Featured post

Gender in the Merovingian World

Wednesday, 6 June 2018

The poetics of politics: Transformations of public space in western Europe around 600

(This is a paper I gave at Durham a few weeks ago.  I had been trying to get back up to date on the archaeology of French towns, partly for the 'never-ending Year 600' project and partly because I perhaps needed to cover a chapter on towns for a forthcoming collective volume.  As it happened that chapter got delivered, and was better than I could have done anyway, so I wasn't called upon in the end, but it was still useful for me to do the work.  I have also been interested for some time in how the focus of politics changes between the sixth and seventh century, with the civitas capital losing its former centrality: a point that - I noticed recently - Simon Loseby had noted as long ago as 1997 (and quite probably other people had too).  All this gave me the excuse to try finally to put something together on the subject.  In the end, for various reasons, it ended up being a bit of a rush job and I think the paper is a bit of a dog's dinner in many ways.  Stitching the elements (about cities and about the social use of space and the scripting of politics) together meant that some ideas got lost in the process (especially about interpellation, in the Althusserian sense).  The last section is too brief and rushed, so a lot of stuff that interests me got lost anyway.  Anyway, here it is.  I present it as a crude, rough first sketch but I think there are ideas here that might be worth developing.)

  In my first book I said that trying to reconstruct urban life from the works of Gregory of Tours was like trying to reconstruct Tudor London from the props box of the Globe Theatre.In this paper I am coming back, after a long interval, to the Merovingian city and pursuing a different, and possibly no less tortured, dramatic metaphor.
What became of classical politics, in the very literal sense of the affairs of the polis, the city? The city, after all, remained the focus of politics throughout the Roman period.  While the municipal sphere had changed dramatically by the fourth century, it can’t be denied that the very form of the town continued to reify that of the Roman political system.  The late imperial period in the West sees, wherever one looks, urban contraction and abandonment to some extent but even in the far north-western imperial provinces the smallest and – by any classical standards – least impressive towns manifest their centrality to the Empire’s political system. 
An extreme case might be Bavay just on the French side of the modern Franco-Belgian border.  At some point around 300 (probably) the town was walled, but those walls, for whatever reason, enclosed no more than the forum of the early Roman town.  By 400 Bavay had lost its status as civitas­-capital of the Nervi to the then-scarcely-more-imposing town of Cambrai, but even so, even in its reduced size, it manifested the urban nature of the politics of the Dominate: a fortified administrative redoubt within which imperial revenues were collected, converted from kind to specie or vice versa; the seat of whatever officials resided there; the focus of the political activities of the Nervian aristocracy. Those aristocrats may have been as relatively unimpressive as their city: former (or current) imperial officials occupying or living off fiscal estates on short-term leases with a smattering of not very wealthy local landowners, mostly living in wooden, thatched- or shingle-rooved farmhouses, rather than the palatial-villa-dwelling Mediterranean grandees who might more readily come to mind when considering the late imperial aristocracy.  The general point, however, remains, mutatis mutandis, whichever town we might be talking about.  Politics, or the political, focused upon imperial service and imperial service remained centred on the town.  It was true in the sticks, at Bavay, and it was true at the heart of politics.  When it mattered Emperors were to be found in towns.  Indeed, even their villas could resemble bigger and more impressive ‘towns’ than poor little Bavay. Diocletian’s palace at Split is the obvious illustration.
Late imperial towns – even in contracting from their Flavian or Aelian heyday – were possibly more political than their predecessors.  The urban readjustments of the late Roman period appear to reflect the cities’ relative economic decline, as centres of markets and production.  Lesser settlements appear by then frequently to have rivalled them in those terms, which might explain some of the perceptible readjustments to the city network: the promotion of settlements like Verdun or Châlons-en-Champagne to the status of civitas-capital with territories hived off from the civitates of Metz and Reims respectively; the transfers of capital like that mentioned from Bavay to Cambrai or (in the civitas menapiorum) from Cassel to Tournai.  In the south, and at a slightly later date, Sidonius might have waxed lyrical about his villa and those of his friends, and sometimes decried the state of some local cities, but the latter were still, clearly, where politics happened.
That continued to be the case well into the next century.  Simon Loseby has set out, in a series of remarkable articles, the continued centrality of the civitas-capital to sixth-century Merovingian politics. Much the same lesson can be drawn from Michael Kulikowski’s studies of the city in late Roman and early Visigothic Spain and the picture is possibly truest of all in Italy, at least in the northern half to two thirds of the peninsula.  The city remained the focus of politics throughout the Lombard period.  And yet, as Loseby noted twenty years ago, something changed between the sixth and the seventh centuries.  The city lost its place as the focus of politics.
Or at least it did in Gaul.  I am going to talk about a peculiarly Gallic problem.  Or what seems to be.  I need to do more work on Spain to examine the extent to which a similar pattern can be traced there; as just noted, something very different appears to be the case in Italy.  Whether an analogous transformation occurred in Britain, or at least the extent to which similar factors could be adduced in forming a hypothetical account, remains to be considered.  The problem, as always, is that we lack the kinds of evidence that would give us a better idea about the earlier periods.  Whatever the case in Spain, Britain or Italy – even if they all turned out to be variations on the same theme – there are, I contend, some specifically Gallo-Frankish features to this process.
The late Roman Gallic town differed from its early Roman precursor in various ways, none of which was unique to Gaul.  They were walled from the late third century, in a probably rather longer process than used to be thought.  There was a huge reduction in private or municipal spending on public monuments, which probably best explains the usually short length of the walled circuits and the common incorporation of existing large monuments within the defences.  They contracted, to varying extents.  The walls certainly do not give a clear guide to the size and extent of the inhabited area, as was once believed, but rare indeed are towns where hitherto-inhabited, now extra-mural, quarters flourished as before.  Public buildings were frequently given over to new uses, often involving manufacture.  None of these features is unusual or specific to Gaul.  All can be observed in Rome itself. 
The other key development, obviously, was the appearance of Christian monuments.  By 400 all the civitas-capitals that had retained their status presumably had a cathedral, even if we know little about it.  As far as we know, such structures tended to be within the walled area; I am not sure that there are any indisputably extra-mural cathedrals in Gaul.  The Christian cult centre was thus incorporated in the same, small urban space as the foci for secular politics.  The phenomenon regularly visible in Italian towns, of a spatial reorientation from a previous civic centre around the forum to a plaza in front of the cathedral cannot therefore be seen very clearly in Gaul. 
Quite how secular politics functioned in spatial terms is not always clear.  As throughout the Empire, fora were often turned over to new purposes in whole or part, or encroached upon by new buildings, and many lay outside the new walled areas. Others, like those at Bavay and Paris, were fortified.  Some sides of Amiens’ forum were incorporated into the city’s defences.  This presumably did not diminish their importance but it must have changed the way in which the space was used.  It is possible that where amphitheatres were incorporated in walled circuits, as at Tours, Périgueux, Amiens, Metz and elsewhere, their arenas were used for public assembly.  The plausibility of this suggestion is perhaps strengthened by the importance of circuses and similar buildings in late imperial palace complexes, most famously at Constantinople.  Where such arenas were now extramural, they might still have functioned in the same way, as might derelict fora now left outside the enceinte.  It may seem logical to suppose that the spatial requirements for urban political gatherings were, in any case, less than before but I would not want to stress that possibility.  One of the key features of the Gallic city is the symbiotic relationship it had with the rural components of the civitas.  Villa-dwelling aristocrats were expected to come to town to participate in public life, however distasteful they claimed the latter was.  The other major possibility, of course, is the use for such gatherings of the interior of larger buildings, such as basilicae (where these had not been partitioned and given over to metalworking or similar, or – and this is will be important – the cathedrals.
The sixth-century Merovingian town continues most of these features. Secular occupation is increasingly difficult to find archaeologically, a feature that must partly be explained by its actual absence as well as, in part, by its ephemeral nature.  There is clear evidence of the further contraction and decay of the old Roman urban centres.  One or two towns in the far north might – like their British contemporaries – have died out entirely, at least as settlements that can meaningfully be described as urban.  Metz and even Trier have proven remarkably barren in terms of early Merovingian evidence, a fact of considerable interest given their secular political importance in the first and second halves of the sixth century respectively. Town walls were evidently not always maintained in a very good state. 
The other well-known component of sixth-century Gaulish urban development was the development of the Christian town.  If the cathedral seems to have been intra-mural, the overwhelming majority of urban churches were not. As foci for burial, they tended to be located in the earlier Roman cemeteries, outside the inhabited area.  Because of the contraction of the latter, these cemeteries could now be somewhat removed from the walled urban core.  To proceed from Tours cathedral, inside the walls, to Bishop Perpetuus’ new, mid-fifth-century church of St Martin, on the fringes of the early Roman town rather than in the cemeteries, is a ten to fifteen-minute walk, across a landscape that, in the sixth century, was a mix of vineyards, derelict areas, workshops and occasional dwellings. Around St Martin’s, though, was something like a new city.  Other churches and monasteries sprang up around that holding the famous bishop’s tomb.  As well as the clerics who tended these churches, the monks and nuns who lived there, there were the staff of these establishments and the people drawn in from the countryside for cures or to receive alms.  Vignettes in Gregory of Tours’ works suggest entertainers and merchants as well.  Thus while on the one hand we can see the continued stagnation of the traditional core of almost every Gallic city, on the other we can trace a sometimes vibrant peripheral community.  Sometimes, as at Tours and elsewhere (Limoges perhaps) a ‘bifocal’ town appeared, with an unwalled Christian town, sometimes called a vicus christianorum as well as the walled redoubt of the old Roman city.  Some towns (Trier; Lyon) were multi-focal.  In others, such as Metz, we might be able to detect a drift of settlement to certain extra-mural areas (in that case around the Great Amphitheatre).
The sixth-century Gallic city was primarily a Christian city, where it could be called a city at all, in anything other than the technical sense.  Through the fifth and sixth centuries, new churches were founded around the edges of the town, and new cults discovered.  The bishops played an increasingly important role in the towns.  Most jealously guarded two key privileges: baptism and preaching (Caesarius of Arles was very much an exception), meaning that Christians had to go to town for their principal spiritual needs.  The great cults were urban; most monasticism was city-focused.  Bishops organised great processions around the various churches of the city: the rogations first instituted at Vienne, the processions from the cathedral to St Martin’s at Tours, the great (three-day) excursion from Clermont to St Julian’s at Brioude, the procession mentioned by Gregory from the walled city to the church of St Remigius outside Metz, that across the Rhône bridge from Arles to Trinquetaille and the shrine of the martyr Genesius, and so on.  These were important social gatherings of the civitas’ inhabitants.  Bishops took over many other urban functions: feeding the poor, maintaining aqueducts and so on. 
All this is pretty well known.  Equally appreciated is the continuing centrality of the civitas in secular politics.  The city-districts were the building blocks of the sixth-century Merovingian kingdoms, forming the basis for tax collection and the levying of armed forces.  Each city – it seems – had its own count responsible for the administration of justice, the imposition of the royal dues and probably leading the civitas’ military contingent.  Some other urban institutions continued, even if we know little of them: the municipal archives, various officers such as the defensor civitatis.  Political competition between civitates was well-attested: sometimes armed conflict; rivalry over saints’ cults; even competing ways of counting the years.
After this descriptive preamble, we come to the focus of my paper: the construction of space and of the political.  Space is not neutral; it is, as Lefebvre said, constructed.  This is not merely a question of the enclosure or partition of spaces, even if these can play a very large part.  They act as cues to behaviour, to the bodily inhabiting of the space, in Bourdieu’s terms to the repeated bodily dispositions – the habitus – that construct categories.  A slightly, glib, simplistic and extreme example might be the physically often barely-delineated difference between road and pavement.  Even that, to continue being flippant, is somewhat more than merely an issue of not getting squashed, as anyone will know who appreciated the recent Daily Mash quiz that began by asking ‘Can you perform the relatively simple tasking of walking down a street without making other pedestrians want to punch you?’  Architecture provides cues; the entrance to a church, to take an obvious example, marks a point at which comportment, at which your bodily occupation of space is expected to change.  There are what Gernot Böhme calls ‘atmospheric architectures’, designed to enhance the nature and use of space.
Clearly, one does not need to go far to see antique and late antique illustrations of these points.  In earlier, classic Roman urban forms, crossing the pomerium provided a cue for different social expectations; the architecture of the forum marked a traveller’s arrival at the political, social and economic heart of the city.  Annabel Wharton’s studies of late classical towns documents shifts in the importance attached to open or ‘optic’ vistas of the urban space, to what she calls a more haptic space, more enclosed, experienced more through the act of passing through it.  The latter is especially true of churches but we can see other examples of the trend as perhaps at the palace of Galerius in Thessalonica.  Analysis of the church of St Martin’s and the various verses placed on its walls is a better illustration still of the ways in which various architectural components combined to influence a visitor’s comportment as he or she approached the tomb of the saint.  As Raymond Van Dam has argued, the architecture worked to impose a sense of awe that inhibited a pilgrim in venturing too close to the shrine unless she was convinced of her worthiness to do so: a key element in Van Dam’s ‘socio-somatic’ interpretation of healing miracles.  The same would obviously be true of the imperial palace and especially the great audience chambers, where the attendees were arranged by very tightly policed rules of status and precedence, and where comportment was of notable importance.  Even the emperor’s bodily comportment was a matter of significance and expectations could change from one setting to another.  In the Merovingian countryside, entrance into the cemetery, still largely removed from the space of the living, provided further cues, essential for the functioning of ritual.
Transposed into the realm of identity, such cues emphasis the question ‘what do they want of me?’ ‘How am I meant to behave?’  The use of space is vitally important in social interaction, as a means of attempting to freeze interactions between different social categories in particular modes.  Take, for example, a large Roman villa.  A visitor enters the complex, perhaps through a gate, after approaching via routes that present the building in a particular way.  Possibly after crossing the more functional ‘rustic’ part of the villa and perhaps going through a second gate, the visitor might cross a more enclosed courtyard, again with the main house as the focus of the gaze.  On entering the main building, the visitor enters a reception chamber, decorated with mosaics and wall-paintings depicting the ideals of rural aristocratic land-owning life.  If a client, or even if a guest of equal or superior standing, the setting emphasizes the expected, formal behaviour of host and visitor: the Latin word for host and guest is the same: hospes.  Both are bound – held hostage by – the rules of hospitality.
This is where the poetics of my pretentiously mock-Aristotelian title come in: the shaping – poesis – of reality, its staging, its scripting.  This helps freeze the interaction between categories of people within a certain, formal range, where people know the correct ways to behave and where transgression can be clearly recognised.  We are all aware of the awkwardness of meeting someone only known from a particular setting in a completely different context.
It is important to consider how the state operates spatially.  The whole space of the Roman Empire was subject to the operation of imperial law.  The Antonine Constitution, making all free inhabitants citizens, considerably simplified the hitherto complex overlapping and interlocking of different legal statuses, especially in relation to towns.  Location within the nested jurisdictions of the Late Empire made certain space into place.  The seat of a provincial governor, acting in his public capacity, within a public audience hall, bestowed a certain legality on his actions, within tight rules.  The occupant of such a spot acted to some extent in the place of the Emperor, as his vice-gerens, as made clear by various cues in the building itself.  Legitimation was brought, limits to it assigned (and acceptable behaviour constrained), by the office, not by the dignity of the individual who held it, although the imperial state recognised the problems that might go with that by assigning particular status to the holders of specific offices.  An official’s jurisdiction operated broadly uniformly within a defined territory.
How might these ideas have played out in a sixth-century urban context?  A key shift concerns the nature of the state and public space.  The Roman Empire, throughout its existence, was fairly clear about the nature of public space: space in other words where politics was enacted.  Book 15 of the Theodosian Code prescribes how such spaces were – ideally – to be maintained.  Quite how such theory related to practice in the contracted and remodelled cities of Gaul is, however, an intriguing problem.  Theodosian Code 15.1 contains over 50 rescripts concerning the maintenance of public buildings, frequently forbidding any privatisation of such space and decreeing that any private structures that encroached onto such spaces or sites be torn down.  And yet, in an interesting insight into of the efficacy of imperial law, two constants of late antique urban development are the dereliction of public buildings and the encroachment of buildings into public space.
There must have been considerable fluidity. What marked urban space in the last centuries of the western Empire: the old pomerium, or the new walls?  A possibly ruined or derelict arch or the new gateway?  Burials began to intrude into formerly inhabited zones, but rarely – except in Paris – in a dramatic way in Gaul, which suggests some confusion among contemporaries too.  We might recall the problem mentioned earlier, of where the public spaces were in the new towns.  Many must have been considerably smaller than in the early Roman period.  The ‘placette’ that, it has been suggested, might have been late Roman Tours’ replacement for the now extramural forum is smaller than the cathedral. 
I referred earlier to the absence of traces of high-status Merovingian occupation in towns, whether of kings or their aristocrats and officials.  This is a conundrum as we know these buildings existed.  Earlier Merovingian palace complexes evidently maintained the same key elements of late antique palaces: an audience chamber, a more general assembly area and a cathedral.  These elements can be seen clearly at Trier but also at Metz, where the Austrasian kings transferred their seat after the mid-sixth century.  It is possible that something similar existed at Soissons.  Frankish kings were still sometimes interested in providing spectacles for their subjects.  Gregory says that Chilperic wanted to provide circuses for the people at Paris and Soissons, though quite what was meant by that (aedificere) is anyone’s guess.  When Childebert II had an otherwise unknown aristocrat called Magnovald murdered at his court in Metz it was while he was watching bear-baiting, presumably in the small amphitheatre. 
One possible reason for the lack of archaeological evidence is that such buildings remained public and in continuous use, eradicating traces of occupation.  Another is that, simply enough, the Frankish rulers or their representatives did not invest significant resources in modifying or adding to them, but just maintained them.  As Simon Loseby once said in a sadly unpublished comment on Merovingian economic policy, the twin pillars of their attitude were ‘if it ain’t broke, don’t fix it’ and ‘if it is broke, forget about it’.  In this, their attitude seems typical of the earlier Merovingian elite, which seems to have spent very little on high-status buildings.  In the past I have argued that this was because of a lack of secure control over surplus and, while I think that that is correct of most of the sixth-century northern Gallic aristocracy, clearly it does not apply to the kings or, probably, to their most powerful service aristocrats.  In that case I think it probably illustrates a particular sixth-century Merovingian attitude towards wealth and its display. 
Of course, the one type of public building that the sixth-century Gallic elite certainly spent money on was churches.  The royal capital at Metz certainly had twenty-three, and possibly as many as thirty-three, churches by 700; it might have had eighteen a century earlier.  By the mid-eighth there were certainly forty-three churches in existence there.  If Metz’s status as a principal urban residence of the Merovingian kings made it exceptional, it was not by very much.  Le Mans had over twenty churches by the time that its bishop Bertram drew up his will in the early seventh century.  All this must have a significant bearing on the nature of public space and its relationship with political power.
Wealth was, however, above all worn in the sixth-century West.  Costume appears to have been even more important than it usually is in marking identity.  Analysis of northern Gaulish and other burial patterns appear to show that costume was one means by which social categories were created.  The wearing of particular forms and arrangements of jewellery marked out different stages of the female life-cycle.  The carrying of weaponry played a similar role among males, where (as with female costume) it might also have had an ethnic significance (provided that we’re clear about what we mean by ‘ethnic’ in this context).  Furthermore, it seems that all social classes followed these guidelines, differing only in the lavishness of their display.  Costume and other bodily markers apparently created the behavioural cues in social interaction. 
As I have argued before, the jewellery worn by younger and married women tended to highlight the parts of the body that were, in Frankish law, not to be touched without incurring penalty.  The cues on how to behave were now worn: costume created social space.  That is, I suggest, related to the general absence of the old, clearer, architectural markers of space.  In those terms, political space was much more fluid than before; it overflowed what had been its normal boundaries as the latter broke down.  This was doubtless particularly important in the processions that were mentioned above, which flowed from the city, through formerly urban space into what had been the city of the dead: now a place where, in Peter Brown’s classic analyses, heaven touched earth and where time stood still.  The limited evidence we have suggests that on these and similar occasions the population could be divided into various sub-groups, on ethnic or other lines.  Such arrangements doubtless underscored expectations about how the different social categories were expected to interrelate with each other.
It may be that assemblies in towns were still on occasion the site of traditional classical urban political displays.  On occasion at least, wills were still read out in the public spaces of the city.  Chilperic for example says that he will – in best Roman tradition – give the citizens of Tours a slogan to chant at Gregory. 
The Magnovald incident raises the point that, whatever behavioural cues there might have been in the imperial palace, they appear to have been significantly altered in sixth-century Gaul, to judge from Gregory’s accounts of things that happened in the various chambers of the court, including murders, tantrums and slanging matches.
A number of cautionary points must be made here. I am not assuming that the socio-political boundary markers, architectural or bodily, are automatically effective; that they cannot permanently freeze a set of attitudes – the habitus – in a given mode is elementary: Bourdieu 101.  They are in a constant state of renegotiation.  Further, one might wonder whether the infractions – the, by Roman norms, bizarre behaviour – are recorded by Gregory because of their significance or unusualness.  That must in part be the case, but there do seem to be an awful lot of them, when you remember that the bulk, the last six books, of Gregory’s Histories covers only 16 years.  Part of the problem must be related to the fact of living – from the middle of the sixth century – in a post-Roman world, where the political itself, the forms of rulership and authority of all sorts, was in the process of reinvention.
When gatherings did not take place outside, it might be that another element of post-Roman confusion entered the equation.  As is well-known, much late antique ecclesiastical architecture followed the example of imperial secular building, notably the great audience chambers.  Obviously, the classic basilical church plan was based upon the great assembly halls – basilicas – of the civic forum and the imperial palace.  Most discussion of which I am aware points out how the altar in the raised apsidal end occupies the space of the imperial throne, or the portrait or statue of the emperor, in straightforward and unproblematic manner.  What I am not aware of is any discussion of the element of semiotic confusion that could be introduced at a time when both secular and ecclesiastical authorities were using the same spatial arrangements in buildings of essentially identical layout, at the same time.  One need not think that late antique people had no idea what to do when going into an imperial or royal audience hall on the one hand, or a church on the other, to wonder whether the varying hierarchies set out in identical spaces did not produce changes in the political script for which the architecture provided cues.  Sometimes the king or emperor – or his image and his representative – occupied the focal space and ecclesiastical as well as secular aristocrats were expected to approach only with deference; sometimes the altar, the image of Christ, and his representative took that position and even emperors and kings were supposed to approach with humility. 
Caesarius’ sermons to the people suggest that the inhabitants of Arles did not always read the spatial behavioural cues of church architecture as assiduously as their bishop thought they ought to: lying down, plaiting their daughters’ hair, fixing their jewellery, chattering away, making a break for the doors when Caesarius got up to give his sermon (only to find that Caesarius, in a shocking infraction of basic health and safety guidelines, had had them locked from the outside).  Was this typical?  Did Caesarius simply lift it from Augustine?  Was it a general problem or only at various points of the liturgy?  Gallican liturgy included a point where the deacon asked the people to be quiet, just before the sermon.  Did such behavior spread from the cathedral to the royal court?  What happened when the person occupying the place of the emperor in a secular court could no longer make a legitimate claim to be the emperor’s representative, or to have imperial legitimation for his position?  Was it worse when they occupied actual former imperial space like the aula palatina at Trier?
Sometimes the confusion was only emphasized by the principal actors.  In Book 7 chapter 7 of Gregory’s Histories, King Guntramn addresses the people of Paris in church.  Gregory says he did this ‘after the deacon had asked the people to be quiet’, and thus at the very point when the bishop was meant to speak.  It wasn’t the only time that Guntramn was described by Gregory as acting ‘like one of the bishops of the Lord’.  Quite what Gregory thought of this is difficult to unravel; he does not seem to me to have thought that it was a straightforwardly Good Thing.  It might be though that there was a lot of that in the air in the late sixth century, quite possibly as a result of the renegotiating of the bases of monarchical power after Justinian launched his wars of ‘reconquest’.  When Guntramn presided over a meeting of his bishops at Chalon-sur-Saône, the preamble he issued makes clear – a decade or so before Gregory the Great wrote the Pastoral Care – that he viewed kingship as a ministry.
It is perhaps not surprising, then, that the Cathedral could be the location for dramatic confrontation.  Gregory’s account of the career of Nicetius of Trier highlights several occasions when the prelate came into conflict with the kings of Austrasia, haranguing them from the pulpit.  Gregory has Chilperic I lament “Behold how poor our fisc is! Behold how our wealth has been transferred to the churches! Really, no one rules other than the bishops. Our honour will perish; it has been transferred to the bishops of the cities.”  Whether or not Chilperic really did ‘often’ say this – or even think it – is impossible to know, but it wasn’t a bad diagnosis of the way things were going. 
Paul Fouracre has pointed out that the killing of bishops was much more common in Merovingian Gaul than anywhere else in the post-imperial West, a phenomenon for which he did not have a ready explanation.  The domination of cities by bishops might nevertheless be an important factor.  The domination of the Church by the senatorial aristocracy from the fifth century onwards is also a peculiarly Gallic phenomenon, that illustrates the importance of the bishop in local and regional politics.  In northern Gaul at least, the bishop was also generally the winner in power-struggles with the local comes civitatis. 
As if this weren’t enough, and possibly also explaining the greater tendency to kill bishops in the regnum Francorum, the Gallic episcopate appears to have taken up with more relish than most the tradition of speaking truth to their secular overlords.  Nicetius of Trier is a particular example, drawing on strong local, Treverian traditions going back to the fourth century, but he was far from alone.  Gregory of Tours himself thought that the duty of bishops to speak out against the actions of kings and keep them on the straight and narrow was especially important.  There are plenty of other examples: Germanus of Paris; Salvius of Albi.  In the early sixth century the entire Burgundian episcopacy went on strike to protest one of King Sigismund’s actions.  It may be significant in these regards that the Anglo-Saxon bishop with the most awkward reputation, Wilfred of Ripon was educated, and chose to be consecrated, in Gaul, and that the only time (surprisingly enough) anyone actually tried to kill him was also in Gaul (though they got the wrong man).
For all of the reasons that I have just set out, I have suggested that the Merovingian kings of Austrasia abandoned their attempts to take over the former imperial capital at Trier and moved to Metz.  Metz had the correct architectural furnishings for a palace complex, but it was not overburdened either by a ghostly imperial presence or by an awkward episcopal tradition.  Unsurprisingly, the kings appear to have kept a tight rein on appointments to the see of Metz throughout the period.  The bishops we know about, even Saint Arnulf, were all former palatine officials whom the kings felt they could trust.  On occasion they might even have been their relatives.
Nonetheless there remained some crucial features.  The count governed the territory of the civitas and exacted taxation and other dues from it, moderated by some exemptions concerning certain duties, and administered the law throughout it, either personally or via his officers, such as the centenarii or hundredsmen.  The title remained an office, appointed by the king, which could be withdrawn: it was neither a job for life nor a hereditary title.  Thus, wherever the count was, in a sense, the authority of the king – the presence of the state – was too.  The count’s court and the hundred mallus might move around but they retained their place, within the nested jurisdictions of the sixth-century state.  This surely remained important even in the fluid and somewhat ambiguous political space of the day. 
Taken together, these factors go a long way towards explaining why the city had lost its place at the centre of Gallic politics from the second decade of the seventh century.  This is an especially peculiar phenomenon given that it was at this time that a more general urban recovery appears to have begun in northern Gaul, and the decline of southern towns does not appear to have really begun until much later in the century at least.
The city lost out to two other territorial units during the seventh century.  On the one hand civitas-identity appears to be replaced by a broader regional, kingdom-based identity.  Changes in ethnic identity might also have had an effect.  At a lower level, the sub-division of the civitas, the pagus became much more important.  Individuals are identified by their pagus of origin, for example, and the pagi acquire their own counts.  The direction in which things were moving is perhaps very clearly indicated by Fredegar’s account of the territorial straits to which Chlothar II was reduced by his cousins, Theudebert II and Theuderic II.  He says that his authority was confined to twelve pagi, rather than three civitates as one assumes Gregory would have said.
The secular political decline of the civitas might be linked to the demise of the royal dues that had been extracted on the basis of the city network.  Frankish taxation certainly went into terminal decline in the early seventh century, with – in my reading – salaries hitherto paid via the delegation of tax-revenues from specific tax-payers extended to ownership of the lands from which those taxes were collected, as well as by increasing grants of immunity. Military organisation by civitas also ended.  The last mention of a civitas contingent in battle is the account of the treachery of the men of Mainz in Sigibert III’s 636 Thuringian campaign.  That same account, however, also refers to a contingent from the Saintois, one of the pagi of the civitas of Toul.  In any case, the army appears to have been made up much more of aristocratic households or retinues rather than raised by general levies of Franks, commanded by royal officers.  Immunities were also exempted from the usual levying of military service.  Counties proliferated and, even if not technically inheritable, seem much more to have passed down within families.
The increasing numbers of charters show that public legal gatherings could took place in local churches and at or in the villae of particular landowners. They do not suggest that the city retained its importance in that sense, in the north at least.  Charters signed at gatherings in old Roman cities tend unsurprisingly to be those involving the bishop.  The great rural monasteries founded in the seventh century also drew from attention away from the cities as central places and the foci of political action, not least because many were the centres of immunities, from episcopal as well as royal control. 
In the seventh century, the Frankish kings are much more likely to found on their rural estates than in the old cities.  The majority of royal assemblies took place there, including church assemblies, as at Clichy in the 620s.  Another location was the great royal abbey of Saint-Denis near Paris.  The kings still held gatherings at cities especially in Burgundy, just as they had held important gatherings at their villas in the sixth century, but the ratio seems to me to have shifted dramatically in the direction of the countryside.
The city had lost its place.  Urban Politics were now very much scripted by the bishops.  The space of the state was increasingly perforated and fragmented as the early Merovingian state broke down so that the town’s administrative role was greatly reduced anyway.  The stage for politics passed to various places in the countryside: royal and aristocratic villas, monasteries, and the elite spent their now more secure resources on these centres, especially from the middle of the seventh century.  How the immunity impacts upon the space of the political, as an interstice within the areas of operation of royal and episcopal power, a zone of exception, is an interesting issue, but one which alas I have no time to discuss.  How did cues for behaviour change in this different seventh-century world.  To what extent did the private permeate the public?  Some of the incidents that took place at royal gatherings in the seventh century certainly suggest a different scripting from before.  Paying more attention to that constitute the other half of this project (not, fortunately, this paper) and perhaps we can discuss that.
Public space and its use had evolved significantly through the Roman and immediately post-imperial period.  But in many ways the rules of the dramas that unfolded there remained within a Roman tradition – even if an observer who teleported from the second century to the early sixth might have had no idea whether he was watching a tragedy or a comedy.  As in so many things, though, the crisis of the mid-sixth century and especially Justinian’s wars and their attendant ideology brought abut a radical rescripting.  Kings and their officers were no longer able to play the parts that they had hitherto known off by heart.  They needed to ad lib and to find new theatres in new locations.  Unsurprisingly, as the one character whose part in urban dramas remained unchanged from the late Roman period, the bishop remained alone on the stage.

Tuesday, 6 March 2018

History: Faith and doubt


[Here is a very rough draft of a section of the book I am - supposedly - writing about the philosophy of history, Why History Doesn't Matter, which has run into the sands somewhat of late.  In this chapter I am trying to explore the fact that there is a core of doubt in all historical endeavour, which can only be (and then only provisionally) be bypassed through an act of faith, and then (not in this section) that that act of faith presupposes an ineluctable, impossible and universal ethical demand that lies at the heart of the historical project.  Conformance to that ethical demand provides a political foundation for critique of historical writing.  Anyway, here is the first tottering step...]

Much of the argument of this book depends upon a theological-looking combination of two concepts: faith and doubt.  The proposition is that there remain kernels of doubt at the heart of systems of meaning, which cannot be bypassed other than through acts of faith, themselves often linked to the historical and geographical context in which they are made.  This idea is often assumed to represent an invitation to compete relativism: a step towards the notion that any interpretation is ultimately only as good as any other.  I do not intend to go into that issue in detail, other than to say that it is a complete misrepresentation, usually based upon an unwillingness to consider, engage with or even read, the ideas in question.  It was partly because of this that it was necessary for me to begin with a defence of method and critical rigour.

My contention is that the historian proceeds with her enquiry, at all sorts of stages and levels, through an inextricable combination of faith (or commitment) and doubt.  Perhaps predictably, I want to use René Descartes’ thought-experiment, set out in his Meditations, as a way of illustrating this.  The experiment itself is well known.  Sitting in his room by the fire, Descartes ponders how it is that he can be sure that the world of his perception is as it seems.  Assuming that he is neither dreaming nor mad,[i] how can he nevertheless establish that everything that he senses is not a huge trick or deception played upon him by an Evil Demon?  Eventually, by the end of the second Meditation, Descartes reaches the conclusion that, whatever he might doubt of his sensory perception, if he is pondering this issue he himself must exist (cogito ergo sum; je pense donc je suis; I think therefore I am: the phrase itself comes from the Discourse on Method) or, perhaps more accurately, that, since there is an ‘I’ that is thinking, the subject ‘I’ must itself exist.  This is all fairly well-known; less frequently remembered is the fact that Descartes reached the cogito en route to an attempted proof of the existence of God.[ii]  In fact, contrary to usual perceptions of ‘Cartesian doubt’, Descartes begins not from a position of complete scepticism but rather from one of faith:

However, there is a certain opinion long fixed in my mind that there is a God who is all-powerful, and by whom I have been created such as I am now. …[N]or shall I ever break the habit of assenting to them [i.e. Descartes’ long-held beliefs] and relying on them, as long as I go on supposing them to be such as they are in truth, that is to say doubtful indeed in some respect, as has been shown just now, and yet nonetheless highly probable, so that it is much more rational to believe than to deny them.  Hence it seems to me that I shall not be acting unwisely if, willing to believe the contrary, I deceive myself and make believe, for some considerable time, that they are altogether false and imaginary until, once the prior judgements on each side have been evenly balanced ion the scales, no evil custom can any longer twist my judgement from the correct perception of things. …
I will therefore suppose that, not God, who is perfectly good and the source of truth, but some evil spirit supremely powerful and cunning, has devoted all his efforts to deceiving me. I will think that the sky, the air, the earth, colours, shapes, sounds, and all external things are no different from the illusions of our dreams and that they are traps he has laid for my credulity.[iii] (My italics)

Descartes’ experiment is, thus, to attempt to establish the veracity of things that he has heard and in which he has always believed.  More importantly, he his enquiry begins from a fixed position: that there is a good god who would not deceive him.  This, clearly, is incapable of empirical verification; it is purely a matter of faith.  It remains, nonetheless, the point which enables his thought experiment to proceed. 

Staying with Descartes, we can move to another, more famous concept associated with him, albeit with less precision: Cartesian coordinates.  The system relies upon the use of two or more axes proceeding from a point of origin, that enables the mapping of an area or space and the establishment of the relative positioning of objects or places within it.  For Descartes, this permitted the representation of geometry in algebraic terms.  It remains fundamental to cartography and forms an entirely coherent system with predictive and verifiable value … provided, of course, that you accept its conceptual starting points, which, as with that of Descartes’ thought experiment, are pragmatic and provisional rather than absolute.  For one thing, its conception of space is Euclidean or, to put it crudely, operates around the notion of flat and perpendicular planes.  Since the nineteenth century it has been recognised that, while Euclidean and thus Cartesian geometry is perfectly acceptable for all kinds of purposes, especially as a starting point, it has crucial practical and theoretical flaws, which I have neither the space nor the competence to describe.  Second, the system of coordinates relies upon the concept of zero as an integer and thus the notion that the difference (or distance) between 0 and 1 is the same as that between 1 and 2 or between -1 and -2.  Again, in practice that is unobjectionable but it was impossible before the introduction of that concept of zero.  In classical and medieval Europe, the difference between one and ‘one minus one’ was infinite: it was the difference between presence and absence, between something and nothing.  That was essential for Zeno’s paradox, for example.  Third, in most cartography the location of the point of origin or zero is purely arbitrary or conventional, like the use of the Greenwich Meridian.  Descartes’ acceptance of the truth of an eternally good God, or the various suppositions underlying Cartesian geometry, are precisely the things that Jacques Derrida labelled aporias: points where empirical observation or verification fails but which are nevertheless essential to the coherence or progression of an argument.

This admittedly superficial detour yields several points that are vital to my argument.  Entirely coherent systems, even ones with enormous scientific and logical value, begin around points that are established not empirically but pragmatically or provisionally, or through faith.  Sometimes their acceptance is a matter of sincerely-held belief (as with Descartes’ concept of God); sometimes those who employ a system are more or less aware of its non-empirical foundations.  In practice, there is no difference between Descartes’ belief in the good God as the starting point for his logic and an eighteenth-century geometrician’s acceptance of the Euclidean and other assumptions behind Cartesian coordinates. Indeed, there is no practical difference between our eighteenth-century mathematician’s conviction about the foundations of the Cartesian system and a modern mathematics teacher’s pragmatic acceptance of them for the purposes of explanation.  I am by no means equating the objects, or the nature, of faith in my three examples but they can nevertheless all be said to be founded upon matters of belief. 

More importantly, given the frequent perception of so-called ‘postmodernism’, unpicking those foundational concepts and showing the empirical failure that lies at their heart does not result in absolute anarchy.  It does not mean that ‘anything goes’.  Quite apart from the ideas that follow from Descartes’ acceptance of a benevolent deity as the starting point of his thought, philosophers have long debated whether the ‘I’ of the cogito provides as secure a basis for his further argument as Descartes imagined.[iv]  Identifying those aporias does not however, authorise us – on that basis alone – to develop any particular stage of the Meditations’ argument in any direction we like.  It does not permit us to identify every step of the Meditations’ argument as illogical, or to claim that Descartes said things that cannot be observed in his writings, or to use it as the basis for arguments that do not proceed from his writings.  Even when considering things that Descartes did not say but seems to have assumed or those which he left unsaid, an argument can only proceed via careful, empirical scrutiny of the text itself.  It needs to be repeated that Derrida’s own philosophy rested on close, empirical, microscopic even, readings of the authors in question: Rousseau, Austin, Marx or whoever. Similarly, recognising that the Greenwich Meridian is a purely conventional starting point for longitudinal measurement, that it was chosen for very specific and contingent reasons, does not enable one to postulate that Edinburgh and London stand in an entirely different geographical relationship, or to deny that the system provides practical measurements between points on the earth’s surface.  It does not justify anyone’s belief in a flat earth, or that the earth is a disc carried by elephants.  It does not – in any way – deny the concepts and procedures of scientific verification.  Nor, indeed, does it imply anything at all that is un-, let alone anti-scientific.  What, one might ask, has driven scientific research if not the on-going desire to question and unpick the aporias of previous generations’ thinking?  This is not a matter of a radical doubt or unbelief.  I would argue that it manifests an ongoing fidelity to the notion that there is something out there that is capable of systematic explanation.  The fact that the existence and origins of such a system have yet to be empirically established, if indeed they are ever capable of such a complete account, is why there is no logical contradiction between faith and science; why many rigorous ‘hard scientists’ are, for example, committed Christians.[v]

The point or the consequence of Derridian deconstruction is not complete relativism, or indeed relativism of any sort, but simply to expose the foundational cores of undecidability that exist within all systems of meaning.  It does not aim at the pulling apart and destruction of systems (this is an egregious misuse of the term deconstruction); rather, deconstruction troubles or unsettles them. It exposes the fact that such systems are not self-contained, self-present truths, helps us understand how this is the case and how they are interrogated.  Deconstruction is therefore neither as irrational, unrigorous, unscientific or nihilistic as its often quite hysterical opponents claim nor as radical and revolutionary as some of its proponents have been wont to imply.

The acceptance of the existence of points where doubt never be entirely negated is in no way disabling for historical research.  For over two centuries, English Law has functioned on the basis of the pragmatic concept of ‘reasonable doubt’.  For a defendant to be proved guilty the prosecution must present evidence and present a case that satisfies a jury, not beyond all doubt, but beyond reasonable doubt.  In the late eighteenth century, English jurists realised that a requirement, bolstered by religious sanction, for the accused to be proven guilty beyond doubt was hindering juries and judges from reaching verdicts.  The distinction between absolute doubt and reasonable doubt is that which (usually) renders law practicable. It is also, I suggest, that which permits historical analysis to proceed to argumentation and conclusions.   Sometimes the scope for reasonable doubt is, as with major political historical events, non-existent.  That does not negate the point that doubt is an integral component of the historian’s engagement with the evidence or that in practice some level of doubt can never be resolved other than through a form of faith. 


[i] Foucault and Derrida.

[ii] S. Blackburn, Think!

[iii] Descartes, Meditations on First Philosophy, 21-22

[iv] References

[v] In this sense and others, Richard Dawkins has done atheists and scientists enormous disservice.


Tuesday, 26 September 2017

Why History Doesn't Matter: An unreachable horizon?

[This is the concluding paragraph - and caveat - of the still-very-much-unfinished chapter 3 ('Those who do not know their history are doomed to repeat it: History as Memory') of Why History Doesn't Matter..  I think I need to get it out there as a rider to other provisional discussions of that book-project and its ideas.]

There is one very important caveat to my general recommendation that the aim of history should be to free people from perpetual commemoration of the past, from identification with past identities, and from the projection onto the past, and thus eternalisation, of modern identities.  It is one which leaves that recommendation with the status of an ideal, a horizon towards which we must work rather than a simple item of business to be put directly in effect.  In some ways, it relates to the areas where there is still much to do in compiling the factual record of the past, discussed in chapter 1.  It is all very well for me, a privileged white male from western Europe, to tell people to let it go, to stop identifying with the battles and oppositions of the past.  I don’t – as I will argue further in the next chapter – have to live with the burden of the past.  The past itself is as light as air, as I have argued, and that is always true in all situations, but the use of the past to justify the present, whether unconsciously or not, can create a very heavy burden for others.  It is not for me to tell African Americans or black or Asian Britons to forget the history of slavery and colonialism, especially when white Americans and Britons celebrate that legacy.  When white Americans and white Britons object to the removal of statues commemorating the figures who created and defended the systems that oppressed the world’s non-European population as an attack on their ‘heritage and history’ they underline and perpetuate their association with those systems and their place in them.  Who then has the right to tell the descendants of the oppressed, who still live in societies that preserve all too many of the elements of those systems, that it was ‘a long time ago’ and that they should just learn to live with it?  To make this clearer, let me use another example: who has the right to tell the Catholic population of Northern Ireland to forget the history of their oppression by the protestants of the province when the latter organise annual parades through Catholic districts to celebrate that history and defend their political supremacy?  In all these cases, it is for the more powerful to forget first.  Only when the groups who gain politically, economically and culturally from a system that has its roots in the past recognise that fact, only when they face up to the lessons of that history sufficiently, only when they learn that an appreciation of the evils of slavery and colonialism does not materially affect them (indeed in most cases it would benefit them), and that the critical appraisal of this history is not a personal attack on them as people, only when they dismantle the residue of those historic systems of oppression: then – and only then – might one be able to think about a different attitude to that history on the part of non-white inhabitants of Europe and the Americas.  Let me also make clear that – as intimated earlier and as I will argue in chapter 8 – this does not mean dealing with historical slavery in an ethically or morally detached sense.  I have no idea how realistic this idea is; there is too much work to be done for me even to envisage it becoming viable in my lifetime.  It’s not an objective that can be reached purely by governmental fiat, though that must surely play a key role.  I remain committed to it, nonetheless, as a horizon towards which we must continually strive.  This argument is developed in the next chapter, on the ‘burden’ of the past.

Monday, 18 September 2017

Gregory of Tours, the 'Angers Consularia' and Clovis 'consulus aut augustus'

Chapters 18 and 19 of Book Two of Gregory of Tours' Histories have long been recognised to represent information copied from some lost source.  Departing from Gregory's usual narrative style, the chapter comprises a pretty bald list of events:
18: Childeric fought a battle at Orléans. Adovacrius came to Angers with the Saxons. A great plague devastated the people. Aegidius died and left a son called Syagrius. This man [Aegidius] having died, Adovacrius received hostages from Angers and other places. The Britons were expelled from Bourges by the Goths, many of them having been killed at Vicus Dolensius [Bourg de Déols]. Count Paul with Romans and Franks waged war on the Goths and took booty [from them]. Adovacrius having come to Angers, Childeric came the next day and, Count Paul having been killed, captured the city. That day the Domus Ecclesiae burnt down in a great fire.
19: These things having been done, a great war was waged between the Saxons and the Romans but the Saxons, turning their backs, with the Romans pursuing, lost many of their men to the sword. Their islands were captured and ravaged by the Franks, many people being killed. In that year in the ninth month there was a great earthquake. Odovacrius [Odoacer, the King of Italy, and not the Adovacrius leading the Saxons in ch.18] made a treaty with Childeric and subjugated the Alamanni who had invaded part of Italy.

If you break the passage up into its sentences it looks like this:

  • Childeric fought a battle at Orléans. 
  • Adovacrius came to Angers with the Saxons. 
  • A great plague devastated the people. 
  • Aegidius died and left a son called Syagrius. 
  • This man [Aegidius] having died, Adovacrius received hostages from Angers and other places. 
  • The Britons were expelled from Bourges by the Goths, many of them having been killed at Vicus Dolensius [Bourg de Déols]. 
  • Count Paul with Romans and Franks waged war on the Goths and took booty [from them]. Adovacrius having come to Angers, Childeric came the next day and, Count Paul having been killed, captured the city. 
  • That day the Domus Ecclesiae burnt down in a great fire.
  • A great war was waged between the Saxons and the Romans but the Saxons, turning their backs, with the Romans pursuing, lost many of their men to the sword. 
  • Their islands were captured and ravaged by the Franks, many people being killed. 
  • In that year in the ninth month there was a great earthquake. 
  • Odovacrius made a treaty with Childeric and subjugated the Alamanni who had invaded part of Italy.

Set out like that, the passage clearly resembles a set of annals and for that reason, and the prominence of Angers in the account, historians have long been accustomed to refer to Histories 2.18 as the Annales Andecavenses: the Annals of Angers.

I don't want to discuss the events described, although these are interesting and important. Instead I want to think a little more about this lost source.   Quantitatively at least, most narrative historical composition in the fifth and sixth centuries took the form of chronica minora, minor chronicles or annals, and one of the most common ways of organising such annalistic sources was by consul. It had long been the custom to identify a year by the individuals who held the consulate at the start of the year.  Annals were thus frequently compiled by taking a list of years/consuls and adding key events that happened in that year.  Several such consularia are known from late antiquity, notably the Consularia Constantinopolitana.  It may, therefore, very well be the case that the 'Annals of Angers' of which Gregory had a copy took this form.

He seems to have taken other information from this source.  In the celebrated chapter 9 of Book Two of the Histories, where he undertakes research into the origins of the kings of the Franks, Gregory tells us that we read 'in the consular lists' (in consolaribus) that Theudemer son of Richimer, king of the Franks, and Ascyla his mother werre executed with the sword.  The event seems to have taken place in the earlier fifth century, if we can judge from where Gregory places it in his account, immediately before his reference to Chlogio (or Chlodio), who is also mentioned by Sidonius and from whom 'they say' (in Gregory's words) Merovech was descended.  The execution of Theudemer and Ascyla is not mentioned in any surviving consularia, as far as I know.  I wonder if this is another piece of information that Gregory drew from his Annales Andecavenses, which I propose that we could rename the Consularia Andecavensia.

That leads me to a final suggestion.  In Book Two, Chapter 38, Gregory tells us that Clovis came to Tours after his victory over the Visigoths at Vouillé. Gregory says that he received letters from Emperor Anastasius conferring the consulate upon him.  After describing a victory procession, Gregory says that from that day on Clovis was called consulus aut augustus.  This is a passage that has often puzzled historians.  As I have written before, itseems to me (as to Michael McCormick in his book on triumphal rulership) that Clovis might indeed have allowed people to call him augustus, as his contemporary Theoderic the Ostrogoth ceratinly did, without actually taking the title.  His consulate, however, as many have pointed out, is attested in no list of consuls ... or at least in no surviving list of consuls.  Officially at least Gregory seems to have been mistaken, but there were often diverging views on these things.  Some eastern consuls were not recognised in the west and vice versa.  A Trier inscription suggests that Constantine 'III' thought he held a consulate, although this has - at best - been expunged from the surviving lists.  I would, especially given the Loire valley location of the episode, like to suggest that in writing his account of Clovis' victory celebration after Vouillé, Gregory perhaps embroidered and interwove with details from other sources information from a document that did indeed list Clovis as consul for 507 or 508, and that that source was the same document from which he took other information concerning the history of the Franks, the lost Consularia Andecavensia.

Disclaimer: For all I know, someone has already suggested this.  I don't know anyone who has done so, or cited such, in recent decades but there could well be some scholarly article in German from the first half of the last century that does.

Friday, 15 September 2017

Barbarian Migrations and the birth of medieval Europe: From unity to diversity

[I was asked to write a chapter on this topic for a volume coming out in Spain in association with an exhibition, for which I think it will be translated. Because of that I offer the English version here. It's essentially a potted narrative of the period between 376 and 533 and not really any sort of big deal, especially if you are an aficionado of the period or familiar with my other works. On the other hand there are some subtle differences /developments between this reading and the narrative in Barbarian Migrations and the Roman West, of which I am working on a second edition, which you may or may not find interesting. Or you may want a short narrative overview for some purpose or other, in which case I hope this is of use.]
Introduction
It is traditional to trace the origins of medieval – and even modern – Europe to the end of the Western Roman Empire and the so-called Barbarian Migrations. The ruling classes of much of Europe claimed descent from the barbarians who were believed to have conquered and supplanted the Roman Empire.  Conquest of the Roman Empire was necessary for the medieval to come into existence: the ‘Middle Age’ lay between the demise of the classical world and its supposed rediscovery in the Italian Renaissance.  For centuries, any debate focused upon whether the collapse of the Empire had been a good thing or a tragedy rather than upon whether Rome’s fall was really caused by barbarian invasion.  The notion that the Middle Ages were created by ‘Germanic migrations’ from east of the Rhine was not much challenged before the modern period and doubts about the issue only gained historiographical momentum from the very end of the nineteenth century.  Historians like N.D. Fustel de Coulanges and of course Henri Pirenne proposed that the ‘Barbarian Migrations’ played no significant role in crucial areas, such as governmental or administrative institutions or economic structures.  These largely remained the same throughout the migration period and sometimes beyond.[1]
Crucial for the modern disciplines of History and Archaeology was the development of the notion of Late Antiquity after the Second World War, culminating in Peter Brown’s classic, The World of Late Antiquity.[2]  The idea of Late Antiquity combined numerous strands of historical research, some with quite deep roots, around the idea that there was much similarity between the third-century Mediterranean World and that of the seventh century.  The fifth-century ‘fall’ of the western Roman Empire and the Barbarian Migrations did not, therefore, mark a great historical rupture.  This view stresses continuity and a much more gradual evolution of the western Roman Empire into we might call ‘Medieval Europe’.
There is much debate about the ‘Late Antique’ interpretation.[3]  Is ‘Late Antiquity’ a Mediterranean as well as an imperial Roman phenomenon?  Important work about Late Antiquity has often concerned Mediterranean, particularly eastern Mediterranean, society?  Do places outside the Empire have a late antique phase of history?[4]  Can they have such a period?  The classic studies of Late Antiquity have also focused on religious and cultural history.  Were changes that took place in the spheres of political, military or economic history, or in the history of social structures unduly neglected?  The late antique paradigm has, moreover, never seemed to know what to do with the phenomenon of barbarian migration that concerns this chapter.

Relationships between Empire and barbaricum

In traditional accounts, the imperium Romanum and the barbarian lands formed two separate worlds in perpetual confrontation.  Barbarians were believed to have an insatiable desire to conquer the Empire so that the limes had constantly to be reinforced to keep them out.  In this view of the Fall of the Roman Empire, the Huns caused this pressure to become irresistible.  The frontiers collapsed and barbarians ‘flooded’ into the provinces.  All these ideas are questionable. 
Wars between Romans and barbarians could be very damaging, for Roman provinces ravaged by barbarian warbands and for areas of barbaricum harried by larger Roman armies in reprisal. Young warriors raided imperial territory to acquire loot and captives for many reasons but warlike relations were not the only ones that existed between the Empire and its northern neighbours and it is unlikely that they formed the most common, let alone the usual, situation.  The frontier was heavily defended, it is true – even in the late fourth century, but this was not simply to hold back hordes of invaders.  Indeed, that might have been one of the less important reasons for the Romans’ concentration of troops, and expenditure of resources, on the frontier.  The military balance of power was overwhelmingly in the Romans’ favour.  With at least 400,000 well-equipped, -armoured and -supplied soldiers to draw upon, with its network of solid, stone fortifications, its expertise in siege warfare and its complex logistical organisation, there was no way that the Empire’s existence could have been threatened by even the largest barbarian confederacy.  Roman reports of barbarian armies numbering tens of thousands of warriors cannot be taken seriously.[5]
The Roman obsession with the limes was produced by imperial ideology and politics.[6]  Good imperial rulership involved devoting time and energy to the frontiers, ensuring that the barbarians were kept out.  It also involved winning victories over neighbouring peoples.  These ideas required the stationing of armies along the frontier.  Keeping a large army at the centre of the Empire was dangerous.  First, such a display would look very much like tyranny to the Empire’s political classes; second, troops on the frontiers, far from direct control, were likely – as second- and third-century history had repeatedly shown – to raise their commanders to the purple; and third, by staying with a large military force in the centre of the Empire an emperor abandoned the notions of ‘good rulership’, which required campaigning on the frontiers.  It made sense, therefore, for fourth-century emperors to move their political capitals to towns like Trier or Sirmium, close to the limes, and to keep a large body of their best troops under their immediate control, ostensibly to fight the barbarians but actually to counter threats to their authority from the within the Empire, like rebellions and usurpations.  This ‘inside-out’ Empire, with the political centres on the fringes of imperial territory, was vital to the cohesion of the imperium.  The ‘barbarian threat’ was thus magnified in imperial ideology to proportions far greater than those it really presented. 
Late antique Romano-Barbarian relations took many forms other than those of raid and counter-attack.  The Romans used the barbarians and vice-versa.  The logistical needs of large concentrations of Roman troops and civil servants on the frontiers were partly met through commerce with the frontier peoples.  This trade was a motor for barbarian social development. The stability of the limes was maintained through alliance networks with barbarian rulers.[7]  The recruitment of barbarian warriors constantly topped up the ranks of the army’s élite regiments, especially, and simultaneously deprived barbarian leaders of manpower.  As the size of the Roman army grew while that of the imperial population did not, this was particularly important.
The Empire was, if anything, even more important to the barbarians.  Roman gifts and diplomatic patronage were vital to barbarian politics.[8]  In fourth-century Germania Magna, Roman culture and political ideas dominated.  Roman material is found in lavish inhumations and elements of Roman uniform in cremation burials; Alemannic leaders apparently copied Roman official metalwork to give to their followers as badges of rank.  An Alamannic king even named his son Serapio!  It is hard to know exactly how the inhabitants of barbaricum used imperial symbols, how they adapted them and what was ‘lost in translation’ but, clearly, Roman expressions of power dominated political discourse.  By AD 300 it is difficult to identify an entirely non-Roman vocabulary of power.  The Empire was a world power which, in political, military, economic or cultural terms, had dominated the lands east of the Rhine for nearly four centuries.  It is entirely to be expected that when barbarians thought of power they thought of the imperium romanum and its ruler. 
The Empire presented other prospects.  In the later fourth century, developments in imperial administration meant that non-Romans could reach the very highest military ranks and play a dominant political role.  The Empire also accepted immigrants from barbaricum and their families, especially when this brought lands back into cultivation and yielded extra fiscal income.
Successful raids produced booty that barbarian leaders could distribute to followers to cement their loyalty but this could be risky; because Roman forces were larger and better equipped, imperial reprisals were far more ferocious and damaging than barbarian raids.  Some raids, therefore, seem to have represented ‘bargaining’.  The Alamans attacked the Empire when they received less valuable diplomatic gifts than expected.[9] A raid demonstrated power or at least ‘nuisance-value’, which the Romans might want to neutralise through diplomatic payments or political support.  Romans encouraged barbarian leaders to attack their Roman rivals during civil wars, authorising the acquisition of the loot that oiled the cogs of barbarian politics.  This was a dangerous strategy nonetheless. Romans always prioritised the threat posed by imperial rivals over that presented by barbarian attackers – a Roman enemy’s military capacities were far greater than any invader’s.  The first task of the victors in civil war, however, was to make an example of any barbarians who had taken the opportunity of that strife to attack imperial territory.  This restored peace and order and demonstrated the winners’ ability to conduct proper imperial activity, in line with the ideology discussed above.  It is most important point to remember that the Roman frontier was emphatically not a zone of constant military tension, wherein the slightest weakening of the imperial defences would lead to the ‘dam’ breaking and the barbarians ‘flooding’ in.  The limes were no more like that than a modern border between nations necessarily is.[10]
A key aspect of frontier relations was that losers in barbarian politics, especially factions supported by the Romans, habitually fled for refuge in the Empire.  This had happened since Republican times.  The Romans accepted their friends’ surrender and settled them inside the limes.  This cycle continued to repeat through late antiquity and beyond.  The so-called ‘Great Migrations’ of the late fourth to sixth centuries can be seen to have fitted the pattern.  There were recognised routes to the limes, points of entry, and organised mechanisms by which appeals to enter were adjudicated.  There were means of receiving immigrants into imperial society and communities to which newcomers could be assigned.  Because much migration into the Empire was cyclical – barbarians entered the empire, served it or worked in it for some time and then returned home – and because of the arteries of trade, information about the possibilities for migration could spread into non-Roman territory.  Inhabitants of Germania Magna knew of specific opportunities and communities into which they could fit inside the Empire.  It is important to remember that there were specific routes and mechanisms for migration.  None of this represented a primeval surge towards the Mediterranean, a general ‘flooding’ over the frontiers or a chaotic set of random population movements, no matter how it may have seemed to the Romans at times.[11]
The Gothic Crisis[12]
The crisis that broke in 376 must be seen in this context.  Traditionally the migrants are supposed to have represented the Tervingian Gothic people, fleeing from the Huns, a terrible new force who had already defeated the Alans and Greuthungian Goths.  The Romans could not keep this horde out and so had to admit them, but in disorganised and corrupt fashion.  The Goths kept their weapons and local Roman officials exploited them for their own ends.  The result was a Gothic uprising.  Joined by other Goths and Huns from across the Danube, they ravaged the northern Balkan provinces until a large Roman army confronted them near Adrianople in 378.  The Eastern Roman army was destroyed and the emperor Valens killed.  The Goths were unable to capture Constantinople but plundered the Balkans for four more years until Valens’ successor Theodosius signed a treaty recognising the Goths as a semi-independent people within Roman territory.
Nearly three decades of scholarship and close analysis of the evidence have questioned almost all the elements of this story.  Rather than a bolt from the blue, a deus ex machina in the drama of Roman-barbarian relationships, the Huns had probably been neighbours of the Greuthungians for a decade or more prior to 376.  Ammianus’ celebrated picture of them is a patchwork of standard Roman ethnographic depictions of ‘extreme barbarians’ rather than an accurate description of new, terrible steppe nomads.[13]  An overlooked cause is Valens’ war against the Goths in 367-9.  Ostensibly a reprisal for Gothic support of a rival three years previously, this was Valens’ attempt to win the great victory over barbarians required by the ideology of good imperial rule.  Valens failed and, in 369, trouble on the Persian frontier compelled him to make peace with Athanaric, the Tervingian iudex (judge) or leading ruler.  This was not the triumphal outcome Valens wanted and in some ways a rather embarrassing failure but it was hardly a victory for the Goths either.  The Romans had ravaged their territory for two of the past three years.  Athanaric had failed to defend his people.  Defeats had been suffered by all the trans-Danubian barbarians: Tervingians, Greuthungians and Alans.  Furthermore, the Romans ceased the traditional diplomatic gifts to the Tervingians and restricted the number of markets where the Goths could trade with the Empire.  This dramatically reduced access to the Roman goods so important in barbarian politics.  Athanaric’s authority was questioned and in response he persecuted the Gothic Christians.  Civil war broke out, with the Romans apparently supporting one faction.[14]  The Greuthungians were probably similarly destabilised by Valens’ war.  This turmoil enabled the Huns to intervene – the sources mention that they allied with some Gothic groups.  Athanaric was defeated and withdrew with his followers into the mountains.  His enemies, grouped around Fritigern and Alaviv, alongside a losing Greuthungian faction, fled to the imperial border and asked for admittance.  This process, as noted, had happened many times over the centuries.  The scale of the Gothic receptio (admitting a barbarian group) was unusual but not unprecedented.
Crucially Valens was far from the scene of the action, trying (and again failing) to find his great victory over barbarians, this time Persians.  Communication was slow and he was badly informed.  Roman forces could have kept the Goths out but could not, as they were evidently ordered to do, admit the Tervingians but exclude the Greuthungians, disarm the former and escort them into the interior.  Far from the emperor’s gaze, local officials were able to exploit the starving Goths.  The situation got out of hand and when Valens extricated himself from his Persian war the Gothic rebellion was in full flow.  The catastrophe at Adrianople also resulted from Valens’ need for a triumphal victory to bolster his rule.  Rather than await the arrival of Gratian, his nephew and co-emperor, and thus share the glory, the eastern army hastened into battle without a plan, to disaster.
Adrianople ended the Eastern Empire’s ability to raise an effective field army for perhaps a decade. The Goths were worn down through attrition and starvation.  The end of the crisis was clearly a relief to Theodosius but there was no triumphal celebration.  Nonetheless, the Goths had been beaten and had surrendered, as all contemporary sources agree.  The first author to mention a foedus or treaty was Jordanes, nearly two hundred years after the event. The Goths were split up and settled.  Many joined the army, where, after Adrianople, the Romans were only too happy to employ them.  The idea of a semi-independent Gothic polity inside the Roman Empire is a myth based upon Jordanes’ questionable account and upon modern historians attempting to reconstruct the terms of the supposed foedus from particular readings of what later Gothic leaders wanted from their dealings with the Romans.  The Goths who crossed the Danube in 376 ceased to exist as a unified group after 382.  The group which emerged in the 390s was quite different. 
The ‘Perfect Storm’: The West, 383-421
In 383, the western emperor Gratian was overthrown by Magnus Maximus, commander of the British army.  Maximus soon controlled all the West except for Italy and North Africa.  Gratian’s young half-brother Valentinian II remained in control there, protected by his brother-in-law Theodosius.  Gratian had moved the court from Trier to Milan in 381, causing resentment amongst the Gallic aristocracy, which had been prominent in his government, and threatening the efficient distribution of patronage that maintained social stability in the north-west.  Maximus ruled at Trier for five years.  In 387, however, he expelled Valentinian II, who fled to Constantinople.  Maximus perhaps thought that his links with Theodosius – they were both Spaniards and had served together – would lead the latter to acquiesce in his invasion.  He was wrong.  Theodosius invaded the West and, after two heavy defeats in the Balkans, Maximus was beheaded.
Imperial government never properly returned to northern Gaul after Maximus’ suppression.  Valentinian II reigned from Vienne in southern Gaul.  In 392, Arbogast, his Frankish Master of the Soldiers, either murdered Valentinian II or drove him to suicide.  He put a rhetorician called Eugenius on the throne and civil war was again inevitable. Theodosius’ army destroyed the western field army at the River Frigidus (5-6 September 394) but died in Milan in January 395, leaving the two halves of the empire to his sons, the East to Arcadius and the West to Honorius.
At this point several factors came together to undermine the ‘inside-out Empire’.  The political history of the period between 395 and 421 is extremely complicated and there is no space here to do more than sketch its principal features.  These two decades were of critical importance, however, not simply for creating the geo-political situation which defined the parameters of fifth-century military and political history but also for establishing crucial precedents.  Perhaps the single most important factor that led to the break-down of the fourth-century system of government was the youth of the emperors Arcadius and Honorius, and their political inability even after they came of age.  The effective operation of the Roman Empire after the third-century crisis had relied upon an able adult emperor to distribute and redistribute imperial patronage and thus keep the Empire’s most important interest-groups content and in balance.  After Theodosius’ death, in the eastern and western courts different factions competed for the regency control of the palace and the emperors.  The courts also squabbled with each other over crucial territories in the Balkans and over the fact that the western regent, Stilicho, claimed to have been given regency over the East as well. Although Stilicho clung to power for thirteen years, the politics of regency required constant attention to court intrigue.  The empire looked inwards and, as a result, the management of imperial patronage and the checks and balances between regional aristocratic groups was neglected.  So too was the maintenance of imperial frontier policy.  This had devastating results, which can be discussed in turn: rebellion and usurpation; barbarian invasion; and federate revolt.  Occurring together, they created a ‘perfect storm’.
Northern Gaul and Britain had relied, in different ways, upon the presence of the heart of imperial government on or near the Rhine frontier.[15]  Involvement in imperial bureaucracy, tax-collection and the supply of the frontier armies appears to have been essential to maintaining Britain’s stability and prosperity.  The northern Gaulish landscape had, after the third century, been harnessed to the supply of the large armies and the concentration of imperial civil servants on the Rhine.  The state and its officials appear to have been the most important landowners north of the Paris basin.[16]  The removal of the heart of imperial government to southern Gaul and then, definitively, to Italy plunged both areas into crisis, visible archaeologically in the abandonment of villas, the decline of towns and signs of local stress and competition in the furnishing of graves.  In southern Gaul, aristocrats had become used, during the later third and fourth centuries, to having an emperor close to them and indeed to taking a prominent role in imperial government.  The imperial court’s departure for Milan and then Ravenna threatened to end this.  It is unsurprising, therefore, that rebellion broke out in Britain in 406 and quickly spread to Gaul, then Spain and, briefly, North Africa.  For the next eight years a series of usurpations created disunity and damaging civil war beyond the Alps.[17]  Apart from underlining and exacerbating the effective withdrawal of government from the areas north of the Loire, this warfare further prevented the recreation of a coherent, effective western field army as different western armies tore each other apart in repeated battles.  This feature cannot be stressed too strongly.
The consequences of the end of properly-managed frontier policy were unsurprising.  The empire’s system of alliances and diplomatic payments was essential to social and political stability throughout barbaricum, but especially in the heart of Germania Magna.  Around 400, archaeological signs of socio-economic crisis are found along the North Sea Coast, occupied by the Saxon confederacy, which might have fragmented.  Political strife broke out in barbaricum.  North of the Danube, destabilised by Valens’ war in the 360s and by the Gothic crisis after that, a new hegemonic power emerged: the Huns.  Leaders who had relied upon their links with Rome found that imperial support for their position was now unreliable, their power was threatened, and their rivals seem to have sided with the Huns in the ensuing civil strife.  The old cycle was repeated with yet more dramatic results.  Another faction of Goths, led by one Radegaisus, invaded Italy before being defeated by Stilicho.  An even worse consequence was the so-called ‘Great Invasion’ by groups from various peoples of central Germania: Vandals, ‘Sueves’ (possibly an Alamannic faction; possibly a group from the interior of Germania Magna) and Alans.  They crossed the Rhine at the end of 405 or 406, followed shortly afterwards by the Burgundians.  The Rhine frontier appears to have been largely unguarded by the Romans (the troops seemingly having been removed to fight in the various civil wars), its defence having been given over to the Frankish kings, who fought hard to stop the invasion.  The invaders rampaged through northern Gaul.  By 420 the surviving elements of the invasion formed two groups: the Sueves, in north-western Iberia, and the Vandals and Alans, at that point operating in southern Spain.
The third element of the ‘perfect storm’ was federate revolt.  As noted, the destruction of the eastern field army at Adrianople had led to the recruitment of large numbers of experienced Gothic warriors.  Entire armies seem to have been composed of Goths, formed into regiments of a new sort, called foederati.[18]  The most famous federate leader, Alaric, was only one of several powerful Gothic commanders active around 400: others were Gaïnas, Fravitta, Tribigild and Sarus.  The number of Goths in the army was clearly a problematic issue for some Roman political factions and could be used as an effective rallying cry – a political manipulation of suspicions about immigrants that is all-too familiar.  There were two massacres of barbarians in the first decade of the fifth century, one in Constantinople in 400, and another, including the wives and children of Gothic recruits, in Italy in 408.  Even faithful soldiers like Fravitta found that in this climate his Gothicness counted for more than his loyalty: he was murdered shortly after defeating a rebel Gothic commander, Gaïnas. Whether Gothic soldiers and their leaders were more mutinous than their Roman counterparts is disputable but the anti-barbarian ethnographic rhetoric projected against them doubtless added a very specific element to their actions.  It certainly raised the stakes involved in any participation in high politics.  Perhaps this made the bond between Gothic soldiers and their commanders even closer than was usually the case in the Roman army. 
The late fourth century had seen several non-Roman army officers rise to the very highest positions in western Roman politics: Arbogast, Bauto and Stilicho are only three examples.  These officers appear, unlike so many powerful generals of Roman birth, to have had no ambition to seize the imperial throne.  Bauto and Stilicho nevertheless married into the imperial royal family; perhaps their imperial ambitions were deferred to their offspring.  In the meantime, the exercise of enormous power at court more than sufficed.  Alaric seems to have wanted to emulate these officers.[19]
As mentioned earlier, the courts in Ravenna and Constantinople disagreed about the control of Illyricum, which passed back and forth between them.  When Alaric acquired an official military command in the region, all too often this was undermined by a change of regime in Constantinople or by the transfer to the other half of the Empire of the region where he was stationed.  Such events removed the legitimacy of his command.  Alaric’s desire for a high command and repeated frustrations at the hands of court factions led him to a series of revolts, first in the Balkans and then in Italy.  When various other strategies failed, he even made common cause with the Roman senate and raised his own emperor to the throne: Priscus Attalus.  As early as 409, then, it was not unthinkable even for the grand old families of the Roman nobility to ally with a ‘barbarian’ general.  This too failed to give Alaric what he wanted, as did deposing Attalus in order to treat with Honorius, and he famously sacked Rome itself (24 August, 410).  Even the fall of the ‘eternal city’ failed to move Honorius and Alaric died the same year.
Importantly, when he had no legitimate Roman title or office, Alaric called himself King of the Goths.  This gave him a title which the Romans viewed as legitimate, for non-Romans, and enabled him to negotiate with the court.  After Alaric’s death, his successor Athaulf pursued most of the same strategies and even briefly married into the imperial family.  The repeated failure of the Goths to gain a stable position within the imperial system led, however, to their use of the title king becoming almost constant.
Eventually, even a storm as perfect as that of Honorius’ reign blew over.  By 416 Honorius’ skilful general Constantius had defeated all of the usurpers, brought the Goths to heel and used them to inflict serious defeats on the Vandals and Alans in Spain.[20]  He then withdrew them from Spain and stationed them on the Garonne in Aquitaine.  Constantius’ armies campaigned successfully in northern Gaul, winning victories over local rebels and Frankish invaders.  When Orosius finished his History Against the Pagans in 417 he thought that the crisis that had reached its nadir with the sack of Rome had been surmounted, like all the others that had beset the Res Publica since its creation.[21]  By 420 that view must have seemed even more convincing and, to cement his successes, Constantius became co-emperor and married into the imperial family, as commanders like Bauto, Stilcho and Athaulf had done before him.  He wed Honorius’ sister (and Athaulf’s widow), Galla Placidia.

The Reign of Valentinian III[22]

Constantius died suddenly in 421 and that promising picture rapidly disintegrated as rival generals competed to succeed him as overall commander and dominant figure at court.  One result was a defeat at the hands of the Vandals in Spain.  When the ineffectual Honorius died in 423 his nephew and successor, Valentinian III, young son of Constantius and Galla, was deposed in yet another usurpation, this time by a certain Johannes.  With Johannes’ suppression in 426, Valentinian was restored to the western throne but, not least because he was still a child, the strife for control of the palace continued, between the main western generals: Felix, Boniface and Aëtius.  By 434, Aetius had disposed of his rivals but the situation for the Ravennate court had deteriorated.[23] 
In imperial politics, Roman rivals were always prioritised over barbarians for good reasons; Roman armies posed a more serious threat.  But the constant civil warfare since the battle of the Frigidus (395), in addition to one or perhaps two serious defeats at the hands of the Vandals, had serious results.  The western field army had never been given the chance to be rebuilt after its late fourth-century defeats.  Forces assembled from remaining field army regiments and troops drawn away from the frontiers had repeatedly fought each other, inflicting heavy losses on each other.  Even Johannes’ suppression had seen the defeat of the Eastern expeditionary force by Aëtius’ Hunnic auxiliaries.  Civil strife meant that imperial government had been unable to be restored in the north-west, although Aëtius campaigned in northern Gaul in the late 420s.  Imperial control over Spain appears to have been lost.  The Vandals, after defeating the Sueves, who had begun to expand from their base in Gaelicia, crossed into Africa. Eventually the Ravenna government recognised their control over Mauretania and Numidia.  With the Vandals’ departure and the defeat of significant Roman forces, the Sueves now seemed likely to dominate the Iberian Peninsula.  The loss of imperial authority in Britain, northern Gaul, Spain and Africa had in every case been brought about to some extent (usually decisively) by Roman civil war and palace politics.  
Valentinian’s reign marked an important shift in western politics.  The failure of the numerous usurpations of Honorius’ reign and of Johannes at the start of Valentinian’s appears to have convinced participants in Roman politics that the imperial dynastic ‘card’ trumped all others.  The competing generals – in this period entirely of Roman origin –  strove for the same things as Stilicho, Alaric and Athaulf had before them: domination of the imperial court, supreme control of the armies and an incorporation, via the marriage of their children, into the royal house.
Aëtius set about restoring governmental power but the shortage of good troops meant that he could not wage war on more than one front. Concentrating upon Gaul, he defeated the rebellious Goths and campaigned successfully in the north against Franks and rebels.  Disaster struck, however, when the Vandals captured Carthage in 439 and, the same year, the Goths defeated and killed Aëtius’ subordinate, Litorius, at Toulouse.  These events changed the picture dramatically.  Aëtius abandoned his northern Gaulish campaigns because his southern base was now threatened.  Troops were also needed to defend Italy against Vandal attacks.  In Spain, the Sueves entered upon a period of domination, under their kings Hermeric and Rechila. 
Aëtius again responded energetically.  A treaty was made with the Goths – possibly the first to recognise some sort of autonomy for the group – bringing them back into the imperial fold.  Soon Gothic troops were employed in attempts to restore imperial authority in Spain. In Gaul, while no territory was abandoned, militarily Aëtius seems to have withdrawn to a line roughly along the Loire.  He settled the Burgundians in Sapaudia in 443 to strengthen that line, alongside an Alan settlement at Auxerre.  It is noteworthy that the shortage of good soldiers was forcing the Empire to rely on ‘barbarian’ armies, like the Goths, Burgundians and Alans.
Nonetheless we must consider what we mean by ‘Barbarian’ when discussing mid-fifth-century wars.  The ‘Barbarian’ groups most heavily involved in imperial politics had all entered the Empire many years earlier: the Aquitanian Goths (we can now call them Visigoths) in 376, the Sueves, Vandals and Alans (now in Gaelicia and North Africa) in 406 and the Rhône-valley Burgundians by Rhby 413.  By 440 the warriors and leaders whom we are accustomed to think of as ‘Barbarians’ were men who had grown up or – increasingly – had been born inside the Roman Empire.  This is crucial: most historians write as though Rechila of the Sueves, Theoderic of the Visigoths and Geiseric of the Vandals were warriors fresh from across the Rhine or Danube.  Yet the culture that impinged upon them as children, adolescents and young adults was not that of the forests or steppes of barbaricum: it was that of the western provinces, especially Gaul and Spain.  The younger men likely had provincial Roman mothers.  These were the ‘barbarians’ who played the most important roles in imperial politics and history.  What differentiated these people from other provincial Romans is difficult to know.  By contrast, throughout this period, actual barbarian invasion from across the frontier (by Franks and Alamans for example) was rare, small-scale and usually swiftly defeated.  The great exception to this was Attila the Hun’s attack on the west in 451.[24]
Attila’s invasion wrecked a situation that was again looking promising for the Roman.  In 445 a treaty with the Vandal leader Geiseric, which betrothed his son Huneric to Valentinian III’s daughter, incorporated the Vandals within the régime.  Once more we see how ‘barbarian’ leaders sought legitimation, involvement in government and marriage into the imperial dynasty.  Offensives involving Gothic and Vandal forces were launched in Spain against the Sueves, although the Sueves had defeated these by 446.  In Gaul, Aëtius’ forces were again waging successful campaigns against local rebels and Frankish invaders, so that a restoration of the Rhine frontier might not have looked impossible.  Attila’s attack ended this possibility.
The Hunnic attack in 451 destroyed the remnants of Roman Rhine defences (if there was anything left at all).  Aëtius’ brought together the different Gallic factions and some barbarian allies to defeat the Hunnic forces but his (possibly deliberate) failure to inflict a decisive reverse on Attila was crucial.  Aëtius could not oppose the Huns’ devastating attack on Italy in 452, which was eventually turned back by disease and by an East Roman offensive across the Danube.  Famously, Pope Leo I garnered contemporary praise for halting Attila’s march on Rome; any military laurels fell to an Eastern general confusingly also called Aëtius.   Attila died in 453 and, with his threat removed, in 454 Valentinian attempted to profit from Aëtius’ temporary military discrediting and rid himself of his mighty generalissimo; he assassinated him in person.  In 455, two of Aëtius’ former bodyguards killed Valentinian in revenge.  The early sixth-century chronicler Marcellinus Comes considered this event to have marked the end of the western Roman Empire (a significance he also, more famously, ascribed to the events of 476) and his assessment is justifiable.[25]

Ephemeral Emperors, 455-76

The extent to which the dynastic ‘card’ had become decisive in Roman politics since 425-6 is more than amply demonstrated by the history of the succeeding twenty-one years, when no fewer than nine emperors attempted to govern the ever-shrinking territory that accepted rule from Ravenna. The last eastern emperor of the Valentianic-Theodosian dynasty, Theodosius II, died in 450. Without dynastic legitimacy, no western ruler could automatically obtain recognition from the factions that held power in other regions of the West.  Indeed, without dynastic legitimacy, eastern emperors also found it difficult to bring the West into line with their policies. All the western regional factions now had armed forces: the Visigoths acted entirely in line with the Aquitanian senatorial nobility; the Rhône valley nobles made common cause with the Burgundians settled in their region from the start; the Roman army of Dalmatia formed another interest group and one reason why the heretical Vandals in Carthage attracted such contemporary opprobrium was, it has been suggested,[26] their success in convincing North African Romans to support their leadership.  The situation looks different in Spain.  The Sueves seem to be the only barbarian group not to have formed a regional alliance with the local Roman nobility, although even by the 450s the difference between ‘provincial Roman’ and ‘barbarian’ was one of nuance.  Any Sueve who could remember life before the ‘Great Invasion’ would have been very old. Nonetheless in 453 a treaty was signed between the Sueves and two Roman commanders, Mansuetus and Fronto, which seems to have returned Carthaginensis to Roman authority while recognising Suevic control over other territories.[27]  That the Suevic king Rechiar had campaigned in Tarraconensis in 449 alongside a Roman rebel (Bacauda) called Basilius further suggests that, despite the impression given by Hydatius, the Sueves also forged alliances with provincial leaders.
Valentinian III’s murder dissolved two critical alliances.  One was with the Sueves, who resumed their attacks on Carthaginensis and Tarraconensis.  The other, more important, was with the Vandals, in many ways the faction with the strongest hand in 455.  Huneric, heir to the Vandal throne, was betrothed to Valentinian III’s daughter; the Vandals were the only group who could play the ‘dynastic card’.  This was made clear immediately, when Geiseric – possibly summoned by Valentinian’s widow – led the Vandal fleet against Rome and Petronius Maximus, who had plotted Valentinian’s murder and had himself made emperor.  Petronius was killed by an angry mob and the Vandal sack of Rome was far more serious than Alaric’s.  Geiseric brought Valentinian’s widow and daughters back to Carthage, where Huneric finally married his fiancée.  Their offspring would be descendants of the Valentinianic-Theodosian dynasty and thus have a strong claim to the imperial throne.  Geiseric wanted one Olybrius, married to Valentinian’s other daughter, the sister of Huneric’s wife, on the imperial throne.  To support this claim, the Vandals waged a destructive twenty-year war of raiding around the Mediterranean.
The history of the final twenty years of the western Empire’s formal existence is tortuous and confusing. [28]  Any emperor faced severe difficulties.  The diminishing extent of the territory under direct control by Ravenna (the empire had now effectively lost control over Britain, most of Gaul north of the Loire, large areas of Spain, and its Balkan and North African provinces) reduced his ability to raise taxes and raising armies.  This led to increasing reliance on recruits from beyond the frontier and to their payment in delegated taxation and settlement on some forms of land.[29]  After two generations, these mechanisms bound troops like the Visigoths and Burgundians closely into provincial society, further strengthening links between regional armies and aristocracies.
Several problems demanded the attention of every occupant of the throne.  One of the most pressing was how to bind the Gallic and Italian aristocracies into the regime.  The Gauls resented their exclusion from the centre of politics while the Italians, having lost their pre-eminence in the third and – especially – fourth centuries, were unlikely to acquiesce in a return to the ‘inside-out Empire’.  The tricky African situation compounded this.  With the Vandals opposed to the Ravennate government, the grain fleet necessary to feed Rome was always in danger of being withheld, causing riots.[30]  The threat of Vandal attack also demanded the presence of a significant army in Italy, giving military muscle to the Italian faction.  Countering the Vandals required one or more of three options: risky naval operations from Sicily, land-based invasion from Tripolitania or attack from across the Straits of Gibraltar.  The second (and in practice probably the first) required the Eastern Empire’s support; the third necessitated the nullification or incorporation of the Sueves, control of the route through Provence and Tarraconensis at least, and ideally of Carthaginensis and Baetica too.  Given the difficulties of military recruitment and waging war on two fronts, campaigning in Spain required the Gothic and/or Burgundian armies, underlining the need to incorporate the Gallic factions into a regime. Command of the Italian army and thus de facto the role of the most important political-military figure at court was held by Ricimer for over fifteen years between 456/7 and 472.[31]  Apart from possibly providing the military backing to the claims of the Italian aristocratic faction, Ricimer, like Aëtius before him, clung viciously to power.  He was opposed to any reduction of the military presence in Italy.  As time passed, in any case, barbarian soldiers came to be as incorporated into Italian society and politics as the Goths and Burgundians in Gaul, the Sueves in Spain or the Vandals in Africa.[32]  Ricimer, unsurprisingly, was wholly or partly responsible for the downfall of three emperors.  Each change of Emperor potentially added to these problems by alienating the individuals or factions that had supported his predecessor.  Thus, when Ricimer had the capable Majorian executed, Aegidius, Majorian’s commander of the (largely Frankish) Roman army on the Loire, took his troops (and thus defence of the Gallic frontier) into rebellion for three years.  Invasion by barbarians from beyond the frontier continued to be the least of an emperors’ worries.  As always, internal politics and warfare dominated the imperial outlook. 
These problems would be difficult for even the best general and politician to surmount, and most of the emperors about whom we have any information, and who reigned for long enough to attempt to accomplish anything, seem to have been capable.  The difficulties got the better of them all.  All the different factions seem to have managed a turn at controlling the throne: the Aquitanians and Goths with Avitus (455-6); the Italians and their army with Majorian (456-61), Libius Severus (461-5) and Romulus (475-6); the Dalmatian army and its supporters with Julius Nepos (474-5), the Vandals/North Africans with Olybrius (472) and (possibly) the Burgundians/Rhône valley senators with Glycerius (473-4).  Anthemius (467-72) was the appointee of the Eastern Empire.  Anthemius could additionally play the dynastic card, being related to the Constantinian and Valentinianic/Theodosian houses.  He worked hard to solve the West’s difficulties for five years before falling victim to them and Ricimer.  A key feature of politics in this period was the use of the title rex (king) by leaders of ‘barbarian’ forces not incorporated in the imperial regime.  Thus, when his imperial candidate Glycerius was deposed by Julius Nepos and the Dalmatians in 473, the general Gundobad left Italy and returned to his people, the Burgundians, among whom he became king.  The rebel general Aegidius appears to have taken the title of rex francorum (King of the Franks) when he rebelled against Ricimer and Libius Severus.[33]  By contrast, the barbarian soldier and emperor-maker Ricimer never took a royal title.

The Undead Empire, 476-533

When Odoacer, the Italian army’s commander, deposed Romulus ‘Augustulus’ (little emperor) in 476, no faction had the power to defeat the others and create a lasting imperial regime.  The West had endured three interregna after 455, after the deaths of Avitus, Libius Severus and Olybrius, totalling over three years (a seventh of the total period between Valentinian III’s murder and Romulus’ deposition).  The different regions could govern themselves via the imperial bureaucracy, especially as even the ‘barbarian’ kings also had legitimate Roman titles, whether there was a western emperor or not.  Odoacer’s return of the western regalia to Constantinople in 476, simply recognised this fact.  After Emperor Zeno refused his request for the title of patrician, Odoacer did what others like Aegidius and Gundobad had done before him: he took the title king. 
The various regional factions of the Western Empire thus crystallised into a network of kingdoms: the Vandals in North Africa, the Burgundians on the Rhône, the Visigoths in Aquitaine and, eventually, Spain, the Sueves in north-west Spain and Odoacer’s kingdom in Italy.  In northern Gaul, the leaders of Aegidius’ old army on the Loire carved out a Frankish realm while Britain and much of southern and eastern Spain saw competition between warlords of all sorts.  All recognised (in theory) imperial rule from Constantinople, legitimising their government.  Although contemporaries clearly recognised that there was no functioning Western Empire or reigning Western Emperor, and that this made a real difference, there is no evidence that they thought the Empire had definitively ended or ‘fallen’.  The one possible exception to this was Hydatius, the Gaelician bishop-chronicler, but it is probably not coincidental that he lived in an area where barbarian rule, unsanctioned by any imperial regime, was being imposed by force.
This new order did not end the fighting and struggles for mastery, however.  The Goths expanded their power into Spain, defeating local leaders, while also fighting the Franks to their north.  Most importantly, in 493 the Ostrogothic king Theoderic completed his imperially-sanctioned conquest of Italy.  Now controlling Italy and the formerly western territories in Illyricum, he defeated the Vandals and drove them out of Sicily.  In 507 the Frank Clovis defeated and killed Alaric II of the Visigoths and conquered Aquitaine.  Theoderic took the opportunity to take over Provence and by 510 he controlled the Visigothic territories in Spain.  At this point he had overlordship over much of the former Western Empire. Meanwhile, Clovis eliminated rival Frankish leaders and extended his domination over the trans-Rhenan peoples – Saxons, Alamans and Thuringians.  It probably seemed as though the old rivalry between Gaulish and Italian factions could explode back into open warfare.  Both Theoderic and Clovis acquiesced in being called augustus by their subjects even if they never took the title.[34]  If one had managed to subdue the other it is reasonable to posit that they might well have revived the western Empire under their own leadership.  Clovis died in 511, ending the threat of war between the two overlords, but Theoderic continued to dominate the west, seemingly bringing Burgundy and Vandal Africa into his web of alliances and promoting an ideology of an ancient Ostrogothic power (virtus) equal to Rome’s.[35] 
This self-confident Gothic ideology provoked a response from Constantinople, where the new Emperor Justin I and his nephew and successor Justinian began to proclaim that the West had been ‘lost’ to barbarians.  The West, they said, had fallen and needed to be reconquered.[36]  It was no longer part of the Roman Empire.  In 533, Justinian turned these words into action and invaded Vandal Africa.  Over the next twenty years, Justinian’s troops put an end to the Vandal and Ostrogothic kingdoms and conquered a swathe of territory in Spain.  The destruction they wrought, however, and the fact that they failed to reconquer the whole of the old pars occidentalis, meant that a new frontier came into being between the imperium romanum and the kingdoms beyond.  Because of Justinian’s ideological pronouncements, those kingdoms could no longer claim even the fiction of being part of the Roman Empire.  They needed new forms of legitimacy and a new phase of western European history came into being.  The imperium that had once unified the west had now formally dissolved into the series of diverse, regional kingdoms of medieval Europe.

Bibliography

Ammianus Marcellinus, Res Gestae: Ammianus Marcellinus. trans. Rolfe, J.C., (London 1935-39) (3 vols.).
Gregory of Tours, Histories: Gregory of Tours. The History of the Franks, trans. Thorpe, L., (Harmondsworth, 1974)
Hydatius, Chronicle: The Chronicle of Hydatius and the Consularia Constantinopolitana. Two Contemporary Accounts of the Final Years of the Roman Empire, ed. & trans. Burgess, R.W., (Oxford, 1993), pp.1-172.
Marcellinus Comes, Chronicle: The Chronicle of Marcellinus, ed. & trans. Croke, B., (Sydney, 1995)
Orosius, History against the pagans: Paulus Orosius. The Seven Books of History against the Pagans, trans. Deferrari, R.J., (Fathers of the Church 50; Washington, DC, 1964)
Sidonius Apollinaris, Poems and Letters vol.1, ed. & trans. Anderson W.B., (London, 1936); Sidonius: Poems and Letters vol.2, ed. & trans. Anderson W.B., with Warmington, E.H., (London, 1965).
Amory, P., 1997: People and Identity in Ostrogothic Italy, 489-554 (Cambridge).
Brown, Peter, 1971: The World of Late Antiquity (London).
Burns, Thomas S., 1994: Barbarians within the Gates of Rome. A Study of Roman Military Policy and the Barbarians, ca.375-425 (Bloomington, Indiana)
Conant, Jonathan, 2012: Staying Roman. Conquest and Identity in Africa and the Mediterranean, 439-700 (Cambridge).
Croke, B., (1983). ‘476: The manufacture of a turning point.’ Chiron 13:81-119 (repr. as B. Croke, Christian Chronicles and Byzantine History, 5th-6th Century (London, 1992), no. V).
Drinkwater, John 1996: ‘“The Germanic threat on the Rhine frontier”: A Romano-Gallic artefact?’, in R.W. Mathisen & Sivan, H. (ed.) (Shifting Frontiers in Late Antiquity (Aldershot), pp.20-30.
Drinkwater, John 1997: ‘Julian and the Franks and Valentinian I and the Alamanni: Ammianus on Roman-German relations.’  Francia 24:1-16.
Elton, Hugh, 1996: Frontiers of the Roman Empire (London)
Esmonde-Cleary, A. Simon, 1989: The Ending of Roman Britain (London)
Gerrard, James, 2013: The Ruin of Roman Britain: An Archaeological Perspective (Cambridge)
Goffart, Walter, 1980. Barbarians and Romans AD 418-585: The Techniques of Accommodation (Princeton).
Goffart, Walter, 2013: ‘Administrative Methods of Barbarian Settlement in the Fifth Century: The Definitive Account’ in Diefenbach, S., & Müller, M., (ed.) Gallien in Spätantike und Frühmittelalter : Kulturgeschichte einer Region (Berlin), pp.45-56.
Halsall, Guy, 2007: Barbarian Migrations and the Roman West, 376-568 (Cambridge)
Halsall, Guy, 2012a: ‘Northern Britain and the Fall of the Roman Empire’ The Mediæval Journal 2:2: 1–25  
Halsall, Guy, 2012b: ‘From Roman fundus to Carolingian Grand Domaine: crucial ruptures between late antiquity and the middle ages.’ Revue Belge de Philologie at d’Histoire 90: 273-98
Halsall, Guy, 2014. ‘Two worlds become one: A counter-intuitive’ view of the Roman Empire and ‘Germanic’ migration’. German History 32.4: 515-32
Halsall, Guy, 2016: ‘The Ostrogothic military’ in Arnold, J.J., Bjornlie, M.S. & Sessa, K., A Companion to Ostrogothic Italy (Leiden), pp.173-99
Heather, Peter, 1991: Goths and Romans, 332-489 (Oxford).
Heather, Peter, 1996: The Goths (Oxford).
Heather, Peter, 2001: ‘The late Roman art of client management: Imperial defence in the fourth-century west’, in Pohl, W., Wood, I.N., & Reimitz, H., (ed.) The Transformation of Frontiers. From Late Antiquity to the Carolingians (Leiden), pp.15-72.
Heather, Peter, 2005: The Fall of Rome: A New History (London).
James, Edward, 2008: ‘The rise and function of the concept ‘Late Antiquity’ Journal of Late Antiquity 1.1: 20-30.
James, Edward, 2009: Europe’s Barbarians, AD 200-600 (Harlow)
Kelly, Christopher, 2008: Attila the Hun. Barbarian Terror and the Fall of the Roman Empire (London)
Kulikowski, Michael, 2000b: ‘Barbarians in Gaul, usurpers in Britain’. Britannia 31:325-45.
Kulikowski, Michael, 2006: Rome’s Gothic Wars from the Third Century to Alaric (Cambridge).
Kulikowski, Michael, 2014: ‘The failure of Roman arms’, in Lipps, J., von Rummel, P., & Machado, C. (ed.), The Sack of Rome in 410 AD. The Event, its Context and its Impact (Wiesbaden), pp.77-83.
Lenski, Noël, 1995: ‘The Gothic civil war and the date of the Gothic conversion.’ Greek, Roman and Byzantine Studies 36:51-87.
Lenski, Noël, 2002: The Failure of Empire: Valens and the Roman State in the Fourth Century A.D. (Berkeley).
Liebeschuetz, Wolfgang 1991: Barbarians and Bishops.  Army, Church and State in the Age of Arcadius and Chrysostom (Oxford).
Liebeschuetz, Wolfgang, 1992: ‘Alaric’s Goths: Nation or army?’ in Drinkwater & Elton (ed.) (1992), pp.75-83.
Lütkenhaus, W., (1998). Constantius III. Studien zu seiner Tätigkeit und Stellung im Westreich 411-421 (Bonn).
Maas, Michael (ed.) 2014. The Cambridge Companion to the Age of Attila (Cambridge)
MacGeorge, Penny, 2003: Late Roman Warlords (Oxford).
McCormick, Michael, 1986: Eternal Victory.  Triumphal rulership in late antiquity, Byzantium and the Early Medieval West (Cambridge)
Rousseau, Philippe (ed.), 2009: The Blackwell Companion to Late Antiquity (Oxford)
Stickler, Timo, (2002). Aëtius: Gestaltungsspielräume eines Heermeisters im ausgehenden Weströmischen Reich (Munich).
Whittaker, C. R., 1994). Frontiers of the Roman Empire (Baltimore, Maryland).
Wolfram, Herwig, 1988: A History of the Goths (Berkeley)
Wood, I.N., 2013. The Modern Origins of the Early Middle Ages (Oxford)




[1] For this historiography, see Wood 2013; James 2009 is the best guide to debates about the barbarians.
[2] Brown 1971.
[3] James 2008 and other articles in the same issue of the journal; Rousseau (ed.) 2009.
[4] Halsall 2012a.
[5] For a general survey of the frontier and Romano-Barbarian relations see Halsall 2007:138-62, with references.
[6] Drinkwater 1996; 1997; Elton 1996; Whittaker 1994.
[7] Heather 2001.
[8] For an overview of barbarian society and politics, see Halsall 2007:118-36.
[9] Ammianus Marcellinus, Res Gestae 26.5.7.
[10] Halsall 2014.
[11] Halsall 2014.
[12] There are many studies of this crisis: Burns 1994; Halsall 2007:170-85; Heather 1991:122-56; 1996:98-102, 130-8; Kulikowski 2006:123-53 ; Lenski 1995; 2002:320-67; Wolfram 1988:
[13] Ammianus, Res Gestae 31.2.1-12; Halsall 2007:171-2.
[14] Lenski 1995.
[15] Two classic studies of late of Roman Britain are Esmonde-Cleary 1989 and Gerrard 2014.
[16] For this interpretation, see Halsall 2012b and references.
[17] Kulikowski 2000.
[18] Liebeschuetz 1991: 34-6.
[19] For studies of Alaric’s career, see Burns 1994; Halsall 2007:194-206, 214-7; Heather 1991:193-218; 1996:138-48; Kulikowski 2006:154-77; 2014; Liebeschuetz 1992.
[20] On Constantius see Lütkenhaus 1998.
[21] Orosius, History against the Pagans 7.43.16-17.
[22] For different narratives of this period, see Halsall 2007:234-56; Heather 2005:251-348, 369-75.
[23] On Aëtius, see Stickler 2002.
[24] On Attila, see Kelly 2008; Maas (ed.) 2014.
[25] Marcellinus Comes, Chronicle, sub anno 454.
[26] Conant 2012:130-86.
[27] Hydatius, Chronicle 147.
[28] For narratives, see Halsall 2007: 257-83; Heather 2005:375-430.
[29] On which see Goffart 1980; 2014; Halsall 2007:422-47 for discussion of the debate; Halsall 2016.
[30] Sidonius Apollinaris’ panegyric to Majorian (Poems 5) makes this very clear.
[31] On Ricimer see MacGeorge 2002.
[32] Halsall 2016.
[33] Gregory of Tours, Histories 2.12.
[34] Theoderic: McCormick 1986:278-80; Clovis: Gregory of Tours, Histories 2.38.
[35] Amory 1997:59-71.
[36] Croke 1983.