Ethnic Transformations: A Rider

Although I had bumbled my own way to the conclusions reached in the previous article about the context for Æthelberht's Laws, Patrick Wormald had already reached them (or ones very similar) 20 years ago.  See The Making of English Law round about p.100.  This will go into a footnote, as will an acknowledgement to Dr Levi Roach for putting me onto it, though I really should have known to look there to start off with if I was any good... What I will make of this point, however, is likely to be rather different from what Wormald did.

The Ethnic Transformations of the Year 600 (A preliminary sketch): Part 2.

[Here is the second instalment of my thinking about ethnic change around 600.  As you will see, it is largely consists of somewhat rambling 'thinking aloud', which tends to be the way I formulate arguments and generate texts in their initial form, but any thoughts are as always welcome.]

One of the most important elements of these changes is the seemingly general demise of prestigious secular ‘Roman’ identities.  The ‘Roman’ dux Chramnelen listed by Fredegar points us at one potential exception, to which we shall return.  This must be one of the more dramatic consequences of Justinian’s wars and accompanying ideology.  As has been noted, in Lex Salica the romani represent a parallel free population to that of the franci within which three different strata can be noted: the tributarii, (tax-payers), the possessores (land-owners) and the members of the convivia regis (‘the kings dining companions’).  These strata have a lower wergild than, but otherwise seem broadly equivalent to, their Frankish counterparts: the francus (for the possessores) and the antrustio or member of the trustis regis (the royal bodyguard, for the members of the convivia regis).[1]  The names of the different strata make clear the functional division between the Franks and Romans, with the former occupying a military role and the latter paying taxes and handling the civil administration.  Outside a few specific cases relating to offences between people of different ethnic identities, however, the general term ingenuus seems to encompass freemen of any ethnicity.[2]  More than that, as in a number of other immediately post-imperial law-codes, ethnic identity seems primarily to apply to mature adult males.[3]  This appears to imply that a fully ethnic legal identity was achieved, presumably when a male married and established a household, with legal dependents. 

The picture given by Ripuarian Law, as described earlier, is quite different.  In the territory within which the law applied, romani were half-free and required a Ripuarian to speak for them at law.  The Ripuarians, moreover, include men, women and children.  Ethnicity now appears to be something ascribed at birth.  The romani who can be tried by their own law, mentioned above, are those – presumably aristocrats – from other areas (mainly, one assumes, Aquitaine or Provence).  There were many reasons why one might adopt a Frankish identity in sixth-century northern Gaul: legal privilege, exemption from some forms of taxation, the greater prestige of military service (enabling one to attend the gatherings of the army, the most significant political assembly of the realm), and so on.  It is therefore not difficult to envisage a gradual drift towards Frankish identity over the period.  The shift seems to be rather more significant than that model would allow, however, accelerating dramatically in the late sixth century.  There had, after all, been cultural and other bases of Roman identity that could be deployed to counter the de facto political and military power of the Franks.  Mid-sixth-century Merovingian kings had continued to bestow patronage upon Roman aristocrats and the Church had remained an area wherein some form of Roman identity was important.  Most sixth-century churchmen, even in the north, have Roman names.[4]  It seems most plausible to see the ideological shifts associated with Justinian’s campaigns of ‘reconquest’ as finally cutting away the cultural or social benefits of Roman identity in much of the West.  As a result, as we have seen, ‘Romans’ in north-east Gaul effectively ceased to be a part of the free population. Another indication of the general collapse of a Roman identity can be seen in name-giving practices.  In the sixth century most churchmen had, or took, Roman names.  The bishop list of Metz is not untypical; after the accession of Agiulf sometime around 590-600 only three other bishops had Roman names during the Merovingian period.  Fredegar’s Roman dux Chramnelen had taken a Frankish name although the adoption of non-Roman names by Gallo-Romans entering the service of the Merovingian kings had begun much earlier.[5]

What, however, of those ‘Romans’ in parts of the Frankish realm that had never experienced significant settlement by people claiming a non-Roman identity, such as Aquitaine?  This area had always been a somewhat unusual part of the regnum.  The land-owners in the region, it would seem, were expected to perform military service as well as pay taxes from their estates.[6]  There was clearly some social cachet to their identity which could be played off against ‘Franks’ administering the region for the kings.[7]  Gregory was not above giving a certain pejorative, ‘barbarian’ sense to the label ‘Frank’ on occasion.  It is interesting, then, that Germanic names seem to have become almost as ubiquitous here as north of the Loire by the seventh century.  It is interesting too that in the Life of Saint Eligius the epithet romanus thrown at the saint by northern countryfolk is clearly derogatory - as indeed it would have been in the north by that date, if we can judge from Lex Ribvaria.  In this context then perhaps it is significant that was from the seventh century that a ‘Gascon’ (Basque) identity began to be important in southern Gaul, associated with a military élite.  Romanness was no longer adequate.

The clause of Ripuarian Law alluded to earlier, about the personality of the law, seems to be indicative of a general trend from around 600.  The prologues to the Pactus of Alamannic Law and the Bavarian Law refer to the creation of laws for these peoples in the early seventh century.  The Alamannic pactus claims to have been issued by a king Chlothar, generally seen as Chlothar II, whereas the preface to Lex Baiwariorum tells of how king Dagobert perfected earlier laws drawn up by earlier kings (Theuderic, Childebert and Chlothar are named[8]) and gave them to each people.[9]  Bavarian Law quotes the words of Isidore of Seville describing the people who gave the laws to the different peoples of history, down to the Theodosian Code, before saying that thereafter ‘each people chose a law for itself from its customs’ and quoting Isidore again on the distinctions between lex, mos, and consuetudo.  It seems very likely, then, that the association of peoples with their own law was a development of the period around 600.

In this connection it is interesting to reconsider some clauses of Chlothar II’s Edict of Paris, issued in October 614.  Clause 12 of the Edict requires that no iudex be appointed from outside the region in which he was to exercise his functions.  The judge would have to have property in the region from which he could compensate claimants in the case of any wrong-doing.  Seeing such an enactment in terms of a concern for (in modern terms) ‘accountable’ local government seems reasonable enough.[10]  The clause makes yet more sense, however, in a context where the different regions of the realm had their own law.  If, for example, a Burgundian was sent into Austrasia as a judge then, by the very terms of Ripuarian Law, he would be entitled to be tried by Burgundian Law rather than the law he was supposed to be administering and this would indeed cause problems in gaining restitution for any offences, especially if he was required to recompense any plaintiffs with property that lay in a different law’s area of jurisdiction.

The Pactus Legis Alemannorum, which claims to have been issued by King Chlothar (presumably Chlothar II, as noted), is an interesting text.  Unlike the near-contemporary Lex Ribvaria it makes no reference to internal ethnic divisions.  The term Alamannus itself appears but rarely; the concern of the law is much more with stratification within the free population.  Of course, the bulk of the area within which the Pactus applied lay outside the late Empire’s boundaries, so the absence of romani is a problem that probably does not arise.  We can at least say that not trace remained in the law of the Roman inhabitants of the agri decumates.  Whether the Pactus applied in Alsace (formerly a part of Germania Prima) of is moot: perhaps unlikely but not impossible.  The promulgation of the Pactus has been linked by some scholars to the presence of the king at Marlenheim near Strasbourg and the relationship between Alsace and the Austrasian court was troubled in the first decades of the seventh century but it is difficult to draw firm conclusions.  The implication nevertheless is that the seventh-century population of Alamannia shared a single ethnicity.

With this in mind, then, it might not be surprising either that the first Anglo-Saxon law-code, that of Æthelberht of Kent, also belongs to the reign of Chlothar II.  The acquisition of a bishop with the pallium, direct from Rome may have been a means of cementing Æthelberht’s new dominance in English politics and an independence from the powerful rulers across the Channel.  The issuing of a set of laws for his people might be another element of such a strategy, or it could be seen in the context of broader developments within the area of Frankish hegemony.  All the ‘peoples’ of the Frankish regnum should have their own law.  However, ethnic labels of any sort are strikingly absent from Æthelberht’s code (as they are from all of the seventh-century Kentish codes), even in the description of the kingdom.[11]  This fact is difficult to interpret.  Given Bede’s famous statement about the origins of the various kingdoms of the English, with Kent being founded by Jutes, the absence of any reference to Jutes in Kentish documents is immediately striking.  Was ‘Jute’ an outside appellation?  It has been suggested plausibly enough that the Eucii who appear alongside the Saxones in Theudebert I’s letter to Justinian could be the ‘Jutes’ of Kent.[12]  Was it rather that the ethnonym disappeared from use in southern England by the early seventh century?  By the time our records appear the Kentish kingdom and its inhabitants seem exclusively to be associated with the Roman civitas of the Cantii.  If so, it is intriguing to speculate on the sort of process that might have brought this about.  Perhaps, for example, only one civitas of a once-larger sixth-century ‘Jutish’ realm remained?  This might explain why the other ‘Jutish’ region in Bede’s view was the Isle of Wight and the coastline opposite, a quite separate region.  Ethnic labels of a higher, ‘gentile’[13] level might have been lost in the process.  Either way, in Æthelberht’s Code the distinctions among the free population are entirely based upon status within a legal hierarchy that makes no reference to ethnic differences.  In this – and in other aspects – the law interestingly resembles its close contemporary, the Pactus Legis Alemannorum.  Unlike the area within which the pactus applied, however, the absence of a Roman population cannot be so easily bracketed. Were we to assume that, instead of simply talking about freemen, the text spoke of ‘freemen of the Cantwara’, the code would show some similarity to Ripuarian Law, which likewise uses a ‘regional’ ethnic signifier for its subjects.[14]  It could be that the local Roman population had – again as in Lex Ribvaria – either adopted the general identity of the free population or sunk into the semi-free classes and so required no separate legislation.  The only indication that this might be the case is the appearance of the term laet, from the Latin laetus, to denote an evidently half-free category.[15]  Quite how or why laeti (originally barbarians captured in war and settled inside the Empire) would come to refer to people of provincial Roman origin is difficult to imagine, however.

---

[1] Notably PLS 42.

[2] Halsall Settlement and Social Organisation, pp.27-29, for more detailed discussion.

[3] Halsall, Barbarian Migrations, pp.

[4] A letter from Gogo, the nutritor of Childebert II, to Bishop Petrus of Metz (c.580) provides an interesting snapshot of local ecclesiastical personnel, listing seven office-holders among the latter’s circle.  Four have Germanic names; three have Roman.  This might indicate the shift towards Germanic names already under way.  One may be a civic official, however.  Theodemund is described as civium praesidium, which is especially interesting as the description is a literal translation of his name.  However one reads it, this suggests some knowledge of Germanic language in the Moselle valley.  Either Gogo was punning on Theodemund’s name, or Theodemund took a name that described his position. Ep.Aust. 22.

[5] Gregory of Tours’ maternal great-uncle, Gundulf, who had been one of the domestici at the Austrasian court, is perhaps the best-known example.  The penetration of the fashion for non-Roman names more generally into southern Gaulish society is visible in Gregory’s works during the 570s-90s.

[6] Halsall, Warfare and Society in the Barbarian West 450-900 (London, 2003), pp.

[7] Seen classically in Gregory of Tours’ dealings with Leudast, the count of Tours.  Leudast, said Gregory, was a low-born Poitevin but had risen in the service of the Frankish kings and had been given, or had adopted, a Frankish name.

[8] This statement was one of the admittedly slender bases for Eckhardt’s identification of different recensions of Salic Law associated with these rulers in the manuscripts of the PLS.

[9] Pactus Legis Alemannorum preface; Lex Baiwariorum preface.

[10] See James, The Origins of France (London 1982), p.140-1.

[11] Æthelberht’s eventual successors and fellow law-makers, Hlothhere and Eadric, and Wihtræd are all styled kings of the Cantwara.  Æthelberht is simply ‘Æthelberht the king’.

[12] I.N. Wood, The Merovingian North Sea (Alingsas)

[13] I have used this admittedly problematic term to denote an ethnic identity which relates to a ‘people’, like the Franks, the Alamans or the Bavarians.

[14] As I have argued elsewhere (Barbarian Migrations), and intimated above, it is analytically mistaken to assume that what might look like regionally- or geographically-based signifiers are less ‘ethnic’ than those that derive from the names of ‘peoples’ – not least because the former can become the latter, as in the case of the Cantwara.

[15] This is one of the few points of similarity between Æthelberht’s Code and the Pactus Legis Salicae, and the latter’s half-free class of liti.  That the laets might have been Romano-British has been suggested by several authorities, such as Whitelock.

The Ethnic Transformations of the Year 600 (A preliminary sketch): Part 1

["Not talkin' 'bout the roots in the name/ talkin' 'bout the roots in the man', sang Lamont Dozier (and, later and more famously, Odyssey). When I started working on the 'Transformations of the Year 600', I thought that there would be little or nothing to say about ethnicity, unlike in its predecessor, Barbarian Migrations and the Roman West, where it is an important element of the analysis.  This was because I accepted the view, which seems general, that that was, if not 'the' period of ethnogenesis, certainly a period when ethnic identity was very important in political transformation.  Now I am wondering whether this was in fact the case.  Thinking about the nature of post-imperial polities either side of 600, and the processes of subjectification, made me reconsider this.  Here are my current, very preliminary thoughts on this issue.  These, clearly, deal only with the Frankish kingdom.  Whether things might be different in the Visigothic, Lombard or Anglo-Saxon realms remains to be seen.]

***

‘It wearies me to record the diverse civil wars that beset the people and kingdom of the Franks [francorum gentem et regnum]...’[1]

One of the biggest areas of research in early medieval history in recent decades has been the study of ethnicity and its role in the disintegration of the Western Roman Empire.  Indeed Walter Goffart, an outspoken opponent of the ‘ethnogenesis’ approach, begins one book with a statement that the Roman Empire ran into a ‘wave of ethnicity’ in Late Antiquity.[2]  The often vitriolic debate has overwhelmingly concerned the period between the late fourth and the sixth centuries, that of the ‘fall of the Roman Empire’; the period covered by the present book has featured very much less prominently except, perhaps, in discussions of the Lombards and Avars.  In this chapter, however, I will suggest that in fact it was precisely in the decades around 600 that ethnic identity really came to the fore as a politically significant factor. 

I have offered a fairly full discussion of the nature of ethnicity before[3] but it will be helpful to offer a résumé of the conclusions reached there. There is no single feature that can be used to define an ethnicity, other than the belief in one’s membership of a group and in the distinctiveness of that group.  Ethnicity is subjective, multi-layered (one can have more than one level of identity that functions in an ‘ethnic’ way), performative (ethnic identity is not immanent but is activated by performance), situational (it, or certain levels of an actor’s ethnicity, is performed in certain situations) and dynamic (what an ethnic identity means, or the effects achieved in its performance, change through time). The principal means by which I would modify this view today would be in which the identity being performed is seen.  In the 2007 book there is insufficient theorisation of what is meant by an ‘identity’, although this flaw is very common across the innumerable studies of identity in early medieval history.[4]  It seems too strongly implied that the identity, however mutable, is some sort of self-present entity or at least a set of unproblematically, mutually-understood roles and images.  The point to be stressed though is that any identity is a mental image of an ideal, comprised not only of what the social actor sees as its advantages and disadvantages but also what they perceive to be others’ expectations of the correct performance of the identity.  That mental image will of course constantly be changing as a result of the myriad ongoing performances of social interactions between people of particular identities.  The mental ideal image associated with the identity can, furthermore, never be attained but is always striven towards.  There is also no guarantee that that image, as perceived by one actor, will correspond exactly to that imagined by other parties in the interaction.  That gives the decision to perform a particular identity in a social interaction some of the characteristics of a wager rather than something of which the outcome can be entirely predicted in terms of the relative prestige or privilege associated with particular identities.

In the fifth and early sixth centuries it can be argued that ethnic identity was particularly fluid and multi-layered.  This should not be assumed to mean that one could pick and choose identities at will.  The ethnic shifts that can be observed are best explained in terms of the acquisition, and reordering through time, of new identities within the ethnic field.  Thus a provincial Roman (and one must remember that ‘Roman’ identity was itself multi-layered, with regional and various civitas identities beneath the general ’Roman’ heading) might acquire a new, ‘barbarian’ ethnicity, such as Goth or Frank, related to various activities or functions within the realm.  Over time that identity might come to be stressed more than the others, until it was his primary means of identification.  There can be no doubt that politics in this period could be seen as the interaction of ethnic identities. 

That said, however, ethnicities at that time appear very often to have been functional.  Famously, for example, barbarian ethnic identities were associated with the military, across the west, while the ‘Roman’ identity was linked more heavily to educational culture, to civic administrative roles, to the Church, and so on.  The implication of this is that ethnicity remained ‘nested’ within and authorised by a particular conception of the Roman or post-imperial state.  What shaped the social role of, for example, Gothic identity was the perception of the military function of the Goths.  This can be seen very clearly in the writings of Cassiodorus and the ‘civilitas’ ideology to which he gave voice.  Goths defended the realm, allowing Romans to maintain their traditional civilised way of life, which included serving in the administrative offices of the kingdom.  The two identities were largely defined by their interrelationship within that structure and any power associated with them was ultimately legitimated by a link to the king.  In many regards, this simply continued the late Roman distinction between the civil and military arms of imperial service.  None of this denies that the relationship between such ethnic identities was dynamic and that the relative standing of each was subject to change.  This was one factor that led to particular social actors adopting other identities.

Andrew Gillett has argued that ethnicity was not politicised in the immediately post-imperial period.[5]  This immediately, of course, begs the question of what is meant by ‘politicised’.  Gillett appears to take the political to mean the level of political units and their leaders, which is too restrictive an interpretation.  Clearly, within any kingdom the interplay of ethnic identities was political.  Nonetheless, at that high level of politics, his argument is forceful.  Terms that associated royal titles or realms with people, like those that appear in the epigram from Gregory of Tours with which this chapter opened, do seem to be rare before the late sixth century.  The preceding discussion perhaps suggests why this was.  The realm itself was not associated with a single dominant ethnicity or ethnic group and the ruler was the ruler of all groups within a territory.  As we have just seen, it was the overarching structure of the territory or state and its government and ruler that legitimised, and created a frame for, the performance of ethnic identities. 

The place of ethnic identity in society and politics may, however, have changed in the later sixth century.  Two documentary references might give us a way into examining this.  The first comes in Fredegar’s Chronicle (IV.78), describing a punitive campaign launched in 636/7 by the Franks against the Basques.  King Dagobert sent an army led by ten dukes: ‘Arinbert, Amalgar, Leudebert, Wandalmar, Wandelbert, Ermen, Barontus, Chairard ex genere francorum, Chramnelen ex genere romano [sic], Willibad the patrician genere burgundionum, Aighyna genere saxsonum [sic]’ and many other counts. 

The second comes from Ripuarian Law.  Lex Ribvaria’s date is difficult to establish with any precision; it lacks an identifying prologue and the manuscripts, as is so often the case, are later than its presumed date of issue.  Nonetheless, the communis opinio, based on cross-reference from the prologues to the late Alamannic and Bavarian codes, seems to be that it was issued in the reign of Dagobert.  Whether the promulgation of the law would date to the beginning of his sole rule (629) or, as I have proposed before, associated with the council of Clichy (626) and the defusing of tensions between Dagobert and his father, is uncertain.  We might nevertheless accept an early seventh-century date.  The law might then belong to more or less the same time as the Basque campaign described by Fredegar.  Clause 40 of the law sets out the penalties to be paid if a Ripuarian Frank kills another freeman of varying ethnic identities: Frank, Burgundian, Roman, Alaman, Frisian, Bavarian and Saxon are listed.  More significantly for the development and historiography of early medieval law, clause 35 says that someone of Ripuarian, Frankish, Burgundian, Alamannic ‘or whatever other’ origin can be tried by the law of the place (or of the people into which) they were born.  This is the first reference to the much discussed ‘personality of the law’ in the Merovingian world.

The two texts appear to suggest that different ethnic identities were important within the ‘kingdom of the Franks’.  The analysis, however, may require us to look at the two levels of kingship within the Merovingian world: the regional Teilreiche and the broader regnum francorum made up of all of the different kingdoms and occasionally unified under a single ruler (such as Chlothar I at the very start of our period or his grandson Chlothar II between 613 and 621). If we start with the lower of these two levels, represented by Ripuarian Law, it is difficult to imagine any ordinary freeman making effective a claim to a non-Ripuarian Frankish identity and thus to be tried by the law of his own people in a typical Frankish rural mallus or court.  The title guaranteeing the right to be tried by one’s own law seems more plausibly to be aimed at the higher ranks of society.   Within the Austrasian Teilreich, then, an aristocrat could effectively employ an ethnic identity from outside the kingdom.

This is underlined by Fredegar’s account, which shows that when the aristocracy of the different Teilreiche came together within the aristocracy in the politics of the broader Frankish kingdom, their different ethnicities were noted.  The context of Fredegar’s description, moreover, makes it clear that aristocrats of all of these different identities (including Roman) were involved in leading the army.[6]  Simple administrative function within the realm’s governmental structures, then, does not explain ethnic difference, as had been the case in the sixth century.  Fredegar’s Chronicle regularly notes which of the various Frankish Teilreiche a particular nobleman came from.  Fredegar particularly mentions those from Burgundy, evidently his own region, for whom he occasionally uses the term Burgundaefarones (loosely: the clans of the Burgundians).  These features are generally absent from the writings of Gregory of Tours, who commonly identifies the actors in his account by the civitas of origin or simply qualifies them as a Frank.  While the Frankish label is clearly ethnic, it can also function, for Gregory, as a description of the highest political levels of the realm. On the other hand, an identity focused on a city-district is more than a geographical denomination;[7] it is a form, or level, of ethnicity.  What seems different in Fredegar’s vocabulary is the increasing stress laid upon broader identities within the ‘kingdom of the Franks’.

However, while this does seem, as mentioned, to be a largely aristocratic phenomenon on the texts, that may be because of the greater ability of the more powerful strata of society to move within the realm and come into contact with each other.  Returning to the level of the regional Teilreiche it may be that the free population came to share the same regional identity.  Ripuarian Law uses the term francus to mean a Frank from outside the region.  Sometimes it employs romanus to mean an outsider as well.[8]  The standard member of the free population (man or woman) is called a Ripuarian.  It may be significant that this regional label is used, rather than a more obviously ‘ethnic’ term or qualifier like ‘Frank’ (or Ripuarian Frank).  That would fit with the use of identities based upon regional kingdom used in Fredegar’s Chronicle.  When discussing society from within the territory of the Ripuarians, however, the ‘Romans’ are no longer an element of the free population, divided into grades roughly equivalent to those of the Franks, as in Pactus Legis Salicae.  Instead, they have slipped down the scale to being a semi-free social category, needing Ripuarians to speak for them at law.  One exception to this rule may have lain in the region of Trier, where some of the local population clung on to a Roman identity.  According to his biographer, Saint Gaugeric (later Bishop of Cambrai) was born in the late sixth century into a family of ‘Romans’ of the middling sort.  In the seventh century, another holy man from the same civitas, Saint Germanus of Grandval, was born into a family of ‘senators’.  It may, however, be suggestive that Germanus’ familial identity was based upon senatorial rank rather than a Roman ethnicity.  Perhaps even here the cachet of Roman identity had been seriously weakened.  An increasing stress on rank rather than ethnicity would tie in with other developments.

The ethnic structure of the Frankish kingdoms seems, therefore, subtly but importantly different in the seventh century.  Regional, ‘kingdom’ ethnicities appear to be important and, within those regions, the free population appears to share that identity.  Furthermore, whereas ethnic identity in the sixth-century Merovingian realm had largely related to function within the realm and could thus be brought together in the kingdom itself, there is no such mechanism that similarly ties Neustrians to Austrasians, to Burgundians within the overarching structure of the regnum francorum.  These look more like regional factions or interest-groups within the ‘kingdom of the Franks’, a feature that seems to be underlined by repeated Austrasian demands to have a king located within their region.  Even the other Teilreich, however, remained divided into Burgundians and Neustrians.  In its broader sense, the seventh-century kingdom seems, in a very real sense, to be significantly more ‘polyethnic’ than its sixth-century predecessor.

---

[1] Gregory of Tours, LH 5.Preface.

[2] W. Goffart, Barbarian Tides

[3] Halsall, Barbarian Migrations, pp.35-45.

[4] Most significantly and ironically, there is, as far as I can see, no sufficient examination or theorisation of what an identity is in the volume Texts and Identities in the Early Middle Ages, ed. R. Corradini, R. Meens, C. Pösel & P. Shaw (Vienna, 2006).

[5] A. Gillett, ‘Was ethnicity politicised?’

[6] Note that the names do not give a very clear indication of origin.  The sole ‘Roman’, Chramnelen, has a Germanic name whereas one of the ‘Franks’, Barontus, has a Roman name.

[7] In the same way that today, if one says that someone ‘comes from’ Jamaica or Germany that is more than a simple account of their geographical origins.  Halsall, Barbarian Migrations, p.39.

[8] For a similar use, to mean someone from the south, see Life of Eligius.