Fourth-century ‘barbarisation’ and fifth-century foederati
The so-called barbarisation of the Roman army has sometimes been argued to have been one reason for the Empire’s fall. The traditional view (one that goes right back to Zosimus, a pagan historian writing in the early sixth century) is that the Empire recruited ever more barbarians, who in some ways acted as a sort of ‘fifth column’ for the ‘barbarian invasions’ and who, by saturating the army with their ‘barbarian’ ideas, drained it of its old military discipline and quality.
The late Empire recruited heavily among the non-Roman peoples beyond its various frontiers; it always had done. Whether it did so any more in the late Roman period is difficult to decide. The army grew but the Empire’s population, it seems, did not (indeed it might have been falling). This possibly increased the need to recruit barbarians. More importantly, the late Roman separation of military and civil branches of imperial service meant that Romans had to choose one or the other. This, inevitably, left more room for non-Romans to rise to higher ranks, giving an impression of a more barbarian-dominated army. One detailed study of the make-up of the imperial guard units (traditionally thought to have been the most ‘barbarised’ regiments) suggests, however, that between half and three quarters continued to be Roman-born. As I said in Part 2, the army continued to be a career path open to and used by Roman aristocrats and others.
Other aspects of the ‘barbarisation’ thesis have been strongly criticised. Back in the 1960s, A.H.M. Jones, undoubtedly correctly, said that there was no evidence that barbarian soldiers were anything less than good, loyal and effective servants of the Empire. The argument that they were suspicious tends to be based upon the nineteenth-century idea that there was some sort of unifying ‘Germanic’ identity and culture among the barbarians east of the Rhine. Alamannic soldiers did not see themselves as having anything in common with Franks or Goths (other than, perhaps, shared service in the Roman army!). The only instance I can think of, of treachery, involved a Roman attack on a soldier’s own tribal homeland. Connection with the Empire was a huge source of pride and prestige among barbarians. In the cremation cemeteries of the Saxon homelands one frequently finds the buckles and brooches used by the Roman army. When laying an old Saxon to rest there was no better way of showing how important he had been than by cremating him in his old Roman uniform. That these people were secretly working from the inside to bring down the Empire simply beggars belief.
The idea that the army was saturated with barbarian customs and other influences has been attacked from various directions. Michel Kazanski, a Russian archaeologist working in France, has demonstrated that Roman military equipment shows very little influence from Germania Magna. The theory that the foulkon formation of early Byzantine times is a ‘Germanic’ influence has been thoroughly demolished by Philip Rance. Other ‘barbarian’ customs, when subjected to critical scrutiny, turn out to be attested far more among the Romans than among the barbarians.
The army was, however, definitely trying to look or act barbarian. When the civil and military services were separated, much of the old, traditional Roman ideology and identity went with the civil service, leaving the army to develop new identities. It seems that, in differentiating itself from the Roman civil bureaucracy the army focused on things that were fierce and non-Roman. The Romans dwelt a lot on how fierce the barbarians were. If we compare depictions of late Roman troops with the pictures of Dacians on Trajan’s Column many similarities become apparent, not least the army’s use of trousers, the barbarian item of clothing par excellence. Late Roman troops seemingly let their hair grow and wore torques, again very ‘barbarian’. The emperors were very concerned about the popularity of this ‘barbarian
chic’ and passed laws banning Romans (including senators) from wearing trousers or military cloaks when in Rome. Clearly Roman officers enjoyed swaggering about in their outlandish costumes. The people who gilded officers’ armour were called barbaricarii. The fashion for things barbarian grew. Not surprisingly, whereas barbarians in the early fourth-century army tend to have adopted Roman names, by the end of the century they openly retained their non-Roman names.
Another support for the barbarisation thesis is the number of units with non-Roman ethnic names in the list of officials, the Notitia Dignitatum, which includes regiments called Sarmatian, Herul, Frank and so on. Some, many perhaps, of these units owed their original name to being recruited from the people in question. Whether they continued to be made up of such people is less likely. A unit of ‘Heruls’ stationed in Egypt for two or three generations would by then largely be made up of Egyptians. Sometimes this can be illustrated by inscriptions. Other reasons exist for not taking these units at face value. Roman ethnography played on the military characteristics of particular barbarians, from whom Rome employed specialist troops, right back to the Balearic slingers of Hannibal’s day. In some cases an ethnic name is clearly shorthand for function. The Parthian horse archers in fourth-century Italy were hardly likely actually to be Parthians, but horse archers were naturally associated with Parthians. The Moors, Illyrians and Dalmatians were also types of light horse. That units of Mauri (Moors) were composed of Romans can actually be shown from the inscription evidence.
More interestingly one must consider the nature of such late Roman ‘ethnic’ units. They are overwhelmingly (especially in the West) found among the Palatine Auxiliaries. Now, in these units we find a very common penchant for non-Roman ethnic names. It is all very well to concentrate on those that were current for the fourth century but they also include ‘Medes’, ‘Cimbrians’, ‘Celts’, ‘Batavi’, and ‘Sabines’. No one is going to argue that the fourth-century (AD) Roman army was recruiting Sabines… These units also include those with animal names – Cornuti (‘horned ones’ – probably goats or bullocks), Leones (Lions) and those with what I call boasting names – Petulantes (Vicious), Feroces (Fierce), Felices (Lucky), Invicti (Undefeated) or Victores. These are all rather different from early Roman unit names, not least in the shift from naming after the unit itself (usually a feminine singular: e.g. Valeria Victrix) to naming after the men in it (thus masculine plural: e.g. Invicti). 80% of western units with ‘barbarian’ ethnic names, and over 70% of western units with ‘warlike’ (animal; boasting) components to their titles, are found amongst the palatine auxiliary regiments. Looked at in the round, there was a vogue for naming units to stress ferocity or martial prowess. The troops are fierce, barbarian, animal even: all things (as indeed with the fact of boasting itself) directly antithetical to traditional Roman ideas of what distinguished a Roman male from a savage.
So, what am I arguing? The late Roman army certainly contained lots of non-Romans in its ranks – let’s make that clear. Crucially, though, instead of these troops becoming Roman (as had hitherto always been the case), the army became ‘barbarian’. As the army had always included lots of non-Romans this must be more about a shift in ideas than about simple barbarian recruitment. The Roman army’s barbarisation was, in my view, much more about the Roman army actively adopting what it thought were barbarian characteristics, to accord with the new ideas pervading the army, about ferocity and non-Roman-ness, than about the passive absorption of influences brought by barbarian recruits. The best analogy, in my opinion, is provided by the nineteenth-century Zouaves. Originally recruited among native Algerians, and named after the ‘zaouia’ or religious brotherhoods, they were soon made up entirely of Frenchmen, and their uniform, rather than being genuinely North African, was a French idea of local costume. Yet these troops, often consciously adopting North African, Berber and Islamic aspects of life, drew a terrific esprit de corps from their ‘non-French’ identity and distinctive ‘ethnic’ dress. This was played off competitively against the more ‘traditional’ French units (guards, chasseurs). When the Zouave troop-type was transferred to North America during the American Civil War the point is only underlined.
This has important implications for the nature of fifth-century armies. Romans were, by then, well used to serving in units that bore barbarian names, alongside men who were actual barbarians, and to strutting about in a uniform which was supposed to look barbarian. This brings me to the question of the foederati. There are two types of foederati. Technically a foederatus was someone bound by a treaty (foedus) with the Empire. Thus traditional foederati were provided by barbarian allies, serving in their own contingents for the duration of a particular campaign. This is often felt to have been blurred by the fact that after 382 federate contingents are thought to have been provided by bodies of non-Romans settled as quasi-autonomous units within the Empire, principally the Goths. This is viewed as a momentous change, as indeed it would have been; an emperor did not enter into treaties with his own subjects. However, this relies on the idea that the Goths signed a foedus in 382, giving them a semi-independent territory in Thrace, in return for troops whenever required, but troops who served under their own commanders. The problem is that no contemporary or near-contemporary writer actually mentions such a treaty; they all say unequivocally that the Goths surrendered to the Empire (became, in technical terms, dediticii). The earliest source to mention the treaty is Jordanes’ Getica (‘On the Origin and Deeds of the Getae [Goths]’), written 170 years after the supposed event. This is an unreliable source at the best of times and it has its own clear reasons for wanting to talk about treaties. Late fourth-century sources have been read of the light of Jordanes’ dubious statement. When looked at for what they actually say, it is clear that late fourth- and fifth-century writers are referring to a new type of foederati.
These foederati were regular units of a novel but nevertheless recognisable type. One source refers to a Gothic officer commanding an ala (cavalry regiment) of foederati. Nevertheless the key thing about foederati was that initially they were recruited from barbarians. Zosimus consistently distinguishes foederati from stratiotai (Roman soldiers), which, given that the stratiotai must include the Roman regiments with barbarian names, must further underline the point that fourth-century ‘barbarian’ units were often, perhaps usually, nothing of the sort. After the Adrianopolis debacle the Romans made much use of these foederati, recruiting very large numbers of Goths into the army. Many came from the Goths defeated in 382, but many more were surely recruited from beyond the Danube. There are Gothic officers all over the place in the decades after 382. Whole armies could be formed of such troops and referred to, consequently, as ‘the Goths’. This fact could present serious problems for foederati as, when things went wrong, they could have all the usual Roman stereotypes and prejudice about barbarians trotted out about them, and some people did not like the idea of such barbarian units (Synesius of Cyrene wrote a whole diatribe about the situation). This, again, insinuates that there was something new about them, differentiating them from earlier ‘barbarian’ soldiers. This could have drastic results: the Gothic officer Fravitta, having just quashed a rebellion of another Goth, Gaïnas, in 401, was murdered soon afterwards. One suspects that his barbarian status made this easier. Nevertheless that and the (briefly) anti-barbarian stance of the Constantinople regime seem to have helped produce the first rebellion (late 401) of the most famous foederatus of all, Alaric the Goth. For reasons that this series will underline, this situation kept repeating itself in the fifth century, and the Roman army became the focus for new identities. The Goths recruited into the units commanded by Alaric probably came from all sorts of groups that hitherto probably did not see themselves as having much in common. Shared service, especially when Alaric’s actions pitted them against ‘legitimate’ forces, helped foster a sense of identity as Goths. However, whatever civilian politicians and writers thought, Roman commanders rarely worried about their foederati.
It is important, though, to remember the earlier discussion of the fourth century. Romans had become used to serving in units with barbarian names, barbarian fellow-soldiers, and barbarised costumes. Although foederati were, at first, new and different, Romans were already accustomed to seeing the army as ‘barbarian’ (‘soldier’ and ‘barbarian’ seem to be synonyms in some documents). The foederati regiments rapidly went the same way as the earlier forms of barbarian regiment (such as some of those listed among the Auxilia Palatina). In about 425, a fascinating historian called Olympiodorus of Thebes said that already, in Honorius’ reign (392-423), the word foederati had come to mean units of men of diverse origin (in the context of the passage that must mean that they included Romans). A century later, Procopius says that while foederatus had, once upon a time, meant barbarians serving the Empire on a basis of equality (the traditional meaning of the term), it had come to be applied to anyone in a unit of foederati. A couple of generations afterwards, in Maurice’s Strategikon, it is clear that the foederati are just élite cavalry units holding the middle of the front line. The fourth-century background allows us to see how ‘Goths’ did not have much difficulty in recruiting local Romans or people from other ethnic groups.
There was change, which continued as the fifth-century progressed, not least in constantly shifting the status and nature of the western foederati in the opposite direction from that taken by their eastern equivalents – away from just being Roman regulars. My point is that it should be seen as a process of evolution rather than a sudden and drastic shift. It began with something looking more like a jolt than a revolution, a jolt stemming from the specific situation after Adrianopolis. Nevertheless, the evolution is firmly rooted in the fourth-century situation. What this means will, I hope, become clearer in the course of the scenarios. For now, let me just reiterate that if we think of fifth-century barbarian foederati as wild barbarian irregulars barely controllable and plotting to create their own kingdoms within the Empire
we will be well wide of the mark.
Traditional views of the fifth century see the Roman Empireas brought down by invading barbarians. There is no doubt that political power in western Europe passed by 476 from Roman Emperors to kings who commanded forces that at least claimed to be of non-Roman origin. To see these as invading or migrating ‘peoples’ is, however, at best an over-simplification and at worst (and more commonly) a gross distortion of the evidence. People certainly moved from barbaricum into the Empire in the fifth century. They did in every century of Roman imperial history. More may well have moved in the fifth century, although we cannot be sure and it is not clear how many more in any case, but what is much more important is not the migration but the facts, firstly, that the mechanisms behind them remained more or less the same as they had since Julius Caesar’s day and, secondly, that they became, unlike in any previous period, foci for important political developments.
Although there is some evidence for large-scale migration in the last century of the Western Empire’s existence, two things are clear: such migrations were always of short duration and they never went very far. The traditional image of the ‘Barbarian Migrations’ with its maps of sinuous, long, windy arrows linking ‘Goths’ in Scandinavia with ‘Goths’ in Spain or ‘Vandals’ in Poland with ‘Vandals’ in Africa can only be produced by joining together quite different movements made in response to specific circumstances and often separated from each other by long periods of time, often decades, and, in the earliest stages, movements that probably never even happened! During the gaps between the movements the nature and composition of the groups in question usually changed profoundly. The narrative of each of these stories (of the Goths, of the Vandals or of whichever other group) is completely disrupted and subverted by these intermissions. The barbarians that moved into the Empire in the fifth century were usually élite military groups and their dependants. Those who originated within the Empire, as foederati, were somewhat different, initially at least, but both types rapidly became rather similar. Their similarity lay in the fact that all became intimately involved, as factions, in Roman politics (as discussed above). This series aims to demonstrate that the history of the fifth century is not about a steady narrative of a shrinking Roman Empire failing to stave off waves of invading barbarians. Instead it is about a swirling, bewildering, series of faction fights, which pitted ‘barbarian’ against ‘barbarian’ and ‘Roman’ against ‘Roman’ and, usually, one coalition of ‘Romans’ and ‘barbarians’ against another. These faction fights were not about defending or conquering the Roman Empire; they were about controlling the Roman Empire. As I have already stated twice, the problem was that no one won and that fact meant that the Empire sprang apart into various regional units, ruled by the ‘barbarian’ leaders of the military wings of these factions (often springing from the ranks of the foederati), but administered by the provincial Roman aristocrats who had usually been their partners in these factions throughout.
Finally, we need to remove the tag of ‘Germanic’ or (worse) ‘German’ often applied to the barbarians. This is enormously misleading and stems from modern politics. In the sixteenth century, Germans, cut off from the Roman culture revived in the Renaissance, wanted an alternative antiquity but the tag mainly derives from the politics of German reunification, when centuries of often bitter rivalry between Saxons and Bavarians or Prussians and Hessians were to be smoothed over by showing that, historically, all Germans were the same, and had been since Tacitus’ day. The barbarians east of the Rhine all spoke a Germanic language; that does not make them fundamentally unified any more than Italians, Spaniards, Portuguese and Frenchmen can be treated as interchangeable or sharing a unified culture because they speak a Romance language. There is no evidence before the eighth century that members of any of these so-called ‘Germanic’ peoples saw themselves as having anything in common with any of the others. Their archaeological remains show considerable regional cultural differences and we can only completely obscure history if we assume that one ‘Germanic warrior’ was interchangeable with another, or assume (without other data) that evidence about, say, the Alamanni can be applied glibly to the Saxons or the Franks. Perhaps my most detested wargames myth is that of ‘the Germanic warrior’, which describes Germanic-speaking barbarians interchangeably as ‘early Germans’. There were no ‘Germans’ before the ninth century, when that tag (Theotisch – Deutsch – ‘the people’) was created in the fragmentation of the Carolingian Empire to differentiate the Germanic-speaking East Franks from the Romance-speaking West Franks. Even German history professors now argue forcibly for the abandonment of the term ‘Germanic’ except when applied to linguistics. Ironically, Germans can distinguish die Deutsche (Germans) from die Germanen (ancient Germani), but English has no such linguistic resource.
Having completed our historical, ground-clearing (thanks for bearing with me), we can turn to the armies and battles of the fifth century.
 E.g. by A. Ferrill, The Fall of the Roman Empire: The Military
1986): not a subtle, sophisticated or especially well informed analysis.
 H. Elton, Warfare in Roman Europe, 350-425 (Oxford, 1996), pp.151-2.
 M. Kazanski in ‘L’Équipement et le
matériel militaires au Bas-Empire en Gaule du nord et de l’est’, Revue du Nord 77 (1995), pp.37-54.
(accessed 1 August 2007). I am grateful to Nik Gaukroger for this
 M. Speidel, ‘The rise of ethnic units
in the Roman imperial army’, Aufstieg und
Niedergang des Römischen Welt 2.3 (1975), pp.202-31, gives some examples.
 I discuss all of these ideas about the
army’s barbarisation at greater length in chapter 3 of my book, Barbarian Migrations and the Roman West,
Migrations and the Roman West, ch.6.
 John Lambshead, in the Fall of the West WAB
supplement, p.11, gets this about right, except that the foederati
don’t seem to have been commanded by ‘Roman’ officers.
 Olympiodorus, fragment 7.4. The
Fragmentary Classicizing Historians of the Later Roman Empire, ed. &
trans. R.C. Blockley (Liverpool 1981), vol.2,
 Procopius, Wars 3.3.11. Procopius, ed. & trans. H.B. Dewing,
vols.1-5 (London, 1914-28).
Strategikon: Handbook of Byzantine Military Strategy trans. G.T. Denis (Philadelphia, 1984).
 This is (though over-simplified) the
basic idea of Peter Heather, the prominent British historian of the
‘migrations’. Heather is a fine
historian and his books are essential reading: Goths and Romans, 332-489 (Oxford,
1991); The Goths (Oxford, 1996); The Fall of Rome: A New
2005). The first of these is the best;
the others repeat the argument with (in my view) less subtlety but more
‘coverage’. I agree with very much of
Heather’s work but not with his basic thesis about barbarian migration bringing
down the Empire. My critique can be
found in ‘Movers and Shakers: The Barbarians
and the Fall of Rome.’ Early Medieval
Europe 8.1 (1999), pp.131-45, reprinted in From Roman Provinces
to Barbarian Kingdoms, ed. T.F.X. Noble (London, 2006), ch.9.