapter 8: The Settlement Pattern (3): The Religious Landscape
Bishoprics; urban churches; urban monasteries; rural churches and monasteries; change through time; pagan religious landscapes; bog deposits; temples
Chapter 9: Trade, exchange and the economy
Local levels: ‘Gift-exchange’; barter. Intermediate: control of local exchange; monetisation. Inter-regional exchange: routes and zones; mechanisms. ?Slave trade; Change through time.
Chapter 10: Death and Burial
Burial rituals; Descriptive account of archaeological data; Interpretations; written sources’ accounts; role in society and economy (linking cc9 & 11)
Chapter 11: Social Structure
Hierarchy: Aristocrats; free peasants; half-free and slaves; social mobility. Community: family and kindred; gender; age; ethnicity
Chapter 12: A provisional materialist explanation
Struggle of kings and aristocrats for control of surplus; war; plague; climate change; regional explanations: Merovingian minorities; Augustinian mission and conversion; dynastic instability; Lombard conquest.  Attempt to unify these.

Part 2 The Ideal: The End of the Roman World

Chapter 13: Introduction: The subject and the world
Chapter 14: Rulership and authority
Ideology; kingship; aristocracy; masculinity; the family; ?abbacy/ministry
Chapter 15: Political Identity and community
The political subject; ethnicity and politics; politics and the political; public space; spaces of the political; the making of community, local and ‘national’
Chapter 16: Religion
Monasticism; the ‘ascetic invasion’; conversion; typology; the miraculous; uniformity over orthodoxy; ministry?; apocalypticism
Chapter 17: Belief
Time; gender; the body; the wild and the cultured; sickness and disease; ?the miraculous; nature; responses to plague and weather
Chapter 20: Conclusion
A world after Rome

Monday, 17 June 2019

Renegotiating Power and Identity in Earlier Merovingian Gaul: A Material Cultural Approach

[This is the keynote paper I gave to the 12 or so people who had stayed to the end of the recent conference on 'Renegotiating Power' held at Christ Church University, Canterbury.  Thanks to the organisers, especially Charlotte Liebelt, for the invitation, Leonie Hicks for chairing, and the audience for insteresting questions and discussion afterwards.  Thanks also to Rob Heffron (Sheffield) for some helpful information about gendered space in Christian basilicas.

The argument looks first at the ways of ordering space via architectural cues, at the breakdown of the distinctive settlements of the social elite - villas - and of the basilica used as a secular political space and at the replacement of both to some extent by the hall.  Then it examines the ways in which costume symbolised identity and in so doing was employed in creating political space - or the space of the political. Throughout, emphasis is placed upon the possibilities for miscommunication and thus renegotiation that inhere in all communication.]


This paper comes out of work I have been doing under the general heading of a project I started (ahem) nine years ago on The Transformations of the Year 600, which I am hoping that I might actually finish within the next couple of years.  One of those areas concerns what became of public space in the sense of the spaces of the political.  Another concerns why high-status sites are curiously absent in this period, or at least are, in the current state of our knowledge, not very archaeologically visible.  That does beg a number of questions to which I will return.  Finally, linking all of this, how political communities change in the period I am looking at, between c.560 and c.650.  I will talk principally about Gaul/France but I will bring in some other areas of western Europe here and there.  I am going to talk about the production of space, whether of politics or of the political (in the distinction made in French thought since the 1950s, between la politique (politics) and le politique (the political)).

Spatial Shifts

The key starting point for my analysis is the sociological studies of Pierre Bourdieu and, in his early work, Anthony Giddens, which, though very different, come together around the idea that social structure is not some extrinsic set of laws that governs social behaviour but is perpetually constituted and reconstituted by social interaction itself.  It is useful to think of it as a cumulative memory bank, an archive if you prefer, of those ways of interacting that people approve, and those of which they disapprove.  Every social interchange between people of particular categories – gender, age, social rank, ethnicity, etc. – has, by adding to that archive, the capacity or potential to renegotiate the limits of the acceptable. In this perspective, change is inevitable; the chance of social structures continually and exactly reproducing themselves over time are pretty thin.

There are nonetheless strategies that attempt to put the brakes on the renegotiation of social identities, or to keep the interplay of social categories within particular limits.  An important one is the use of space.  Space sets the tone for the exchange.  It sets up cues about how one deals with the particular people or classes of people that one might be expected to encounter.  Most of us are familiar with the awkwardness involved in meeting someone in an unexpected location or setting.  Location sets up a range of expectations, a script.  It literally sets the scene.  Obvious, though in many respects this is, it is actually fundamental to rethinking some points about the interplay of identities.

In the Roman world, different types of space were quite clearly delineated.  To give a couple of political examples we could cite, first, the reception rooms of villas.  The approaches to villas, as with later castles, were carefully devised to present a particular view of the house, passing along which, through gateways and into ornately-decorated reception rooms, set the tone, or the stage, for the encounter with the estate’s dominus.  Whether the visitor was a guest of more or less equal or superior status to the villa-owner, or a tenant or client coming to pay rent or beg a favour, the expectations of behaviour were clearly set up, framed and limits set upon the range of acceptable outcomes.

Equally, the public spaces of the classical city functioned in similar ways, whether we are talking of the for a, the civil basilicas, the baths.  Again, in many well-studied cities, urban planning made use of the possibilities of vistas and lines of approach.  These are cues; they establish the expectations of how to speak and how to behave: of bodily posture.  Bourdieu said that a component of the habitus was repeated, learned, bodily dispositions and uses of space.  This seems quite a good illustration of the concept.  What I want to add, though, and it is something to which I will return throughout this paper, is the possibility of slippage and miscommunication – the mis-cue – that inheres within visual cues precisely because they function ‘textually’ in the sense that I will outline later.

From Villa to Hall

The fairly traditional classical forms of space had undergone or were undergoing profound change by the middle of the sixth century.  By that time, the villa pattern across western Europe had disappeared or was in its final throes.  Wherever one looks in the former western provinces, there is no new class of settlement that replaces the villa as a separate elite residence and focus for display and consumption – no class of settlement that creates social space and distance in the same way.  Across western Europe, from the mid-sixth century onwards, the settlements we know about are much less clearly distinguished – whether hilltop sites in southern Gaul, Spain and Italy, the communal-looking remodelling of villa-sites in Spain, the villages of Italy, new settlements in Spain, or the rural settlements of northern Gaul and England.  It may be that some more obviously elite settlements were coming into existence in Anglo-Saxon England around 600 but such sites are generally not archaeologically visible in Gaul until the middle of the seventh century.

One common feature of settlements is the hall.  Clearly there are all sorts of spatial, hierarchical cues in the hall but they are of an importantly different variety from those of the villa and there is a key theme of commensality as the nexus of social interchange.  This needs more work and I would be glad of any thoughts or recommendations but it seems to me that there are some very important differences, in terms of the experience of space, between Roman public assembly or reception spaces and the halls of the post-imperial period.  One might start from the location of the entrances and the perception of spatial hierarchy as a subject entered the space.  At least when used as a reception chamber, one entered the space from the opposite end of the building’s long axis from the seat of the dominus.  One entered facing the lord and furthest from him (or her), behind an audience facing away from you.  The experience of space was one of approaching as close to the focus at the front as one felt one was worthy.  The main entrances of post-imperial halls, by contrast, were on the long sides of the building.  It might be that some of these opened on to a corridor and that the main reception hall was thereafter entered, analogously to the basilica, opposite the lord’s seat at the far end of the room.  Where this was not the case, though, one entered from the side, some way between the lord and those seated furthest from him, and one entered in the gaze of most of the people present.  The decision of where one should or could sit, whether to move towards or away from the Lord’s seat, was thus made and enacted in front of an audience.  This was all the more true, given how one imagines the benches were laid out, if one entered opposite the Lord, though the movement would concern how far towards him one moved.  The arrangement of the tables means, however, that the lord’s seat was not the sole possible visual focus of the space.  Another key shift, alluded to earlier in the references to benches and tables, is to the seating of the community.  Other than in the senate, the Roman political community stood, with the exception of the dominus (whether Emperor or local lord) who remained seated.  This is but one instance of the shifts in the political gaze that occurred between the disintegration of the western Roman Empire and the early seventh century.  Add to this the different sensory and emotional architecture of basilica, on the one hand, and the hall, on the other, and I think one can gain an impression of a real shift in the experience of enclosed political space between the fifth and the seventh centuries.

How this shift might have come about is an intriguing problem and very difficult to answer. Most of the arguments usually proffered stumble on the same block.  A move away from the old villa-focused uses of social space to the kind of hall just described has been variously ascribed to ‘Germanic’ influence, a rejection of Romanitas, or the militarization of society.  All of these have something to be said for them, even the allusion to ‘Germanic’ influence – and I don’t often say that! – but they all run into trouble in dealing with the fact that the highest rank of the Roman population of early Merovingian northern Gaul were the ‘dining friends of the king’ (the Convivia Regis) whereas the Frankish equivalent were the members of the Trustis Regis – the Antrustiones – the senior members of the royal bodyguard.  So, the group defined, in a sense, by its involvement in commensality is defined by its Romanness and in opposition to ‘Germanic’, barbarian, military identity.  One could of course object that this was a different form of dining culture from that of the hall and the ‘mead-benches’ but it is difficult to see the continuation of the context for the old sort of Roman dining in the Gaul where that law was drafted. 

Clearly halls are important in the settlement architecture of Germania Magna.  Architecturally it seems very likely, at least in some areas, that at least part of the influence came from there, but the simple ethnic ascription won’t suffice.  The phenomenon is too ubiquitous and the origins of the Germani who eventually settled in the different parts of the former Empire too diverse.  More to the point, the aisled hall had plenty of antecedents in the Roman world, from various forms of settlement.  One was the typical ‘cross-hall’ of the principia found at the centre of every Roman fortress.  Roman military buildings had, however, undergone considerable change in the later imperial period and are famously less well-known or understood, and more diverse, than their precursors.  Halls are nevertheless known from forts – perhaps most famously in Britain from Birdoswald on Hadrian’s Wall.

The search for origins, though, probably misses the point.  The type of social interaction for which the hall set the scene is probably itself symptomatic of the socio-economic changes that brought about the demise of the old villas.  I would like to suggest that the kind of relationship between lord and follower implicit in the feasting hall is crucially different from that signified in the audience chamber or the basilica.  The provision of food in the format of the shared meal is indicative of a very different form of reciprocity from that of the old aristocrat-client or landlord-tenant relationship.  A clientship of sorts is produced of course but a closer, personal bond, in a smaller, more face-to-face arena.  That shift in relationship between an aristocrat and a follower seems to me to be central to the demise of the old Roman country house and its hierarchical spaces. 

The gradual disintegration of the Western Roman Empire undermined much of the local and regional security that kept local aristocrats in their position.  This happened early and quickly in the north-west; the process was slower elsewhere.  The top tiers of the Roman aristocracy lost access to lands overseas and the revenues from them and had to focus their efforts on a particular diocese: Gaul, Spain, Italy, Africa, or the East.  Even within these regions political change, fragmentation and uncertainty probably led to the loss of outlying estates and a concentration upon lands in only one or two neighbouring civitates.  The importance of the civitas as the centre of political identity and allegiance in Gregory of Tours’ Gaul is well-known. 

As well as the reduction in wealth, however, the restriction of effectively-managed estates to much smaller geographical zones meant the reduction of the social distance between the upper and lower tiers of the aristocracy and a new, more evenly-matched competition for local and regional authority and status.  In this context the need to acquire local support increased and it is not difficult to see the cost of doing so decreasing the amount of wealth available for the upkeep of villas of the old style.  At the same time, however, the spaces delineated in that old architecture would become less useful in the creation and maintenance of local power.  Previous explanations for the demise of the villa, including my own, have invoked too simple a cause-and-effect model, whether the cause be economic contraction, an abandonment of traditional Romanitas or the militarization of the provincial aristocracy.  The argument I am proposing here envisages economic constraint, for the simple reason that I cannot see why the Roman country house would not have been maintained by the aristocracy had it the economic wherewithal to do so.  It does not imply a necessary decrease in the productivity of the land; what is at stake here is the control of surplus, not the capacity to produce surplus in the first place.  But the model I advance also accounts (or attempts to account) for the precise architectural or structural changes involved.

The end of the civic basilica as a political space

There might be a further reason for the changes away from traditional Roman reception areas.  Now, as Derrida argued over 50 years ago, all communication works according to the same general principles as written text.  In order to convey information, each sign – each grapheme in his term – must be capable of iterability: repetition in a context where one or both of the parties to the communication, transmitter or receiver, are not present.  Once any sort of signifying grapheme is understood to convey a particular signified, then it is capable or reproduction outside its original context.  Indeed, one of Derrida’s key points is that there can never really be an original context; the capacity for iterability that separates sign from context was always already present.  This applies to everything, including buildings.  A type of building, once recognised as such, acts as a sign, a combination of signified and signifier.  This applies even to the ‘unique’.  Once a structure is recognised as a particular building it acquires a meaning, a signified content, and that signifier can be employed outside its original context.  Take the Eiffel Tower: a unique building but capable of endless repetition in new contexts, as in Las Vegas or on a key ring.  Indeed, it occurs to me that many of the best-known buildings of Las Vegas stand as an architectural illustration of Derrida’s concept of iterability. 

The concept is equally well-illustrated by the basilica.  At some point in classical antiquity a particular form of building was understood as meaning an assembly hall.  Within its semantic range was the audience chamber, in which the emperor, his representative, his statue, occupied the focal point within the apse at the far end of the central nave.  When Christianity, permitted to build its own structures and now the favoured, then official, eventually exclusive religion of the Empire, built churches it did so, as is well known, on the basilical plan.  The ‘sign’ of the basilica was essentially repeated in a different context.  If you like, the semantic range of the sign widened further.  Could a stranger tell which was a civil basilica and which a church?  Location would be a clue: Christian basilicas tended generally to be located on the edges of towns; the civil ones in the old municipal centre. I am not suggesting that late antique westerners habitually bumbled in and out basilical structures at random, taking a wild guess at whether it was a church or an audience chamber.  Nonetheless it is interesting to think how the iteration of the basilical form might have created a space in which power and identity were renegotiated.

Basilicas had always had a range of functions; what interests me is the wholesale reproduction of the hierarchical spatial organisation of the civic basilica.  The space occupied by the emperor, his image or his representative becomes occupied by the altar and the officiating priest.  This means quite a reshuffling of the usual hierarchical arrangements.  In the palace/audience chamber the emperor or secular leader occupies the key space and nearness to or distance from him – or occasionally her – was determined by secular worldly status.  Those at the front would be the highest-ranking and clergy would be expected to respect that hierarchy.  If one moved next door to the cathedral the bishop would occupy the centre of the space and secular officials, even emperors and kings, would take their place relative to that.  From one building to another, who was or was not permitted entry was dependent upon different people, and different considerations.  It is very likely that there were significant readjustments in the gendering of space between the civic and religious basilicas.  Women were allowed into churches but how many women rubbed shoulders with the men in the main aisles of civic basilicas?  Doubtless there were innumerable local variations, not least dependent upon architecture, such as the presence or absence of galleries. 

This must, given the similarities in spatial layout, have given rise to myriad interactions and renegotiations, infractions and reactions.  You can get a sense of some of these from sermons of Caesarius of Arles.  Caesarius berates his flock for conducting business in church and general chatter, quite apart from trying to leave the building before he could give his sermon!  Caesarius says a lot about posture and comportment.  Don’t lie down as though you’re in bed, he says; sitting is fine if you are old or infirm.  Stand or prostrate yourself to pray; bow your head or genuflect to the Host.  Matters went beyond that though.  One of Caesarius’ repeated themes was self-control and concentration.  Keep your mind on God and on prayer; don’t be distracted by other thoughts.  Idle speech and impure thought offered a way in for the demonic.

How do these ideas and instructions contrast with the usual bodily dispositions?  What were the restrictions on talk and posture in the civil basilica?  Could you lie down at the back if you were tired?  As mentioned, though, a dominus, local or imperial, sat when he granted an audience, and his petitioners, counsellors and the rest stood.  In church all stood or bowed, regardless of worldly status.  What did it cost a lord to bend the knee or prostrate himself with everyone else and was it a price freely granted?  At the highest levels, perhaps not.  There are some pretty fraught confrontations in churches between bishops and emperors, empresses and kings.  One of the more interesting is that between bishop Nicetius of Trier and King Theudebert I of the Austrasian Franks, related by Gregory of Tours in his Life of the Fathers.  This showdown concerns Theudebert’s entrance into church with a number of his senior aristocrats or leudes, whom Nicetius had excommunicated.  Nicetius declared that he would not continue mass until these men had left his cathedral; the king refused to send them away.  Who was in charge in this space?  In other cases the palace is the location for the confrontation, as in the Life of Saint Martin, where Emperor Maximus is compelled to stand to receive the holy man, or in Gregory’s account, again in the Life of the Fathers, where King Chilperic of the Burgundians feels his throne tremble as if there was an earthquake when the fearsome abbot Lupicinus enters the palace.  Whether this forced him to stand up is not specified but it seems reasonable.  One interesting point about that story, though, is that Chilperic is described as being seated at table with his courtiers.

An intriguing reverse example can be found in Book VII of Gregory’s Histories.  Gregory tells us that in 585 in Paris – he does not say where but probably one of the Cathedral basilicas on the Ile de la Cité – the deacon asked the congregation to be quiet to that the mass could take place.  Apart from providing a glimpse into the realities of a Merovingian church, this is actually a part of the Gallican liturgy.  It precedes the address by the bishop.  Yet it was not Bishop Ragnemod who spoke next but King Guntramn of Burgundy.  Guntramn essentially made a plea for loyalty to the Parisians, at this point effectively under siege by an Austrasia army.  This was not the only time that Guntramn played the part of a bishop in Gregory’s Histories and in the Edict that he issued in conjunction with the Council of Mâcon that same year he espouses, a decade or so before Gregory the Great’s Pastoral Care, the idea that kingship is a ministry.

So, in a church the bishop takes the space usually occupied by the king but, in a church, sometimes a king might speak in the place of the bishop.  Below that level there were countless other shifts in disposition and in the relative positioning of people of differing status and gender.  The verses composed by Venantius Fortunatus for the basilica of Saint Martin in Tours are designed to impose upon the visitor the sense that one ought to approach no nearer the front than one was worthy but, on the other hand, the surviving wall mosaics at the back of the nave at Sant’Apollinare Nuovo in Ravenna suggest that the decorations in the secular and religious buildings might not always have been very different.  In buildings that were organised, in terms of spatial hierarchy, pretty much identically, where were the semiotic cues?  Who was in charge of this space, ultimately?  Who controlled the terms of the discourse? 

There are yet more points to add into this mix.  One is that, as has become increasingly clear to me over the past decade or so, the fifth century was really characterised by the Christianisation of political discourse.  One of the many ways in which this is shown is in the building of churches.  This continued into the sixth century, when in some parts of southern Europe especially churches were built on villa sites.  One of the upshots of this was surely that secular rulers and leaders legitimised their position through public appearance and devotion in church; that this in turn became a new means of demonstrating leadership in the community, in a sort of spiritual commensality.  I suspect nonetheless that this might have been a further factor that made the traditional secular basilica, as an architectural form, a contested space, an arena for the renegotiation of power.

Furthermore, the authority that a secular lord positioned in front of the apse at the end of a basilica was, as mentioned earlier, largely sanctioned and bolstered, symbolically, by the fact that he occupied the place of the emperor, as his agent or representative.  After the western wars of Justinian (533-65) this symbolic support was cut away.  Justinian based his wars of reconquest upon the idea that the western Empire had been conquered by barbarians and thus was no longer a part of the Roman Empire.  This was news to the occupants of the western provinces who, while clearly aware that the pars occidentalis currently had no emperor, certainly did not feel that the Empire itself had come to an end.  Most of the rulers of those regions thought that they held an official title as an imperial official, legitimising their rule over Roman citizens.  Indeed, even their title of king was essentially one adopted to facilitate relationships with the Emperor and to legitimise power in his eyes.  The imperial declaration that the western provinces were not part of the Empire cut the traditional ways in which power was legitimised from underneath western rulers and, in turn, their officials and commanders.  It is possible that this sort of cultural shift played a part in the end of the villas

I would like to argue that if one put all of these factors together one might be able to see why the basilical form drops out of the repertoire of secular political spaces, even though it is clear that at least some aristocrats continued to have the wherewithal to build them.  In the eighth century, the Carolingians seem to have brought them back, but that would be a different story. 

Making space. The Materiality of identity

Public space had become quite different by 600 AD.  The clearly demarcated political arenas had atrophied.  Aristocrats and others, men and women, were more likely to rub shoulders in a far less structured fashion in all kinds of spaces, whether settlements, churches, religious processions.  How could one attempt to restrict the free renegotiation of status and power in this setting, without the old architectural or spatial cues?  I want to discuss some ways in which identity was materialised and, in so doing, produced a particular space, or distance; created a spatial structure for social interaction.

To do this I want to think about the ways in which the subject is presented/presents itself.  By way of a metaphor, it might be worth considering someone coming into one of the halls I discussed earlier, coming in, like Bede’s sparrow, from the dark into the warmth and glow of the fire.  In one of his early works, Time and the Other, Emmanuel Levinas first introduced his concept of the ‘il y a’, the ‘there is’: the notion that there is always something and someone ‘out there’; the ineffable sense of shared existence.  Levinas uses the metaphor of sleeplessness, lying awake in the dark, sensing that other shapeless existence, but sitting alone in a fire-lit hall, looking at the door, might provisionally serve almost as well. Levinas discusses the sense of solitude, of being ultimately alone in your own being, but within that shared existence.  The moment of presentation, for which Levinas used the term hypostasis – let us envisage it as the moment where someone steps into the light from the darkness outside, before which we only sensed their presence – is the instance where that solitude is made material.  It is a moment at the extremes where a being touches being in general.  At that moment though, that solitude becomes dispersed into various categories which are shared with others, identities.  One might want to think this phenomenon with Jean-Luc Nancy’s discussion of community.  He describes what he calls, using his own neologism, as comparution, translated by the equally neologistic ‘compearance’, a shared appearance with others, appearance together.  This is a simultaneous appearance and withdrawal in Nancy’s view: the appearance of someone or some category/identity that is familiar, simultaneous with a withdrawal: the interiority, the secret thoughts of the subject.  In Nancy’s thought, it is a hesitation on this moment that keeps community, in his terms, ‘unworked’, ‘désoeuvrée’.

If we, like Nancy, pause at this moment, how is the subject to be comprehended in social interaction?  How is the subject identified ascribed an identity, categorised?  How does the subject present itself to its audience, to those in whose gaze it finds itself, to those amongst whom it finds itself thrown?  As noted, we are thinking of a moment and a circumstance where spatial cues are of no help.  This is where the archaeology of earlier Merovingian Gaul is of interest.  By the end of the first quarter of the sixth century, across Gaul north of the river Loire, whole communities had adopted the custom of burying its dead with grave-goods.  Increasingly, the study of these goods and other aspects of the burial ritual has shown – in Gaul and its northern neighbouring regions – the correlation between particular types of grave-goods and the age and gender of the deceased.  One of the great unknowables, of course, is the degree of correlation between the association of particular classes of people with types of costume and artefacts in death, and the relationships between such objects and costumes and those particular categories of people in life.  In the Merovingian context at least, there is sufficient evidence to support the hypothesis that funerary costume at least bore a reasonable relationship to formal dress.  Indeed, one might go further and suggest that the very degree to which Merovingian people lived their lives in the gaze of the community suggests that even ‘everyday’ costume may have born some sort of relationship to the formal and stylised construction of social categories in death.  If one ran the risk of meeting people in fairly random or unstructured settings then one needed some other way of keeping interactions within an acceptable set of parameters. 

If we are thinking about the renegotiation of power, we need to think more deeply about what identity is, what we mean by it, how it functions in social interaction.  Identity is a word that is ubiquitous in medieval studies – in paper-, book-, article-, chapter- and conference-titles – but there is hardly any serious theorisation of what identity is at all, even in the area of ethnicity.  Generally, what is discussed under the heading is the issue of groups, identifiers and labels, or it acts as some sort of vague ontological place-holder.

My earliest discussions of this topic (1995/1997) were based around the contingent, active interplay of different identities and the stressing of links and barriers in social relations or encounters between different people.  Much of this model was sociological in its inspiration and formulation and was concerned with how people achieve aims vis-à-vis other people.  It was concerned with power and principally a theory of status, value, worth and social roles.  The model worked according to the idea that identity was a stable entity that could be communicated more or less unproblematically.  It implied that identities were not only things that you had but also things that you were in a straightforward way. It envisaged a sort of free choice in the deployment of identity.  You picked an identity and invoked the power that went with to achieve your aims.  This now seems hopelessly naïve

However, all identities are categories: means of organising the world. As such, they are constructed as signs or groups of signs. Even where they are based upon differences that are, or might be, naturally-occurring or visible regardless (hair-, skin- or eye-colour for example; differences in genitalia; physiological stages of ageing), the choice to use them as categories, their precise definition, the way in which they are employed and therefore the ways in which the people of the categories so created experience their lives, depend upon their position in a contingent system of signs. As such they function textually (in the Derridean sense), within chains of presence and absence, similarity and difference. Because no concept can be understood separately from those signifying chains, or comprehended apart from its relationship with other signs, there is always something of the ‘different’ within the ‘same’ and that is very important to remember. To be Derridian about it, the first time anyone said ‘I am a Goth’ to someone else (and was understood), the term ‘Goth’ already had an iterable place in a signifying chain. Logically, if not temporally, the identity must be prior to its instantiation. It already related to an ideal, which was never coextensive with that which instantiated it, and to its constitutive outside (all the things which, ideally, it was not).

Identities function in the imaginary as well as the symbolic registers. That is to say that there remained (as with all signifiers) a notion of the ideal member of the category. Normally that was structured by some of the aspects which helped define the category (social and ritual mores, etc.) to create concepts of the ideal member of a sub-group within it (young woman, male elder, monk, king etc.). This has important implications. Social identities are constituted in citation and in performance. Even more crucially, identity is itself a motion towards an ideal. The ideal can never be attained, because it never had a pure, originary existence. It’s a motion of desire: what do I want to be, but also, crucially, what do they want me to be? As Lacan famously said, a fool who thinks he is a king is no crazier than a king who thinks he’s a king.  (He might better have said that a fool who thinks he is a president is no crazier than a president who thinks he’s a president.)  In any interaction there are at least two sets of signifieds in play: both parties’ ideals of what their status and identity and that of the other person means.  These might, of course, not coincide.  The performative citation of an identity is always, to some extent, a risk, a wager.  That is one of the most important things I want to stress.

Those ideals, moreover, are always themselves changing in the course of social practice. They can never be entirely recreated. It is thus critically mistaken to talk of the maintenance of a Gothic or Frankish identity by a particular group, whether the guardians of the Traditionskern or an equally mythical group of Gothic Königsfreie; no such thing had ever existed that was capable of maintenance in the first place.  It was always already in a state of renegotiation and reinvention.

I must underline the textual and discursive elements that are central to identity, and the inescapable fluidity that that implies. I also want to link identity to speech, subject and authority.  To deploy, perform or cite an identity is to give an account of yourself – to borrow a phrase from a recent book by Judith Butler – but it’s also, as I said, a wager on recognition: of the identity-ideal, the signifier, and of the right to speak/act from that subject-position.  It is in the element of risk or wager that I differ from Butler.  That links identity to subject-position, and indeed to subjectification.  One of the most important ways in which an identity or subject position was made manifest in late antiquity was through costume, broadly defined (including the artefacts carried with it, buckled on to belts and so on).  It conveyed information about the person sporting it, and the social category to which they belonged.  The repeated patterns of association within the sixth-century Merovingian cemetery record suggests that costume was capable of transmitting fairly precise information, about adolescent boys, young women, old men and so on.  As such it provided cues as to how one might expect such a person to behave, how one might judge their speech, how one would be expected to behave towards them.  This provided the cues that could create social space or distance. 

We can read some of this from Merovingian written sources such as the laws, which penalise touching of women’s bodies.  These parts of the body are generally those highlighted by Merovingian jewellery.  The laws’ system of wergilds also set out various levels of legal protection or esteem for particular people: women of child-bearing age; young boys; Franks; royal officers, and so on: all categories that seem to have been visible from the costume of the person in question. 

As we have seen, to be capable of communicating any sort of information, any concept must be capable of iteration, that is able to refer not simply and exclusively to that specific instance but to others too.  This implies the ever-present chance of misunderstanding or miscommunication in the interplay of identities.  This is a key support of Judith Butler’s work on, for example, performative gender identity and drag.  We can see iterability illustrated with the figure of Zercon the Moor, the “jester” at Attila’s court whose “act”, so to speak, involved dressing up (or being dressed up) as a warrior.  Because Zercon was a dwarf, the Huns, for once living up to their stereotype, found this incongruity hugely entertaining.

An example a little closer to Butler’s might be found in the Poitevin who appears in Gregory of Tours’ account of the tribunal that ended the Nuns’ Revolt at Poitiers.  In Gregory’s description, this was a man who in Gregory’s report of the exchange dressed as a woman because he was ‘were incapable of manly work’.  This is a complex text to read in terms of that person’s identity, and how the semiotics of their practice worked is difficult to disentangle.  This difficulty is only magnified by another iteration of feminine costume.  Several late antique texts notionally about pagan behaviour refer to and condemn the practice of dressing up as an old woman on the Kalends of January (a harsh law, as I have always thought, if you actually were an old woman…  Iterability again).  This alone gives us a range of different possible ways of reading feminine costume: different signifieds.  There is always, thanks to iterability, the potential for slippage from one to another; of miscommunication.  This is the space of deconstruction: in our terms, a space of constant renegotiation: the remaking of the bases of power.

The relationship of costume to person is worth more consideration as it will lead us further into thinking about the practice of negotiation.  You might have noticed that I have avoided the term individual in my paper.  I have done so for many reasons but not the least of these is that the subject is the meeting place of a number of categories or identifications: gender, age, family, ethnicity, religion, and so on: an assemblage if you prefer.  In that sense the category expressed in costume rarely conveys more than one or two, considered to be the most important at a particular moment.

At this point it is important to think about the social body.  Jeffrey Jerome Cohen importantly talked about how the construction of identity blurred the edge of the human body: hybridised it with the objects – and animals – that conveyed the image of the identified category.  This was part of Cohen’s ongoing post-humanist project and a very important contribution.  I want to push back a very little against this, however, partly because I find quite problematic some of the political implications of post-humanism and related approaches that stress the agency of objects, and partly because I don’t find the reading entirely satisfactory.

To be brief, I want to uncouple the elements of desire and queering, in Cohen’s account, which I find more interesting, from the probably lesser element of the hybridisation and blurring of the body.  I am not sure that costume and the accessories intrinsic to the signification, embodiment and the very inhabitation of an identity really do blur the boundaries of the body in the way envisaged.  Leaving aside the slippages of communication that have been my theme and which, I think are inherent in Cohen’s examples, I would rather read the assemblage from the outside in, as layers of social skin.  Does one really ever get beyond layers of social skin, back to an entirely pre-social human body?  Again, in my view, there is an absent centre.

The final point that this too brief consideration leads me to is how one could get out of the situations where a miscue, misfire or miscommunication had occurred.  One way out here can be thought in terms of Deleuze and Guattari’s concept of lines of flight.  (I can’t as yet claim to have read that much of or be very well versed in DeleuzoGuattarian thought.) Social actors, as I have said, can be seen as assemblages.  Even if elaborate costumes or layers of social skin aim to convey one identity, thought to be most important, those layers can still be peeled back to reveal others.  Laying aside the weaponry that might have conveyed Frankishness or a particular age-grade, could strip that persona back to a layer of general masculinity, for instance, that expressed a shared identity; buckling on such items could remake distance. The sheer multiplicity of identities that converge in the social actor make this sort of thing possible.  The other ‘line of flight’ is humour, which plays on the very possibilities for miscommunication that inhere in interaction.


In the early Merovingian world, the space of the political was up for grabs.  Old architectural cues broke down, were renegotiated; new, different ones were tried.  A greater relative investment in costume, the social skin, was one response to this.  Wherever we look, we can see, in my reading, the interaction of decentred subjects, fraught with potential miscues, miscommunications, and scrambles to remake or reconfigure social space: social structure was a chaotic, constantly reordered, teleologically re-read archive of precedent.  ‘Negotiating power’ is thus, in a way, a tautology.  Power does not, and cannot, exist other than in its constant negotiation.