Pages

Sunday, 27 October 2013

Transformations of Romanness

[I am off to Vienna to a conference with the same title as this post, which is also the title of my contribution.  This is a much longer version of the paper I will give, which will dwell on the changes after 476, but it is a draft of the version I hope to publish.

Update 25/11/2014: I have updated this post now so that is the same as that submitted for publication, rather than the text that was delivered in Vienna.]

***
 
No identity can remain fixed and stable.  Any perspective informed by modern continental philosophy cannot fail to see any primordialist position on ethnic identity as fundamentally wrong-headed.  Identity is a question of desire, of a ‘motion towards’ an ideal.  The ideal, naturally, is unattainable.  No identity is coextensive with itself, nor can it be, except among the dead (which opens a different perspective on the study of identity in early medieval cemeteries).  What makes an identity is, furthermore, always something more than itself; something in other words that – in elementary Lacanian – operates at least as much in the realm of the imaginary as in that of the symbolic.  This is why external markers or signs of distinction never suffice.  An identity is always constructed in part by what it is not, however that negation is articulated.  In this purely symbolic sense it is no different from any other sign, operating within a chain of metaphor and difference.  It is impossible to separate an identity from its alterity, from its negations.  This is one reason why it makes no sense to study ‘othering’ or alterity as some sort of process distinct from identity-construction.[1]  A strategy of distinction is always at the self-same time a strategy of identification, and vice versa, bound together like the two sides of a Möbius Strip.  To raise a common identity in a social interaction is simultaneously to raise those things that both actors share in not being and those things that they do not share.  The things held to constitute an identity, those things which are ‘in it more than itself’ are contingent, ever changing, and yet, at any one moment, always constructed as timeless and essential.  In any given context, an identity is always already what it is.  This is yet another reason to mistrust views that portray something like Gothic identity as an unchanging monolith.
In my 2007 book Barbarian Migrations and the Roman West I included a long discussion of ethnicity in which I made a number of points which, sadly, I do not think have made much impact on the study of the topic.[2]  One is the that ethnicity is multi-layered, so that ethnic change is much more about the acquisition of new layers and the reshuffling of old ones through time than it is about swapping one for another.  It is this misguided notion that is at stake, for example, in Patrick Amory’s work and in Peter Heather’s critiques of it.[3]  I argued that everything that we might say about what we consider to be an unproblematic ‘ethnic’ identity at the level to which I will, here, provisionally (with full recognition of the problems of the term) refer as ‘gentile’ (i.e. as a gens ‘people’) can be applied to identities that come at perhaps lower levels, which might loosely (but misleadingly) be described as residential, local or regional.[4]  I then made the point that what we think of as the appropriately ‘ethnic’ (i.e. ‘gentile’, as above) level within this arc of the spectrum of identities is contingent upon historical circumstance.  Ethnicity is dynamic.  I hypothesized that a break-up of the UK into regional units might lead to a situation where a Yorkshire identity was considered more important than an English or British one.  Again, the point concerns layers of ethnicity and their contingent reordering as part of the dynamics of historical change.  Hierarchies of identity might be reordered in the opposite direction too.  Some of us would rather be seen as British than English and as European rather than British.  The study of Roman identity, its many levels and its change through time, is an excellent case study of these points.  There are other reasons why the topic is important.

Deconstructing the tortured historiography of Roman ethnicity

Derrida said that ‘deconstruction is what happens’ (‘ce qui arrive’) in a reading.[5]  Let me offer a little deconstruction of my own previous writing.  In Gregory of Tours’ account of the various diplomatic comings and goings in late sixth-century Gaul, he describes one embassy as including ‘Warinar [or Warmar] the Frank and Firminus the Arvernian’.[6]  Another comprised ‘Bodegisil the son of Mummolenus from Soissons, Evantius the son of Dynamius from Arles, and Grippo the Frank.’[7]  In my own work,[8] I have used these references to talk, first, about what a Frank was in Gregory’s writings and, second, about the northern Gallic aristocracy and why Gregory did not think there was a Frankish nobility.  Yet what one might term the ‘repressed other’ of the discussion is represented precisely by the two non-Frankish characters, Firminus and Evantius.  What of their identity, which Gregory specifically tells us about in terms that seem to mark its structural equivalence to francus: sessionicus and arelatensis?[9]  What passes without discussion is what it meant to be an Arvernian or from Arles, or – if Mummolenus’ sons weren’t identified as Franks – from Soissons.   
I am, however, in good company.  A similar deconstructive reading of Walter Pohl’s writings illustrates, as I hope to show, how a crucial misunderstanding has come about and how the present volume might deal with and ameliorate the results of that, to move the debate on in productive fashion.  The debate – maybe dispute would be better – between Walter Pohl and Walter Goffart is well known.  It causes me some distress.  I have friends on both sides – I like to think that I get on well with both of the principals – and no one likes to see their friends arguing so bitterly.  Most of the rudeness has come from the western shore of the Atlantic, but rudeness is not the only, and certainly not the most effective, form of academic aggression.  This confrontation has long perplexed me, largely because I have strained to see exactly what the ‘Toronto School’s’ objection to the ‘Vienna School’ was.
I should not have to make clear that Walter Pohl’s works have been immensely valuable and important to me, or that I am almost entirely in agreement with its principal conclusions.  Both points should be visible from my own previous writings. It must furthermore be stressed that deconstruction is not in and of itself a hostile move.  As any aficionado of Derrida’s writing knows, it is a recognition of a text’s, or a body of writing’s, quality, importance and value.  Having thus made clear that what follows is motivated not by hostility or confrontation but by respect and friendship I should like to discuss what seemed to me to emerge from a deconstructive re-reading of as much of the Pohl oeuvre as was available to me, especially the classics of the corpus.[10] 
Throughout this work, the analysis of what makes (or does not make) and what distinguishes (or does not) a people remains at the level of those groups which have always been considered to be peoples: Franks, Goths, Lombards, Burgundians and the rest, for whom the constitutive outside is indeed formed by the Romans, the Roman Empire.  For example, in his classic article, ‘Telling the difference’, Walter Pohl asked what it meant to be (inter alia) a citizen of the civitas of Tours.[11]  That part of the question, however, was never answered.  That type of identity, Wir-Gefühl or whatever played no further role in the discussion.  The ‘repressed other’ throughout the text is the non-ethnic group, whatever (if ever) that was or might have been.  Put another way, the analysis (however subtle, persuasive and brilliant) remains at the level of the ethnic groups that have always been considered to be ethnic groups and which have always been the ethnic players in the story: the barbarians.  It will become clear below that this is in fact commonplace in discussions of early medieval ethnicity, from all historiographical camps, but the inevitable point is that if one limits the discussion in this way, to those social units which have always been held, a priori, to be ethnic groups, the argument runs a strong risk of circularity, or at least of simply reaffirming its initial premises.
What has therefore escaped the discussion thus far is how the Turoni, Arverni and Bituriges, or the Gauls, Spaniards and Italians, differed from each other, or what the difference was between those ethnic taxonomies and that of the Germani.  Was a Batavian any more different from a Frank than from a Treverian?  I doubt it, but then I think we have not sufficiently carefully kept the analysis of the Roman-Barbarian dichotomy separate from that of the taxonomic ethnography that pervades our sources.  And yet, the deconstructive reader of the Pohl oeuvre (as it stood before the 2013 Vienna conference) will see an effective division between Germani – who have ethnic groups – and Romans – who, within the explicit discussion of the texts published thus far, seem not to have.  Also visible is the persistence of the type of identity, which earlier I provisionally labelled gentile (identity as a people), as something somehow ‘special’, vis-à-vis other types of identity that could and should be positioned within the same, ‘ethnic’ part of any map of identities.  On that basis, whether one likes it or not, seeing ethnicity in the Pohl oeuvre as a means of clothing old-style invading barbarians in new garb is one possible legitimate reading.  This is permitted by two aporias: silences, or rather by the spaces left by those silences, which represent a ‘blockage’ in the reading of a text where interpretation can proceed no further but must follow the reader’s choice.[12]  These are points of undecidability.  One point is the silence about what distinguishes an ethnic group from a non-ethnic group within the same part of the spectrum, or layering, of identities.  The other is the silence about what distinguishes such identities or layers of identity among the barbarians from their equivalents among the Romans, or more accurately among the inhabitants of the Roman Empire.
Now, as it happens, it is vital to point out that a deconstructive reader will encounter precisely the same aporetic silences – if anything – more easily, in the works of some of Pohl’s critics, and indeed with more serious consequences for their arguments, as we shall see.  Leaving that aside for the moment, however, I would like to propose, in the interests of attempting to bridge the historiographical divide, that it is in the silent spaces of these aporias that the misunderstanding between the Toronto and Vienna ‘schools’ originates.  The Toronto school views the silences as constitutive, integral to the argument.  By contrast, in my own previous readings of the corpus Pohlianus, I never have assigned them any significance.  I admit some positive curiosity about them but I perceived them more as a sign of incompletion, as a space yet to be filled, as an area where the argument could be advanced.  The Toronto School’s critique, by contrast, is – as I see it – based on a reading of those silences that suggests that they imply a genealogical link to the more traditional Lehre of ‘Germanic’ history.  Now, just as long as one sees the critique as based upon extant silences within the Pohl argument, and just as long as one acknowledges that an argument is composed of its silences as well as its statements, one must concede that the Toronto critique is based on a valid reading.  If one does not accept this, one must continue either, as I did, to be unable to recognise the works of Walter Pohl in the Toronto critiques thereof and – in consequence – to be simply bemused about what was going on, or more actively to see them as a malicious and wilful distortion.  I do not think that the last option is justified either.  There may have been malice involved – certainly the critique could have been expressed in less hostile and offensive language – but I do not think there has been deliberate distortion.
This, I think, suggests why the two sides have continued to talk past each other and how such bitterness and anger has arisen.  To repeat: what is at stake is silence and, when silence is at stake, ‘I never said that’ will never be an adequate riposte.  Hence, as I see it, one side’s frustration with the other for not dealing with what is actually written on the page and the other side’s frustration that their opponents seem to refuse to answer its criticism.  When the space exposed by those silences opens onto the traditional Lehre of the Völkerwanderung and thence – inevitably – to Nazis, and where the two principals are of the precise respective heritages of those involved here, unsurprisingly tempers will flare and would have done, I suspect, even if the critique had been made in less deliberately (or carelessly) provocative terms.
The point about an aporia, or an aporetic silence, as here, is that it is, in Derridian terms, a space of différance, where a choice between two undecidable options can only be made on the basis of a purely political decision.  Put another way, as intimated above, the text itself provides no empirical pointer so one must decide for oneself, for one’s own reasons, what the silence means.  Thus the debate becomes as tribal as it has done: almost ethnic (meta-ethnic?) in itself.  That is why we have made so little progress in resolving the issue, for all the debate’s heat.  Not dealing with those aporia in the argument, once pointed out, though, will not merely not close up that space; it will actively keep it open.  That is why this volume and the conference upon which it was based are so important.  They overtly address – or should do – both of the silent spaces I have mentioned.  I am not the sort to argue for cosily artificial rapprochement or, worse, consensus but one of the many tributes one can pay to Walter Pohl is that, in spite of the calumniation he has received, he has continued to talk to the Toronto historians.  What I hope for from addressing those aporias is – one way or another – the provision of something concrete and decidable, on the basis of which those of us with no tribal affiliation can make a choice.

Roman Identity

Roman identity serves as a particularly good case study of the multi-layered, situational and dynamic nature of ethnic identity.  Let us return to Gregory of Tours.  Edward James and Walter Goffart have both argued that ethnicity was not important for Gregory.[13]  Why not?  Because he rarely ascribes a ‘gentile’ identity to the people in his stories.  He does not talk all that often about franci; saxones crop up a couple of times, once famously or infamously cutting their hair and dressing in the Breton style; a Goth makes an appearance here and there; and that is about that.  But the Historiae are full of people identified by civitas.  James says that Gregory identifies himself not as a Roman but as an Arvernian.[14]  That sort of identity, according to civitas or in some cases, as with the men of Champagne, ducatus, was – very obviously – something that mattered a lot in Gregory’s world.[15]  There is nothing that allows us analytically to distinguish these types of identity from the ‘gentile’ level. 
Let us pause here to note the implicit assumption within James’ and Goffart’s articles: that ethnicity is a level of identity equating with ‘people’ generally (that is to say it operates, in the term provisionally adopted here, at the ‘gentile’ level) and with Germanic people specifically.  Presumably, ethnicity can only have mattered to Gregory if his works were filled with descriptions of the characters in his tales as ‘Franks’, ‘Goths’ or ‘Saxons’ or, for James, if he had self-identified as ‘Roman’ (apparently in opposition to Frankish).  This is a point of considerable interest, not least because it marks a point at which Pohl’s and Goffart’s writings come together and indeed join those of many other writers on the topic.  For Pohl and Goffart equally, as for James (and the early Halsall), ethnicity and its importance is to be judged according to the usage of Germanic ethnonyms as markers of identity.  The implication of Goffart’s article is that people had no ethnic identity of any significance if that was not ‘Germanic’.  The fact that the Libri Historiarum are replete with Arverni, Turoni, Lemovici, Turnacenses, Bituriges and the rest is a point which, for Goffart, seems to have no bearing at all upon the question of the political importance of ethnic identity in Gregory’s Gaul.  Indeed, close reading and comparison of the Goffart and Pohl corpora reveals that ‘Germanic’ ethnicity is far more real for Goffart (and his followers, especially Callander Murray[16]) than for Pohl.  The crucial issues are these: first, in Goffart’s view these Germanic ethnic groups did not bring down the Roman Empire and were of no historical significance whereas, for Pohl, whether or not they brought down the Empire, ethnically-named political groups were of central importance in the political changes of the fifth and sixth centuries; and second, Goffart reads Pohl’s argument (as above) as a refiguring of the old view of the conquest of the Roman Empire by Germanic peoples.[17]  Indeed – ironically – it may actually be his own view of ethnicity (especially ‘Germanic’) and its reality that predisposes Goffart to read Pohl in this way.
Returning to Gregory of Tours and the precise problem of Roman identity, one must ask why the bishop of Tours avoids the designation of Romani for those people he identifies by civitas or ducatus.  Any answer must acknowledge first of all that this was nothing new; Roman ethnicity had always worked at multiple levels.[18]   Another key problem in so much discussion of late antique ethnicity is the failure to tease out or consistently analyse these different levels.  Especially important – and perhaps confusing for the issue – is the fact that the concept of ‘Roman’ functioned at a structural as well as a taxonomic level and that these two levels could sometimes be run together.  By the structural level I mean the use of the terms ‘Roman’ and ‘barbarian’ to mark an important organisational, cultural difference between civilised and savage.  By the taxonomic level I mean the way in which the world was described as divided up into the territories of different peoples.[19]  The two different levels are well illustrated by the two parts of Tacitus’ Germania.  The first section differentiates the Germani, qua barbarians, from the Romans in a way that, as has long been noted, cannot really be read other than as a critique of Roman society and politics under Domitian; the second half gives a taxonomy of the Germanic peoples, with few or no points of contact with the first section.[20]  Like any identity, Roman identity operated in the symbolic and imaginary realms (as above)[21].  The crucial point is that the structural level of Roman identity – that which has barbarian as its opposite or other – functions almost entirely in the imaginary register and even then in a mainly self-referential way; the taxonomic level works in the symbolic to a much greater degree.  The term Romanus may confuse the issue by (unlike barbarus) being capable of being used in both levels but it is analytically vital to keep them distinct.  Failure to do that has bedevilled much study of late antique ethnicity.  The opposite of barbarus is romanus; the opposite of, say, Francus is not.
The structural level mapped onto the taxonomic in historically-contingent ways.  Caesar, for example, described the Gallic and Germanic peoples but there is no sense from Tacitus’ historical writings that the movement of the imperial limes in the intervening century and a half had much altered this taxonomy.[22]  When Ammianus launched into his periodic ethnic excursus it makes little difference whether he is talking of areas inside or outside the imperial frontiers.[23]  The whole world was made up of different peoples with their own characteristics.  Plenty of fourth-century evidence backs up the idea that the Romans thought of the world within the limes as a mosaic of different ethne.[24]  The late Roman popularity of works on the origines of those people, within which genre post-imperial origines gentium are surely to be located, makes this clearer still.[25] 
Thus, contrary to what Walter Goffart has written,[26] one of the many interesting things that happened to the Roman Empire in the fifth century was not that it ran into a wave of ethnicity.  Ethnic identity was alive and well throughout imperial history.  The problem, noted above, is that Goffart’s work contains the same aporia, the same silence concerning what differentiates an ethnic from a non-ethnic identity and about what differentiated an intra-imperial ethnic group or identity from an extra-imperial one.  Throughout Goffart’s work they remain tacitly present and unchallenged, and with a much more serious impact upon his argument.

The Development of Roman Identity to c.476

The questions I wish to examine in the remainder of this paper are, first, how and especially why Roman ethnicity should have defaulted to the civitas level by Gregory’s day; and, second, why it then got worse.  In Salic Law, the Romans are clearly a parallel population to the Franks, even if legally disadvantaged in some ways.[27]  A century or so later, the Romani of Lex Ribvaria are just one of several semi-free categories, all of which must have a Ripuarian speak for them at law.[28]  That situation would have been unthinkable even a century earlier, let alone in 400. 
An answer to this question requires us to consider the role of Romanness, as a supra-regional, imperial identity in the process of subjectivization.  Here in important ways, the taxonomic and the structural come together with the political and cultural.  Romanness was central to the formation of the political subject.  What was held to distinguish man from woman also distinguished Roman and barbarian, and human and animal: moderation, control of the emotions, reason.  These aspects learned during socialisation, paideia, enabled participation in legitimate government and rendered Roman forms of government superior to others.  Movement towards the ideal legitimised behaviour and authority of all sorts.  Movement away – real or alleged – had the opposite effect.  Control over the political centre, the imperial court, enabled one to define who was and who was not behaving in the correct legitimate fashion or moving in the right direction.  Thus Roman identity, as something moved towards, was central to the sex-gender system and to political legitimacy.  This transcended taxonomy.  As is well known, a barbarian could behave in such a ‘Roman’ fashion that his non-Roman origins were held of no account or – more correctly – were held simply in the taxonomic register, just as we may suppose were the origins of a Spaniard, Gaul or African at the imperial court.  This should not be controversial.[29]
Nonetheless, the point just made is worth stressing.  In the taxonomic sense, Romans had always had multiple layers of identity: as a citizen of the Empire; as originating in one of the major imperial regions, whether or not fossilised as dioceses in the Diocletianic Empire (Britain, Gaul, Spain, Africa, etc.); as the inhabitant of a particular province or civitas.  The writings of the Gallic authors of the late Empire make this last level very clear.[30]  Someone could claim different civitates as bases of identity from paternal and maternal descent.[31]  Perhaps there were lower, nested levels too, based around lesser settlements or communities, or pagi.[32] 
None of these levels differs fundamentally from a ‘gentile’ or similar identity.[33]  The tribal origins of the Gallic and British civitates in any case made them ’gentile’ identities of a sort.  Classical ethnography provided a resource for the assigning of characteristics to such identities.  We can see this in Ausonius’ jibes about a British rival or in Ammianus’ comments on the Gauls, and their differences from the Italians, or on the Pannonians who came to pre-eminence under Valentinian, or, earlier, in Cassius Dio’s ascriptions of Caracalla’s diverse personal defects to his family’s origins in different regions.[34]
It is by no means clear that any of these identities functioned differently in social relations within the Empire – that is at a level below a shared Roman political identity – from barbarian confederate or tribal identities or origins: a resource for differentiation, assimilation or other relationships.  Classical ethnography played a part of course as can be seen in Ammianus’ criticism – or praise – of the inhabitants of different regions.[35]  That makes it difficult to assume a different treatment at this level of identity, of people from within or without the limes.  The late Roman army’s élite auxilia palatina included regiments named after Celts and Batavians as well as after Franks and Saxons.[36]  That sort of ethnographic taxonomy played on the kinds of bio-geographical pseudo-science that was held to explain the civilised:barbarian dichotomy, so the two aspects bled into one another.[37]  This further emphasises my point about the contingency and analytical interchangeability of different levels of ethnicity.
Crucial to the development of Roman identities are the changes in the Roman army in the fourth century and its so-called barbarization.  The extent to which this was an actual matter of real non-Roman influence can be debated, as in Barbarian Migrations, where I argued suggested the existence of a certain ‘barbarian chic’ that might usefully be considered as an analogue for the nineteenth-century French Zouaves: French troops who wore a French idea of North African native dress and who adopted a number of other North African cultural practices, all of which gave then a tremendous esprit de corps.[38]  It nevertheless seems clear that following the division of civil and military services the army began to create a set of new identities that centred on the very antitheses of the civic Roman masculine ideal: animal, ferocious, braggart, barbarian.[39]  Such would of course form only one level of identity, nested within and as contingent as the others.  Nonetheless that represents a crucial development that provided a hugely important resource within the political and social developments of the fifth century.
The key feature of fifth-century politics, especially after Valentinian III’s assassination, was faction fighting between groups made up of Romans and barbarians.  A failure to control the centre, or a defeat by those who did, led factions cut off from traditional legitimation of status to seek other forms of legitimate political authority.   In this context, the barbarized military model of Romanness was very valuable to an aristocracy used to serving in or working alongside a ‘barbarised’ army.    This surely eased the transition to the state of affairs around the 470s, where a series of regional factions existed, each grouped around a particular army, none of which was able both to gain lasting and secure control of the political centre and defeat, and establish legitimate dominance over, the others.
In the course of the fifth century, barbarian military leaders had, as Michael Kulikowski has very clearly shown, using insights from postcolonial theory, developed a particular pattern of behaviour with regard to the imperial court.[40]  This involved stressing the stereotypical, threatening role of the barbarian when outside the circles of legitimate authority, as a means of being reintegrated into the latter, when such identities were dropped.  Crucially, though, Roman military commanders seem to have picked up this model of behaviour too in the mid- to late fifth century.  This, as I have argued, involved the adoption of the quintessential gentile title of rex as a basis for an authority that could be dealt with legitimately by Romans.[41]

The Development of Roman Identity c.476-c.550

So we arrive at the situation that seems to have predominated between about 476 and the middle of the sixth century, when a Roman civil aristocracy and administration served alongside a barbarian army.  This situation was importantly different from that which existed 100 years earlier but nevertheless was clearly descended and developed from the latter.  It is important to stress that this was not a situation of straightforward binary oppositions but one of nested levels of identity.  In the bizarre political situation that existed in the half century between Romulus Augustulus’ deposition and the death of Theodoric, when I would argue that people were aware that the Western Roman Empire was no longer functioning but not that it had ended, it is unsurprising that discussions of Roman identity largely took the form of discussions of legal relationships between Roman citizens and barbarian soldiers.  This can be seen in the famous texts of Ostrogothic Italy, or in Salic or Burgundian Law.[42]  Otherwise, as with the slightly earlier writings of Sidonius Apollinaris, they concerned the traditional underpinnings of Roman identity: culture, education and so on.[43]  Yet at the same time the Roman aristocracy’s militarisation continued, as is well known.  Simultaneously, the Church was adopting classical aspects of Roman civic masculinity and was, in some areas, such as Gaul, becoming a focus of Roman aristocratic competition.  Here too, though, there was change and an opposition to a very un-Roman competitive asceticism.[44]
It is difficult to see how this situation could fail to cause the renegotiation of Roman identity.  As I stated at the beginning, identity is a motion towards, an issue of desire.  Any identity depends upon a set of ideal images and commensurate oppositions.  In the situation that was emerging in the course of the fifth century the oppositions were neither as pronounced nor as negative.  The political advantages of Romanness were lesser, too.  That did not mean that there were no attractions.  In the strange ‘sleepwalking’ period after 476 the emperor remained the ultimate political reference point, in whom barbarian soldier and Roman civilian, and the legitimation of the forms of authority invested in both, came together.[45]
The situation is perhaps very well illustrated in the Pactus Legis Salicae, where Romani and Franci have seemingly well-defined functions but both have access to the king.[46]  The Franks have legal privilege, which is hugely significant, but the image is of two parallel populations.  We should not teleologically assume that this situation was destined to develop along particular lines.  The world after 476 contained many possibilities.
One of these possibilities was the re-establishment of unity by military action.  In c.510 there existed a situation wherein two kings, Clovis and Theodoric, having between them established dominance over almost the whole western Empire, faced off against each other.  Both were evidently happy to be addressed as augustus by Roman subjects, even if neither formally adopted the title.[47]  To contemporaries, this situation may well have seemed simply like the next, perhaps decisive, round in the struggle between the Gallic and Italian factions that had dominated fifth-century western politics.  Had such a play-off come about, and been as decisive as many other battles of the period had been,[48] it is likely that a western Empire would have been re-established, however permanently or impermanently, under Amal or Meroving rule.  Who can guess what might have become of Roman identity in that event?  I suspect that something closer to the fourth-century situation may have emerged, although it was unlikely to have represented a re-establishment of or reversion to precisely that state of affairs.
Such a decisive confrontation, of course, never took place.  Instead, possibly motivated by the developments around 510, the Constantinopolitan court began to emphasize its exclusive, Roman legitimacy and in time to attempt to re-impose political unity through its own military actions.  What made a crucial difference to these campaigns was the well-known Justinianic ideological offensive.  Famously this involved a rewriting of fifth-century history to portray the West as lost to barbarian invasions, which can be seen in works from Marcellinus Comes’ Chronicle, through at least the early books (though I would say all, to some extent) of Procopius Wars.[49]

The development of Roman Identity, c.550-c.625

The impact of the Justinianic Wars, and especially of the fact that they did not result in the West’s military domination by the eastern Emperor, cannot be overestimated.  After twenty years of brutal destructive warfare waged to make the point, no one could be in any doubt that the areas beyond actual imperial authority were not part of the Empire any more.  They remained lost to barbarians; the frontier between imperial Roman inside and outside had formally been redrawn.  As far as Roman identity in the West was concerned, this completely changed the game.  It did so for all sorts of identities, the traditional bases for which had to be redefined.  As is well-known, the Old Testament became a new source of models and ideals.[50]  What could be done with Roman identity though?
In this context it is not surprising to see the dramatic decline in Roman identity at the end of the sixth century in Gaul.  The parallel societies of Lex Salica disappeared.  In the sixth century the personnel of the Gallic church was dominated by people with Roman names.  Around 600, that changed so that bishops overwhelmingly had Frankish names.  The episcopal list of Metz, for instance, reveals only a couple of non-Roman or non-biblical names before about 600.  After that the situation is reversed.  This is fairly typical for northern Gaul.[51]  One might read that change in several ways.  The families who provided members of the episcopate changed their naming practices; the people entering the episcopate ceased to adopt Roman names as more appropriate to their status; or the Roman families that had provided the bishops dropped in status.  Either way, the significance of this change for Roman status remains and cannot be ignored.
This is also the period when Gregory of Tours was writing his Histories.  In this context I think it is unsurprising that Roman identity is conspicuous by its absence.  As Edward James says, Gregory does not self-identify as Romanus but as Arvernus.[52]  What might have been seen as the imaginary element of classical Roman identity is displaced into senatorial noble identity and into Christian behaviour.[53]  Otherwise Roman identity has defaulted to the level of the civitas identity as in the case of the embassies mentioned earlier.  It is no surprise that civitas identifiers are mostly confined to the south.[54] This identity is no less ethnic than that of Frank.  Like all such identities, it could be the object of violence, as with the killing that broke out along the Loire after the death of Chilperic.[55] 
Gregory’s own rather sneering view of the men of Bourges further illustrates the point.[56]  A comparison of Gregory’s story of the foundation of the see of Bourges with that of his home town of Clermont reveals the men of Bourges to have been much more unwilling than the Auvergnats to receive the word of God.[57]  Only the intervention of a distant relative of Gregory’s even enables the embryonic church to acquire a place of worship.  After their deaths the burial places of the first bishops of both cities are forgotten but whereas the grave of Stremonius was revealed by a vision received by a future bishop of Clermont and his body translated in fairly standard fashion, at Bourges, Ursinus’ grave was only revealed after a member of the bishop’s staff received a cure (at St Martin’s , Tours, significantly) and even this revelation was disputed by the local bishop.  Only the intervention of St Germanus of Paris and further visions led to a translation.[58]
In contrast to the numerous saintly figures of Tours and Clermont catalogued by Gregory, the Berruyard holy men are fairly nondescript, and manifest a very frequent association with Tours, Clermont or St Martin.  Otherwise they are faintly ridiculous.  Witness Saint Marianus:[59] Marianus, a recluse, was found dead under an apple tree and consequently was rumoured to have died by falling out of a tree.  ‘But it was not known for certain because no one had been an eye-witness.’  Whatever the case, it was hardly the most dignified form of death for a holy man, and the locals, perhaps understandably enough from a modern point of view, were not over-impressed, in spite of unspecified healing miracles.  One local, rebuked for working on St Marianus’ feast day, angrily replied ‘Do you think that a man who slipped from a tree whilst satisfying his appetite has been included in the company of angels, so that he ought to be venerated as a saint?’  Needless to say, his house burnt down.  Only after another miracle, where some stolen oxen miraculously wandered home on their own, does Gregory say ‘after these events, the people of Bourges began to honour this confessor of God with more diligent concern.’  What better part of the world for a bogus holy man to emerge?  Towards the end of the Histories a man from Bourges is attacked by a swarm of flies and as a result goes mad.  Eventually, after a career which exactly parodies that of a proper holy man, he was killed by the pueri of the bishop of Le Puy.[60]  Clearly, identities like these, could operate in the register of the imaginary as well as the symbolic.  There is no way, analytically, of distinguishing these identities as somehow less ‘ethnic’ than those associated with the recognised ‘peoples’ of late antiquity.  Their status as a rung below more ‘gentile’ identities was only contingent upon the nature of fifth and sixth-century politics and the larger size of western kingdoms at that time. In some ways this was the golden age of civitas identity.[61] 
Gregory is famously tacit about the end of the Roman Empire in Gaul.  What is less often remarked upon is the fact that he is at least as reticent about beginning of the Roman Empire in Gaul.[62]  Roman history, as one might expect, has been displaced in favour of Christian history.  The eschatological implications of this are unclear.  Obviously, after several centuries of Christian linkage of the Empire with the Sixth Age the end of the Roman Empire should have produced a great deal of concern about the end of the world, and in my view it did.  But Gregory’s precise position on this is vague.  His most overt statement on the issue can be read in diametrically opposed ways.[63]  Nonetheless he certainly had concerns, as the Preface to Book 5 of the Histories makes very clear.[64]
By around 600, then, it is difficult to see Roman identity in Gaul as a pole of attraction.  Much of its component elements had been displaced into other areas.  The ideal behaviour associated with legitimate political authority was no longer exclusively associated with Roman education and subjectivization.  One has to recall the oppositions and differences inherent in all identities.  For Roman identity in Gaul the key opposition involved the legal privilege, tax exemption and military-political avenues for advancement associated with Frankish identity.[65]  Such issues had, I suggest, less important implications for civitas identity.  For one thing, there was no binary opposition between arvernus and Francus any more than there was between arvernus and arelatensis in the embassies mentioned by Gregory.  For another, civitas identity might coexist with Frankish in a nested way, as perhaps with Bobo and Bodegisel or the franci tornacenses of Book 10 of the Histories.[66]  For a third, in at least the southern parts of Merovingian Gaul, civitas identity was part and parcel of political and military activity.  Military service was structured differently.[67] 
If the components of identity are perpetually renegotiated but always already in existence, then the disadvantages of Roman identity in northern Gaul around 600 would be perceived as natural.  In that context it is not surprising that those who could had laid aside this level of identity and that those who could not had sunk to the level of a legally-dependent stratum of society.  It was possibly not until the category of the half-free Romanus had been absorbed within a general economically class of the dependent, perhaps by the later eighth century or perhaps earlier,[68] that Roman identity could again emerge as something to be stressed, created or fought for at high levels.
Clearly, the transformations of Romanness in late sixth-century northern Gaul were varied.  I have not mentioned the peculiarly Roman population of Trier.[69]  Nor have I mentioned the attempts by Chilperic I to incorporate traditional elements of legitimate Roman rule within the image of Frankish monarchy.[70]  Across the West that diversity would be magnified.  Seventh-century Spain for example shows some similarities with the Frankish situation and perhaps a more sustained attempt to adopt a solution similar to Chilperic’s.[71]  For Lombard Italy and Anglo-Saxon England we simply do not have the relevant data but general similarities might be suggested.

Conclusion

Roman identity had never been an immutable or monolithic identity – like any other identity it never could have been – and it is important that we early medievalists remember that.  Roman identity survived the supposed barbarian invasions of the fifth century in the West as perhaps diminished – temporarily inconvenienced – but nonetheless as an important resource in political activity.  The mid-sixth-century crisis associated with Justinian’s wars put an end to that.  However it was responded to, after that, Roman identity could not survive in anything like the old way.  In this as in so many other areas it seems correct to say that the post-Justinianic transformations that took place in the West around 600 marked the end of the Roman world.[72]


[1] See also G. Halsall, ‘Identity and otherness in the Merovingian cemetery’, in Identity and Otherness among the Barbarians in Europe during the Late Antiquity and the Early Middle Ages: The problematic relationship between Texts and Archaeology ed. J. López Quiroga, M. Kazanski & V. Ivanisevic (forthcoming).
 
[2] I had actually made them in 1995 in Early Medieval Cemeteries (Glasgow, 1995), pp.56-58, but they are more fully worked through in Barbarian Migrations and the Roman West, 376-568 (Cambridge, 2007), pp.35-45.
 
[3] P. Amory, People and Identity in Ostrogothic Italy, 489-554 (Cambridge 1997); P. Heather, ‘Gens and regnum among the Ostrogoths’ in (ed.) (2003). Regna and Gentes: The Relationship between late Antique and Early Medieval Peoples and Kingdoms in the Transformation of the Roman World ed. H.-W. Goetz, J. Jarnut, J., & W. Pohl (Leiden, 2003), pp.85-133; id. ‘‘Merely an ideology? Gothic identity in Ostrogothic Italy’, in The Ostrogoths from the Migration Period to the Sixth Century ed. S.J.B. Barnish &  F. Marazzi (Woodbridge, 2007, pp.31-60.
 
[4] The problem with the term is that it resonates with earlier uses of gentilismus in the literature.  However, it was felt that a term was necessary to specify the level of ethnicity concerned with named ‘peoples’, especially as the subtleties and many-layered nature of ‘ethnicity’ have become widely recognised.  For the problems of gentilismus, see W. Pohl, ‘Zur Bedeutung ethnischer Unterscheidungen in der frühen Karolingerzeit.’ Studien zur Sachsenforschung 12 (1999): 193-208 at pp.195-6.
 
[5] Derrida’s most accessible comments on these key issues may be found in J. Derrida & H. Ronse, Positions. (Paris, 1972) [English translation A. Bass, Positions (London, 2002).  Useful introductions to Derrida’s thought include: M. Dooley & L. Kavanagh, The philosophy of Derrida (Stocksfield, 2007); S. Glendinning, Derrida: A Very Short Introduction (Oxford, 2011); C. Howells, Derrida: Deconstruction from phenomenology to ethics (Cambridge, 1998); B. Stocker, The Routledge Guide to Derrida and Deconstruction (London 2006).
 
[6] LH IV.40. Monumenta Germaniae Historica. Scriptores Rerum Merovingicarum 1.1, ed. B. Krusch & W. Levison (Hanover, 1951); English translation: Gregory of Tours. History of the Franks trans. L. Thorpe (Harmondsworth, 1974).
 
[7] LH X.2
 
[8] Guy Halsall, Settlement and social Organisation: The Merovingian Region of Metz (Cambridge, 1995), pp.31, 39.
 
[9] I assume that Firminus, Evantius and (perhaps) Bodegisil were the people who in the Pactus Legis Salicae would be termed Romani – perhaps belonging to the evidently Roman category of conviviales regi.  Pactus Legis Salicae 41.8. Monumenta Germaniae Historica: Legum, Section 1, Vol.4, Pt 1, Pactus Legis Salicae ed. K.-A. Eckhardt (Hanover, 1962); The Laws of the Salian Franks, trans. K. Fischer Drew (Philadelphia 1991).
 
[10] These works included: W. Pohl, ‘Telling the Difference: Signs of ethnic identity’, in (1998). Strategies of Distinction.  The Construction of Ethnic Communities, 300-800 ed. W. Pohl & H. Reimitz (Leiden, 1998), pp.17-69; id. ‘Conceptions of ethnicity in early medieval studies’, in Debating the Middle Ages, ed. L.K. Little & B.H. Rosenwein  (Oxford, 1998), pp.15-24; id. ‘Zur Bedeutung ethnischer Unterscheidungen in der frühen Karolingerzeit.’ (above, n.3); id. Die Völkerwanderung: Eroberung und Integration (Stuttgart, 2002); id., ‘Invasions and ethnic identity’ in Italy in the Early Middle Ages, ed. C. La Rocca (Oxford, 2002), pp.20-33; id., ‘Aux origins d’une Europe ethnique. Transformations d’identités entre Antiquité et Moyen Âge.’ Annales HSS (2005):188-208;  id. ‘Rome and the barbarians in the fifth century.’ Antiquité Tardive 16 (2008): 93-101.  I am grateful to Walter Pohl not only for copies of most of these works but also for tolerating this close reading of his oeuvre with good humour.  I must underline that, as any reader of Derrida’s work will know, a deconstructive reading is a mark of respect, of a text’s importance and influence.
 
[11] Pohl, ‘Telling the difference’, p.22.
 
[12] On the aporia in Derrida’s thinking, see N. Royle, Jacques Derrida (London, 2003), pp.92-93.
 
[13] W. Goffart, ‘Foreigners in the Histories of Gregory of Tours.’ Florilegium 4 (1982): 80-99. Reprinted in id., Rome’s Fall and After (London, 1989), no.11.  E. James, ‘Gregory of Tours and the Franks’, in After Rome’s Fall: Narrators and sources of Early Medieval History ed. A. Callander Murray (Toronto, 1998), pp.51-66.
 
[14] James, ‘Gregory of Tours and the Franks’, p.66.  James (p.60) says that Gregory writes of the identities ‘Frank’ and ‘Arvernian’ ‘as if they were equivalent ethnic terms’ but the implications of that point are left unexplored.
 
[15] The men of Champagne: Histories 5.3, 5.14, 8.18 (implicitly), 10.3, 10.27  There is a good discussion of Gregory’s references to civitates in P. Heather, ‘State, lordship and community in the west (c. AD 400-600)’, in Cambridge Ancient History vol.14: Empire and Successors, A.D.425-600, ed. A.M. Cameron, B. Ward-Perkins & M. Whitby (Cambridge, 2000), pp.437-68, at pp.441-3, 456. C. Lewis, ‘Gallic identity and the Gallic civitas from Caesar to Gregory of Tours’, in Ethnicity and Culture in Late Antiquity, ed. S. Mitchell & G. Greatrex (London, 2000), pp.69-81.
 
[16] A.C. Murray, ‘Reinhard Wenskus on ‘Ethnogenesis’, ethnicity and the origin of the Franks’, in On Barbarian Identity: critical Approaches to Ethnicity in the Early Middle Ages, ed. A. Gillett (Turnhout, 2002), pp.39-68.
 
[17] W. Goffart, Barbarian Tides. The Migration Age and the Later Roman Empire (Philadelphia, 2006), pp.47, 198 etc.
 
[18] P. Geary, The Myth of Nations. The Medieval Origins of Europe (Princeton NJ, 2002), pp.63-73; Constructing Identities in Late Antiquity, ed. R. Miles (London, 1999); Ethnicity and Culture in Late Antiquity ed. Mitchell & Greatrex; G. Woolf, Becoming Roman. The Origins of Provincial Civilization in Gaul (Cambridge, 1998); recently, J. Conant, Staying Roman. Conquest and Identity in Africa and the Mediterranean, 439-700 (Cambridge, 2012)
 
[19] I owe the term ‘taxonomic’ to Michael Kulikowski.  As late Roman examples of the genre, one can cite Ammianus’ ethnographic excursus on the people who live beyond the Danube (Res Gestae 22.8, 31.2), the Arabs (Res Gestae 14.4), and so on. Ammianus Marcellinus, trans. Rolfe, J.C., (London 1935-39) (3 vols.).
 
[20] Germania, 1-27 (structural discussion) and 28-46 (taxonomy).  Tacitus. Germania trans. J.B. Rives (Oxford, 1999)
 
[21] See above, p.000.
 
[22] Caesar, Gallic War, 6.21-24: Caesar: The Gallic War, ed. & trans. Edwards, H.J., (London, 1917).
 
[23] Thus Ammianus, Res Gestae 15.11-12, on the Gauls, or 23.15-16, on Egypt, differ little if at all from the excursus referred to at n.18.
 
[24] For example Ammianus Marcellinus’ excursus on the Gauls (above, n.18).
 
[25] See Ps.-Aurelius Victor, Origo Gentis Romanae: Pseudo-Aurelius Victor, Les Origines du People Romain ed. & French trans.  J.-C. Richard (Paris, 1983); English translation: http://www.tertullian.org/fathers/origo_00_intro.htm (accessed 24 Nov., 2014).
 
[26] Goffart, Barbarian Tides p.1.
 
[27] E.g. PLS 14, 32, 41, 42.4,
 
[28] Lex Ribvaria 61.10-11; 61.19; 68.2-3; 69.  The necessity for a Frank to speak for Romans at law is attested in Lex Ribv. 68.3: Monumenta Germaniae Historica. Legum, Sect.1, Vol.3, Lex Ribvaria, ed. F. Beyerle & R. Buchner (Hanover, 1951); Engl. Trans. T.J. Rivers, The Laws of the Ripuarian Franks (New York, 1987)
 
[29] For an important discussion of some of these issues, see M. Kulikowski, ‘The failure of Roman arms’ in The Sack of Rome in 410 AD. The Event, its Context and its Impact. Proceedings of the Conference held at the German Archaeological Institute at Rome, 04-06 November 2010, ed. J. Lipps, C. Machado & P. von Rummel (Wiesbaden, 2013), pp.77-83.  On paideia, a good introduction is P.R.L. Brown, Power and Persuasion in Late Antiquity. Towards a Christian Empire (Madison, 1992), pp.37-41.
 
[30] See, for example, Ausonius, Carmina 108-113, where he mocks the notion of a good Briton: Monumenta Germaniae Historica. Auctores Antiquissimi, ed. C. Schenkl (Berlin 1883), p.225.  Rutilius Namatianus also suggests a view of Britons as barbarous (ferox): De Reditu Suo, line 500: Minor Latin Poets, Volume II, ed. & trans. J. Wight Duff and A.M. Duff, (London, 1934), pp753‑829.  Sidonius Apollinaris’ pride in the history of his fellow Arvernians is clear at Carmina 7, line 139ff) and his resentment at the Gauls’ exclusion from the centre of politics, to the benefit of the Italians, is visible at Carmina 5, line 349 ff.  Sidonius: Poems and Letters, ed. & trans. Anderson W.B., vol.1 (London, 1936).  The non-Gallic Ammianus famously criticises Italians with reference to the Gauls at Res Gestae 15.12.3.
 
[31] Ausonius is an example: J.F. Matthews, Western Aristocracies and Imperial Court AD 364-425 (reprint: Oxford, 1990), pp.81-82.   See also the case of Sidonius: J. Harries, Sidonius Apollinaris and the Fall of Rome (Oxford 1994), pp.27-35.  A classic statement of multi-layered Roman identity is Ausonius’ Order of Famous Cities, lines 40-41, about Ausonius’ two homelands: Bordeaux and Rome.
 
[32] In the seventh century, Fredegar identified a Frank as homo scarponnensis (i.e. from the pagus scarponensis, on the Moselle above Metz) Chron. IV.52.  The Fourth Book of the Chronicle of Fredegar with its Continuations ed. & trans. J.M. Wallace Hadrill (Oxford, 1960), p.43.
 
[33] Halsall, Barbarian Migrations, pp.39-40.
 
[34] See above, n.29 for Ausonius’ jibes.  Cassius Dio, Roman History 78.6.1a: Dio’s Roman History, trans. E. Cary (9 vols.; London, 1914-27).
 
[35] Above, nn.18, 22.
 
[36] Halsall, Barbarian Migrations, pp. 106-8.
 
[37] The classic accounts of the geo-biological reasons for the Roman:barbarian difference are at Pliny, Natural History 2.80.190; Vegetius, De Re Militari, 1.2; Vitruvius Architecture 6.1. Pliny: Natural History, ed. & trans. H. Rackham, W.J.S. Jones & D.E. Eichholz (10 vols.; London, 1938-62); Vegetius. Epitome of Military Science, trans. N.P. Milner (Liverpool, 1993). Vitruvius: On Architecture, ed. & trans. F. Granger (2 vols.; London, 1931-34).
 
[38] Halsall, Barbarian Migrations, p.109.
 
[39] Halsall, Barbarian Migrations, pp.101-10.
 
[40] Kulikowski, ‘The failure of Roman arms’.
 
[41] Halsall, Barbarian Migrations 202-6, 266-7, 281, 408-9.
 
[42] Halsall, Barbarian Migrations, pp.475-6, 485-7, with references.
 
[43] Harries, Sidonius Apollinaris, pp.243-51. R.W. Mathisen, Roman Aristocrats in Barbarian Gaul. Strategies for Survival in an Age of Transition (Austin Tx, 1993).
 
[44] K. Cooper & C. Leyser, ‘The gender of grace’, in Gendering the Middle Ages ed. P. Stafford & A.B. Mulder-Bakker A.B., (Oxford, 2001), pp.6-21
 
[45] On the continuity of Roman titles, see A.H.M. Jones, The Later Roman Empire 284-602 (Oxford, 1964), pp.238-65; P.S. Barnwell, Emperor, Prefects and Kings. The Roman West, 395-565 (London, 1982). On the ‘Romanness’ of ‘barbarian’ rulers, see Halsall, Barbarian Migrations, pp.488-94.
 
[46] E.g. PLS 41.5, 41.8.
 
[47] Gregory, Histories, 2.38 for the acclamation of Clovis as augustus, which historians have generally rejected but on no clear grounds.  M. McCormick wisely leaves the issue open: ‘Clovis at Tours, Byzantine public ritual and the origins of medieval ruler symbolism’ in Das Reich und die Barbaren, ed. E. Chrysos & A. Schwarcz  (Vienna, 1989), pp.155-180.  On Theodosius’ description as augustus, see M. McCormick, Eternal Victory. Triumphal rulership in late antiquity, Byzantium and the early medieval West (Cambridge 1986), pp.278-80.
 
[48] Such as the battle of Vouillé (507).
 
[49] B. Croke, ‘476: The manufacture of a turning point.’ Chiron 13 (1983):81-119 (repr. in B. Croke, Christian Chronicles and Byzantine History, 5th-6th Century (London, 1992), no. V).  The historiographical tradition is most recently represented in the writings of Peter Heather, notably the somewhat ironically-subtitled The Fall of Rome: A New History (London, 2005).
 
[50] Y. Hen, ‘The uses of the Bible and the perception of kingship in Merovingian Gaul.’ Early Medieval Europe 7.3 (1998):277-90.
 
[51] Halsall, Settlement and Social Organization, pp.14-17, 29.  N. Gauthier, L’Évangélisation des Pays de la Moselle. La Province Romaine de Premiere Belgique entre Antiquité et Moyen-Âge (IIIe-VIIIe siècles) (Paris, 1980).
 
[52] James, ‘Gregory of Tours and the Franks’, p.66.
 
[53] A classic study is R. Van Dam, Leadership and Community in Late Antique Gaul (Berkeley, 1985).
 
[54] Heather, ‘State, lordship and community in the west’.
 
[55] Gregory, Histories, 7.2.
 
[56] Gregory, Histories, 1.30-31.  Gregory’s oppositional attitude doubtless relates to the fact that Bourges was the metropolitan see of Aquitanica Prima, the province in which Gregory’s home town of Clermont was located.
 
[57] Gregory, Glory of the Confessors 29. Monumenta Germaniae Historica. Scriptores Rerum Merovingicarum 1, Part 2 ed. B. Krusch &W. Levison (Hannover, 1969), pp. 744–820; R. Van Dam (trans.), Gregory of Tours: Glory of the Confessors (Liverpool, 1988).
 
[58] Gregory, Glory of the Confessors 79.
 
[59] Gregory, Glory of the Confessors 80.
 
[60] Gregory, Histories 10.25.
 
[61] Lewis, ‘Gallic identity and the Gallic civitas from Caesar to Gregory of Tours’.  See also M. Handley, ‘Inscribing time and identity in the kingdom of Burgundy’, in Ethnicity and Culture in Late Antiquity, ed. Mitchell & Greatrex, pp.83-102; Halsall, Barbarian Migrations, pp.480-2.
 
[62] Cp. Gregory, Histories, 1.18 claims that Julius Caesar was the first emperor and mentions the foundation of Lyon but there is no overt statement about the conquest of Gaul.
 
[63] Gregory, Histories, 1.praef. G. de Nie, Views from a Many-Windowed Tower. Studies of Imagination in the Work of Gregory of Tours (Amsterdam, 1987), pp.57-59.
 
[64] Gregory, Histories V.praef. ‘The preface to Book V of Gregory of Tours’ Histories: Its form, context and significance.’ English Historical Review 122 (April, 2007), pp.297-317.
 
[65] Halsall, Settlement and Social Organization, pp.26-31, 258; G. Halsall Warfare and Society in the Barbarian West 450-900 (London, 2003), pp.46-47.
 
[66] Gregory, Histories X.27
 
[67] Halsall, Warfare and Society in the Barbarian West, c.450-900 (London, 2003), p.48.
 
[68] Halsall, Settlement and Social Organisation, p.59
 
[69] Halsall, Cemeteries and Society in Merovingian Gaul. Selected Studies in History and Archaeology, 1992-2009 (Leiden, 2010), pp.225-229, 258.  See also Jamie Kreimer’s paper in this volume.
 
[70] Histories 5.17, 5.44, 6.2, 6.46
 
[71] On Isidore of Seville, see J. Wood, The Politics of Identity in Visigothic Spain. Religion and Power in the Histories of Isidore of Seville (Leiden, 2012).
 
[72] This is the title of a project I am currently working on, for which I received generous support in the form of a Leverhulme Major Research Fellowship in 2009-12.