It seems clear that things had changed somewhat by the time that authors like Gregory of Tours wrote their works, and this makes it difficult to project their testimony back into the fifth century. The dynamics for change would appear, not surprisingly, to have lain in the ethnic nature of the armies, and in the late and immediately post-Roman system of paying the troops. Gradually, whatever the original details of the settlement, the soldiers paid in ‘thirds’ began to settle on the land, either converting an earlier right to revenue into one of ownership, or, perhaps, an initial grant of possessio into rights of dominium. One might expect this to be particularly common as the soldier in question grew older, married and began to think of settling down. In the late Roman Empire, military service had usually been an hereditary profession. In the early fifth century the empire had enacted that in some cases subsistence payments, made directly from tax-payers, might also be inherited by the sons of senior bodyguards. When military service became associated with an ethnic identity, this was underlined. The sons of Franks or Goths, for example, would be considered Frankish or Gothic too, and thus heirs to their fathers’ status, lands, obligations and privileges. Thus troops probably regarded their tertia as heritable too.
Soldiers also began to buy or otherwise acquire other lands to live on which they regarded as subject to the same exemptions. The tertia granted to barbarians, whether land or revenue, were tax exempt. Roman soldiers had been exempt from the capita, the poll-tax, in any case. The ‘barbarians’ who had become the army in the post-Roman west, regarded these exemptions not only as heritable but also as applicable to all their landed possessions. It would seem that a common political conflict in the sixth century arose between the army (whether it be known as ‘Goths’, ‘Franks’ or whatever) who wished to preserve the ‘logical’ extension of their privileges, and the kings, who attempted to retain the old system in its strict sense, and thus avoid the gradual dissipation of their tax base.
Hand in hand with these developments was the gradual adoption of ‘barbarian’ ethnicities by the free population. The acquisition of Frankish ethnicity in northern Gaul brought increasingly thorough exemption from taxation. It also brought legal privilege. Although a Frank would be liable to military service, this was, to all intents and purposes, a privilege, as the army was the political assembly. Good service brought rewards in royal patronage: grants of land or tax revenue, treasures, and offices in royal service. By about AD 600, Frankish ethnicity had become more or less universal north of the Loire. Those of the free population who had not managed to acquire this ethnic identity lost their free status as well, so that Romans in seventh-century Ripuarian Law are equated with the half-free, and are required to have a free Ripuarian (a Frank) speak for them at law. A similar process seems to have taken place in Visigothic Spain too – by the seventh century, Gothic identity became the norm.
The result of these processes was that by the later sixth century the social group from which the army was raised had become a class of landholders. The nature of service was not defined by landholding, and landholding was often a reward rather than a prerequisite for service, but the members of the army were nevertheless, on the whole, landowners. In some areas they were those who claimed a particular ethnic identity: Frankish in northern Gaul; Gothic in Spain.
A similar situation may have pertained in Italy after the Lombard invasion. The period immediately following the initial Lombard attack is very badly documented but it would seem from a rather garbled account in Paul the Deacon’s Historia Langobardorum that the Lombards adopted the system which had been in widespread use in western Europe up to that point; the Romans seem to have been divided amongst Lombard warriors as tributarii (tax-payers). However, this system would appear to have been short lived. There is no very clear indication of it in the earliest Lombard law-code (Rothari’s Edict of 643), although the king does describe the political community of the realm as the army.
This situation was not universal, however. A later sixth-century northern Gallic capitulary refers to Roman soldiers and some Romans at least continued to serve in Visigothic armies. In Aquitaine, Roman landholders had, by the sixth century, become subject to military service, this clearly being a development of the situation at the very end of the empire. Gregory of Tours’ Histories abound with references to the ‘men’ of particular civitates under arms. He refers to king Guntramn of Burgundy calling up those liable for military service, and punishing those who failed to turn up. The implication of his account is that landowners had become liable to military service whereas tenants were exempt. This body of landowners is probably what Gregory of Tours means when he talks of the populus of a civitas. It also seems that populus acquired a military sense in his writing, which is interesting, as the vernacular languages also came to use their words for ‘the people’ as a synonym for ‘the army’. Minor populus (‘the lesser people’) is used on one occasion to mean the rank and file of an army.
The army, as implicit in the foregoing discussion, was still raised by royal officers from old civil administrative units, normally the civitates but also its sub-divisions, the pagi. Gregory, in his Life of the Fathers, seems to refer to a man from the pagi on campaign. In some parts of Gaul, at least by the seventh century, the pagi had replaced the civitates as the basic units of administration.
As before, bodyguards remained the core of military forces. Sources refer to the personal retainers not only of kings but also of dukes, counts (and their wives) and even bishops. Such troops are, as mentioned, called bucellarii in Visigothic Spain, although the term saiones is also used. Once granted his weapons, a Gothic bucellarius was permitted to keep them for good. Any landed property he received, however, remained ultimately in the dominium of his patron. In Lombard Italy, by the seventh century at least, members of the royal retinue were known as gasindii. Such royal bodyguards were known as antrustiones in Merovingian Francia and as gardingi in Visigothic Spain, and are, in all cases, a slightly more complex category than might at first be expected. Although some at least may always have been in attendance on the royal person, by the time anything substantive can be said about them (in the seventh century) they seem to have been high-ranking, probably fairly experienced soldiers, in many cases found not at court but elsewhere in the kingdoms, commanding their own forces. By this time they seem to be a category of service aristocrat. Whether this had always been the case, or whether it was a transformation which had taken place since the sixth century, is difficult to establish. Merovingian evidence suggests the former alternative. In the sixth-century Gallic narrative sources such ‘old soldiers’, close to the king but only in attendance upon him intermittently, are referred to by a different term, leudes, and there are two Frankish references to spatharii – one of the terms for late Roman imperial guardsmen – who are certainly such palatine officers. This last term is also used in an earlier Burgundian source. However, early sixth-century Salic Law refers to two types of royal retainer: the member of the royal trustis or retinue (the antrustio), and the puer regis (king’s boy). The former is a powerful figure, with a high wergild, as a result of his connection with the king, but seems already to be commonly found away from the court. Often assumed to be a class of unfree or semi-free warriors, the pueri regis clearly simply represent a junior, lesser level of the royal bodyguard. The general picture would seem to be that young men were sent, or promoted, to serve in the royal bodyguard, where they would, in addition to serving as bodyguards, receive a military and probably administrative education. They would occasionally be sent out to enforce royal orders or administration in the kingdom. Gradually, good service would lead to more substantial rewards in land, titles and other valuables and the guardsman would move out from the court either to take up a military/administrative office (as a count or duke, for example) or settle on his lands as an antrustio (or gardingus or gasindius), returning to court for major occasions or for shorter spells of service, perhaps commanding and teaching the junior pueri. This pattern, of permanent service at court gradually giving way to a position as a landed aristocrat, remains common throughout the early medieval west. In Spain, King Theudis took the interesting step of raising a personal army of 2,000 slaves from the estates of his wealthy wife. Quite apart from providing him with a handy independent force, this seems to have been the forerunner of the extensive use, in Visigothic Spain, of personal retinues raised from slaves and other lowly inhabitants of one’s estates.
Discussion of the pueri introduces the point that age and position in the life cycle played a very significant role in the raising of early medieval armed forces. In the immediately post-Roman period it seems to have been particularly important. The process of male socialisation was a long one which began with legal majority between about twelve and fifteen, and did not end until the man married and began a family of his own, perhaps in his late twenties. The Latin term puer, as well as meaning ‘boy’ in the sense in which we would understand it today, could be used to mean any male who had not yet married and settled down. This is the sense in which it is used in the context of the pueri regis. Once he reached legal majority, it seems that a Frankish boy would be sent as an apprentice to an older male from another family, perhaps a royal officer or an ecclesiastical potentate. From there the puer might come to the notice of someone more powerful, even the king, and be transferred to their household. Such apprenticeships could clearly be hard on the young apprentice, who was expected to get his master’s permission before he could marry, an act which symbolised his change of status and release from the bonds of his apprenticeship. This career path, and military service, was thus an important engine of social mobility and opportunity.
As well as forming the retinues of older and more powerful figures, bound by what might be termed vertical ties to the latter, young warriors might also be bound by ‘horizontal’ ties into bands of contemporaries. Such warrior bands are probably those called contubernia in Salic Law, envisaged as involved in house-breakings and kidnappings, precisely the sort of errand on which pueri are often engaged in the narrative sources. The bands of young men within a royal or aristocratic household could be called a contubernium. Membership of such a band created long-lasting bonds and a group of young men could be called contubernales. Thus it seems likely that in many sixth-century armies, there was a division into older and younger warriors. Whether such warriors were formed into separate units, as in other age-based societies, is perhaps unlikely but far from impossible. This would be more likely in those areas north of the Loire where the social role of age appears to have been more significant.
The development of military service in Britain is even more difficult to establish. Here we must work largely from inferences from the archaeological evidence. Excavated cemeteries provide the primary form of evidence, but, as in Northern Gaul, furnished inhumation with weapons does not become common until the sixth century. In the fifth century, weapon burials seem to represent displays of prestige by the local aristocracy; they are fewer than would be the case in the sixth century. Clearly, military power was an important factor in local leadership. What is perhaps more interesting for our purposes is that in the middle quarters of the fifth century, south of the Thames, a style of metalwork became common which was descended from the official metalwork of the late Roman Empire, and in particular the late Roman army. This ‘Quoit Brooch Style’ is found on belt buckles, clearly continuing the Late Roman fashion for displaying power and rank through elaborate belt sets, and on some weaponry. This insular development of late Roman military fashions, and its use in burial rituals to demonstrate power, suggests that in Britain we may also be faced with a situation where a late Roman field army gradually evolved into the basis of a post-Roman kingdom, as in Northern Gaul, Aquitaine and Italy.
By the sixth century furnished inhumation had become much more common in lowland Britain. Recent work has shown that weapons, the primary symbols of masculinity, were used as symbols of age. Unlike in Gaul, male children were commonly buried with a spear. However, shields and other weapons tended to be acquired in the late teens and early twenties. Again, this implies that only at this age was it deemed appropriate to symbolise a man’s male identity with a full set of weaponry; the symbolic category of ‘warrior’ seems to have begun at this age, as in Gaul. The overall picture suggests that by the sixth century a system similar to that in Gaul was in place, where armies were levied from broadly defined categories of adult males. The right to symbolise one’s status through weapon-bearing seems to have been held quite widely. It seems likely that this weapon-bearing class may have been that which claimed an English identity. The construction of several substantial linear earthworks at this time implies that rulers could levy significant numbers of men through this system.
In the west and north of Britain, there is less evidence to help us. Gildas mentions, in connection with the employment of Saxon federates, hospites and the payment of annonae (food and other supplies raised as a tax), which would indicate a survival of Roman commissariat systems. However, the chronology of Gildas’ ‘historical section’ is notoriously difficult to establish, and he may actually be describing a late fourth-century situation. Though it is difficult to be sure of when Gildas was writing (estimates range from the late fifth to the mid-sixth centuries), we might at least state that when he did write he and his audience were familiar with such institutions. The construction or refortification of numerous hillforts, and the evidence found inside them of craft specialisation and industry suggests that magnates had considerable control of surplus and the ability to mobilise significant amounts of manpower. It would seem that, in the fifth and sixth centuries, the rulers of these areas were rather more powerful than those in the lowland zones. How armies were raised is difficult to know in any detail. From the available evidence we might suggest either that kings raised troops through an intermediary tier of lesser royal officers based in local hillforts, or conversely, that kings mustered armies by calling upon lesser local magnates whose power centred on these fortifications. The two alternatives are equally plausible from the data available, but have rather different implications for the relative strength of the kings; the first suggests powerful, centralised states with effective royal administration of local areas, whereas the second implies a more fragile situation where powerful independent lords would need to be bound into the polity through strategies such as the bestowal of prestigious gifts.
In the former ‘Free Germany’, similar points may be made from the cemetery evidence as were made for Anglo-Saxon England. In Alamannia, the archaeology suggests change between the fifth and sixth centuries. In the fifth century a number of fortified hilltop centres exist, similar to those in western and northern Britain. These largely appear to have been abandoned around AD 500. At the same time there is a change in the nature of the cemetery evidence. As elsewhere, furnished burial becomes more common within cemeteries. Some sixth-century sites imply that age played an important role in the levying of armies. Children are sometimes buried with arrowheads, and adolescents sometimes have a scramasax or a spear. The full weapon set is reserved for adults above the age of twenty, however. These changes in the archaeological record probably indicate a change in the nature of local authority in Alamannia, from a network of powerful independent local rulers to one wherein power was held by officers appointed by the Merovingian kings and their representatives.
Further north, it is difficult to say very much. The Saxons and Thuringians proved significant foes for Frankish armies in the fifth and sixth centuries. The Thuringians indeed established a fairly large kingdom and were deemed worthy of alliance by Theoderic of Italy until Clovis’ successors eliminated the realm in the earlier sixth century. This may have led to a situation analogous to that in Alamannia, with the removal of earlier power structures and the introduction of a state of affairs in which appeal to Merovingian authority was used as a basis for local power.
In Scandinavia we have only archaeological data to help us but these have been carefully recovered and studied. The Danish bog finds, which cease to be deposited around AD 500, have been studied and shown to represent armies largely made up of lightly armed infantrymen, led by a smaller number of better-equipped men, who possessed horses as part of their equipment. This suggests some stratification within the armies. However, one must be cautious in using this evidence in too straightforward a fashion. Throughout Scandinavia at this date, fortifications show an ability to organise significant amounts of manpower. In the west of Norway, a detailed study of large boathouses has revealed concentrations of such buildings which could house ten or more ships. Such numbers of vessels would need more manpower than the local population could furnish to put to sea. Taken in conjunction with neighbouring high-status settlements and burials, it seems that these clusters of prestigious sites reveal the centres of magnates who could call upon the manpower of more extended areas to raise naval forces. The mechanisms whereby this claim on manpower was articulated are not known.
Overall, in much of the west, the sixth century is a period when kings seem to have raised armies by a general levy on certain types of free landowner. In many areas this landowning class was defined by a claimed ethnicity (Frankish, Gothic, Lombard, Anglo-Saxon), which also represented a direct link with the king. Elsewhere the levy may simply have been on all landowners. Such a ‘horizontal’ levy was carried out by royal officers within the kingdom’s administrative districts, and focused on the household retainers of those royal officers. Thus in this situation the king retained a significant amount of power. The army was mobilised according to a general obligation to the king, an obligation complied with because attendance at the muster was one way in which claimed ethnic and other statuses were justified. Furthermore, the officers mobilising and commanding the troops were in many cases service aristocrats whose power and authority over other freemen depended entirely upon their tenure of royally bestowed office, and receipt of other forms of royal patronage. Even where, as for example in Aquitaine, the aristocrats who led the regional contingents came from independently wealthy noble families, they still competed for office, for it was largely their participation in royal administration which cemented and safeguarded their families’ local standing. In this context, the king held most of the aces, and could use the army effectively as a royal coercive force. This can be seen very clearly in sixth-century Merovingian Gaul, where the kings frequently called out armies to crush aristocratic dissent.
(Source: “Warfare and society in the barbarian West, 450–900”, by Guy Halsall)
Research-Selection for NovoScriptorium: Anastasius Philoponus