The Genesis of the Frankish Aristocracy (Part 4 of 4)

The earliest Merovingian aristocrats: the ‘Flonheim-Gültlingen Horizon’
[Part 3 of this article is here]

The period between the final demise of the Western Empire and the end of the first quarter of the sixth century has long been considered to form a fairly discrete chronological unit, in its first phase largely defined by a series of prestigious graves. From two of the best-known, this archaeological horizon is often known as the ‘Flonheim-Gültlingen Horizon’ or, from a specific form of metalwork and ornamentation, defined by the ‘polychrome’ style. When chronologies of objects were developed through typology and patterns of association, the last quarter of the fifth century and the first quarter of the sixth became known as Böhner’s Stufe I. The coherence of this phase has been confirmed by subsequent studies of other regions, refined through the use of computerised techniques.

These burials have long been the subject of discussion. The traditional reading is that they represent the leudes of Clovis, settled and granted lands in conquered territory. The dates of the burials thus represent the stages of the Frankish conquest of northern Gaul, the latter seen very much in line with the ‘moving front’ model. There is no reason to dispose of this interpretation, provided that one is clear about how one understands the power relationships involved, both between the king and the aristocrats and between the aristocrats and the remainder of the free population.

The burials of the first half of Böhner’s Stufe I are – classically – those of adult males, lavishly furnished with grave-goods: weaponry above all, but also with belt-sets (like the sword hilts decorated in particular styles), ceramics and bronze and glass vessels. Furnished inhumation (especially of male graves) had been less common in the middle quarters of the fifth century (a decline which I have associated with an uncertainty over how one displayed legitimate status in the de facto absence of the Empire but with no clear successors to that power either). With the burials of the last quarter of the fifth century, it burst back into fashion. Many of the graves under discussion are, furthermore, ‘founder graves’. Some may have been accompanied by above-ground monuments. Burial no.319 at Lavoye (département of Meuse) possibly had a barrow built above it; Childeric I’s famous grave in Tournai almost certainly did. Indeed, the relationship between burials like Lavoye 319 and Childeric’s grave has lain at the heart of the problem. Classically, the aristocratic graves, tombes de chef or Adelsgräber are seen to be copying the royal ‘mode funéraire’ established with Clovis’ interment of his father. Thus, runs the argument, when Clovis arranged for his own burial in a church (Holy Apostles, later Ste Geneviève) in Paris in 511, the Frankish aristocracy imitated this too and were interred under their own newly-founded Eigenkirche.

There are important problems with this interpretation. One is that of relative chronology. It is by no means clear that Childeric’s burial predates its ‘imitators’ – if we leave aside the universally-cited but entirely unreliable date given for Childeric’s death by Gregory of Tours, the Tournai burial has exactly the same absolute chronological indicators as Lavoye 319, for example: 474x491. Thus, when he buried his father, Clovis could have been developing a form of burial that was already being employed by local aristocrats to demonstrate their standing, but taking that demonstration to an extreme degree – as one might expect. The relative chronology of these burials and their use in establishing a political history of northern Gaul is also blurred, first, by a reliance upon Gregory of Tours’ dubious chronology of Clovis’ reign and, second, by circular arguments. Burials south of the Somme, for example, must date to after 480 because that was the date at which the Franks (to whom these burials are attributed) acquired this region; the Franks only occupied the lands south of the Somme after 480 because Frankish burials in those lands are all later than that date. When one adds the facts that the historical record provides no support for the vision of a steady, north-to-south advance of the Franks (as discussed above) and that there is no good a priori reason to assign these burials necessarily to invading Franks in any case, it becomes apparent that the traditional reading is no more than an elaborate construct. Whether Clovis started a new fashion by burying his father in this way or whether (as seems to me to be rather more likely) he employed an existing fashion in an unusually elaborate fashion to make a particularly important statement at a moment of dynastic crisis, is impossible to establish from the evidence as we have it.

Apart from issues of chronology, the ‘emulative’ nature of the late fifth- and sixth-century ‘tombes de chef’ is further questioned by the fact that such burials continue through the sixth century and even beyond, whereas Childeric’s grave is, as far as we know, unique amongst Frankish royal burials (with neither predecessors nor successors), especially if one looks more closely at cemeterial context. To found a new cemetery or a new part of a cemetery one might expect unusually lavish displays of grave-goods; that such displays were not exactly replicated in succeeding generations is not really surprising. The absence of such graves does not imply that the local aristocrats had taken to interring their dead under churches. The building of churches to serve as aristocratic mausolea does not, in any case, really take off until the seventh century whether in town or country. In the sixth century, if aristocratic burials took place under churches, they did so in the civitas-capitals and in the castra and vici, on which the pagi were centred, in other words in the administrative nodes of the kingdom.

Associated with this suggestion is the fact that, at least In Picardie, the ‘tombes de chef’ associated with the ‘Frankish conquest’ are actually away from the civitas-capitals and other centres of the region (figure 2 [This is just a net of Thiessen Polygons based on the towns of the area and laid over the distribution of burials, in the best traditions of the New Archaeology of the 1970s]). Taken together, these points shed an important light upon the nature of these early ‘aristocratic burials’. The rite of burial with grave-goods is essentially a transient display of status (or a claim to status) that requires an audience present either at the grave-side or along the route of the funerary procession. Burials in the countryside, like Lavoye 319 (in Lorraine rather than Picardie but likewise located some way from a civitas-capital) would thus represent demonstrations of status to other members of a rural community. Interestingly Françoise Vallet showed many years ago that these late fifth-century graves are, while often on different sites, found in roughly the same (thus peripheral) parts of the countryside as the late fourth- and early fifth-century lavish burials.

There are some key differences nevertheless. The changes of site or the foundation of new areas within cemeteries is one; the different forms of material culture are another. Rather than being imported from barbaricum, ultimately from the Danubian Hunnic realm, it now seems most likely that the costume represented by the grave-goods of the Flonheim-Gültlingen burials was of a more widespread later fifth-century type, using artefacts made in the Mediterranean and associated with the ‘barbarian’ late Roman armies. The ‘barbarian’ import of the costume is thus of a rather more subtle and specific form than was once thought. In a northern Gallic context, especially when one remembers the intimate connections between the Merovingians and the Loire army (demonstrated clearly enough in Childeric’s burial), such a ‘Roman-barbarian military’ identity would be associated with the Franks. Nonetheless, the material’s association with a new political order seems clear, as is further underlined by the fact that this material tends to appear in burials only from c.475, and thus the precise moment that the western Empire was dissolved, de facto if not de jure. Another difference between the graves of ‘Stufe I’ and those of the late fourth-/early fifth-century ‘horizon’ is that the former are overwhelmingly the burials of adult males.

If we take all of these features together a significant change is suggested. The rite itself makes clear that local instability continued. However, a rupture with the preceding situation is evident. The bases of local pre-eminence had changed; claiming legitimate imperial Roman status in the old way was now ruled out. Local authority was associated with the Franks and their military forces (once again demonstrating the Merovingians’ political advance from the south rather than the north). This might have involved changes in the local ‘pecking order’; new families might (as in the traditional interpretation) have been brought in and settled on lands by Childeric and Clovis. Alternatively, a local family might have sided with the Franks, adopted their identity and been given power in the community. Such a family could well have been that which had held sway since the fourth century or before, or it could have been an opportunistic rival. In each case, a dependence upon the Franks is clear enough. So too are the risks involved in the gamble. Unlike twentieth- and twenty-first-century historians, the inhabitants of the late fifth century did not know that this move towards the Franks would turn out to be the ‘right’ one. Furnished burials of Stufe I are not especially numerous, especially in its first half (the Flonheim-Gültlingen horizon proper); only after about 500 are women and children found in any numbers. At Lavoye itself only a dozen or so burials belong to the half-century between 475 and 525, compared with about 200 from the succeeding fifty to seventy-five years – although the group is characterised by well-furnished burials with clearly-marked gender differences. Even the probable family of the ‘chieftain’ of grave 319 do not appear to have associated themselves with him in death until the second quarter of the fifth century. Evidently it took time to convince the people of northern Gaul of what Hegel might have called the Reason of History! In such a situation the dependence of the subjects of these burials upon Frankish backing is underlined. The developments of the remainder of the sixth century (not the subject of this paper) only serve to emphasise this.

Closer to the urban or sub-urban nodes of the region, it may be that the families who supported the Franks demonstrated this link to their rivals and the potential distributors of patronage in burials in the civitas-capitals, the castra and the vici themselves, perhaps under any churches that existed at that time. Some of the burials of the period, grave 1760 at Krefeld-Gellep for instance, are found close by such nodal points. Again, though, a link to the new powers is clear. In defence of the old view, however, it must be repeated that the bulk of locally-important families simply (in terms of their archaeologically-visible traces) sat on the fence for much of this period. That said, the dynamics of the situation are clear. The economic stagnation of the period is amply demonstrated. The craft-specialisation visible in the polychrome objects of Stufe I soon disappears and the further decline of Argonne-Ware production by c.540 has been mentioned. Rural settlements and any urban occupation remain notable chiefly for their archaeological invisibility. Acquiring the resources to establish local leadership on a more solid footing, let alone breaking out of very local political arenas, was fast becoming a matter of association with the Frankish kingdom as it became ever more powerfully established in Gaul. This dynamic is even visible in the Triererland, where (in contrast to the so-called Föderatengräber horizon, c.375-c.450) some (albeit not many) Stufe I burials are known and where, as in the rest of northern Gaul, furnished burial had become common across rural communities by c.525, even if the epigraphic habit of the ‘senators’ continued in the civitas-capital itself.

When the Merovingians established their rule in Gaul north of the Loire, then, it can be fairly solidly established on the basis of the evidence we have that, even by the last quarter of the fifth century, they held the ‘whip hand’ in any dealings with the local aristocracy. Outside Trier, no independently powerful aristocracy existed (whether in socio-economic or cultural terms). There was no need for widespread purges, exiles or expropriations across the region. What made the core of the Merovingian kingdom so resilient was the fact that, by 500 at the latest, local leaders came to the kings for the basis of their authority. The creation of an independently wealthy, powerful Frankish nobility would, in the context I have outlined, have been a more difficult task than the acceptance of the existing situation. The class of landed magnates that we know so much about by the late seventh century and which was indeed (or at least could be), by the late eighth century at the latest, unusually wealthy in pan-European terms (richer indeed than many a king in the British Isles), was a much more recent creation.

The Genesis of the Frankish Aristocracy (Part 2 of 4

Part 1 of this piece can be found here

Transformations around 400


Thus far, the evidence points overwhelmingly to the facts that the northern Gallic social élite was, outside the Triererland, not especially wealthy and that, along with the region’s economy and most of its social structures, it was intimately connected to the imperial state focused upon Trier. Considering the main issue with which this article is concerned, we can conclude that, even were the Merovingians handed control of northern Gaul via treaty in a smooth transfer of political authority, they would not have inherited a powerful, independently wealthy regional aristocracy.

The preceding discussion renders almost predictable the effects on the region produced by political changes after c.380. In 381 Emperor Gratian moved the imperial court from northern Gaul to Italy and Milan. A series of changes is then visible across the region. In most areas villas enter a final phase of desertion, being abandoned by the second quarter of the fifth century. There were exceptions of course. In more southerly areas, around Paris, there is much better evidence of villa survival until rather later in the century, something that seems also be true in the Triererland. The picture, as before, is rarely a straightforward one of economic decline. In the south of the modern Netherlands, around 400 there is a late phase of construction on some rural sites, such as at Gennep. However, these are not villas of the old type. Even with these caveats, however, the impression cannot be avoided that the very late fourth and earlier fifth centuries constituted a period of profound change in the northern Gallic rural settlement pattern and economy. The development is again incompatible with an explanation in terms of an alleged ‘de-Romanisation’ (as is made clear by the continuing – indeed the increased – usage of Roman symbols in burials) or simply through a new military culture. The fifth-century end of the villas makes the latter explanation unlikely. The fourth-century aristocracy of the region had already, as we have seen, been very largely militarised and even those not involved in the army directly were linked to it economically. We cannot be sure that fifth-century aristocrats were markedly more militarised than most of their predecessors. Indeed, sixth-century Frankish law suggests that civic, Roman aristocrats were still a feature of the area’s social structures. If the fifth-century insecurity led to more fighting and greater (and more violent) competition for local leadership, then this affected the survival of villas not through a shift towards a more military state of mind, allegedly eschewing elaborate building, but through the need to spend surplus on local alliances and the equipment of a retinue, leaving little for the upkeep of stone buildings. Aristocrats had to choose where to spend their limited resources and the times ultimately demanded that they choose politics and security over architectural embellishment. Ultimately, the final demise of the northern Gallic villa is an economic issue, not one of a shift in mentalities – even if the latter can reasonably be postulated.

This impression is underlined by study of the region’s towns, which underwent further dramatic contraction and in one or two cases died out completely. There is little trace of occupation on the intermediate settlements, the vici and castra. This is partly related to the problems of dating very late Roman occupation. The two principal supports for such chronologies are coins and finewares and both are problematic after c.400. The latest developments of Argonne Ware pottery, to which we shall shortly return, were not recognised as such until about 1990, which probably means that traces of fifth-century occupation had earlier been wrongly assigned to the fourth century instead. Additionally, the region’s coin supply dried up early in the fifth century after the closure of the Trier mint. While undoubtedly making the identification of late Roman levels very difficult, these changes are themselves significant. The end of coinage and the failure of local powers to mint replacements, after the end of a series of silver imitation solidi in the middle quarters of the century, imply a significant reduction in the scale and complexity of the economy.

Truly monetized commerce requires a neutral medium of exchange and a guarantee of a coin’s value, accepted by both parties to a transaction. The government of a state or polity has the power to provide such a guarantee, moreover one which can be accepted across large distances. With the crisis of the imperial state in northern Gaul around 400, such guarantees disappeared and the areas over which objects were traded shrank accordingly. In whose name the silver imitation solidi, already mentioned, were struck remains mysterious but these coins nevertheless enabled some monetary transactions to take place across a reasonable distance in the middle quarter of the century. Their face value was nevertheless fairly high and the absence of small change is a crucial index of a downturn in the extent of the economy’s monetization. Such coins possibly served other purposes than the strictly commercial, as was the case with the gold solidi. When these silver coins, which are not numerous in any case, ceased to be struck, coinage in the region was limited to imported Eastern Roman solidi until the Frankish rulers began to strike solidi themselves in the sixth century. The function of this type of high-value coinage (1/72 lb. of gold) might very well have been more political than economic. Small denomination coinage remained absent until the seventh century.

Other coins were available nonetheless. The frequency with which Roman coins are found in the pouches buried with sixth-century Merovingian males suggests that such coins continued to serve as handy units of bullion. Their use was more limited than that of a properly minted and guaranteed currency. It has long been known that, in the sixth century, scales or balances are known in northern Gallic burials. Frequently found in lavishly-furnished graves, their symbolism seems to refer to a role in vouchsafing ‘weights and measures’ and this might (though there are other interpretations) have been related to determining the correct quantity of precious metal in old coins. Although these data come from a later period than that which under consideration, they seem suggestive of mechanisms that could have existed as the late imperial monetary economy collapsed. If we combine this evidence with the conclusions just reached about the relative power of the local aristocracy, it is clear that the word of such a local leader would not be recognised by both parties to a transaction over wide areas: another feature in restricting the distances over which commercial exchanges might be made. Some evidence, to which we will return, suggests that the standing of northern Gallic aristocratic families might have been somewhat more secure in the early fifth century than it was a hundred years later and this could have extended the zones over which their word was held to be good, but the general point will surely stand. With the collapse of monetary exchange, the only other mechanism for long distance movement of goods was that associated with the imperial economy but, in the context we have outlined, this too was fading fast.

It is here that Wickham’s attention to the ceramic data is important. His account is as follows:

In northern Gaul around 400 by far the commonest fine ware in the sigillata tradition was Argonne ware … often quite elaborately decorated with a roller wheel … with a 400km radius of distribution from the Rhine to well south of the Loire … [I]t continued into the late sixth century; it reached less than 200km by now … but survived a century into the Merovingian period as a production on a substantial scale.

If this conclusion can be reached from this evidence, then there must – clearly – be something wrong with the model I have sketched. Something about the other evidence, whether of the rural and urban settlement sites or of the burials, to which I will shortly return, must conceal a crucial element in the equation or else the way we read such data is fundamentally mistaken. The picture of imperial crisis and collapse in the region, after c.380, that I have drawn from the written sources must also be wide of the mark. Wickham has (as we have seen) ways of explaining the exiguous settlement evidence in terms of a shift in aristocratic culture to a more military model, which would fit with the idea of the region’s militarisation. This latter proposal is not entirely satisfactory for reasons that have been discussed, but the main point is that Wickham presents a coherent, rounded argument.

Whether intended this way or not, a fair and straightforward reading of the passage quoted is as follows: this pottery was distributed over an area in excess of 500,000 km2 and continued to be produced ‘on a substantial scale’ through the fifth century to the end of the sixth century, even if the area over which it was distributed had shrunk by half by then. The image presented by such a reading is, however, misleading. If we examine Didier Bayard’s study of this form of ceramics, a rather different picture emerges. We find (figure 1) that in his early fifth-century Phase 2, almost all Argonne ware is found within a 300km-radius of the kilns (6 sites yielding such pottery beyond that radius compared with 66 within it) and within in a ‘box’ 500km (east-west) by 300km (north-south). That is an impressive area of 150,000 km2, but still considerably less than that implied by Wickham’s statement. More to the point, by the time of Bayard’s Phase 3 (roughly 440s-460s) this had contracted further. All of the finds he catalogued from the middle decades of the fifth century lay within a 300-km radius and most (67-79%) of them within 200km. Most lie in a box covering 120,000 km2, less than half the area calculated on the basis of the 300km radius of distribution. Thus, this contraction, which a straightforward reading of Wickham’s account implies was something that happened slowly over the fifth and sixth centuries, actually happened quite suddenly around the middle of the fifth, with the abandonment of the Rhine forts. By the time of the political end of the western Empire in the late fifth century (Bayard’s Phase 4) the distribution of Argonne ware had contracted so that 98% of it was found within a 200km radius – in fact within a 200kmx200km box (a considerably smaller surface area) – though fairly evenly distributed within that zone. Argonne ware does continue into the late sixth century but it is important to clarify that the last decorated phase dies out around 540 and that thereafter only standard undecorated forms were produced. So, rather than being distributed across somewhere between 125,000 and over half a million square kilometres during the late fifth and sixth centuries, the impression easily gained from Wickham’s statement, this pottery was in fact only traded across 40,000km2 during this period. In comparative terms, nevertheless, that might represent a widespread distribution of material, but how it relates to other post-imperial ceramics needs to be reassessed. It is now, for example, suggested that some post-imperial wares made in Leicestershire were distributed over an area ranging from the Channel coast to Yorkshire, a not dissimilar reach. It is also important to note the end of decoration in the early sixth century and the restriction in the range of forms, both of which features underline an economic change not unfairly characterised as decline.

The last phases of occupation on the Rhine forts are shadowy and a sophisticated reinterpretation, pondering whether they were still bases for regular troops or, moving away slightly from the usual narrative, centres for local warlords, is overdue. Either way, it seems clear that after the middle of the fifth century whoever did control these forts was no longer in a position to be able to guarantee a market for the products of the Argonne kilns on anything like the old scale. Overall, the link between the collapse of the state and severe economic contraction could not be clearer.

The archaeological cemetery evidence fits this picture of crisis. From about the time that Gratian moved the court back to Italy, the number of lavishly furnished burials in northern Gaul increases steadily. In these burials, men are interred with weapons and, more frequently the belt-sets and brooches that were the insignia of imperial office. In some cases they were accompanied by burials of women and children, the former buried with a wide range of new jewellery forms, notably brooches. The latter and the desire to fit this change in the record into the old narrative of barbarian conquest led to the assignment of these graves to incoming ‘Germanic’ settlers. A closer examination of the archaeological data (the rite itself and the artefacts deposited), freed from these assumptions, combines with the lack of any documentary historical support for the notion to compel a more subtle reading. This sees the subjects of these burials, as yet comparatively few in number and found in small clusters, often on larger cemeteries, as representing locally powerful families whose status was called into question by the death of a member. Given what has been said about the bases of the northern Gallic aristocracy’s power, so closely related to the presence and legitimation of the Roman state it should be no surprise that their local standing should have been jeopardised by the removal of effective, regular governmental presence. It should equally be unsurprising that the Moselle valley, where the wealthiest nobles seem to have been concentrated, is largely free from such burials at this time. The choice of items, and their symbolism, also makes sense in the context described. In the absence of effective imperial presence, the bases of a family’s legitimate authority were proclaimed, especially when an adult male member died, questioning the inheritance of such authority. In this situation, legitimate power was proclaimed by the use of badges that made a link with imperial power. Otherwise they stressed traditional Roman aristocratic virtues and pastimes, such as hunting. The women’s costume, one imagines, made a comment about their status as a chaste wife, a good mother, and so on. As imperial presence grew ever more distant, the use of the badges of office waned accordingly, although other symbolism persisted. Nonetheless, examination of the ritual in comparative perspective suggests that, as yet, the power of these families was not decisively threatened. In a slightly later period, the distribution of furnished burials was far more widespread across communities, and the choice (and number) of goods related to the life-cycle and gender. Rather than being concentrated in the burials of a particular kindred, but spread across subjects of both sexes and all ages, grave-goods were focussed upon mature adult males and younger women.

On the eve of incorporation into the Merovingian kingdom, the northern Gallic aristocracy was even less wealthy than it had been before and its status within local communities was more under threat as the effective legitimacy of a claimed link with the Empire faded. Although this does not imply that many aristocratic families had necessarily lost their local pre-eminence, it seems to be the case that the social, political and economic arenas within which they lived had shrunk considerably. It is against this backdrop that the famous passages of Salvian’s De Gubernatio Dei should be understood. Long taken, doubtless wrongly, as the paradigm for late Roman western aristocracy, Salvian’s comments must be placed in a very specific chronological and geographical context. The assumption that his tirades against the corrupt aristocracy of his times were aimed at the magnates of the Trier region whence he hailed (and whence he had fled, not least as a result of the actions of these rapacious individuals) is not certain but is a reasonable working hypothesis. We have already seen that the Triererland was an exceptional region of northern Gaul. Archaeological data make it clear that we should have no reason at all to generalise from the aristocrats of the lower Moselle valley. What has perhaps been less fully discussed is the precise moment that Salvian was describing. Writing in the 440s, it is reasonable to suppose that his account of the tyrannical curiales belongs to the 430s or perhaps slightly earlier; the issue turns on how recent one supposes that Salvian’s arrival in the south was at the time of his writing. If the picture he painted does belong to the 430s then it is quite instructive when viewed against the archaeological evidence.

In our current state of knowledge, this decade would lie towards the end of the period of occupation of the Triererland’s villas. The sharp decline in the distribution of Argonne ware and the end of occupation of the Rhine forts in the 440s have also been mentioned, and the politically-active generation of the region would largely have been children (at most) when even a usurper emperor last ruled at Trier. The area was fast approaching a severe, critical point and it is unsurprising that it had become a political hot-house. There was no imperial presence to regulate those who claimed to wield power in its name and none of the usual rotation of offices that was part of the efficient management of patronage. Thus those who could continued to cling onto their ‘legitimate’ power, and in the critical situation of the second quarter of the fifth century they exploited it to the maximum. Without the opportunity to share this power, their opponents could only adopt the strategies mentioned by Salvian: either to wield local authority without formal imperial legitimation, that is to say to become rebels or bagaudae (in the eyes of the imperial government or of those who claimed to act in its name) or to turn to the barbarians for support. The three responses to crisis described by Salvian (claiming legitimate power; claiming power without allegedly imperial legitimation; and turning to the barbarians for support) were, in general, the options available to the political classes throughout the fifth century west, but around Trier they took on a particular form and intensity. The fourth option, the one taken by Salvian, was to flee to areas where the Empire’s writ still ran, and he does not seem to have been the only one to have chosen this course of action. In the Triererland of the 430s-440s, this must have seemed an attractive choice, especially as (unlike us) contemporaries did not know that the Empire would not return. Indeed their knowledge of history doubtless suggested that, eventually, inevitably, it would. For these reasons, the decisions to join the barbarians or to follow the ‘bagaudic’ course – those that seem to modern observers to be the ‘far-sighted’ or ‘realistic’ options – must have been taken by contemporaries very much in extremis. On their periodic forays back into northern Gaul (fizzling out in the 440s – as we know but contemporaries did not), the representatives of the Empire dealt equally harshly with bagaudae and barbarians. As well as creating these risks, turning one’s back on the traditional bases of political power brought all sorts of other identities into question, not least one’s masculinity. That the depredations of those who claimed a legitimate imperial basis for their power should have driven their rivals to take these actions is a graphic indication of how critical the situation on the lower Moselle had become.

Part 3 here.

The Genesis of the Frankish Aristocracy (Part 1 of 4)

[Here - broken down into four parts for - comparative - ease of reading, is a piece I have been working on of late.  It is still very much in first draft form and thus very woolly, un-foot-noted and unchecked.  When it is foot-noted, checked and edited down, I will send it off to a journal and take it down from the blog, so if you wish to comment or otherwise have any input - which will duly be acknowledged - now is the time. 

Essentially the piece confronts a key element in the debate over whether or not the Merovingian aristocracy was an independently wealthy magnate class or - essentially - a service aristocracy dependent upon the kings.  It deals with the assumption, which would be very important, if not fatal, for the second interpretation (which has generally been my own), that there were established Gallo-Roman and Frankish aristocracies/magnate classes existing in northern Gaul when the Merovingians took over.  Using a full range of evidence and paying close attention to regional diversity, the paper demonstrates (I hope!), first, that outside the Triererland the late Roman aristocracy was not an independently wealthy nobility of the type known elsewhere, second (Part 2) that the crisis of c.400 reduced the Gallo-Roman aristocracy's wealth and power further, third (Part 3) that, via a close study of Frankish migration and its mechanisms, we can see that the Frankish leaders who settled in Gaul in the fifth century were already closely dependent upon the Merovingians by c.500 at least, and fourth (Part 4) that a consideration of the archaeological traces of the northern Gallic aristocracy between 475 and 525 underline the point established in the rest of the article: that neither the Franks nor the Gallo-Romans presented the Merovingians with a powerful landed aristocracy with which to contend.  The mighty landed aristocrats of the seventh, eighth and especially ninth centuries were members of a class that was a later creation (of c.600).]

The nature of the Frankish aristocracy has featured heavily in the historiography of Merovingian Gaul. One of the main problems has been whether or not the northern Gaulish aristocracy, which we might term Frankish (as opposed to the more Gallo-Roman magnates of Aquitaine and Provence or the hybrid Gallo-Roman/Burgundian/Frankish élite of Burgundy), was formed of families whose wealth, local standing and power existed independently of the patronage of the Merovingian royal dynasty. This debate has never been resolved to everyone’s satisfaction. Perhaps, one hopes, it never will be (at least in its entirety) but a new contribution is nevertheless required. Chris Wickham’s monumental Framing of the Middle Ages recently presented a strong case for the wealth and independence, indeed for what he suggested was, in comparative perspective, the quite unusual wealth and independence, of the northern Gallic Frankish aristocracy throughout the early medieval era.

Wickham’s contribution is unusual in that it made well-informed use of archaeological as well as documentary sources and, perhaps more importantly, excavated material other than that normally employed for this kind of enquiry. Most previous studies have drawn mainly upon the evidence of excavated cemeteries for an archaeological insight into local social structures, giving at best a partial image. Wickham however employed ceramic data, which now exists in sufficient quantity and quality for reasonable observations to be made – something that was not true even twenty years ago. This would itself be reason enough for a re-examination of the economic bases of the Frankish aristocracy.

Wickham’s case is solid and well-argued and, for the second half the period covered by his survey (thus c.600-c.800), his conclusions seem entirely valid. Moreover, they are in harmony with what has, perhaps, always been the most common interpretation and with other recent scholarly work which has suggested a more direct continuity between the Gallo-Roman nobility of the region and the land-owning magnates of the Carolingian world. This paper argues against this trend. In a companion essay I deal with the nature of the sixth-century northern Gallic aristocracy and its transformation around the turn of the sixth and seventh centuries. Here, I address a fundamental basis of the ‘established aristocracy’ view. That is the nature of the social élite that existed in northern Gaul at the time of the establishment of the Merovingian kingdom. This paper questions the validity of Wickham’s (and others’) assumption that the Merovingian kings had to confront already-established Gallo-Roman and Frankish aristocracies in the creation of their realm.

Like Wickham’s discussion, it will use the whole range of data available to us, documentary, archaeological, epigraphic and numismatic. As well as chronological change, geographical diversity will be noted. One problem with the period covered by this paper, the late fourth and fifth centuries, is the general absence of written evidence. This means that the overwhelming bulk of the evidence used will be archaeological. Nevertheless, a great advantage of this form of data, and of the cemetery material in particular, is that, as well as being increasingly voluminous, it is securely anchored in time and space, allowing us to explore change through time, something which, as noted, has not always been recognised in previous studies, even Wickham’s. It will also allow us to shed a critical light upon the handful of well-worn fragments of written evidence which do seem to treat with the social structures of the region, most notably Salvian’s On the Governance of God.

Late Roman Background

The obvious preliminary to any discussion of the Merovingian aristocracy in northern Gaul is to examine the regional élite during the late Roman Empire. If a magnate stratum can be shown to have existed there, with secure local pre-eminence and secure control of extensive estates, how the Frankish kings could reduce such a group to the level of a service aristocracy would constitute a sizable historical problem, though not an insurmountable one. After all, such a reduction need not involve forcible displacement or dispossession, let alone the widespread slaughter, of such aristocrats. In fact, however, this problem does not, on closer inspection, really present itself. Across most of the region, the late Roman aristocracy was not composed of independently wealthy, powerful landowners. It seems to have been every bit as dependent upon the state as I would argue that their Merovingian descendants were. There were variations in the degree to which this was the case and at least one region where it certainly was not true but, as a broad categorisation, it will suffice.

This conclusion is based principally upon the archaeological evidence. The northern Gallic aristocracy had never been the wealthiest in Gaul. Studies have suggested that it was less locally dominant than its counterparts in southern Gaul even at the time of the Roman conquest. An exception to this rule might have been found in the civitas of the Treveri, in the lower Moselle valley, and the unusual character of the aristocracy in this area persisted throughout our period and beyond. In the prosperous early imperial era, northern Gaul saw the creation of numerous villas, but these tended to be fairly small establishments. A crucial change occurred after the third-century turmoil, when these settlements were either abandoned or (probably to an even greater degree than is currently known) changed decisively in their form and archaeological visibility. Increasing stress has recently been laid on the fact that better quality excavation and analysis reveal that many fewer of these sites were abandoned than had been believed. This has been vital in reassessing the settlement pattern and economy of late Roman northern Gaul. Nonetheless, even the more subtle analyses suggest that the rates of abandonment were very high, frequently in the region of 50%.

Concentration upon continuity of occupation also ignores a very important aspect of the problem – the change in the character of the settlement itself. The classical villa, stone-built with tiled roofs and, frequently, mosaic floors, often with under-floor hypocausts, manifests a particular set of social and economic relationships between the owner of the villa buildings and other inhabitants of the locality. Whatever the precise function of the site, whether working farm, ‘country house’ or hunting lodge (and this surely varied from site to site and from one phase of occupation to another), the villas reveal an ability to concentrate surplus and spend it upon the construction of a building that made a permanent mark upon the landscape. Such a building made a claim for the owner’s active subscription to a particular set of cultural attitudes associated with the Roman Empire, however those attitudes were played with and modified in local context. Stone-quarrying, tile-manufacture and mosaic-construction all required specialist manufactures and industries, organised transport networks and so on. Though the potential sophistication of timber architecture should not be neglected, it remains the case that the construction and maintenance of stone buildings necessitated a more complex matrix of specialist skills and industries. The change from these structures to timber, thatch- or shingle-roofed halls therefore marks a vitally important change in the nature of the local social élite and of the northern Gallic economy.

Debate on this change has hitherto tended to focus upon whether the shifts involved implied an economic decline. It is very difficult to avoid the conclusion that, in some ways at least, this must have been the case, but the best explanation of the transformation of the northern Gallic countryside in late antiquity seems to be to link it to the nature of the Roman Empire that emerged after the ‘third-century crisis’. It has for some time been noted that the evident change in the region’s settlement actually post-dates the conventional dates of the ‘crisis’, coming after Gaul’s reincorporation into the legitimate Roman Empire by Aurelian, who crushed the separate ‘Gallic Empire’ in the 270s. This precise political historical context can be combined with what we know of the Late Empire’s political economy to produce the following reconstruction.

One effect of the third century’s economic difficulties was, as is well known, an increase in the levying of taxes, and in the payment of state employees, in kind. Another well-known late imperial characteristic was the residence of the emperors on the frontiers, in the west most often at Trier (former home of the Gallic emperors). This presence was associated with a concentration of higher-grade troops near the Emperor himself and renewed emphasis on frontier defence and military operations against the barbarians (classic signs of imperial good management). This change, the shift in the system of taxation and payment, and other reorganisations of the hierarchy of troop-types within the army resulted in troops being concentrated in northern Gaul but spread over a wider expanse of the countryside. These features were crucial to the survival of the ensuing ‘inside-out’ late Roman Empire.

These points can be combined with the suppression of the Gallic Empire to suggest that the transformation of the northern Gaulish countryside resulted from the harnessing of the region’s surplus to the maintenance of the huge imperial presence now more or less permanently stationed in the area. Confiscation of the lands belonging to supporters of the Gallic imperial regime may have been associated with other expropriations or ‘compulsory purchases’ to ensure that the production of foodstuffs and other materials was geared to the supply of the large number of troops and civil servants. This hypothesis sidesteps the old argument about economic decline. Production could have continued on a scale commensurate with, or even greater than, that which existed before, but far less of the surplus would have accrued to the region’s local élite. This could nonetheless have meant a decline of sorts – in aristocratic wealth and in some of the industries that had hitherto existed to maintain high-status dwellings and way of life. Some slack was nevertheless taken up by the imperial presence in the region. The regional distribution of the late Roman northern Gallic fine-ware, Argonne Ware, seems geared to the supply of the frontier bases and the Rhenish glass industry flourished as well. The scale of production created by this demand might have stimulated the export of these classes of material further afield, to Britain and elsewhere. Indeed northern Gaulish products were traded far beyond the imperial frontier. Bronze bowls known as the ‘Vestland Type’, because of the concentration of find-spots in Norway, seem for example to have been produced in the Meuse valley. The precise form of exchange represented by this ‘export’ doubtless varied between a possibly ‘normal’ market exchange with barbarian communities immediately beyond the limes, through exchange with what resemble ‘gateway communities’ around the Baltic coasts, to diplomatic payments to the interior of Germania Magna. To whom any profits from the commercial transactions included within this range fell is impossible to establish. Some of it doubtless went to the imperial treasury, other elements probably ‘piggy-backed’ upon the empire’s ‘command economy’ and ended up with those to whom this was entrusted, and some surely represented independent commercial enterprises conducted by such producers, perhaps generated by the possibilities attendant upon large-scale imperially-sponsored manufacture. Some, therefore, must have ended up in the hands of the northern Gallic aristocracy, probably explaining the continued existence, as such, and refurbishment of at least some of the region’s villas.

Overall, though, the nature of this élite very likely changed too. If the majority of northern Gallic land was now imperially-owned, it could nevertheless have been leased to palatine aristocrats on long-term emphyteutic leases. Other parcels of land might have been used to reward civil and military service and still others could have been used for the traditional retirement gifts to soldiers. In all of these cases it is easy to see how insufficient surplus remained for estate-owners to continue the construction and maintenance of old-style villa buildings. Either possession of the territories and revenues was temporary or the land was not owned on a very large scale. One imagines that the managers of fiscal or imperial estates would not, furthermore, be retaining more of the fruits of the lands for which they were responsible than was necessary for the upkeep of ordinary dwellings and storage facilities. In this context the observable developments in the late Roman landscape of northern Gaul are easily explicable, without resort of ideas of economic decline or a rejection of Romanitas.

The latter concept, although much promoted in recent years, is particularly unlikely. Building a villa in the fourth century did not make a political statement in the same way as the construction of such a building in the centuries around the birth of Christ, in that it did not represent a new form of monumentalisation that differed from that which went before and proclaimed an adherence to new social, cultural and political modes of life. Nonetheless, the continued centring of Roman aristocratic culture on the villa cannot be denied. Furthermore, if one looks at the symbolism of the grave goods deposited in burials during the time of crisis around 400 (to be discussed shortly) one can see that they focus overwhelmingly upon the traditional symbols of Roman aristocratic life, the bases of social distinction and élite pastimes. A continuing subscription to precisely these well-established modes of Roman aristocratic life is clearly attested. The idea that the regional élite engaged in some form of active ‘de-Romanisation’, let alone ‘anti-Romanisation’ is extremely improbable.

Wickham presents a more interesting and subtle modification of the idea under the heading of the region’s militarisation, a process which is certainly demonstrable. Buildings within late Roman fortifications have frequently proved difficult to detect, leading Wickham to suggest that the late imperial military did not invest resources in lavishly-appointed dwellings. Therefore, runs the argument, a more militarised aristocratic culture would involve much less attention devoted to lavish building. This is an interesting possibility but it seems less plausible if scrutinised closely. The absence of buildings within late Roman forts is sometimes possibly explained by a lack of excavation within the walls, and on occasion, doubtless, by the failure to detect structures on top of more elaborate and indeed more sought-after early imperial phases. Nonetheless a relative reduction in the permanence and elaboration of officers’ quarters is a hypothesis worth testing and, if accurate, could be significant in the ways that Wickham suggests. It might be the case that the Roman army’s officer corps had (when compared with civilian private building) never invested in especially elaborate dwellings within their forts. If so, though, even if there were not a decline in the relative sophistication of their quarters, a militarisation hypothesis could explain a shift from investment in buildings to an investment in other forms of display. The Roman army’s officer corps was especially fond of its ‘awe-inspiring’ costume and adornment. State factories, interestingly known as barbaricaria, were devoted to the gilding and ornamentation of officers’ armour. All these points support Wickham’s idea.

However, other points counter it. One is that the late Roman ‘soldier emperors’ were very far from neglecting their private quarters and constructed new and very elaborate palaces (most notably, perhaps, at Salona in Dalmatia). The lavish villas in the immediate hinterland of Trier, the imperial capital, also suggest that imperial officers, at the higher end at least (some of whom, surely, were military), still desired impressive homes. A vital factor might have been late Roman aristocratic culture, which tended to avoid spending private money on public building, something which doubtless goes far to explaining the changes on urban sites. The early imperial army had sponsored local markets and private craftsmen to a considerable degree, in a way that paralleled the use of private money on civic projects in the civilian sphere. Given the fluidity that existed between civil and military careers at this time, this is not surprising. The later army, by contrast (and again unsurprisingly given the contemporary nature of civilian political life), was supplied overwhelmingly via the state and its workshops. Roman officers might, like their civilian counterparts, have spent money on private dwellings (urban and rural) but not on the public structures of their barracks. Given Late Roman law’s frequent enactments about officers who stayed away from their regiments for long periods of time, this suggestion gains further credibility. These laws demonstrate that military service continued to be an option for the Roman aristocracy and that it was encompassed within the usual aristocratic norms of otium and negotium. Roman officers seem to have spent large amounts of time living away from their barracks. Further, when, during the fifth century, we can examine areas of Gaul where the Roman aristocracy became militarised but remained wealthy, we see that they continued to spend money on their villae (even if, admittedly, less than before) as well as maintaining retinues and taking part in campaigns. A simple militarisation of the northern Gallic aristocratic culture seems, then, to be unsatisfactory as an explanation for the reduction in the number and scale of villas.

This thesis proposed here, relating changes in the region’s settlement pattern and economy to changes in the late imperial state, accounts for the developments on urban as well as rural sites. Late Roman northern Gallic towns famously underwent serious contraction. This has largely been judged from the length of the late imperial walled circuits, a blunt method which must give misleading results in at least some cases, such as that of Bavay, where the town walls only enclosed the forum. Even so, recent excavation confirms the abandonment of large areas of early Roman towns. These restricted urban ‘enceintes’ probably do not mark economic decline as such. Where details of their above-ground appearance are known, they often show care and attempts at decoration. That the foundations (frequently the only element retrieved archaeologically, of course) include large amounts of reused masonry from demolished early Roman buildings and grave monuments permits no extrapolations about the haste or emergency conditions in which the walls were built. The period of these walls’ construction is much less well established than was once thought. Rather than having been thrown up rapidly in the last quarter of the third century in response to barbarian invasions, their construction could have extended through the fourth century and possibly beyond. Nonetheless the use, in the walls’ foundations, of large quantities of often good-quality sculpture from cemeteries and public buildings makes a statement about urban change that cannot be lightly dismissed.

These fortifications and their scale have valuable things to tell us nonetheless. I suggest that their short length results partly from the well-documented unwillingness of late Roman local aristocrats to spend their money upon public building projects. It might also stem, to some extent, from a lack of independently wealthy aristocrats in the towns’ hinterlands. This decline in the prosperity of the northern Gallic elite probably also lies behind the reduction in the scale of occupation in towns as the small landowners, those not employed to manage imperial estates, seem, as we have noted, to have generally lacked the secure control of surplus necessary to sponsor urban development and manufactures, town houses and so on. A further factor, quite well understood in Britain, would be the rise of lesser settlements, the so-called vici (now often themselves fortified as castra or castella). These settlements often arose as markets and by the late imperial period had probably done much to erode the civitas-capitals’ dominance as economic central places, in economic terms. The Gallic urban hierarchy underwent significant changes in the late imperial era, with some sites (like Verdun, in northern Gaul) gaining civitas status. Slightly further south, by the sixth century (when Gregory of Tours famously remarked upon the fact) the castrum of Dijon had become notably more important than its civitas-capital, Langres. Against such a back-drop it should not be surprising if intermediate settlements rose in importance relative to that of the cities.

The general thesis proposed here to explain late imperial change in northern Gaul is confirmed to some extent by exceptions to what seems to be the general rule. One is the region around Trier, which, although quite badly hit in the third century in terms of the rate of villa-abandonment, nevertheless has the largest and most lavish examples of fourth-century northern Gallic villas (as intimated earlier). Trier itself, of course, saw a rash of public building focused on the imperial presence there and that is the crucial issue. With the emperors and their senior palatine aristocrats resident there for much of the late third and fourth century, it is no surprise that resources could be spent on the city, or that wealth found its way into Trier’s immediate surroundings. When one considers that the prosperity of the aristocrats of the Treveri had been unusual in the early and even pre-Roman eras, the fact that the Triererland bucks the general trend should scarcely be unexpected. The idea that imperial patronage was the main factor determining variations from the norm gains further support at Metz, the next city on the Moselle, to the south of Trier. If Trier’s abnormally long walled circuit probably belongs to the early Roman period, then the 72 hectares enclosed at Metz make that city’s walls by far the longest late Roman enceinte in northern Gaul, perhaps in all of Gaul. The castra of the civitas of Metz also have unusually long walled circuits. This exceptional scale of fortification surely relates to the presence around Trier of troops available for such building projects. An inscription documents the construction of the Langmauer, a long wall presumably enclosing an estate, by a unit of primani.

Even if it does not explain the decline in the number and elaboration of villas, the militarisation of northern Gaul is nevertheless a feature of the region’s archaeology which supports the general hypothesis. The fortification of granaries and other rural sites seems to be associated (as indeed may the walling of urban administrative centres and some villages) with a concern to ensure the safety of supplies ear-marked for the army. The deployment of symbols in inhumations in the region, especially when these become more lavish in the century’s last decades, underlines the close link between imperial (probably especially military) service and local status. For most of the period before 400, official belt-sets and brooches, the badges of rank in the army and civil service, are the most common grave-goods in male burials.

Part 2 Here

Battle in the Early Medieval West

[The following is the text of a piece I wrote for an Encyclopaedia of Classical Battle or some such, which - for various reasons to do with its coverage - the editors decided was not wanted, in this form at least, so I thought I'd post it here instead.]

The Characteristics of the Period

One of the most evocative pictures of
an early medieval battle I know of.
The Battle of Edington as envisaged by
John Kenney in The Ladybird 'Adventures
in History' Book of Alfred the Great.
Never mind the horned and winged
helmets!

The study of battle in western Europe between the final dissolution of the western Roman Empire in the fifth century and the beginnings of the world of ‘knights and castles’ in the tenth, is not easy.  Partly this stems from the problems of the sources for the study themselves, and partly it originates in the historiography of medieval military history.  It will probably be more useful to begin with the latter. 

The most important difficulty in comprehending warfare, and battle in particular, during this period is the fact that the era is very rarely understood on its own terms.  This general problem takes two more specific forms.  The first is the debate over the inheritance of Rome and the impact of the ‘Germanic’ barbarians.  On one side there is the old idea that the barbarians brought with them a ‘Germanic’, ‘tribal’ style of warfare, which replaced, by right of conquest, the scientific warfare of the Roman Empire: we might call this the ‘barbarization’ hypothesis.  At the other extreme is the argument that the armed forces of the Roman Empire and the type of warfare they practised were perpetuated directly into the early Middle Ages: the ‘continuity’ hypothesis.  The barbarization hypothesis is the older of the two and finds, perhaps, its apogee in the work of the British historian Sir Charles Oman, whose view of medieval warfare was as a rather mournful interlude between the Vegetian warfare of the Romans and the rebirth of scientific warfare in the Renaissance.  In his view the fifth century did indeed witness a triumph of barbarism and warfare for the next millennium consisted of little more than mindless clashes between disorganised forces of armoured knights.  Indeed Oman saw the battle of Adrianople (378) as something of an emblematic turning point, as he envisaged the battle as the decisive triumph of mounted Gothic ‘knights’ over the foot-slogging legionaries of Rome. 

This is an argument long realised to have been completely mistaken.  Nevertheless, there were less extreme variations on this theme and indeed it can be seen as only one manifestation of the still-widespread misunderstanding of the post-imperial period as one wherein ‘Germanic’ culture swamped western Europe – as a result of the military conquests that the complex phenomenon of the Barbarian Migrations were (and are sadly still) mistakenly seen as being.  One variation on the theme was the dominance of dismounted infantry warfare imported by largely cavalry-free western ‘Germanic’ tribal armies.  This style of warfare persisted, runs a traditional argument, until the eighth century, when the Arab incursions into the west, allegedly bringing with them a new, more mobile form of warfare, compelled the creation of large Frankish cavalry armies.  The equipment of a mounted armoured warrior in these new-style forces was much costlier than hitherto and, it was said, required the creation of new means of rewarding armed followers.  The latter, in this view, were therefore rewarded with lands, especially taken from the church, held on a form of lease in return for continued good service.  This lease-hold agreement was sealed with the swearing of oaths.  Thus, in this vision of the early Middle Ages, the famous ‘feudal system’, supposedly characteristic of medieval warfare was born.  This system stayed in place for most of the next 800 years until replaced by the new ordonnance armies of the fifteenth century.  Because of that persistence, it is still often thought, there was little change in the practice of medieval warfare.

In fully-fledged form, this ‘barbarization’ hypothesis is nowadays held by few if any historians.  The idea of how feudalism was created in the eighth century has been challenged on all fronts, from the specifically military – the mounted warrior did not suddenly appear or dramatically become more dominant in the 730s, and the Arab armies in the west did not wage a type of war that was radically different from that practised across much of the rest of western Europe (many ideas here came from back-projecting the stock idea of ‘Saracen’ warfare from the age of Saladin) – to the institutional – the novelty or chronology of the benefice and the oath of vassalage and the systematic or formal nature of either.  Nonetheless it might still be the case that important changes did take place around 700, even if their nature has been misunderstood.  All that having been said, there are many areas where the inheritance of the barbarization hypothesis remains in force, often sub-consciously, in historians’ and archaeologists’ assumptions about this period.  It is also the case that an attempt was made recently to revive the ‘Germanic’ nature of early medieval warfare in all its glory.  This attempt suffered, however, from a very crude use of sources and has had little or no impact.

The problems with the ‘barbarization’ or ‘Germanic’ hypotheses are insurmountable.  On the one hand they assume that there was some sort of ‘Germanic’ cultural ethos that unified all the peoples that lived between Scandinavia and the Rhine and between the North Sea Coast and the Ukraine.  This both more and less than a continuation of Graeco-Roman ethnographic chauvinism, which lumped all Germanic-speakers together as fundamentally the same.  More than this, in the sense that the Romans did not regard the Goths as Germani; that is essentially a modern assumption growing out of moves towards German unity from the Renaissance onwards (although the common linguistic roots of Franks and Goths had been noted in the eighth century).  It is less than Roman chauvinism in that the Romans in fact described the Germani as an extension of the same general features as the Gaul – in other words as just one manifestation of the standard ‘northern barbarian’.  Suffice it to say that although this remains a pernicious and deep-rooted idea, there is no evidential basis that allows us to treat the Germanic-speaking barbarians as fundamentally interchangeable exponents of ‘Germanic’ culture. 

The other problem with the ‘barbarization’ hypothesis is that it is more or less impossible to know what the norms of ‘Germanic’ warfare were before the fifth century.  Archaeology gives us some interesting insights here and there about practice, armament and organisation.  The Graeco-Roman authors give us other hints but these are so mired in classical ethnographic stereotyping that they are difficult to make use of and seem often to contradict the implications of the archaeology.  That apart, the idea of ‘Germanic’ warfare have actually been drawn from the sources of the western European early Middle Ages, after the settlement of Germanic-speakers in the provinces of the Empire.  Here the argument, obviously, becomes circular: the nature of early medieval western warfare was ‘Germanic’ because, as we can see in the sources of the period, it conforms to what we know ‘Germanic’ warfare was like; but our idea of what ‘Germanic’ warfare was like is drawn from the sources of the early medieval West.  Even the ‘barbarization’ of the Late Roman army can be argued to represent a very Roman process, based around Roman ideas of what barbarians were like, rather than the importing of actual barbarian practices.

One of the reasons why the ‘barbarization hypothesis’ is no longer as common as it was is that the degree of continuity from the Roman world into the post-imperial has increasingly been appreciated.  From the 1960s and especially from the publication of Peter Brown’s classic and brilliant The World of Late Antiquity (1971), the extent to which there was any really profound break during the fifth century has been put under scrutiny.  Most of this work has been very fine indeed and has radically reshaped the way in which most historians think about the period.  It has, however, had some unfortunate side-effects, one being a view of the early Middle Ages in which nothing at all had changed from the Roman-period.  In the sphere of military history this has been manifested in a large corpus of work by Bernard S. Bachrach, which argues that the armies of the sixth century and later, right on into the central Middle Ages, were simple continuations of Roman armies and waged warfare in the same way as the late Roman army had done. In spite of its volume, this work has had little effect on the serious study of the period, but in numerical terms it does dominate the historiography of the specific subject of early medieval battle, so some further discussion is necessary.

The ‘bottom line’ that must be borne in mind is that there is no evidence at all to support Bachrach’s arguments.  The latter have proceeded essentially by reading uniform, technical meanings (out of several possible even in classical Latin texts) into particular words used in early medieval sources.  A general and unremarkable descriptive word meaning ‘scouts’ or ‘outriders’, for example, becomes the title of a specific regular unit of the Frankish army.  This ignores the demands of early medieval genre and style, which required writers to employ vocabulary ‘ratified’ by its use in the models of Greek and Roman historical writing.  Bachrach is quite right that early medieval writers did not use terms willy-nilly.  They surely chose words that were generally applicable and conjured the right sort of image although it is also clear that they could play around with such terms for stylistic, argumentative and ironic effect).  It is quite a jump from there to the argument that the appearance of this vocabulary reflects direct factual description.  Bachrach’s approach fails, furthermore, to account for the clear variations in which these terms are used within sources or the ways in which they are obviously used as synonyms for other words with technically different meanings.  An example can be seen in the way that Bachrach takes the use of the words acies and cuneus in the Carolingian poem Waltharius to mean two different military formations, a line and a column respectively.  Indeed acies does mean ‘battle-line’ and one of the many meanings of cuneus (literally a wedge) was a dense, deeply drawn-up battle-formation, perhaps like a column.  But in early medieval sources it is clear that both words were used interchangeably for the close-packed formations which (as we shall see) were usually the norm on the battlefield.  This is clear from Waltharius itself, where the poet alternates between one and the other.  Bachrach reads this as a unit of cavalry deploying from line to column and then back again, all while within spear-throwing distance of the other side.  Proper textual analysis makes this unlikely.  So too, though, does comparative military history, which would demonstrate that such formation changes within close range of the enemy would be nearly impossible to perform.  Therein lies another problem with Bachrach’s interpretations; he envisages a scale of warfare in the early Middle Ages that was not matched before the eighteenth century, and a discipline (on the march and in battle) and tactical complexity that study of battlefield actuality in better-documented eras suggests has never existed in practice.

One reason for that is Bachrach’s concentration on the continued copying of the late Roman tactical and strategic treatise De Re Militari (roughly ‘On Matters Military’) by Vegetius.  This work was indeed very popular in the Middle Ages and existed in hundreds of manuscripts in many different forms.  Bachrach has pointed out that medieval authors sometimes made amendments to the text to bring it more into line with the world in which they lived.  That is a good point and shows (as all medievalists know in any case) that post-imperial authors did not simply copy their sources without thinking about them – any more than they used words from the classical past without thinking about how to use them in their own terms.  Indeed the argument that classical texts could be adapted to current actuality by early medieval writers but that the technical meanings of vocabulary remained unaltered is something of an aporia in Bachrach’s thesis.  It is also perfectly true that (in the later Middle Ages especially) warriors are known to have owned copies of De Re Militari.  However, it is another unjustifiable argumentative leap from these (fundamentally sensible) points to Bachrach’s argument that Vegetius’ popularity means that the early medieval (and later) readers of Vegetius used his tactical and organisational recommendations in practice.  Again, comparative military history would argue that, even in eras where they were widely issued to junior officers and even NCOs (in the eighteenth and nineteenth centuries, for example), there was a considerable gap between the diktats of tactical manuals and battlefield reality.  The multiple varieties of forming a square from a line in mid-eighteenth-century handbooks seem, for instance, almost never to have been followed in practice during the Seven Years’ War, when infantry lines either stopped cavalry attacks with volleys of musketry or lay down and let the cavalry ride over them.  One need not proceed so far from the period under review, however.  Close study of Vegetius will soon reveal that this was (overall) an armchair theorist’s idea of how things ought to be – indeed the whole point of the book is pretty clearly that things aren’t like they were in the ‘good old days’ – and not a description of practice even to the extent that a nineteenth-century manual was.  Vegetius’ book, therefore, is not even a reliable description of late Roman, let alone post-imperial, warfare; indeed Vegetius never claimed that it was. 

Another point is that the copies of Vegetius that circulated in the early Middle Ages, even while adding in comments here and there about ‘modern practice’, usually left out the most specifically military elements of De Re Militari.  This is because most circulated within monasteries and it has not implausibly been argued that the point the work was there intended to serve was allegorical: as a ‘tactical manual’ for monks battling sin and evil.  Even the (later medieval) copies of the work that were owned by soldiers were demonstrably acquired by those warriors late in life.  Thus rather than being an educational tool, to be followed in practice when one took the field later on, the preferable argument is that old soldiers liked Vegetius’ book because it gave a respectable classical imprimatur to what they had already learnt by dint of hard practice in the field.  Most of Vegetius’ recommendations are, indeed, of the ‘common sense’ variety (e.g. you should post guards around a camp to prevent surprise attacks) that good generals would know whether or not they read De Re Militari.  A further problem with Bachrach’s thesis is that he tends to misrepresent any historian who rejects the widespread practical use of classical treatises in the Middle Ages as arguing for the existence of primitive tribal warfare, with commanders simply bumbling about, hitting anyone who got in their way (the Oman view of medieval war).  Obviously, such a view in no way follows from the argument that medieval warfare was not a clinical Vegetian science (as Roman warfare had not been either) – which is nevertheless not to say that a lack of strategic and tactical good sense cannot be seen on occasion during the early Middle Ages.  It can, as it can throughout military history.

All this is in fact fairly typical of the early medieval attitude to the works of the classical past.  Such tracts were of huge interest and their possession, and often knowledge of their contents too, were good markers of learning and status.  It has been thoroughly demonstrated that Carolingian writers about geography paid enormous respect to works from antiquity and often repeated their comments even whilst knowing full well that they had nothing at all to do with the ninth-century world in which they lived.  The idea, however often repeated, that early medieval warfare was a continuation of Roman practice must, therefore be rejected at the outset.  This is important to make clear, because no further reference to the idea will be made in this chapter.

Bachrach also portrays any historian who does not subscribe to his view of extreme continuity as holding to the ‘barbarization’ hypothesis: the implanting of barbarian, specifically ‘Germanic’ military practices in post-imperial Europe.  In this, though, he is not alone.  It does seem to have proven enormously difficult for the study of post-imperial warfare to break free of the binary opposition between ‘continuity’ and ‘barbarization’.  The idea that the warfare of the period might be characterised by neither term has not managed to take root.  As an illustration of this point we can cite a recent encyclopaedia entry on the topic, which portrays my own earlier arguments as postulating the ‘barbarization’ of warfare before the end of the Roman Empire.  They do not and, more to the point, it is impossible to read them in this way unless one is incapable of seeing fifth- to tenth-century warfare other than as either ‘barbarized’ or ‘Roman’.  Admittedly this is not the only area of the history of this period that has been incapable of breaking free from this dualism.  Nonetheless unless we break free of it we shall never understand warfare in the West between c.450 and c.900, for one of the most important points that this chapter tries to make is that warfare in this era must be understood on its own terms, not as a simple continuation of Roman warfare or as a simple importation of equally pre-existing, supposedly Germanic norms.  Warfare in this period clearly developed from that of the late Roman period but it was essentially, neither ‘Roman’ nor ‘Germanic’: just ‘new’ and ‘different’.  To persist with the traditional alternatives is akin to arguing that sixteenth century warfare must either be a direct continuation of medieval warfare or a simple importing of Renaissance Italian norms.

This conclusion has a bearing on the other key mistake in envisaging early medieval warfare, which is to assume that it was the same as warfare in the later eras of the Middle Ages, say from the tenth or, especially, the eleventh century onwards.  It has been assumed that norms of medieval warfare existed, which meant that battle was rare and that the principal means by which warfare was prosecuted was via sieges of enemy strongholds.  This is a fine description of the warfare of the age of ‘knights and castles’ but we must think twice before we assume that it applied in an age where neither knights nor castles (as either element usually is understood) actually existed.  Battle seems to have been more common in the period that concerns this chapter than it was in the Central and Later Middle Ages, for reasons that I will return to discuss in more detail – reasons that concern the social and economic structures of the period, which were quite different from those after the economic upsurge of the ninth and tenth centuries, (especially as these concern urbanism).  Thus we must try and understand early medieval battle on its own terms, as it comes down to us in the sources of the period, rather than trying to hammer it into a framework created through the study of other, admittedly better-documented eras.

The Source Material

This, however, is no easy task, for the source materials available to us are infuriatingly uninformative about battlefield reality – it is this fact alone that has rendered it so easy to make early medieval warfare conform to templates created from Roman or Central Medieval evidence.  The silence of early medieval sources on the nature of battle is a very difficult problem to unravel.  There is no doubt that warfare was a very frequent and very real part of early medieval life, whether we are talking about society, economics or politics.  The annalistic sources are full of mentions of wars and there are very good reasons to suppose that violence that would be regarded as warfare in modern international politics was in fact considerably more common even than our written evidence indicates. 

Yet, that evidence uniformly reduces warfare to banalities of one form or another.  The standard western early medieval annalistic account of a campaign and battle habitually contains the information that King X led an army to place Y and fought King Z, with one or other being victorious and the following notable personages being killed – and nothing more.  Often even that information is not all present.  Several battles in the Anglo-Saxon Chronicle, for example, lack a recorded winner.  This might have been was because the ‘wrong’ side won (that is to say not the West Saxons), or because the battle was, effectively, a draw, or because the Chronicler’s own sources did not agree on the victor, but we cannot now decide which of these explanations is the correct one.  On other occasions the location of the fight is omitted; occasionally even the opponent is missing from the account!  Manuscript E of the Anglo-Saxon Chronicle, for example, simply says, under the year 937, that ‘in this year King Æthelstan led an army to Brunanburh’.  Were it not for the preservation, in the other manuscripts of the Chronicle, of the famous ‘Song of Brunanburh’ and its account of Æthelstan’s defeat of a coalition of Welsh, Scots and Vikings, even this, his greatest victory, would possibly have fallen victim to the curious damnatio memoriae visited on Æthelstan by his successors.

There are of course exceptions.  Occasionally another snippet of information is included and towards the end of the period under review, in the ninth century, some of the Frankish annals – especially the Chronicle of Regino of Prüm – begin to include considerably more detail, even if still tacit by comparison with the sources of other periods.  Sometimes other information about the routes taken by armies can be winnowed out of sources like charters, issued by a king en route.  Nonetheless, before the later ninth century the annalistic sources are very terse on battle and its nature.

This is rarely compensated for by the other forms of history available to the student of early medieval western battle.  Saints’ Lives, unsurprisingly, have little to say on the subject although many of the occasional additional pieces information about battlefield tactics that make it through the capricious evidential filter come from hagiographies.  One is the fact that an army led against the Picts by King Ecgfrith of Northumbria in the 670s was equitatus (mounted on horses).  This is the only reference to whether or not Anglo-Saxon armies used horses to appear in a source earlier than the ninth century.  Histories written in the tradition of classical narrative are, as has been alluded to, mired in the demands of genre, to write as closely to the Greek and Roman models as possible.  Thus Asser, biographer of King Alfred, refers to Viking and Anglo-Saxon armies as forming testudines; we should not envisage ninth century warriors drawn up in a testudo.  Asser meant us simply to see them as packed in, shoulder to shoulder, behind their shields.  When Byzantine historians like Procopius and Agathias write about western European armies as, for example, in Procopius’ accounts of the campaigns by the general Belisarius against the Goths in Italy, these problems can be even greater.  These writers were even more constrained by expectations not to use new-fangled terms.  Thus Procopius talks of Eastern Roman (Byzantine) guards using the words hypaspistai and doryphoroi and, although there is an interesting consistency in his use of the words which suggests a two-tiered nature to Byzantine guards units similar to that found in the households of western ‘barbarians’, we have to pause to remember that both terms are found in classical Greek works.  The hypaspistai were the dismounted bodyguards of Alexander the Great, for example.  Whether sixth-century Byzantine guardsmen were really called by these titles is difficult to say; scepticism seems prudent when one notes the tortuous circumlocution used elsewhere by Procopius when he refers to the ‘excubatores (for thus the Romans call their guards)’.  Elsewhere we encounter Gothic ‘Hoplites’ – possibly an instance where the use of ‘attic’ vocabulary was played with for ironic effect, although Gothic close-fighters with round shield and spear would indeed bear a passing resemblance to the Hoplites of ancient Greece.

These examples act as a neat bridge between the formulaic terseness of the annalistic accounts and poetic formulas about battle.  Battle is common subject matter in early medieval poetry, whether in the vernacular (such as late Anglo-Saxon poetry, or the Ludwigslied, about the battle of Saucourt in 882) or Latin (as with Waltharius or Angilbert’s poem on the battle of Fontenoy in 841).  In the Old English poetic corpus we can also find accounts of battles in poems about Christian History (Elene) or biblical topics (Judith).  Sadly for the historian, these accounts rarely tell us very much.  The passage about cavalry fighting in Waltharius, already alluded to, is an important but all-too-rare exception.  The accounts are again reduced to a series of commonplaces, that the brave warriors of one side defeated those of another, giving their people little cause to rejoice, and the carrion beasts of the forest (wolves, ravens, etc.) feasted on their dead.  The tone can be unpleasant and triumphalist in victory (as with the Old English Song of Brunanburh) or elegiac in defeat (as with Angilbert’s poems or some of the British battle poetry that might date to our period, like Y Gododdin) but in neither context is very much of substance relayed to us.

Why there is this unremitting near-silence on the experience of battle in the early Middle Ages is a problem which is difficult to resolve.  Comparison with other eras makes the undoubted horror of close-fighting with edged weapons unsatisfactory as an explanation and the militancy of the early medieval Church precludes an account based on the exigencies of New Testament theology.  The involvement of the Church in warfare rules out an explanation based on the overwhelmingly ecclesiastical provenance of our sources, as do the facts that sources written by laymen (such as Nithards’ Histories) are just as tacit while, conversely, some of the most detailed accounts of battles (like Regino’s) come from the pens of churchmen.  Currently the closest we might come to an answer might be via contemporary views on the reasons for victory and defeat.  It was widely, if not universally, held that victory was granted by God and that defeat was the consequence of sin.  Thus the tactics of clever generals (or of stupid ones) were irrelevant.  That the scheming of an imaginative commander like Charles the Bald, attacking via a night march and under cover of peace-feelers, did not prevent his disastrous defeat in the ensuing battle of Andernach by his inexperienced nephew Louis the Younger only underlined the point, as did the fact that while Charles was generally regarded as impious, Louis had undertaken various rituals to place the outcome in the hands of God.  These factors, presumably, explain why Hincmar of Rheims told us more about the campaign than usual.  The more tactically-interested Romans used to say that, for a disaster like Cannae, you needed not just a genius like Hannibal on one side but an idiot like Varro on the other.  Hincmar’s account suggests that, had they had an equivalent maxim, late ninth-century people would have said that, for a disaster like Andernach, you needed not just a pious king like Louis on one side but an impious wretch like Charles on the other.  Such a supposition might go some way towards explaining why early medieval writers were less interested in tactics than their Roman precursors had been.  It does not, of course, imply that early medieval commanders had no interest in tactics!

The other fundamental sources for early medieval battle are furnished by archaeology.  At one level the period does very well out of this source of evidence as the custom of furnished inhumation (burial with grave-goods), which was widely practised in England, northern France and Germany in the sixth and seventh centuries (and into the eighth century in Germany), less commonly employed in Spain and Italy in the same period and used with fluctuating frequency and uneven distribution across Scandinavia throughout the early Middle Ages, has yielded a corpus of thousands of early medieval weapons.  Other items of military apparel, whether for the warrior or his horse, are also attested.  As stated, these data are unevenly distributed in time, place and, thanks to the ritual nature of the formation of this evidence, form as well.  The custom could involve greater or lesser investment in the deposition of weapons at certain times and in specific areas.  The sample of early medieval helmets and armour is overwhelmingly skewed towards south-western Germany where the furnished inhumation rite could be very lavish indeed.  This does not mean that the Alamans were the most heavily armoured warriors in the West.  Similarly, shields are very well known from Anglo-Saxon burials but rare in Frankish ones; this does not imply that Frankish warriors went into battle unshielded.  These problems of sample and of moving from the ritual deposit of objects in a grave to the availability or otherwise of items in everyday life are very significant.  So too is extrapolating tactical practice from the form of weapons.  Nonetheless, with these difficulties taken into account, an attempt at some tentative conclusions can still be made.

The Role of Battle

Earlier I intimated that battle was more frequent in the early medieval period than at other points in history.  This is quite a controversial statement to take, not least because the good reasons why commanders avoided battle in other times and places – above all the fact that (as Charles the Bald found to his cost at Andernach) battle was a lottery played for the highest of stakes – surely applied in this period too. Indeed one contemporary, Sedulius Scotus, wrote as much. ‘There is much uncertainty’, he said, ‘in the rumblings of war.’  If this was the case, as it seems to have been, why should early medieval western commanders have been more willing to commit themselves and their forces to battle than their predecessors and successors?  Were they mad?

First of all, though, the argument that battle was more frequent needs to be substantiated.  Needless to say, this is not easy given the sources at our disposal but the recorded numbers of what one assumes were significant battlefield encounters are nevertheless quite high.  Perhaps the extreme example is the indecisive winter campaign fought between the West Saxons and the Viking Great army in 870-71 – the ‘Year of Battles’ – when seven or nine encounters were recorded, besides other skirmishes.  It must be conceded that many of the battles mentioned were probably small-scale affairs by the standards of other periods but one ought to judge battles by the political stakes played for and the significance of victory or defeat rather than by the numbers of troops involved.  In this perspective the numbers of important leaders who lost their lives in battle is very significant.  In seventh-century England, for example, no fewer than twelve battles were fought that resulted in the death of at least one king or other royal.

There are two reasons, I suggest, why commanders were readier to trust their lives and those of their followers to the lottery of battle in this period.  The first is economic.  The acquisition of loot and treasure was a very important reason for waging war, in this period as in others, although its importance has perhaps been overstated in comparison with other, less tangible rewards in patronage, office and so on that could be acquired through good service on campaign and in battle.  The crucial fact is that in western Europe for most of the period from the fifth century onwards, settlements do not seem to have been very large or to have been the locus for much wealth.  Towns in particular remained small throughout the period covered.  The fifth and sixth centuries were a period of urban decline across north-western Europe, which began to be reversed in the seventh century.  A real upsurge in the fortunes of towns did not occur until the ninth century and later.  In the Mediterranean, towns were small in the fifth and sixth century and declined later, before a later revival towards the end of the period.  The besieging and sacking of settlements was thus not a means by which large quantities of wealth would be obtained.  The only obvious exceptions to this rule were churches and monasteries and – although – pre-Viking sacking of religious sites by Christian forces are far from unknown, there were important spiritual sanctions against taking this military course of action.  By contrast, armies were concentrations of moveable wealth.  People tended, throughout this period, to wear, and thus to display their wealth.  As the recently discovered Staffordshire Hoard has made abundantly clear, military equipment was a major locus for the investment of wealth in gold and other precious metal, and costly craftsmanship.  Weapons and armour did not come cheap, even if their expense has sometimes been overplayed.  As political gatherings of the powerful, the army was also a site for competitive displays of this sort.  More than that, horses, the sine qua non of the early medieval warrior, were expensive.  The price of a warhorse remained pretty stable at about ten solidi throughout the early medieval West, although especially good mounts might fetch twice that amount.  What this meant in practice is difficult to say, given that a solidus was generally a unit of account rather than an actual coin in western Europe at this time.  Nevertheless reasonably-sized parcels of land could be exchanged for a horse.   Kings, furthermore, took their treasuries on campaign, not to mention evidently luxurious tents and other items.  By the ninth century, when the western economy was becoming more complex, it is clear too that merchants followed armies.  In Charles the Bald’s defeat at Andernach, for example, shield sellers are mentioned as being caught up in the rout.  Thus defeating the enemy army in battle was liable to yield far more booty and disposable treasure than the sacking of any number of rural settlements – unless a king, his senior aristocrats and their followers happened to have been cornered within it.

The economic history of the period, and the study of its settlements also elucidates the other half of the equation.  In later periods, battle was less common an action in deciding a campaign than siege.  For most of the period after the dissolution of the western Empire, fortification is notable by its absence from many classes of settlement (especially high-status or aristocratic rural settlements and the new economic foundations, the emporia), throughout Europe.  Roman walls remained, of course.  Indeed some ‘late Roman’ defensive circuits might in fact belong to our period.  Older walls were subject to renovation and repair, as is clear.  It is also clear, however, that other circuits were in very bad repair, and that the inhabitants of other towns regarded their walls as a hindrance and were not averse to demolishing sections were necessary, whether to extend a church, as in Reims or to facilitate access to a trading zone, as at Mainz.  Many early medieval towns indeed developed their most vital areas in undefended extra-mural zones.  This situation only began to change significantly in the ninth century and, especially, the tenth.  With this in mind it is not surprising that for most of our period we have little evidence of sophisticated siege techniques, or of besiegers taking fortified places other than by direct attack.  Sieges did take place but, at least before the ninth century, these do not seem to have been the decisive set-pieces that they became later.  Where a defending army took to its fortifications, attacking armies usually ended up going home again, often ravaged by disease.  Charlemagne’s capture of Pavia after a long siege, lasting through winter, is an exceptional case where an unusually skilled organiser and commander, in effect, refused to be bound by the usual norms of warfare, and the Lombard kingdom’s existence was terminated as a result.  Otherwise a defending army could allow settlements to be taken and remain unbeaten in the field, without this striking much of a blow to its ability to win the war.  However, most of the time, an army would not simply remain unbeaten in the field; it would seek a decisive encounter.

Apart from economic advantage, the other reason for this concerns the fluidity of status and identity and the role that military status played within defining those identities.  The bearing of weapons, at various times and places, associated with masculinity, ethnicity, age and social rank but in all cases that bearing of arms symbolised, it seems, not just the ability to inflict violence upon other people but the right to participate in specific types or scales of violence, that is to say warfare, the actions of the army.  This equated, throughout the period, with involvement in high-level politics.  As, for much of the period, this right was, or could be, open to competition from rivals, the right had to be made clear on a regular basis.  Thus the calling out of armies, even if only to form political assemblies, was a frequent occurrence.  In Francia in the sixth century and in Lombard Italy in the earlier eighth, it appears to have been associated with March 1^st – the Marchfield, a specific early medieval adaptation of the Campus Martius.  The three Edicts of the Frankish King Childebert II, in the 590s, and all eighth-century Lombard legislation are issued on that date, when the army came together.  Such assemblies were occasions when a king could promote and reward those who had performed well and (at least in some times and places, especially in the sixth century) demote or punish those who had not.

With status open to competition and bound up with one’s identity as a warrior, more or less to the exclusion of all else, outside the Church, it is not surprising that honour amongst warriors was a very touchy subject.  Accusations of cowardice were among the worst one could throw at a man.  Consequently, when it came to questions of honour and its demonstration, battle was frequently sought, regardless of tactical niceties.  Two Frankish defeats came about as the result of commanders not wishing to share any glory obtained and rushing into action before the arrival of another nobleman, the subject of some envy, placed in overall command.  A Lombard disaster against the Slavs was produced when the army’s commanders quarrelled about an accusation of cowardice, which eventually produced a charge up a steep hill against an entrenched enemy.  The rank and file warriors followed their leaders because ‘they considered it base not to’.

Clearly, then, the refusal of battle could be seen as an admission of a lack of personal prowess or bravery.  This could affect kings, fatally, regardless of the demands of a sensible strategy.  The Slavic ruler Liudewit successfully faced down several Frankish invasions of his territory in the early ninth century by holing up in his hilltop fortress and waiting for disease or logistical problems to compel the enemy to return home.  This, alas, was not good enough for his nobles, who rose up and murdered him.  Liudewit’s successful strategy had not, as was necessary, convinced his own followers that he was the best warrior in his lands, or at least that he had the military ability to keep them in their place.  A similar fate befell the emperor Charles III ‘the Fat’ in the late 880s.  Modern analysis has shown that Charles’ strategies against the Viking attackers – essentially involving the avoidance of battle and the payment of subsidies -  was no different from, and no less successful than, those adopted by his predecessors.  Unfortunately, his political opponents were able to use his absence from battle to their advantage, helped by the fact that Charles’ most important rival, Count Odo of Paris, was being lauded for a valiant defence of Paris in a siege which Charles failed to raise through similar virile military action.  Charles could be portrayed as ineffective, lazy, cowardly and thus as a bad king, and within a year he had been deposed, Odo becoming king in his stead.  There were, therefore, several factors that came together in driving early medieval commanders into battle when wiser heads might have stayed their hand.

Bringing About Battle

The preceding discussion raises the issue of how early medieval commanders brought about battle.  Scouting does not seem to have been extensive in this period.  It certainly existed.  In the Pyrenees certainly, and perhaps on other frontiers, the locals were required to perform what seem to have been scouting or guide duties, and there are occasional mentions of outriders in other sources.  Yet there were occasions also when armies were surprised by their enemies in situations that suggest that few or no vedettes or pickets had been posted, such as when a Spanish army managed to attack a Frankish force while the latter was having dinner.  This alone tells us that the requirements of Vegetian warfare were not always met!  In an age, therefore before cavalry scouting screens were widespread or reconnaissance a sophisticated part of the art of war, when the network of route-ways was not dense and maps as we understand them more or less unknown, the most effective means of bringing about a battle was to station one’s army by a well-known landmark or settlement or otherwise across a major line of march, and wait for the enemy to arrive.  This was sometimes also accompanied by the issuing of a formal challenge to battle – although it must be said that most of the challenges we know about were not followed by an actual battle.  Sometimes armies did blunder into each other, as when a Neustrian and a Burgundian army encountered one another late in the day in the 580s.  As we have seen, such a public and obvious statement of the willingness of one side to put the issue to the judgement of battle was likely to produce a response from the other.  Sometimes – perhaps most of the time – ravaging an enemy territory was not aimed at the acquisition of plunder but at shaming the opposing ruler by showing him to be incapable of fulfilling his duty to defend his subjects.  This would very likely bring him out into the field.

With this in mind, it is not surprising to learn that the overwhelming majority of recorded early medieval battles are by the crossings of rivers or other waterways and, less frequently, by monuments or settlements.  Such locations account for nearly all of the known locations of Anglo-Saxon battles between c.600 and c.850.  Most battles in the north of Britain are described as sieges, possibly indicating a regional exception to the general rule discussed above.  However, a close consideration of the sites in question casts some doubt on this as most are very small and do not seem capable of resisting a large attacking force.  This leaves open other possibilities.  One is that a ‘siege’ was actually a rather more formal business wherein an attacking army stationed itself before a defender’s stronghold.  The latter either than came out to fight, producing a battle (a possibility supported by the fact that the mysterious ‘battle’ of Mount Badon in post-imperial Britain, later associated with the even more mysterious figure of ‘Arthur’ is originally referred to as a ‘siege’) or he submitted and paid tribute.  The ‘Distribution of Iudeu’, where Penda took tribute from Oswy of Northumbria, might be a case in point, as Iudeu was a fortified site of some sort (often identified with Stirling but on very uncertain grounds).  Indeed the Northumbrian version of events, wherein Oswy offered to buy Penda off with tribute but the latter refused, might be an attempt to cover up the fact that, having actually paid tribute to Penda, Oswy was breaking his word by attacking him at the battle of Winwaed (tellingly, fought by a river).

There might be other manoeuvres immediately before a battle.  A side might move off early, so as to ensure that it had the high ground, as Charles the Bald’s army did at the battle of Fontenoy (841).  The Earlier Annals of Metz tell us that Charles Martel (Charles the Bald’s great-great-grandfather) moved his army so that its enemies would have the sun in their eyes before the battle of Vinchy (Now Les Rues-des-Vignes, south of Cambrai) in 717.  At the battle of Heavenfield, King Oswald of Northumbria attacked his Welsh enemies at dawn.

Battlefield Tactics

Once the two sides were on the field there may have been other opening moves.  Single combats between champions are occasionally mentioned.  Such fights might have been associated with a phase wherein one side attempted to hold a ford or bridge across a river, either to delay the enemy and ready itself, or perhaps to establish a moral advantage before allowing the enemy to cross  and fight it out (whereupon it would of course have the river to its backs; this might have been earl Byrhtnoth of Essex’s strategy at the Battle of Maldon (991), although in that instance it was his army that lost, leading to criticism of Byrhtnoth ever since).

These preliminaries aside, early medieval encounters do not seem to have involved very much tactical finesse, at least as usually envisaged.  That is to say that our evidence gives us few grounds to envisage sweeping manoeuvres, turning movements and so on, in the Napoleonic mould.  It would seem that battlefield tactics were simple, involving an approach into close range, the exchange of missiles and a charge into contact.  This was then followed by a brutal hand-to-hand fight until one side or the other gave way.  That said, there were doubtless important skills involved in commanding an early medieval force, concerned with maintaining morale, formation and momentum.  Keeping one’s force steady in the face of the enemy, setting a courageous example and knowing when to commit one’s side to the charge were doubtless arts that required some learning in a very tough school that allowed little leeway to those who failed to grasp the lesson. 

There might nevertheless have been variations on this general theme across time and place.  In the first two centuries of the period that concerns me (that is the fifth and sixth) the archaeology of weapons suggests that, as with the recruitment of armies, warfare was still recognisably practised within the general framework set up in the very late Roman period (which is not to say that it represented straightforward continuity).  The weaponry recovered from burials and other deposits in that period appears to suggest a slightly more open and fluid form of fighting.  Throwing weapons such as the francisca (a throwing-axe) and the ango (a heavy javelin descended from the Roman pilum) are known, as are more common arrow-heads and spears that seem adapted to throwing.  Shields, as far as we can tell, were quite small and their bosses appear to have been designed for use in a more open, fencing style of fighting.  Many have a disc at their apex, for which the best explanation is that it was aimed at catching an enemy’s blade.  To see these shields as easily-handled bucklers does not seem unreasonable.  It appears that an emphasis was placed upon on volleys of missiles before the charge, as was the case with the late Roman infantry.  Some of the accounts of western warriors in the fifth and sixth centuries – particularly the Franks – also lay some stress on these aspects.

After the decades on either side of 600, however, a significant change seems to have taken place.  Again, our source for this is essentially the corpus of weapons located in graves.  O some extent this is in itself problematic as the number and variety of grave-goods become much less in the course of the seventh century before largely disappearing around 700.  Nevertheless it is still possible to say that the heavy throwing weapons of the sixth century (the ango and the francisca) drop out of the record, as do some of the other missile weapons, such as the ‘corrugated’ javelin-heads known from Anglo-Saxon England.  At the same time, shields become larger and heavier and their bosses longer, perhaps more designed to punch the enemy, if not just to be held to cover a larger area of the body.  The scramasax, the one–edged dagger, which was short and narrow in the sixth century, becomes progressively heavier – broader and longer – until by the eighth century it is effectively a one-edged short sword.   Spearheads too develop to be longer and heavier thrusting weapons.  Swords, though still known (as they are, of course, throughout the period) become rarer in burials, where the scramasax is often the only weapon deposited.  This could be a simple function of the standardisation and decline of the furnished inhumation ritual, and the large number of ornately-decorated swords represented in the seventh-century Staffordshire Hoard certainly suggests that caution is required before moving from the percentages of weapon types in graves to the percentages of weapon-types in actual use.  Nonetheless, when the development in the form of the scramasax is considered, it does not seem unreasonable to suppose that its popularity may have stemmed from the fact that it married some features of the sword – its point and long edge – and the single-handed battle-axe (which also disappears from the record outside Scandinavia between c.600 and the appearance of the Vikings) – its chopping ability.  This combination would be most valuable in the more closely-packed, ‘shield-wall’ fighting that battle, according to the indications we have, became at this time.  Less likely, but possible, is an increase in the frequency of defensive armour and of the possession of horses by warriors.  Both types of equipment were (as has been argued) expensive and in the social and economic context of the period, which saw an economic upsurge partly at least associated with an increase in the power and stability of the local aristocracy, this suggestion might be more plausible.

Explaining this putative change in tactics is not easy.  It does seem to be the case that at this time the means of recruitment of armies changed, towards a greater reliance on aristocratic retinues, outside those connected with royal service.  Linking this change to the precise shift in the nature of fighting is not straightforward.  It might be that ‘shield-wall’ fighting required a level of mutual trust and coordination that was best forged within the context of an aristocratic household and retinue than in periodically-assembled bodies of warriors.  The economic upturn in the period around 600 and the consolidation of local aristocratic power may, as already intimated, have facilitated a greater ability to purchase body armour and helmets, which would give such forces a significant edge over more lightly equipped warriors in such close-fighting.  Ultimately, however, the link between a change in the nature of warfare and the transformations visible in the economy and social structure remain only a matter of hypothesis and other explanations are possible.  What this important phase does imply, however, is that it is quite impossible to link up, via a simple chain of continuity, the warfare of the Carolingian period and that in the late Roman era.

As far as we can tell, however, after c.600 battlefield tactics revolved around the clash of closely packed ‘shield-walls’.  These shield-walls might have comprised the entire army, or they may have been split up into several ‘divisions’.  It appears that the Frankish armies at Fontenoy each formed three divisions, whereas both the English (West Saxon) and Viking armies at Ashdown (871) drew up in two divisions.  Where the army was broken up into smaller units, two or three such divisions seems to have been normal although there are very occasional references to more.

Whether or not fighting was carried out on foot or on horseback is very difficult to say.  One reason for this is the terseness of our sources’ descriptions or battle, as has already been mentioned.  Another, however, is that where details are given, there seems to be very little to differentiate battles fought on horseback from those waged on foot – something which may further have impaired contemporaries’ willingness to specify about such details.  Another factor was that early medieval warriors were capable of fighting on foot and on horseback as the occasion demanded, and of changing from one to the other in mid-encounter.  King Arnulf ordered his troops to attack a Viking army on foot at the battle of the Dyle (891), whereas King Louis commanded his men to dismount and fight on foot to counter a successful Viking counter attack at Saucourt (882).  Byzantine manuals from the late sixth century had also commented on westerners’ reprehensible ability to dismount and fight on foot in mid-battle and knights were still more than capable of this in the eleventh and twelfth centuries.  ‘Cavalry’ and ‘infantry’ seem anachronistic terms for use in early medieval Western warfare, a fact which, alongside the other points just made about the nature of the evidence, renders equally pointless academic debates about whether the Anglo-Saxons had ‘cavalry’ or not.  The only real distinction was between warriors who owned horses and those who did not.  Experienced warriors right across the West, from Ireland to Italy and from Scandinavia to Spain, owned horses and used them in warfare.  They rode them on campaign and whilst raiding and they fought from them in battle.  Sometimes, however, it was considered more effective, especially in more serious encounters, to dismount.  Such an action proclaimed that this was a fight to the death, as escape would be that much more difficult, and it simultaneously demonstrated unity and a willingness to share the risks, with the dismounted warriors who did not own horses.  These decisions as to whether or not to fight mounted are noted throughout the period, but they make their way unevenly through our capricious evidential filter.  This has made it a methodological mistake of the first order to base discussions of Anglo-Saxon fighting on the handful of snippets of information about battlefield reality, when compared to the much greater array of evidence linking Anglo-Saxon warriors (like their mainland contemporaries, whom no one doubts fought on horseback when the occasion demanded) with their horses.

Fighting between these bodies of men, whether on horseback or on foot, must have been brutal, and yet sources refer to day-long battles or even to encounters that lasted for several days.  The physical and psychological strain imposed by such fighting surely prevents us from accepting these accounts as depicting unbroken fighting for long periods.  We should probably envisage short spells of close fighting followed by periods of rest (perhaps covered by skirmishing and missile exchanges).  It might be that the nerve of one side could snap after a long day’s fighting, even when it had been generally successful.  The resilience of Viking armies, which seem to have been able to be pushed back for long spells, appears to have sapped and broken the will to fight of their opponents, perhaps used to obtaining the victory sooner in these circumstances.  The Anglo-Saxon Chronicle, for example, records English defeats even after ‘being victorious far on into the day’.

The Vikings’ experience was surely crucial here, as was the fact that the ‘Great Armies’ recorded from the 850s onwards stayed together all-year round and thus, as well as being battle hardened, they were accustomed to fighting alongside each other in different contingents.  Ironic though it sounds, the Viking Great Armies were probably one of the closest approximations to regular troops that one can find in this era.  The Vikings do seem to have been very important in the development of early medieval western warfare.  One reason for that was their tactical ability.  A capacity to absorb punishment without breaking – experienced soldiers would know that continuing to fight, even while retreating, was safer than running away – was one element of this.  Another was their ability to withdraw into field fortifications (as at Brissarthe in 866, Reading in 871 and at Saucourt in 882).  Finally they were capable of launching sudden counter-attacks from out of these bases, often with dramatic results.  Two Northumbrian kings were killed and their army shattered by such Viking use of fortifications at York in 869, Count Robert the Strong (ancestor of the Capetians) was killed in similar circumstances at Brissarthe, and ealdorman Æthelwulf lost his life in the Viking counter-attack at Reading.  These fortifications are another noteworthy element..  Not for nothing has Eric Christianssen listed the spade as one of the key weapons of Viking warfare.  Frequently using a pre-existing villa/estate-centre or a church or monastery as a focus, the Vikings were adept at throwing up a bank and ditch around their camps.  The best-excavated example is probably the camp at Repton, which uses the Anglo-Saxon church there as a redoubt-cum-gatehouse.  This use of fortifications was doubtless at least partly a response to the Christian armies’ greater numbers of mounted warriors – although the Vikings too could fight on horseback, as Regino’s account of the Frankish disaster at the River Gueule (891) makes abundantly clear.  Another response to their deficiency in horsemen might have been a more common decision to fight on foot.  Before the eleventh century, mounted troops had few means of breaking a steady formation of foot-soldiers, which usually resulted in their dismounting too.  Indeed all of the mentions of Anglo-Saxon armies dismounting to fight come from the Viking period, which might suggest a change in this era.

The other ‘outsiders’ sometimes thought to have introduced change into the way in which war was waged are the Arabs, who invaded and conquered Gothic Spain in 711.  However, a close examination of the strategy and tactics of western Umayyad armies does not seem to show any significant differences in these areas from the norms of warfare in western Europe.  Warriors were for the most part mounted, armoured and focused upon hand-to-hand fighting after an exchange of missiles.  A famous passage in Nithard’s History of the Sons of Louis the Pious, describing military exercises at Worms, shows that Frankish horsemen were well capable of mobile fighting styles, throwing javelins and executing feigned retreats, at least in skirmishing warfare (there is little or no reference to such practice in set-piece battles).  Only after the development of the heavily armoured, lance-armed knight in the later Middle Ages does a clear difference seem to emerge between Christian warriors and their Andalusian counterparts. 

More unusual, it seems, were the battlefield tactics of the Bretons and Basques on the fringes of the Frankish kingdom.  Both peoples, it seems made a much heavier use of high-speed mounted attacks, involving the throwing of javelins.  The Basques employed their mountainous home terrain to good effect, launching hit-and-run raids rather than accepting battle in the accepted, close-fighting fashion.  Against the Basques, moaned one Frank after the heavy defeat at Roncesvalles (778), battle ‘was not fought fairly’.  The Bretons might have employed the difficult terrain in their homeland to similarly good effect, wearing down Frankish armies by repeated attacks.

Conclusion

Battle in the early medieval west, then, was sui generis, neither ‘just like’ Roman warfare nor a pale precursor of later medieval battle.  There are features which unify it, in battlefield practice, the flexibility of a warrior’s functions, and in battle’s greater importance in strategy.  Nevertheless, things were not entirely the same across the whole of Europe for the entire period under consideration, as we have seen.  Nonetheless, with the reappearance of mercenaries in the shape of Viking armies fighting for money, and the attendant commodification of violence, in the greater emphasis on fortification and in a shift towards a greater dominance of cavalry warfare, developments were evident in the ninth century that, developing through the tenth century, would eventually lead to the ‘classic’ medieval wars of knights and castles in the eleventh.