Featured post

Gender in the Merovingian World

Monday, 2 December 2013

Warfare in the Early Medieval Western European Kingdoms, 600-1000 (the director's cut)

 [This is a long version of the chapter I contributed to the Cambridge History of War, ed. A. Curry.  I cut this by a third before submission.  I did that over two years ago - this version was written three years ago.  As it happens I have heard nothing about this project for some time.  Indeed I have never really received acknowledgement of the final version of the chapter, which you might regard as a tad shoddy. Be that as it may, it gives me no cause to reconsider posting the fuller 'director's cut' here for your enjoyment or scrutiny.  
We have discussed (here and here) one of the elements of this that I cut out.  
Be warned though: some cock from the Society for Creative Anachronism claims here that I am not a specialist on the subject and have nothing new to say on the topic; perhaps Routledge (publishers of Warfare and Society) and the editors of this putative Cambridge volume should have asked him instead.  This is exactly the sort of arrogance on the part of amateur military enthusiasts that I have discussed before.  OK, he spent money on the book, he's entitled to write what he thinks of it on the web, but by the same token I am entitled to say what I think about that!  If only I'd written to him.  I could have saved myself all that time and effort reading early medieval charters digging out references to warfare and military organisation, putting together a volume which most silly old academic medievalist reviewers thought was the first good book on the subject, with arguments that have been much discussed (equally by people who should simply have asked our man, who knew it all already).  I could also have asked him to fill me in on all that extra Scandinavian material that I could have used.  Durrr.
Anyway, as it is 3 years old, I don't guarantee that I still agree with all of this but mostly I guess it's roughly where such thinking I have on this subject these days is.]

Part 1: c.600-800[1]
Change around 600 AD: a military revolution?
This chapter opens during a period of important change.  Between the late fifth and late sixth centuries, Western European military forces had recognisably been descendants of the last manifestations of the western Roman army, as a brief comparison will demonstrate.[2]  Late Roman troops had sometimes been paid via the delegation of fiscal revenues from particular tax-payers and, as earlier, received allotments of land on retirement.[3]  Their service was inherited by their children[4]and they were, furthermore, exempt from certain taxes.  Between the fourth and fifth centuries the imperial armies evolved a series of identities based around oppositions to traditional civic Roman ideals of masculinity.  Much of this turned on ideas of barbarism, something enhanced by possibly increased recruitment beyond the frontiers and the commensurately greater opportunities for non-Roman soldiers to rise to higher command.[5]  As the size of the territory effectively governed from the imperial capital at Ravennashrank during the fifth century, and with it the available taxation and recruiting bases, the enlistment of warriors from outside the Empire and their political pre-eminence grew further.[6] 

Similarly, post-imperial armies were frequently raised from people claiming a non-Roman identity: Franks in the north and Burgundians in the south-east of Gaul, Goths in Spainand Italy.[7]  Such warriors seem to have been paid, partly at least, by the delegation of taxation levied on the civic ‘Roman’ population.[8]  The lands they held were tax-exempt, as had been those of Roman soldiers, and their military responsibilities seem to have been hereditary, linked to the inheritance of their ethnic identity.  Military service was structured around the life-cycle with young men serving as pueriin the households of older warriors, royal officers or even those of the kings themselves (the trustis regis or schola palatina).[9]  If they were successful in this service, then in their later twenties they left the service, married and established their own households.  At this age, apparently, their ethnic identity was formally acquired.  Such older warriors would then serve when summoned, along with any young men who had in turn taken apprenticeships with them.  Some held posts within the royal administration and others doubtless continued at the palace, training the younger pueri regis. [10]  In some areas, such as Aquitaine or possibly southern Italy, the basis of military service appears to have been different, turning on land-ownership rather than ethnic identity.[11]  Quite how this related to the payment of taxes is currently unclear.

In the period c.500-c.600 warriors were raised according to kingdom-wide principles of service, levied by administrative district under the command of royal officers, and the holding of the relevant offices was generally in the gift of the king, rather than hereditary.  Because, in most post-imperial regions, the aristocracy relied upon the favour or patronage of the king, either for its local pre-eminence or position within the ‘pecking orders’ of the class, the king could raise the army to impose his will within the realm, including upon rebel or otherwise recalcitrant aristocrats.  This is very clear from our best documented kingdom, Merovingian Gaul, but the point appears to be generally applicable across most of the West.  This is noteworthy because it is thus evident that, although the armies of the later fifth and sixth centuries were by no means regular forces of the old Roman type, they still functioned as independent coercive forces (and because of that it fact it remains perfectly justified to see these realms as states, up to about 600). 

Nevertheless, this situation contained within it the seeds of change.  The first was the common ethnic nature of military service.  Ethnicity was not fixed and, as alluded to, seems frequently to have been something acquired at the age where a male established a household.  Further, unsurprisingly, it appears to have been gendered in the immediately post-imperial centuries: ethnic identities are ascribed overwhelmingly to men.  However, the attractions of non-Roman ethnicity[12]inevitably meant that it became steadily more common and, by 600, more or less universal among the aristocracy and free landholding classes in most areas of the West.  Simultaneously, it appears to have been more universally assigned within families of such ethnicity (to children and women as well as adult males).  Given the frequent tax-exemption of people subject to military service, this development will have meant the spread of liability for duty in the army, and one presumes a concomitant inclusion within their ranks, of freemen less able to bear the costs of such service.  It will also have reduced the amount of land subject to royal taxation and, through the inheritance of ‘military lands’ by women of non-Roman ethnicity, probably led to its tenure by people considered incapable of performing military service.  These points alone would probably have produced significant change in the raising of armies but they constituted only one group of factors among many that made the decades around 600 a period of intense transformation, detectable across most areas of western European society and economy.  Another key development in many regions was a growth in the power of aristocracies, vis-à-vis the monarchs, and thus it is unsurprising that during the seventh century we can detect significant changes in the nature of armies and the ways in which they were levied.[13]

By the mid-seventh century it is very difficult to see any trace of the post-imperial means of raising armies.  Mentions of the ‘men of Mainz’ and of the pagenses of Saintes in a campaign of 639 may be the last allusions to levies based on the old administrative districts (the civitates and pagi), such as had been common in the sixth century.[14]  Instead, references to armies take us in two complimentary directions.  On the one hand we begin to encounter what might be termed ‘select’ levies – scarae.  The word is cognate with the English words ‘shear’ and ‘share’ and implies a select band, cut (or sheared) off from the mass.[15]  With the presumed spread of liability for military service (if assessed according to ethnic identity) this is unsurprising; a levy of the free population would be entirely impractical.  In the social and economic context of the early middle ages, armies larger than a few thousand men were unfeasible.  The term ‘select’ is unfortunate, though.  The ‘ethnic’ military land-owners of the immediately post-imperial era presumably represented (initially at least) an equally select body and one imagines that even during the sixth century, royal officers – counts, dukes and centenarii – had to select those best suited to serve from steadily larger numbers of theoretically eligible men.  What changed between the sixth and the seventh century was not whether or not warriors were drawn from a select group but the means of selecting such troops.[16]   

How this ‘shearing’ was carried out is suggested by the other prevailing aspect of seventh-century military organisation: the aristocratic retinue.  Sixth-century military households, especially those of royal officers, were mentioned above but they seem to have dominated seventh-century army composition.  Their precise nature also appears to have changed.  Seventh-century sources widely acknowledge the existence of freemen dependent upon their more powerful fellows, and a class of aristocrats whose power was (unlike in the sixth century) quite independent of royal service.  Evidence of more secure tenure of large estates by such magnates is more easily found and these lands were used to reward followers.  This seems to have meant that, whereas in the sixth century the older warrior performed his military service to the king according to general systems of obligation, in the seventh, even after leaving the age-group of the pueri, the warrior still performed military service in his lord’s retinue.  In diverse parts of western Europe such nobles appear to have interposed themselves between the king and the remainder of the free population.[17]  Around 600 the last vestiges of Roman taxation disappeared, largely because these imposts had passed into the control of estate owners.  We find legislation concerned with the frequent intrusion of magnates into the operation of royal justice, protecting their ‘satellites’ from the sentence of judges.  At the same time, aristocratic dynasties become more visible, frequently monopolising administrative offices whose bestowal had hitherto been entirely within the royal gift.[18]  We are as yet some way from the situation where counties or duchies were hereditary but the detectable sequences of counts in particular areas from the same families strongly suggest that, although such a title had no value without the legitimation of royal appointment, these kin-groups had a clear expectation that the king would appoint one of their members when the post became vacant. 

Juxtaposing these general developments, it seems that during the seventh century the magnates were able to insert themselves into the means of levying the army.  When a military force was required, even if legitimised by royal summons and nominally employing old ideas of liability for military service, in practice the local counts (now, as noted, most often members of the most important local families) would select or ‘shear off’, from all those theoretically liable for military service, the most politically and socially important land-holders and their allies and dependents.  As the army was the most important political assembly within a kingdom (see below), the choice of whom to summon and whom to leave behind was a significant source of local power and patronage.  What happened in most of the seventh-century West might, then, not unreasonably be termed a ‘privatisation’ of the army.  I have distinguished the different means of levying an army by the terms ‘horizontal’ (levied according to a kingdom-wide ‘flat rate’ by royal agents working within specific royal administrative districts: the sixth-century model) and ‘vertical’ (raised down chains of dependence within aristocratic estates and dependencies as in the seventh century); this may yet suffice as a crude short-hand.

This is very significant in determining how we classify polities of this period.  It was argued above that, alongside the general continuation of taxation, the ability of rulers between 400 and 600 to summon and employ a coercive military force allows us to think of their kingdoms as ‘states’.  With the change in the raising of armies just described, this situation altered profoundly.  From c.600, the mustering of armies was increasingly (with temporal and geographical exceptions) a matter of negotiation between the royal court and local and regional aristocracies.  This in turn meant that in practice it was concomitantly difficult for a monarch to summon an army to resolve internal difficulties such as recalcitrant or rebellious noblemen.  The army ceased to be classifiable as an independent governmental coercive force, which must have profoundly affected kings’ ability to harness the surplus of their realms beyond their own private estates.  Against recent historiographical fashion, this makes it unjustifiable to classify as states the kingdoms of the period after c.600.[19]  None of this, it cannot be emphasised too strongly, implies that these polities were not cohesive or that kings did not wield considerable authority.  Dynastic legitimacy and other ideological strategies could mean that the royal court remained the essential focus of politics.  Nevertheless, without the ability to raise armed forces independently of regional and local magnates in order to impose their will upon the diverse localities of their realms, it is unhelpful to call these kingdoms states.

Intriguingly, a significant change in armament seems to have occurred at this time.  We can detect this in various regions from the practice of burying weapons in graves, something that provides, by early medieval standards, an enormous sample (many thousands of items) of contemporary weaponry, although unevenly distributed geographically and temporally.[20]  Between c.575 and c.625, several hitherto common items disappear from the archaeological record.  In Francia especially, these include the throwing axe (francisca), and the heavy, iron-shafted barbed javelin (ango).  One-handed battle-axes and, in England, certain types of javelin also cease to be found.  Simultaneously, a change in defensive weaponry occurs.  Fifth- and sixth-century shields were quite small and their bosses frequently ended in a disc, probably used for catching an opponent’s blade in a ‘fencing’ style of warfare.  In the seventh century, however, these ‘bucklers’ are replaced by larger shields with longer and heavier bosses, perhaps more suited simply to shoving or punching an enemy.  Contemporary with these changes, the sword becomes less frequent in the record while the one-edged dagger (scramasax) becomes longer, broader and weightier, resembling a machete.  This is perhaps paralleled in developments in spearheads which likewise become larger and heavier.  It is naturally very risky to deduce a shift in tactics from a change in weaponry but the transformation of armament between c.575 and c.625 seems to point in one general direction: from a possibly faster, more open, type of warfare with small, easily mobile shields and much use of specialised missile weapons (angones, franciscae, ‘corrugated’ javelins) towards combat centred on close-packed hand-to-hand fighting.  The larger, heavier shields and spears seem adapted to this type of warfare and the broad, chopping, single-edged scramasaxmore suited to it than the long, slashing, two-edged broadsword.  Indeed the scramasax apparently combines the best features of the sword and battle axe, the latter of which disappears from the record until the Viking era.  Surviving evidence unfortunately provides no real clues as to whether defensive armour became more common, as one might expect.  Helmets and body-armour are proportionately more frequent in the seventh century than the sixth but we cannot deduce much from this.  The interment of armour was geographically very restricted, in a way that cannot reflect its actual frequency, and thus probably more affected by ritual demands than the burial of other items.  Furthermore, outside southern Germany most surviving examples come from entirely untypical burials (such as the ship burials at Sutton Hoo in East Angliaand at Vendel and Valsgärde in Sweden).[21] 

Relating these developments in armament and the putative tactical change deduced from them to the transformation in the raising of armies is difficult.  Some of the weaponry which dropped out of use – especially the francisca – apparently required specialist training to use effectively, whereas one might, at least superficially, wonder whether close-fighting ‘shieldwall’ tactics were more suitable to larger, comparatively less well-trained forces.  However, as noted, seventh-century armies were perhaps no more select than sixth-century; the means of selection changed.  Further, and this might be crucial, close-fighting techniques probably required more expensive protective equipment, notably helmets and body-armour, which could have restricted participation in it to those with the economic wherewithal to so furnish themselves and their followers.  This would tally with the growth of aristocratic power discussed above.  It is also possible that the coherent employment of close-fighting techniques required the elements of a ‘shieldwall’ to have more frequent training as a body.  This might be more feasible within an aristocratic retinue than in an irregularly assembled conglomeration of land-owners.  In the absence of more evidence, however, this suggestion can only remain as such and there are arguments that one might present in opposition to it.

In much of north-western Europe the period from c.600 onwards was one of economic expansion, which might well have enabled slightly larger armies to be mobilised, at least for large-scale conflict.[22]  This and the increased private resources in the hands of local aristocrats could also support a growth in the frequency of metal body-armour.  It might also have been more feasible for aristocrats to equip larger followings with probably the most expensive item of a warrior’s equipment, his horse.  However logically likely this might seem from the period’s general economic developments, and although there are some indications of this in the data, an increase in the proportion of an army’s mounted troops is as impossible to confirm from our evidence as a growth in the frequency of armour and helmets.  As with the other suggestions linking social and military developments in the period after 600, it might nevertheless be worth retaining as a working hypothesis.

The evolution of systems of military service and the rise of the Carolingians (c.650-c.800)
Just as the military forces of the sixth century can be seen as the last, if possibly rather shadowy, incarnations of a Roman type of army, seventh-century armed forces might be recognisable as the first, equally blurry, manifestations of a general type of ‘feudal’ (and one must emphasise those inverted commas!) army familiar from most of the remainder of the ‘medieval millennium’, perhaps until the creation of the fifteenth-century ordonnance companies and their like.  The late seventh and early eighth centuries, however, saw further important developments which might make armies after that point even more recognisable to students of the central Middle Ages and indeed the ways in which armies were raised through the remainder of the period covered by this chapter, and beyond, are best discussed as developments of the seventh-century ‘template’.[23]

In traditional historiography the early eighth century saw the introduction of ‘feudalism’ in its classical sense: military service performed to a lord in return for land, confirmed by the ritual swearing of oaths of allegiance or ‘fealty’.  This service was, in traditional reconstructions, performed on horseback as armoured cavalry, the increased cost of which required the warrior to be given the necessary estates from which to support himself and bear the cost of his weaponry.  In classic formulations, this involved Frankish rulers, in particular, confiscating Church lands to support their horse-borne warriors, but sweetening the pill by having the land only ‘held’ by the soldier, who paid rent to the Church which owned the land.  Thus we have the tenure of estates, whose ownership remained ultimately in the hands of a greater lord, secular or ecclesiastical, in return for the payment of rent or performance of usually military service, the arrangement being solidified by the swearing of oaths.  All in all, this established the principal elements of the famous so-called ‘feudal system’.[24]

Recent scholarship has called into doubt almost all of this traditional reading.[25]  Nevertheless, even if its precise details can be shown to be incorrect, in its general outlines there may yet be something to be said for the old interpretation.  Something important took place around 700, something indeed related to the ways in which land was held and to the relationships between warriors and their lords on the one hand, and between powerful aristocrats and the king on the other.  The early seventh-century changes described above centred on the creation of a powerful aristocratic stratum with more secure control of landed resources than hitherto, and with other freemen dependent upon them.  This led to a change in the precise nature of the aristocratic armed following and in the composition, and means of summoning, armies.  However, from the surviving evidence, the aristocrats’ means of rewarding their followers seemingly remained the outright gift of land or other resources.  Although ties of dependence or allegiance probably remained, such gifts had the concomitant effect of reducing the resources controlled by the aristocrat.  The common existence of partible inheritance, requiring all sons (at least) to be bequeathed more-or-less equal shares of the patrimony, and the requirements of land-grants for dower and dowry, as well as other gifts, further meant the probable break-up of an aristocratic estate every generation.

This in turn was probably an important reason why politics remained focused on the royal courts in spite of, in Francia, the frequent (indeed usual) minority of the Merovingian kings and, in Visigothic Spain, serious dynastic instability.  Aristocrats had to return to court to acquire new lands to replace those bestowed upon followers and to have titles, the other key support of local and regional pre-eminence, confirmed and legitimised.  Aristocratic families tried many means of combating this situation.  The Roman system of testamentary disposition (by will) was one way by which a greater share of the estate could be left in the hands of one son.  Simultaneously other documents begin to survive or be alluded to, through which other demands of customary law could be circumvented.  Across Europe aristocratic rural ‘family monasteries’ more frequently begin to be attested in the seventh century.  By bestowing lands upon a monastic foundation, over which control was retained, the resources could also be kept intact.  Monastic lands were not subject to partible inheritance but could be loaned, by the abbot (usually a family member or protégé), to one member of the dynasty.[26]

However, in factional politics, such as dominated Spainand, from the middle of the century, Franciathe situation described above spelled serious problems for aristocrats excluded from access to the royal court.[27]  Such individuals found themselves unable to receive royal grants to replace lands given to followers and prevented from receiving the royal legitimation of titles essential for the acceptance of authority.  Most of the means, just discussed, of circumventing the divided inheritance and other aspects of customary law leading to the fragmentation of estates were of only limited utility.  In factional civil war, moreover, the winners showed themselves only too ready to depose the abbots of family monasteries and replace them with their own adherents, thus acquiring the institution’s landed resources.  It is therefore probably not surprising that aristocratic families evolved new means of dealing with the dangers of this political situation.  These means clearly evolved from the previous state of affairs.  The terms precariaand beneficium are both encountered in Roman law, even if neither carried quite the same sense as it was to have in early medieval land-holding, to describe situations where a gift was not regarded as necessarily permanent, and could be revoked later.  In early medieval terms, there appears to have been little real difference between lands held as precaria or beneficium, the difference simply relating, on the one hand, to the way in which the gift had been obtained (through a plea or precaria) and, on the other, to the nature of the grant (as a benefit, or beneficium).  In the post-imperial world, precaria are found in relationship to the stipends of church officials.  As the Church’s patrimony was not supposed to be dissipated through gift or sale the only real means of giving ecclesiastical servants land from which to draw a livelihood, was through lending it to them, as precaria.  As ecclesiastical landed endowments grew and became too large to manage directly, the result was that similar precaria were granted to laymen, in return for rent.[28]

What seems – crucially – to happen from c.700 is the extension of these types of land-grant to secular estates.  We begin to encounter references to lands held by lesser freemen, as precaria or beneficium, from more powerful secular aristocrats.[29]  This is important because it meant that noble families could now reward their followers with lands while not actually diminishing their patrimony, as the land remained (at least theoretically) ultimately under the grantor’s control.  This in turn meant that such aristocrats’ power could be independent of the royal court.  No longer did they need to retain access to court patronage to maintain the extent of their landholdings.  Even the loss of control over bishoprics and monasteries might be survived without implying a concomitant loss of the ability to reward or maintain followers.  This was decisive in the relationships between central, royal government and local or regional power.  In Gaul the Merovingian royal house only became obsolete from this point.[30]  Regions such as Aquitaine and Provence drifted away into effective semi-detachment as local and regional pre-eminence could be maintained without reference to the control of the central court.[31]  Thus, the traditional view of the creation of ‘feudalism’ had correctly identified an important – indeed crucial – period of change but in detail its understanding and explanation of those transformations were flawed.  In purely military terms, the aristocratic retinue’s importance was further underlined and the links that bound such followings were further reinforced.

As Charles Martel’s control over Francia became more secure, the use of Church lands to reward followers seems to have become more widespread.  Indeed by the middle of the eighth century something like a regulated system was introduced, setting out the rents involved and some checks and balances to avoid abuse.[32]  This earned Charles Martel ecclesiastical opprobrium but provided a solid basis for the support of his well-equipped and successful armies.

From the early eighth century, the term vassus(‘vassal’) is frequently used to describe warriors in an oath-based relationship with a lord.[33]  The word itself appears to have been old, possibly of Celtic origin, but its pre-eminence may, it has been suggested, have stemmed from a need for new terminology in a novel political situation.  Old words describing these sorts of links, such as antrustio, had too many connections with the royal house so a replacement was needed.  It must, however, be stressed that at this stage such relationships by no means represented any sort of ‘system’ (if indeed they ever did).  Precarial ownership of estates, held from someone else, existed, as did oaths of vassalage but the extent to which the one was directly linked to the other (if at all) is most unclear.[34]  There is even less trace in the data from this period of any specification of the types or amount of service expected.

This was the type of army that participated in the Franks’ successful wars of aggression during the eighth century.  The frequency with which these campaigns took place bred a class of very experienced warriors, with experience grew military competence and the increased chance of battlefield success and survival.  In turn the latter gave rise to ever greater confidence and high morale.  In these circumstances it is unsurprising that the various aristocratic factions in Francia gradually lost out to the victorious Carolingians as warriors joined Charles Martel, or that, under Martel and his sons and grandsons, the Franks’ frequently less experienced opponents so rarely put up effective battlefield opposition.[35]

Spainand England
The general pattern of developments sketched above has been based squarely upon the evidence from Gaul, or Francia as it had come to be called.  Nevertheless, the general outlines seem to pertain to most areas of western Europe, even if in detail the precise explanations must differ.  Visigothic Spainwitnessed an analogous shift from the sort of royal army raised from administrative units and commanded by the king’s officials towards one based much more squarely around the personal followings of aristocrats.[36] Kings legislated to attempt to reserve the control of these retinues to themselves.  As hinted above, some dynamics of the change might have been analogous to those in Francia.  One distinctive feature of the Visigothic situation, however, was the military service of servi (slaves).  It is not impossible that the slaves that the kings required their landlord-warriors to bring to the army were intended to serve as baggage attendants and to guard the army’s camp rather than fulfil a formal battlefield function.  However, Egica called up all former slaves freed by the king on pain of their return to servile status.  This might have been an attempt to create a military force directly dependent upon the kings, as might Egica’s predecessor Ervig’s summoning of servi fiscals(fiscal slaves).[37]  An army raised from royal estates might form the nucleus of an independent royal coercive force.  Indeed there seems to be something of an Iberian tradition of slave-soldiers during the earlier Middle Ages, beginning with the sixth-century king Theudis’ levying of a force of (allegedly) 2,000 men from his Hispano-Roman wife’s estate slaves through to the use of slave soldiers by the tenth-century Caliphs of Cordoba.[38]  However, the ‘problem’ of Visigothic ‘slave soldiers’ may be more apparent than real, as more recent studies have suggested that Spanish servi, rather than being slaves as such, were simply the tenants of aristocratic estates.[39]

Even Anglo-Saxon England, so often considered in isolation from the European mainland, the same general outlines can be observed.[40]  In the immediately post-imperial era, it can be suggested from archaeological burial data that military service, as elsewhere, might have been linked to ethnicity.  When written records begin to survive, although they are so scarce as to permit alternative readings, a development can nevertheless be suggested that would be recognisable from the Frankish account.  Earlier in the seventh century it seems that lands were granted to warriors who had successfully served an apprenticeship in a retinue.  Lands were granted in perpetuity but subject to partible inheritance and other customary demands that led to the fragmentation of estates each generation.  In the course of that century, therefore, we can detect the same strategies being employed to circumvent this situation as we have seen in Francia: the creation of aristocratic religious foundations and so on.  Around 700, as in Visigothic Spain, kings legislated to ensure that retinues based around lines of dependence were still ultimately liable to royal military service.[41]  Such laws are often assumed to reflect long-standing traditions but it is surely more plausible that they were active attempts to create accepted norms in periods of change.  Kings began to specify, when they granted land to their nobles, that they were still expected to perform military service and this might be reflected in the apparent concern of lesser aristocrats to clarify what ultimately remained their property in gifts to their followers.  In the raising of armies, the Anglo-Saxon kingdoms do not seem dramatically different from those of the Franks.

The Lombard exception
The principal exception to the general pattern just established, of the gradual evolution of armies from royally-controlled coercive forces into conglomerates of aristocratic retinues, whose service was not entirely reliably assured by the monarchs comes not from England but from Italy.[42]  Evidence about military service from seventh-century Lombard Italy is not very forthcoming.  Historians have often assumed that the picture of the Lombardkingdom that becomes visible in the eighth century is applicable in the seventh but this seems a shaky assumption.[43]  If anything, the picture one might draw from the archaeological and other data from the seventh century might be one of a weaker central authority and greater power held by the regional dukes.  What little information one can glean about the seventh-century army suggests one not very different from its contemporaries in Francia, Spainand Britain, based around aristocratic retinues.  This, however, appears to change in the eighth century when, especially in the long reign of Liutprand (712-44), royal power grew significantly, not least in association with military expansion and the absorption of hitherto independent territories.  The eighth-century Lombard army shows some similarities with those of the sixth century in other areas of the West.  Through various mechanisms, the king established a link with the lesser free land-holder, making social and political advancement dependent upon royal patronage.[44]  In this he may have been helped by the fact that Lombard nobles do not seem to have been as wealthy and powerful as their Frankish counterparts.[45]  Free status became commensurate with the title of exercitalis or arimannus (army-man, a social group that, by the eighth century appears to have included Romans and Lombards) and the king instituted regular, annual Marchfields at which laws were promulgated and other political business enacted.[46]  A general oath of loyalty was sworn by the arimanni to the king and the slightly later legal requirements that lesser free landowners serve in the army might well be an attempt to maintain the direct links between the king and this social stratum, ensuring that the army did not just become an agglomeration of aristocratic retinues.  As in sixth-century Francia, the army seems to have served as an independent royal coercive force.  The eighth-century Lombardroyal army was successful within Italy, absorbing most of the remaining East Roman (Byzantine) and independent Lombardenclaves.  This success of course posed a threat to the Pope who, when the inability of the emperor in Constantinopleto provide effective military support became clear, turned to the Franks for help.  The latter eventually conquered the Lombard realm in 774.[47]  The period immediately after Liutprand’s death saw internal dissension within the Lombardkingdom and, unsurprisingly it is from this period of usurpation that almost all of the legislation referring to military service comes.  Often assumed to represent age-old, traditional military organisation, it is better viewed in context, as attempts by insecure rulers to ensure the service of their aristocrats and maintain links with more humble arimanni in a period of crisis.  The Lombardswere unable to put up very serious resistance to the Frankish invasions under Pippin I or his son Charlemagne and this has sometimes been seen as symptomatic of various structural weaknesses.  A crisis in morale has been proposed on the basis of the fact that some Lombards made their wills before setting out to join the army, but this seems a little hasty.[48]  One of the will-makers was a bishop (Walprand of Lucca) and wills or similar arrangements for inheritance are known to have been made in armies across the early medieval west, including those of Charlemagne at the height of his success.  Wickham suggests more interestingly that the military weakness of the Lombards derived from the relative poverty of its aristocracy.[49]  Yet, the law-codes (like the cemetery evidence of the seventh century) suggest that the wealthier rungs of the arimanni were as well equipped as their Frankish opponents.  It is possible that they were outnumbered by their Frankish opposite numbers, and that they were less able to raise large, well equipped retinues, but the principle reason for their military failures would seem to me to be a difference in military experience.  The Frankish armies were involved, as we have seen, in annual, successful campaigning since the second decade of the century ensuring a continuous stream of battle-hardened warriors through the ranks, training their juniors.  By contrast the last decades of the Lombardrealm, especially in the reign of its last ruler, Desiderius, were generally peaceful.  The comparatively inexperienced Lombard army eventually collapsed before the attack of a well-equipped, battle-hardened force commanded by one of the most strategically adept commanders of the whole Middle Ages.

Part 2: Intermission: Early medieval warfare, politics and identity in theory and practice
War and kingship
Before considering the impact of the Vikings upon the conduct of warfare and the raising of armed forces, and the development of military service through the ninth and tenth centuries, a brief ‘intermission’ is necessary to explore other aspects of western European military history in this era, principally the ways in which warfare related to particular forms of power and identity and the nature of strategy and tactics.

Kingship remained intimately bound up with military success.[50]  This was nothing new, although sometimes mistakenly believed to be a barbarian ‘Germanic’ introduction.  Roman emperors, particularly in the later imperial period, had had to be successful soldiers.  What seems to have changed around 600, as part of the general readjustment discussed above, was how martial rulership was discussed.  Even if the ideals remained, there seems to have been a shift of emphasis away from classical towards Old Testament exemplars.[51]  The reasons for the shift might stem from a combination of increasing self-confidence on the part of the kings of western Europe and a growing unease in proclaiming Roman-ness in a period after Emperor Justinian’s attempts to reconquer the Roman Empire’s western territories.  This military effort, which destroyed two kingdoms and helped plunge a third into crisis, was associated with an ideological offensive, propagating the idea that the western kingdoms were ruled not by legitimate kings who (as they had tended to claim in the fifth century) represented the old empire under new management, but by barbarians who had conquered these territories.  In this climate, even if the fundamental ideals of good rulership (piety, justice, generosity and military prowess) remained the same, new touchstones were needed.  Thus Solomon and David began to edge out Trajan, Marcus Aurelius and even Constantine as the model king.

Nor was it only a question of ideas.  Throughout this period the penalties for military failure were high.[52]Indeed a mark of the Merovingian dynasty’s success is that it retained the throne in spite of the fact that these rulers rarely took the field after the earlier seventh century.  The powerful kings of the dynasty had not often participated in military campaigns in the third quarter of the sixth century but from the 590s, possibly as a response to the changes discussed earlier, the Merovingians returned to leading their armies in person up until the 630s.  Nevertheless, Sigibert III’s defeat by Radulf, dux of the Thuringians resulted in the Franks losing their earlier hegemony over the peoples east of the Rhine.  A defeat at the hands of the West Saxons at Burford (752) appears to have caused crisis in Mercia.  Its king, Æthelbald, briefly lost his hegemony over southern Englandand a few years later was done to death by his own bodyguard.  Towards the end of our period, the downfall of the emperor Charles III ‘the Fat’ resulted in no small part from his enemies’ ability to exploit perceived military failures against the Vikings.  Examples can be multiplied from across the early medieval west of kings whose regimes were called into crisis by their failure to perform the role of military leader with adequate success.  Not surprisingly, therefore, leading a campaign was one of the first tasks of any new king.

Concomitant with this military kingship was the army’s dominance of politics.  Early Frankish Gaul and eighth-century Lombard Italy held annual military gatherings on 1 March (the ‘Marchfield’).  In the eighth century in Francia this muster was pushed back to 1 May to ensure better availability of fodder for the warriors’ horses.  This was far from being the only time that armies were mustered; campaigns took place through much of the year.[53]  Nevertheless, these were politically important gatherings.  In the last decade of the sixth century, Childebert II of Austrasia (575-96) held three such Marchfields at which he promulgated edicts.[54]  The many edicts of the Lombardking Liutprand and his successors are all dated to 1 March.[55]  This illustrates how between the seventh and tenth centuries the army, as a gathering of the powerful, remained the principal political assembly, at which a ruler established and underlined the (however fictive) consensus within his kingdom.  Laws are often said to have been issued with the approval of the army.  Royal charters, too, are frequently found being issued while the king was with the army.[56]  In the sixth century it seems that gatherings of the army were important fora at which the king subjected the army to his ideology, rewarded warriors who had done well and punished those who had not.  Even in the Carolingian Empire, when the nature of the army and its precise relationship to the king had (as described above) changed considerably, these assemblies remained enormously important in successfully establishing dynastic legitimacy.

War and Masculinity
Closely linked to the political dominance of the army is another constant of early medieval western European history: the linkage between warfare and masculinity.  Throughout the Roman period, a civic model of masculinity had managed to hold sway even when loyal service to the state in its armies was a component of that ideal.  At the end of the imperial era, however, a rival martial model began to appear, its existence being one means by which the inhabitants of Roman Europe were able to negotiate the end of the empire.[57]  By the start of the period covered by this chapter this model had become completely dominant.  The last traces of a secular civic model of masculinity disappeared during the sixth century.  What had earlier been parallel career patterns and lifestyles – even if the civic model was coming to be valued less – had evolved into a situation where the martial model was the only one available within free, secular life; the other had sunk to an equation with the unfree classes. The church offered the only alternative form of masculinity and the repeated ecclesiastical concern with banning clergy from carrying arms illustrates how even among churchmen this removal from the usual construction of manliness could create serious anxieties.  The association of masculinity with warfare can be seen in a wide array of evidential forms.  One is archaeological, where, throughout the seventh century in much of Europe, weapons remained the masculine grave-goods par excellence.[58]  This seems to have continued in parts of Scandinavia and elsewhere through the Viking Age.  That these weapons do not simply symbolise a right to participate in violence can be argued from the changing, but nevertheless comparatively restricted, distribution of these items, whereas the written sources suggest that legitimate participation in low-level violence was not limited to particular ethnic identities or classes, or either sex, or even necessarily restricted to the free.  Involvement in warfare was, however, a much more regulated matter.  Partly this was because raising the army was a formal, even ritualised business.  The army’s political importance, as described, served to limit participation further.   Thus weaponry, I suggest, represents the right to take part in that more limited form of violence.  That such rights could be jealously guarded is shown graphically by the fact that in 859 the Frankish aristocracy took up arms and cut down peasants and other lowly folk who had presumed to form armed associations to defend their regions from Viking depredations (something the ‘élite’ was signally failing to do).[59]  Charles II ‘the Bald’ even passed a law to ensure that poorer freemen, whom he encouraged to join the army (see below), could do so without fear of attack by their betters.[60]

The Practice of War
It is often thought that the early medieval era was a ‘heroic age’.[61]  It is difficult to see this in the period’s more prosaic records, although it must be conceded that seventh- to tenth-century sources are hardly rich in strategic and tactical detail.  What is noticeable is how frequently depictions of heroic warfare relate to eras before the composition of the, usually poetic, sources.  Let us take, for example, the early Welsh poetry purporting to describe the northern British kings’ wars against ‘Saxon’ invaders.[62]  Although the characters and events supposedly belong to the period around 600 AD, these poems were composed long afterwards.  The earliest recently-proposed dating of this corpus of material sees them as products of the mid-seventh century.[63]  This has not commanded consent but even were the date accepted it is significant that we would nevertheless have to see these poems as being divided from the events they describe by the ‘military revolution’ around 600.  It was suggested that this ‘revolution’ made warfare more of a close-packed, shield-wall ‘slogging match’, with little scope for heroic individuals.  Perhaps the poets were creating a ‘golden age’ for their audience, one which had, conceivably, never existed.  The same could be true a fortiori for the warfare described in the Irish sagas’ accounts of the conflicts between the long-gone realms of Leinsterand Ulster, and Carolingian epics about the warriors of the migration period, such as Waltharius.[64]  Nor is it impossible, given the location of these works’ composition and transmission and the sometimes absurd feats depicted, that they represent monkish satire on secular martial boasting.  To suggest, as has been done, that post-imperial Europeans held martial valour or heroism in little regard[65]is clearly erroneous, but epic or poetic depictions of heroic warfare between bands of individuals do not seem to have much of a basis in reality.

If we leave this ‘heroic’ notion aside, what can be said about the practice of warfare between c.600 and c.1000?  It is vital to note that in scale it varied not only through time and place but also according to its type; there seems to have been a gradation of scales of warfare.  Most appears to have been comparatively small-scale, reasonably described as endemic and even, given its apparent government by norms, as ritual.  Given the dominance in early medieval western society and politics of identities closely bound up with participation in the activities of the army, this should not be surprising.  A reasonably frequent summoning of the army was necessary to underpin these identities.  Yet such warfare could not be carried out with such regularity on a large scale without risking the fatal undermining of the social, political and economic system.  Thus there seem to have been particular norms and codes, enshrined by repeated observation rather than through any actual codification, restricting the scale of conflict.  The frequency of small-scale warfare can be seen in areas such as eighth-century Francia, where we have unusually detailed narrative sources (and even these can be shown not to record all military campaigns!) and also in the concern of some legislation, seemingly drawn up as part of a peace-making process, to limit cross-border raiding.[66]

Yet frequent small-scale endemic warfare could cause tensions to build up.  If military supremacy was established, raids could turn into simple tribute takings and the domination of weaker realms.  This situation might be challenged by an outbreak of more serious war in which the usual norms were ignored.  This dual pattern – a background of small-scale conflict punctuated by periodic outbursts of major warfare – can be observed in several areas.  In Anglo-Saxon England, many sources make clear that war was endemic and, as discussed, vital to social structure.  Yet the great, if limited, narrative sources for the ‘middle Saxon’ period (c.600-c.800) only record wars between the island’s major kingdoms about once per generation.[67] Given that these narratives tend to be written later, it seems clear that what have been remembered are the major outbursts of warfare.  The usual small-scale raiding and counter-raiding, although its existence is acknowledged by these narratives, has simply not been recorded.  Although the sources are much fuller, making the different types of conflict more difficult to distinguish, Frankish warfare seems to have followed similar patterns.  The eighth-century campaigns in Aquitaine, for instance, were composed largely of raids or the insertion of garrisons into strategic points, with only a few major, set-piece encounters.[68]  An analogous pattern can be observed in Charlemagne’s Saxon campaigns.  Round about the start of the era covered by this chapter, the stylised, limited warfare conducted between rival members of the Merovingian dynasty led to a build-up of tensions that resulted in two very serious outbreaks of ‘no holds barred’ warfare, in 574-5 and 612-13.  Other such crises were averted in 587-8 and 625-6 by the intervention of bishops and the creation of peace treaties and new law-codes.[69]

This conclusion matters because it permits us to circumnavigate some controversies about early medieval warfare.  One such debate has concerned the size and composition of armies and has frequently been characterised by the participants’ desire to see their interpretation apply ‘across the board’.[70]  Some sources suggest that armies could be quite small (although the extreme ‘36-man army’ point of view cannot really be maintained), numbered in hundreds; yet others seem reliably to discuss armies numbering several thousand men.  Equally, while some evidence appears to attest that participation in the army was tightly restricted to the powerful; other data seems to show the involvement of poorer freemen.  Another debate concerns whether warfare was conducted primarily on foot or on horseback.  In the area of Anglo-Saxon warfare these debates have often been marred by the selection of sources from diverse points in time, regardless of context.  However, if we accept that war occurred on different scales we can resolve many of these evident contradictions.  Low-level warfare could have been conducted with small armies of noblemen, fighting on horseback and perhaps using particular tactics.  By contrast, major conflicts would see the recruitment of larger forces, including warriors from further down the social scale.  In addition to the possibility that such inclusive forces would not be composed entirely of horsemen and that the ‘rank and file’ might (as is attested) need stiffening by front ranks of dismounted aristocrats, the fact that such warfare was aimed at producing decisive results might further plead for the conduct of such ‘set pieces’ on foot.

The scale of warfare, as mentioned, was also determined by time and place.  The immediately post-imperial period in much of the west saw economic contraction and decline, which must have drastically affected the scale at which war could be waged.  It is unlikely that armies ever numbered more than 4-5,000 men even in major outbursts.  However, as has been noted, at the opening of our period an important economic change took place which led to the revival of the north-west of Europe (northern Gaul, Anglo-Saxon England and other territories bordering the North Sea), even as more southerly, Mediterranean (and, in Britain, western) areas declined. This may have been important in enabling the general military dominance of those areas throughout the centuries that concern us.  Further, steady economic growth took place during the period which surely permitted a concomitant increase in warfare’s scale.  In particular the ninth and, especially, the tenth centuries witnessed significant economic expansion and a growth in the size of settlements.  This, as will be suggested, might well have changed the nature of campaigning; it also meant that whereas between c.600 and c.800 large armies were probably usually numbered in the region of 5,000 men, perhaps more on occasion, after 800 armies of 10,000 and, by the end of our period, more were (at least occasionally and for short periods) feasible.  When making these statements, however, it should be borne in mind that one is thinking principally of the major western realms.  In areas such as Ireland, the north of Great Britain, Scandinavia and the Breton and Basque fringes of the Frankish world, armies would rarely if ever be of this order of magnitude.  Even when the socio-political sophistication of such areas is rightly acknowledged, the economic (or even ecological) bases for warfare would be unlikely to enable warfare involving more than a few thousand men, even when conducted on a major scale at the end of our period.

Unsurprisingly, campaign objectives varied with the nature of warfare.  Endemic war, perhaps surprisingly, seems generally to have been aimed at maintaining the status quo.  As mentioned, participation in the activities of the army was the underpinning of much of the period’s social and political structures.  Small-scale warfare reinforced this.  Loot taken on raids was redistributed to followers, reaffirming ties of dependence; warriors had the opportunity to impress their superiors and receive other forms of patronage, promotion, or titles.  When, as would often be the case, the other side retaliated by launching a counter-raid, the same factors would apply.  Material recently taken as booty would be taken back, in its turn reinforcing social and political ties and identities within that kingdom.  The complete destruction of the enemy would, in this type of war, be counter-productive.

Early medieval western European strategy seems to have been quite distinctive, differing in important ways from that in other parts of the middle ages.[71]  Battleappears to have been an object of campaigns to a much greater degree than it was to be after c.1000.  Most campaigns appear to have focused upon looting enemy territory, which, while important in oiling the cogs of politics within the realm, simultaneously hit the opposing ruler’s claim to be a good lord, protecting his people and performing his royal duties as war-leader.  Raids could be, and were, bought off or ignored from behind the shelter of walls until the attackers went home or succumbed to disease, but a failure actively to confront invaders or, at the very least, conduct a successful punitive raid produced internal tensions.  Even kings who successfully faced down invaders from within fortified centres were overthrown by disgruntled nobles.[72] 

Thus, set-piece encounters (even if small in scale) were comparatively frequent in the early middle ages.  Battles, furthermore, were risky so that even minor engagements could have serious political results, when leading political figures were slain, and many early medieval kings, princes and high nobles died in battle.  Why forces should have been committed to battle so readily, when the outcome, as contemporaries well knew, was so uncertain might be understood by reference to the importance of warfare in underpinning of social and political identities in an era when social hierarchies were potentially quite fluid.  This, it can be suggested, led to a need for frequent battle.  The comparative regularity of battlefield engagement might also be explained by consideration of the economy and settlement pattern.  As has been mentioned, even if the centuries between 600 and 1000 saw steady economic growth, for most of the period towns were small, with inhabitants, even late in the period, numbering only a few thousand.  Estimates of the populations of major seventh- to ninth-century trading centres have placed them only in the region of 1000 souls.[73]  These sites were not, therefore, sources of enormous wealth until quite late within our period.  At the same time, there was an absence of wealthy high-status fortified settlements, such as would later be represented by castles.  The principal exceptions to the rule of general absence of wealthy settlements were monasteries – one reason why they were so frequently targeted by Viking attacks and why the latter produced such outrage. Furthermore, fortification and siege warfare were, for most of the era, fairly rudimentary.[74]  Thus high-ranking secular noblemen tended to carry their wealth with them, including when on campaign.  For this reason, in warfare significant economic benefits were most likely to accrue from defeating the enemy in the field and plundering their baggage, rather than through besieging settlements.  This was beginning to change by the tenth century, however, and by 1000 we have entered the world of knights and castles, with quite different strategic practices.

Nevertheless, we must not over-emphasise booty’s importance in early medieval warfare.[75]  It has frequently been argued that the acquisition of loot was the principal motor for campaigning, with the inability to continue to take rich pickings producing stress within realms.[76]  However, although booty was important, it was not the over-riding object of warfare and given that this was, comparatively, not a period of great economic prosperity we should not be surprised by this.  Warfare yielded other, possibly more important benefits, many of which took quite intangible form.  War offered the opportunity for warriors to come to their superiors’ attention and receive their patronage as a result.  Lands, titles, involvement at the heart of politics and the management of a patronage network of one’s own might ensue.  Even if one did not move far up the social ladder, the backing of an important noble or royal figure could be enormously important within local politics.  Sometimes the booty taken on campaign was passed on to more powerful figures, precisely to obtain these forms of backing and promotion.  In sum, warfare was vital to the machinery of early medieval politics in many ways, of which the acquisition of loot was only one, and probably not the most important.

The evidence for early medieval tactics is notoriously thin.[77]  In the first part of this period, as has been discussed, we are largely reduced to using the forms of weaponry to suggest how battles were conducted, with all the inherent risks that that involves.  One problem is a simple shortage of actual descriptions of battles; another is the enormously stylised nature of those that we do have.  From most written accounts we cannot conclude with certainty even whether the participants fought on foot or on horseback.  The evidential filter is so capricious that we cannot know what weight to place on those snippets of information that do make it down to us.  Some tentative suggestions can be made nevertheless.  Perhaps the most important is that tactics do not appear to have differed much whether the opposing armies were mounted or dismounted.  The two sides seem to have formed large, close-packed blocks, which advanced slowly, paying careful attention to order and cohesion.  Contact was preceded by volleys of missiles and rear ranks continued to shoot or lob missiles over the heads of the forward ranks into the enemy rear.[78]  Usually, one side would eventually break and be pursued and cut down; these were the stages of battle that produced most casualties.  Alternatively, it appears that the two sides would, if neither side broke, draw apart, and a lull would ensue while the armies tended wounds and attempted to recover sufficient energy and nerve to rejoin battle.  Battles could, therefore, be long, bloody and indecisive.  On the fringes of western Europe there were exceptions to the tactical norms.  Breton, Basque and Irish warfare might have involved a more open, skirmishing kind of fighting probably more suited to the terrain in which they fought.  Frankish accounts of the defeat at Roncesvalles bemoan the fact that the Basques did not fight fair, refusing to stand and engage in the usual slogging match.

Part 3: c.800-c.1000
The end of Carolingian expansion[79]
There is little surviving legislation about military service from eighth-century Franciaand in many ways this is unsurprising.  As described, the Carolingians’ successful campaigns brought enormous wealth to the members of the army, making participation in warfare attractive to the land-owning classes.  Ensuring that people performed their military service was rarely if ever a concern to the early Carolingians.  By contrast, the period of Charlemagne’s reign after his imperial coronation in 800 produced considerably more legislation on the subject than had the whole eighth century.  This clearly indicates change and illustrates graphically the problems caused by the end of Frankish military expansion.  Indeed, the growth of royal or imperial pronouncements about who should perform military service, in what form and with what equipment is matched by a directly contemporary rise in the number of legal enactments about those who failto carry out such obligations.

In recent decades the dominant explanation for this shift has been provided by two important articles by the late Timothy Reuter, arguing that the end of the opportunity to take booty produced stress within the Frankish Empire.[80]  This might have been associated with a shift from offensive warfare to defensive campaigns against invaders like the Vikings.  Although Reuter’s explanation is preferable to earlier analyses, it does not seem entirely satisfactory.  What seem instead to be crucial are the size of the Empire and the lack of royal or imperial foci for campaigning as Charlemagne grew old and two of his three sons (all of whom appear to have been able commanders) predeceased him.  As we have seen, warfare was not purely important for material booty but for political advancement and the obtaining of patronage.  With the focus of political activity, the imperial court, remaining at the geographical core of the kingdom there was little incentive to campaign on the frontiers.  Indeed, Charlemagne’s legislation reveals that offensive campaigns were as difficult to recruit for as defensive ones, probably a fatal problem for complete acceptance of Reuter’s thesis.  Ever more distant from the heartlands of the Frankish empire, the frontiers were more costly to get to, reducing the profit margin even of successful warfare.  More to the point, perhaps, if the army were not accompanied by the emperor or one of his sons, campaigning on the frontier represented removal from the centre of politics, in effect a sort of exile.  The first crisis of the reign of Charlemagne’s heir, Louis I ‘the Pious’, was indeed brought about by his despatch of two important noblemen from one palace faction, focused on Louis’ troublesome eldest son, Lothar, to campaign on the Spanish march, while he brought their rival, Bernard of Septimania, from that frontier to the court.  Clearly the opportunities to take part in aggressive warfare had not ceased but were not decisive in governing continued willing participation in the activities of the army.

The Vikings’ appearance, as well as that of a new wave of Muslim raiders and, later, Magyar (or Hungarian) attackers, nevertheless meant that defensive warfare became more of a concern of ninth-century rulers.  It should also be noted that there were qualitative differences in the type of warfare that these attacks involved.[81]  Defensive wars were less popular amongst the military elite; they presented far less opportunity for rewards, whether in loot or in political advancement or prestige.  For these and other reasons, then, the ninth century saw, across Christian western Europe, a constant concern with the nature of military obligation.

The ways in which armies were raised apparently differed between defensive wars and other sorts of conflict, such as the warfare between the various Frankish kingdoms.  In the latter, as in the offensive wars that continued to take place on the eastern frontier of the Empire at least, the essential template established in the seventh and eighth centuries remained.  Armies were raised from the retinues of those of social classes generally held liable for military service.  If the precise choice of who served and who remained at home could be moderated heavily by specific local social relationships and patronage networks, it should not be forgotten that general notions of liability to military service according to wealth or status remained.  Ninth-century legislation and exemptions from military service make that abundantly clear.

Nevertheless, a major problem was presented by the Carolingians’ ultimate reliance upon their nobles’ military followings.  It posed a real threat to the effectiveness of royal government in the localities and increased the extent to which royal action was circumscribed by the need for negotiation with the magnates.  As many of the upper stratum of the latter had relatives, and often held lands and honours, in more than one kingdom, it was frequently a simple matter to transfer allegiance from one king to another.  Haemorrhages of political authority were not unknown and their results could be dramatic.  In winter 858 Charles ‘the Bald’ of West Francia was driven out of his realm when his nobles sided with his half-brother Louis ‘the German’.[82]  Earlier, Charles’ father, Louis ‘the Pious’, had once been deposed when his men deserted him for his sons, Lothar, Pippin and Louis.[83]  The various Carolingians therefore attempted to create armies which they could use as coercive forces without having to rely on the good will of their nobility.  Probably the most effective means was the use of church land, continuing the precedent set by Charles Martel and Pippin I in the preceding century.  Churches and monasteries held large estates and, by the appointment of abbots or bishops, kings were able to ensure that warriors supported by benefices on such lands were available for use in royal armies.  During his 866 Lotharingian campaign, Charles the Bald’s army was described as ‘mostly composed of the bishops’ contingents.’[84]  Through the ninth century, few churches were exempted from the obligation to provide warriors (as far as I can see, Charles the Bald granted only one such immunity, and his nephew Louis II of Italy none; the much-maligned Charles III ‘the Fat’ even altered the privileges of the great abbey of Korvei so that it had to furnish troops in cases of emergency).  The monastery of Prüm in Lotharingia is one of the few regularly granted such exemptions.[85] 

The ninth-century Carolingians also attempted to ensure that troops were available for defensive wars through a system first established by Charlemagne, called the adiutorium(loosely a system of ‘assistance’).[86]  This envisaged a ‘flat rate’ whereby each unit of land of a particular size (or possibly equal estimate of productivity) would furnish one warrior.  Charlemagne had initially envisaged this unit as of three mansi (roughly, farms) but soon changed his mind in favour of a lighter obligation of one warrior from four mansi(these figures are put into some sort of context by the fact that Charlemagne only expected a holder of twelve mansi to serve with full mail armour).  Each owner of four such mansi would be required to provide a warrior (holders of multiples of this figure would provide one for each group of four).  Those owning less land would be grouped together to provide one warrior.  Charlemagne and his successors also legislated to ensure that those not serving in person provided supplies and other logistical equipment (such as carts) for those who were.  Again, the large monastic foundations were crucial.  Fines were regularly (and the death penalty occasionally) envisaged for failure to attend the muster.[87]  Whether all this legislation was especially effective is open to question; descriptive accounts of campaigns still seem to stress the usual type of army, raised from royal and aristocratic retinues.  Nevertheless the attempts to impose general ‘cadastral’ systems of military obligation are important, and we shall return to them later.

Otherwise the Frankish rulers seem, pragmatically, to have focused their efforts upon trying to ensure that aristocratic retinues were available to them when armies were needed, whether for defensive or offensive operations.[88]  This was especially the case in Carolingian Italy, where, by 866 at least, it was envisaged that anyone with less than half of the wherewithal to provide a warrior (the bharigild as it was called) was to be left behind.[89]  Although Louis II of Italy never exempted any churches from military service, he did exempt individual laymen.[90]  The Italian Carolingians seem to have lost interest in a general levy in favour of an army based on the forces of the greater royal vassals and churches.  At the end of his reign, Charles II ‘the Bald’, having apparently failed in his attempts to create a royal army by other means, fell back on the same sort of system.  The king enacted that all great landlords were to draw up lists of their own honores and vassals, and how many mansi they held.  These retinues, thus declared, were envisaged to be at royal disposal.[91]

As earlier, the core of the army was provided by the royal bodyguard based at the palace and, as throughout the early Middle Ages, envisaged as made up of two groups: young warriors serving full time in the palace guard, and older warriors who had received their own lands as a reward for service and who now attended the guard more intermittently or as the officers and trainers of the younger warriors.  The different grades of guard are described in the later ninth century by Hincmar of Reims in his On the Governance of the Palace.[92]

As before, the developments in the Frankish realms (now of course including Germany and Italy as well as much of the Christian territory in Spain) can be paralleled by those in England.  Anglo-Saxon kings seem, equally, to have begun to establish a cadastral system of military service.[93]  The vague indications are that the rate was envisaged as one man from five (or possibly six) hides (a unit of land or productivity similar to the Frankish mansus).  Certainly, the one man to five hidesrate is that which seems to have been of general application by the time of the Domesday Book.  Around the middle of the ninth century, the West Saxon king Æthelwulf seems to have been concerned to provide lands for the upkeep of his warriors (thegns) from royal estates and made a generous donation of one tenth of his lands to the church which, some claimed, was exempt from all duties, including military service.  This is not something any self-respecting Carolingian would have done.

The Vikings[94]
As noted, the stimulus for much of the development in the systems of raising armies in the ninth century was provided by attacks on Christian Europe by non-Christian raiders: Scandinavian, Muslim and Magyar.  It is the first of these groups, the Vikings, that concerns us here.  Scandinavian armies at home and abroad appear, like those of other areas, to have been conglomerations of military retinues.  Sometimes these were assembled for raids and perhaps only for the duration of a particular season.  By the mid-ninth century, though, large assemblies had emerged that stayed in the field all year round, wintering in Christian territory and extracting protection money from local rulers.  This permanent service and extensive military experience made the Viking ‘Great Armies’ very difficult foes to beat, so that even fairly minor successes such as Ashdown (871), Saucourt (881) and the Dyle (891) were celebrated in Christian sources as great triumphs.[95]  When defeated, Viking armies frequently dissolved into their constituent groupings, which might then reconfigure elsewhere as a different army under another leader or leaders.  The dynamics of Viking warfare were much the same as those of barbarian raids during the Roman Empire.  A leader would establish reputation through successful campaigning, building a following in the process.  At the same time, this warfare enabled the accumulation of treasure which could be used to buy support.  After some years of campaigning, a Scandinavian warlord would return to attempt to establish himself as a king in Denmark, especially, or Norway.  The intermittent attempts – more successful in the tenth century – to establish lasting royal authority in Scandinavia in turn frequently led the losing factions in such politics to seek to restore their fortunes in ‘Viking’ campaigning in the Christian kingdoms of Francia, England or Ireland.  Only at the very end of the period covered by this chapter were armies regularly led abroad by the kings of Denmark – most famously Swein and his son Cnut.  These armies appear to have been quite different and belong better in the discussion of tenth-century western European changes.

Viking strategies employed cunning and knowledge of local custom.  Attacks were timed to catch enemies during particular festivals, especially during winter after the harvest was in.  In such circumstances, the assumption – frequently justified – was that the defending ruler would find it more difficult to assemble an army and might thus acquiesce in the payment of tribute and an understanding that the Vikings would move on in spring.[96]  In the British Isles, rulers who attempted to resist in these circumstances frequently lost their kingdom – and often their life as well.[97]  These strategies aimed to make the maximum profit at the minimum risk.  On the battlefield the Scandinavians were similarly crafty.  They used difficult terrain when confronted by armies with more powerful contingents of cavalrymen, and employed field fortifications with notable success.  The Vikings repeatedly withdrew into such defences only to burst out in a sudden counter-attack just when their opponents thought the battle won.[98]  Yet, it does not appear that the Vikings were interested in actual conquest for most of the ninth century.  In defeated realms, new rulers were installed to better facilitate the payment of tribute.  The crucial moment in the shift from ‘land piracy’ to actual settlement and the creation of kingdoms appears, as with so many aspects of early medieval warfare, to have been related to age.  When the Danes settled in Anglo-Saxon Northumbria in 875-6, it seems that it was the length of time that the force had been in the field that was decisive.  The ‘Great Army’ split and those that headed north to settle appear to have been those bands that had been campaigning in Englandfor the previous ten years.  The more recent arrivals, the ‘Great Summer Hosting’ (micel sumerlida) of 871, seemingly provided the core of those who went on fighting until defeated by Alfred of Wessex in 878, when they settled northern and eastern Mercia and East Anglia (later called the ‘Danelaw’).  Even then, others took themselves abroad and continued their freebooting lifestyle in Francia.[99]  When the reality of political control presented itself after the collapse of three of the great English kingdoms, it must have seemed attractive to warriors who had been in the field for many long years to take this option rather than continue with the probably more exciting but nevertheless riskier life of the roving marauder followed by their younger fellows.

Responses to the Vikings
In many important regards, the Vikings were no different from any other military forces of the early medieval west.  Yet, there does seem to have been something qualitatively different about them.  Religious differences meant that they and their Christian enemies frequently had different understandings and expectations of the conduct of warfare.  Without shared beliefs, oaths were difficult to underpin with shared supernatural sanctions.  While Christian armies were certainly not above attacking ecclesiastical establishments, accepted norms of behaviour regarding such churches and monasteries did not apply to the Vikings.[100]  Other features made the Vikings ‘difficult’.  As noted, their armies campaigned all year round (even in winter) and their strategies, just outlined, employed surprise and exploitation of the Christian calendar.  A further novel and disturbing feature of Viking warfare was its commodification.  Between the fifth century and the ninth, mercenaries are generally conspicuous by their absence in Western Europe.[101]  This is probably unsurprising in a largely non-monetary economy where soldiers were rewarded in land. As we have seen.  In these circumstances, the true mercenary, serving for monetary payment for a defined period and cutting all ties with his employer at the termination of the contract, simply could not exist.  Alongside the economic expansion of the ninth century (in which the Viking armies may have played a part) and the concomitant growth in the use of money, the Scandinavians seem to have been happy to sell their loyalty and their services to the highest bidder.  This must have been another shocking aspect of Viking warfare to Christian contemporaries.

These dangerous, mobile and unorthodox foes required specific responses and we can see these on both sides of the English Channel.  Although the popular image of the Vikings sees them as coastal raiders, the great Viking armies were as often as not to be found travelling by land, frequently mounted.  If not left at the coast, their accompanying fleets plied the great river systems of Europe.  One important anti-Viking strategy, therefore, had to be to limit this mobility and this was especially effectively done by the fortification of river-crossings.  These not only barred the rivers to Viking fleets (as barriers, these river crossings are most effectively envisaged as pontoon bridges) they also denied the crossing points to Viking armies travelling overland.  Charles the Bald tried repeatedly, but with mixed success,[102]to create fortified bridges barring the Loireand the Seine to Scandinavian invaders.  After finally defeating the Great Army in 878-9, Alfred of Wessex put into operation a more extensive and, as it turned out, more effective defensive scheme, a network of forts or burhs.[103]  Stationed so that no one in his kingdom would live more than a day’s journey from such a fort, they were sited to control the nodes of riverine and road communication networks.  When the Viking Great Army returned from Francia in the 890s, following a defeat at the hands of King Arnulf of East Francia, these forts proved their worth.  The Danes were rapidly compelled to leave Alfred’s kingdom, hounded by the West Saxon ruler’s field army.

The other component of Alfred’s defensive scheme was indeed to create a permanent armed force.  He did this through the use of the cadastral system of raising the army that, as we have seen, was emerging in the ninth century.  In England that scheme, as noted, seems to have been one warrior from every five hides of land.[104]  Alfred arranged his levy so that they served in two, or perhaps more probably three, rotations.  One contingent would serve with the king, while another stayed at home.  It seems that the third contingent manned the burhs.  This scheme makes more sense as a demand on the time of the land-owning classes.  After their time on their estates, they would man the forts, thus remaining close to their lands and able to manage them if necessary, before leaving the burh and marching as a contingent to the royal army when the next contingent arrived to relieve them.  After their spell in the field force, they were stood down for a third of the year, as he next rotation arrived from the burhs to replace them.  This armed force was the basis for the West Saxon conquest of the midlands and the north of England in the tenth century.  The successful aggressive warfare waged by the West Saxon kings of England joined with the rotating, cadastral basis of military service, bringing all sorts of landholders – not just the most powerful – into the presence of the king and offering them chances to earn rewards in land, office and patronage, to make the late ninth- and tenth-century English army a royal coercive force of a type that had not been seen in the west of Europe since the sixth century.  The system of burhs was expanded as the realm grew, these forts being visible marks of royal government on the local landscape and foci for the performance of the military obligations that underpinned it.

Across the Channel, analogous attempts to respond to the Viking threat in ways that would strengthen royal power were less successful.  Charles the Bald’s fortified bridges were left incomplete and soon abandoned.  Charles also attempted to use cadastral military obligation to create a royal army.[105]  Poorer landholders were encouraged to attend the muster and their more powerful fellows prevented from molesting them when they did so.  Yet these efforts too failed and by the end of his reign Charles had, as noted, adopted the solution of his Italian nephew Louis II and acted simply to ensure that aristocratic and ecclesiastical armed followings were ultimately at his disposal.  Another imaginative attempt to use the Vikings to bolster his rule was Charles’ move to employ one of the Viking armies themselves.  He would take the fiscal levy raised from across the realm to pay off the Danes (such as would be known in England as Danegeld) but pay it to the Vikings not to go away but to act in his service, creating a mercenary force independent of Frankish aristocratic involvement.  Other magnates, Frankish and Breton had already taken to enlisting the support of Viking contingents.[106]  Unfortunately for Charles, the Viking leader he selected, Weland, was killed by one of his fellow commanders in a personal duel not long afterwards and the plan fizzled out, to be replaced first by his Edict of Pîtres (Charles’ most elaborate attempt to employ a cadastral system) and eventually by his move to control military followings.[107]

Tenth-century changes[108]
The novel features of Viking warfare and the late ninth-century Christian responses to it point the way towards developments during the tenth century: the employment of foreign mercenaries, the greater use of fortifications and attempts at cadastrally-organised military service.  The principal outlines of tenth-century English military service have already been mentioned.  One important document is the Burghal Hidage of c.918, which sets down a possibly more systematised version of the reforms established by Alfred in Wessex.[109]  This document tells us that there were 27,000 hides assessed for military service and the maintenance of the burhs.  It envisages one man from each such hide for service in the fortresses (upkeep and repair and probably defence in times of attack).  The usual ratio of one man per five hides suggests that this kingdom could furnish a standing army of about 5,500 men.  Up until the 950s the English army, like that of the eighth-century Carolingians was repeatedly in action in successful expansionist warfare, which made it a difficult foe to beat, as shown by an almost unbroken record of success.  Thereafter, prolonged peace, broken only by smaller-scale punitive raids, appears (as with the later eighth-century Lombards) to have meant a decline in the army’s quality and possibly even a tendency to abandon body armour.  These had drastic effects when the kingdom was attacked by the new royal Danish armies of Swein and Cnut. 

As far as the more ambitious attempts to employ external threats as a basis for increasing royal power was concerned, the baton was taken up by the new rulers of the East Frankish kingdom, the Ottonians (who revived the western empire when Otto I was crowned emperor in 962).  The principal strategy of the Ottonians, like that of the later ninth-century Carolingians, was to ensure their control of armed retinues.  A document, the indiculus loricatorum of 981, represents the Ottonians’ attempt to register the greater land-holdings of the realm and how many armoured horsemen (loricati) they could furnish.[110]  This shows that the German rulers could, in theory, call upon about 20,000 men.  It is extremely unlikely for logistical reasons that this sort of number was ever raised – indeed throughout the Middle Ages these sorts of lists have a tendency to give inflated views of the numbers of troops available to royal rulers.  Nevertheless, it suggests that when the need arose the Ottonians ought to have had little difficulty – even if the constraints of logistics and ‘consensual politics’ cut the numbers to a quarter of all those theoretically available – in levying a force of armoured horsemen that at least matched the maximum size of the West Saxon field army envisaged in the Burghal Hidage.  That said, one story suggests that East Frankish armies were not large.  When, in 955, a contingent of fifty horsemen was slaughtered by the Slavs it was felt that the army had suffered a colossal setback.[111] 

Earlier in the century, the Ottonians had attempted to parallel the Alfredian response to Viking attack (although the East Franks were much more concerned with the raids of the Magyars).  Henry I ‘the Fowler’ had tried to create border fortifications in newly cleared lands.  Farmers attached to these works were divided into groups of ten, one of whom was always present in its garrison.[112]

However, although the Ottonian kings fought numerous successful foreign wars (and some not so successful, such as Otto II’s catastrophic defeat by the Italian Saracens at Cap Colonne in Italy, in 982), proportionately most of their military activity during the century was concerned with internal revolt, by rivals in Franconia, Bavaria or Lotharingia or by disgruntled members of their own family (such as Otto I’s brothers Thankmar and Henry, or his son Liudolf).[113]  When Otto I’s brother Henry became Duke of Bavaria, the two problems combined!  As elsewhere, in this sort of fighting, royal decrees about military service were of little value and armies were composed of the retinues of different factions, in the old way.  Another strategy of the Ottonians, therefore, parallels earlier Carolingian efforts and that is the much debated imperial ‘church system’, whereby members of the ruling dynasty were appointed to particularly powerful bishoprics.  Whether this was really a ‘system’ has rightly been doubted; it could only be an ad hoc response and it was far from universally successful.  Nevertheless the aim was clear enough.  In controlling the great bishoprics (and eventually archbishoprics) the kings would have access to powerful contingents of warriors maintained from ecclesiastical estates.

What one might term the Alfredian model – the combination of royal fortress and permanent army – was probably most successfully emulated at the very end of our period in the newly reunified kingdom of Denmark.  Here King Swein and his son Cnut constructed a series of famous circular, geometric fortresses which housed permanent garrisons (including women and children).  These are to be found outside the political heartlands of the dynasty and serve a similar purpose to the English burhs as visible marks of royal presence.  Attempts to create a cadastrally-organised army are also visible, which eventually culminated in the Leidangor levy.[114]  On another fringe of Christian Europe, the north of Spain, one can also trace attempts to impose general military obligations, all the more necessary in the context of almost perpetual warfare against the emirate, now caliphate, of Cordoba (in the ninety-two years between 791 and 883, for example, forty wars are recorded).  In the Kingdom of the Asturias this obligation was called fosato or fonsado.  Yet, like the Carolingians, the Asturian kings still granted exemptions to certain churches, forbidding their military officials – who continued to be known by the old Gothic term saiones– from entering such estates to levy troops.[115]

Tenth-century France witnessed no such attempts.  Although the once sharp and teleological outlines of the story have been much nuanced in recent decades, the history of western Francia in this period can nevertheless still be characterised by dynastic conflict, especially but by no means exclusively between the Robertians (later the Capetians) and the last Carolingians, and a steady weakening of effective royal control throughout the kingdom.  At the same time took place the growth of what Dhondt called the territorial principalities (such as Anjou, Aquitaine and Normandy) under their own ruling dynasties.[116]  The last western Carolingians were no ciphers and the story of gradual fragmentation of political power was by far from inevitable.  Nevertheless, however one explains them, the outlines of that tale remain the same with the effective political fragmentation first of the West Frankish kingdom and then of the territorial principalities themselves.  The resulting ‘Feudal Revolution’ of the decades around 1000 has been endlessly debated and its very existence called into question[117]but, fortunately for us, we can skirt these tricky issues to some extent by looking at the nature of armies and warfare and indeed these can quickly be sketched.  The general unit of military analysis remained, perhaps more than ever before, the aristocratic retinue, composed of armoured mounted warriors.  Fortifications were also important but in different ways from those in England and Denmark.  Whatever their origins in defence against marauders, ‘small forts’ (castella in Latin, whence – obviously – our word castle) fairly quickly became in practice hereditary, in much the same way as the honoresof which they were the physical focihad become understood as familial possessions, even if originating in, and sometimes still legitimised by, royally-bestowed titles.[118]  These made the negotiation between local military elites and central authorities (royal or ducal), so necessary for the summoning of armies even in the earlier era, that bit more difficult still.  Without doubt, just as the Carolingians had (in many ways very effectively) countered these issues through the employment of royal ritual and ideology, the fluidity of these relations was countered by the growing social and cultural (if not legal) concentration on oaths of vassalage and concepts of fealty.   Here we are fully in the world of knights and castles that characterises the central middle ages.

This survey of warfare and military service across four centuries of western European history has necessarily had to exclude much, and with it a great deal of the nuance and variety (I have said nothing of the distinctive warfare in the so-called ‘Celtic fringe’, for example).  Nevertheless, it can still be seen that this period was neither one of simple continuity from Rome, nor of the triumph of (probably mythical) ‘Germanic’ warrior cultures.  Nor can it be seen as a half-hearted precursor to the ‘fully-formed’ feudalism of the central middle ages.  There is a dynamic to be seen in the development of armed forces between c.600 and c.1000, one which lies in the relationships between central powers and local, increasingly militarised, elites.  The use of military service to explore this dynamic contributes greatly to the understanding of early medieval politics.  But just as this suggests how military history has an important place in the comprehension of early medieval politics, this chapter illustrates clearly, too, how warfare and armies can only be understood in the broader context of contemporary social and economic structures.  The warfare of this period also has its own distinctive characteristics, with a greater emphasis on battle and (at least before the tenth century) less attention to sieges than in later eras.  These characteristics can only really be explained by study of the economy and settlement pattern of the period and of the central place – indeed one that almost excluded alternatives – of military activity in early medieval society.

[1] Halsall, Warfare and Society in the Barbarian West, pp.53-89.

[2] They were not, however, straightforward continuations of the Roman regular army, as has been claimed repeatedly by Bernard S. Bachrach, in spite of the absence of any evidence, written or otherwise, to support this notion.  That said, the traditional idea in response to which Bachrach evolved his own extreme hypotheses, that early medieval armies were simple tribal, ‘Germanic’ warbands, is equally devoid of evidential support.  For a strange revival of this idea, see M. Speidel, Ancient Germanic Warriors: Warrior Styles from Trajan’s Column to Icelandic Sagas (London, 2004).

[3] H. Elton, Warfare in Roman Europe, 350-425 (Oxford, 1996); A.D. Lee, ‘The army’, in The CambridgeAncient History, vol.13: The Late Empire, A.D. 337-425, ed. A.M. Cameron & P. Garnsey (Cambridge, 1998), 211-37; id., War in Late Antiquity: A Social History(Oxford, 2002); P. Southern &. K.R. Dixon, The Late Roman Army (London, 1996); R.S.O. Tomlin, ‘The army of the late empire’, in The Roman World, ed. J. Wacher (ed.) (London, 1987), pp.107-23.

[4] See, e.g., the well-known case of Saint Martin.  Sulpicius Severus, Life of Martin, 2.  F.R. Hoare (trans.),, ‘Sulpicius Severus: The Life of Saint Martin of Tours’, in T.F.X. Noble & T. Head (ed.) Soldiers of Christ: Saints and Saints’ Lives from Late Antiquity and the Early Middle ages (Pennsylvania, 1995), pp.1-29

[5] G. Halsall, Barbarian Migrations and the Roman West, 376-568 (Cambridge, 2007), pp.101-10, and references.

[6] Halsall, Barbarian Migrations and the Roman West, pp.242-83 passim; J.H.W.G. Liebeschuetz, ‘The end of the Roman army in the western Empire’, in War and Society in the Roman World, ed. J. Rich & G. Shipley (London, 1993), pp.265-76.

[7] Halsall, Warfare and Society in the Barbarian West, pp.40-46, and references.  Similarly ‘ethnic’ military service may be claimed for North Africa, in both Vandal and Moorish areas, and Anglo-Saxon England. Halsall, Barbarian Migrations, p.476.

[8] The debate on this was started (or rather re-started) by W. Goffart, Barbarians and Romans AD 418-585: The Techniques of Accommodation (Princeton, 1980).  See Halsall, Barbarian Migrations, pp.422-47, and references, for a summary of the debate up to about 2006.  Goffart makes important refinements to his argument in his Barbarian Tides: The Migration Age and the Later Roman Empire(Philadelphia, 2006).

[9] Halsall, Warfare and Society in the Barbarian West, pp.35-36, 49-50, 56, 58.

[10] Even this age-based structure to a military career might have a Roman origin.  Many late Roman units attested in the western section of the Notitia Dignitatum are divided into iuniores and seniores.  Attempts to explain this division have generally failed to convince and it may be that a straightforward reading in terms of the relative age of the recruits has much to offer.  The iuniores would indeed be the younger, new recruits formed around a cadre of ‘NCOs’ and officers drawn from the older, more experienced seniores.  After successful service the soldier might be promoted to serve in the seniores, possibly eventually returning to the iunioresas an officer.  A parallel with Napoleon’s Imperial Guard might be instructive.  Only about a fifth to a quarter of the definitely identifiable ‘brigades’ of iuniores and seniores recorded in the Notitia are found together in the same army (though most are at least in the same half of the Empire), but equally, for example, the ‘Voltigeur Grenadiers’ of Napoleon’s Jeune Garde were often serving far from their elder brothers in the Moyenne Garde’s ‘Fusilier Grenadiers’ and the Grenadiers of the Vieille Garde (the corps’ ‘parent regiments’). 

[11] Halsall, Warfare and Society in the Barbarian West, pp.44-45, 56-57.

[12] These included tax-exemption, legal privilege, greater involvement in politics and ability to come to royal attention through participation in the activities of the army.

[13] W. Goffart has not entirely unfairly described my previous account of the evolution of armed forces as seeing them as ‘somehow’ evolving into a stratum of landholders.  W. Goffart, ‘Frankish military duty and the fate of Roman taxation’, Early Medieval Europe 16 (2008), pp.166-90, at p.178, n.39.  The following section aims to make the nature of that evolution a little clearer.

[14] Fredegar IV.87: The Fourth Book of the Chronicle of Fredegar, with its Continuations, trans. Wallace-Hadrill, J.M., (London, 1960).

[15] The sense of the word is crystal clear in its earliest appearance – in Fredegar’s Chronicle, IV.74 (dealing with Dagobert I’s 631 campaign in Thuringia) – where the Frankish force is described as ‘the army of the kingdom of the Austrasians’, plus scarae de electis viris (scarae of chosen men) from Neustria and Burgundy, under their counts and dukes. 

[16] E. Renard, ‘La politique militaire de Charlemagne at la paysannerie franque’, Francia36 (2009), pp.1-33, at p.21, n.102, says that the hypothesis of a change in organisation is proposed ‘sans la moindre justification’ other than a terminological shift.  Apart from failing to interrogate whether a terminological shift is to be so lightly dismissed, this also neglects the significance of the precise meaning of the word scara when first used, and all the other changes that are contemporary with it, sketched here, and which point in the same general direction.

[17] Here I summarise my conclusions in Warfare and Society in the Barbarian West, pp.53-70.

[18] H. Ebling, Prosopographie der Amsträger des Merowingerreiches, von Chlothar II (613) bis Karl Martell (741) (Beihefte der Francia 2), (Munich, 1974).  Good studies of aristocratic politics in Francia at this date are P. Fouracre, The Age of Charles Martell (Manchester, 2000), and R. Gerberding,The Liber Historiae Francorum and the Rise of the Carolingians (Oxford, 1987).

[19] I see the crucial definitive element of a state as the ability to impose the will of the governmental centre on the communities of the polity.  This stress on coercive force, of course, stands in a tradition going back to Weber.

[20] A well-illustrated survey of this subject is J. Lebedynsky, Armes et Guerriéres Barbares au temps des grandes invasions (IVeau IVe siècles)(Paris, 2001).

[21] On Vendel and Valsgärde, a good guide is Statens Historiska Museum, Studies 2: Vendel Period, ed. J.P. Lamm & H.A. Nordstrom (Stockholm, 1983).  For Sutton Hoo, essential is M.O.H. Carver, Sutton Hoo: Burial Ground of Kings?(London, 1998)

[22] Economic expansion is well-attested.  G. Duby, Guerrières et Paysans, VIIe-XIIe siècle: premier essor de l’économie européenne(Paris, 1962); R. Hodges, Dark Age Economics: Origins of Towns and Trade, AD 600-1000 (London, 1982); The Long Eighth Century, ed. I.L. Hansen & C.J. Wickham (Leiden, 2000).

[23] Here I summarise my conclusions in Warfare and Society in the Barbarian West, pp.71-89.

[24] The best of the traditional views of ‘the feudal system’ is probably F.L. Ganshof, Feudalism (3rd edition; London, 1964).  M. Bloch, Feudal Society (2 vols.; London 1962) is much more subtle a reading than was produced by most of his contemporaries and successors, and not really concerned with ‘system’ at all.

[25] Classic now (though controversial) is S. Reynolds, Fiefs and Vassals: The Medieval Evidence Reinterpreted (Oxford, 1994).  See also Fouracre, The Age of Charles Martel.

[26] Halsall, Warfare and Society in the Barbarian West, pp.71-74, and references.

[27] For Spain, see R. Collins, Early Medieval Spain (2ndedition, London 1991), modified somewhat in id.,Visigothic Spain, 409-711 (Oxford, 2004), pp.69-116; for Francia, see P. Fouracre, The Age of Chares Martel; Gerberding, The Liber Historiae Francorum and the Rise of the Carolignians; I.N. Wood, The Merovingian Kingdoms 450-751 (London, 1994), pp.221-72

[28] For the history of precaria, see I.N. Wood, ‘Tetsind, Witlaic and the history of Merovingian precaria’, in Property and Power in the Early Middle Ages, ed. W. Davies & P. Fouracre, (Cambridge, 1995), pp.31-52; See also Fouracre, The Age of Charles Martel, pp.137-45.

[29] See, for example, two charters of the Abbey of Wissembourg in Alsace, issued in 735/6 and 739: Traditiones Wizenburgenses: Die Urkunden des Klosters Weissenburg, 661-864, ed. K. Glöckner & A. Doll (Darmstadt, 1979), nos. 9-10.

[30] Fouracre, The Age of Charles Martel, pp.155-74; Wood, The Merovingian Kingdoms, pp.286-92.

[31] A similar process of ‘drift’ into semi-independence had occurred in the former areas of Frankish hegemony beyond the Rhine when adult-to-adult succession among the Merovingians came to an end in the middle of the seventh century.  As with the southern regions of Gaul, however, the rulers of, for example, Thuringia, Hesse and Bavaria continued to maintain links with the aristocratic lineages of the core regions of Franciaand studiously avoided the adoption of royal titles so that they could retain their wider political horizons by involving themselves in the politics of Francia as and when it suited them.

[32] Fouracre, The Age of Charles Martel, pp.138-41.

[33] E. Magnou-Nortier, Foi et fidelité: Réchereches sur l’evolution des liens personnels chez les francs du VIIe au IXe siècle (Toulouse, 1976).

[34] On oaths, see M. Becher, Eid und Herrschaft: Untersuchung zum Herrscherethos Karls der Grossen (Sigmaringen, 1993).

[35] For this dynamic, see Fouracre, The Age of Charles Martel, pp.118-20.

[36] For Visigothic Spain see, above all, D. Pérez Sánchez, El Ejército en la Sociedad Visigoda (Salamanca, 1989).  Halsall, Warfare and Society in the Barbarian West, pp. 59-63.

[37] Visigothic Laws, 9.2.8-9: Monumenta Germaniae Historica Legum Sectio I, vol.1, Leges Visigothorum, ed. K. Zeumer (Hanover, 1902).  S.P. Scott’s early twentieth-century translation is available Online at http://libro.uca.edu/vcode/visigoths.htm(accessed 16/06/2010).  Note, however, that the numbering of clauses sometimes differs from that in the MGH edition.

[38] Procopius, Wars 5.12.50-54: Procopius, ed. & trans. H.B. Dewing (vol.3; London, 1919) Cordoban slave soldiers: R.A. Fletcher, Moorish Spain (London, 1992), p.61; H. Kennedy, Muslim Spain and Portugal: A Political History of al-Andalus (London, 1996), pp.85-87.

[39] C.J. Wickham, Framing the Early Middle Ages: Europe and the Mediterranean, 400-800(Oxford, 2005, rev. ed. 2006) p.231.

[40] For Anglo-Saxon military service, see, above all, R. Abels, Lordship and Military Obligation in Anglo-Saxon England (London, 1988).  See also N.P. Brooks, Communities and Warfare 700-1400 (London, 2000); Halsall, Warfare and Society in the Barbarian West, pp.50-51, 57-59,

[41] Ine’s Laws 5, 51. English Historical Documents vol.1, c.500-1042 ed. & trans. D.M. Whitelock (2ndedition; London 1979), document 32

[42] On Italian military service see S. Gasparri, ‘Strutture militari e legami i dipendenza in età longobarda e carolingia’, Rivista Storica Italiana98 (1986), pp.664-726.  In English: Halsall, Warfare and Society in the Barbarian West, pp.64-65, 81-84.

[43] This is the principal flaw in D. Harrison, The Early State and the Towns: Forms of Integration in Lombard Italy, 568-774 (Lund, 1993).

[44] An extremely good study of this is C.J. Wickham, ‘Aristocratic power in eighth-century Lombard Italy’, in After Rome’s Fall: Narrators and Sources of Early Medieval History: Essays presented to Walter Goffart, ed. A. Callander Murray (Toronto, 1998), pp.153-70.

[45] Wickham, ‘Aristocratic power in eighth-century Lombard Italy’; id. Framing the Early Middle Ages, pp.203-219 (though Wickham’s estimate of Frankish aristocratic wealth before 600 is greatly exaggerated).

[46] On ‘army-men’, an exhaustive study is O. Bertolini, ‘Ordinamenti militari e strutture sociali dei Longobardi in Italia’, Ordinamenti Militari in Occidente Nell’Alto Medioevo (Settimane di Studio 15) (Spoleto, 1968), pp.429-607.  Update with Gasparri, ‘Strutture military.’  All fourteen dated legal pronouncements of Liutprand are dated 1 March, as are the laws of Ratchis and both codes of Ratchis’ brother Aistulf.  The Lombard Laws, trans. K. Fischer Drew (Philadelphia, 1973).

[47] For this campaign, see Royal Frankish Annals, s.a. 773-774; Revised Annals, s.a.773-774.  Both sources (and others of relevance to this campaign) can be found in Charlemagne: Translated Sources, trans. P.D. King (Lancaster, 1986).

[48] C.J. Wickham, Early Medieval Italy. Local Society and Central Power, 400-1000(London, 1981), p.46; The wills in question are found in L. Schiaparelli (ed.), Codice Diplomatico Longobardo ed. L.Schiaparelli (2 vols., Fonti per la Storia d’Italia 62-63; Rome, 1929, 1933), nos. 114 (Lucca, July 754) and 117 (August 755).

[49] Wickham, Framing the Early Middle Ages, p.218.

[50] The classic work on the subject is M. McCormick, Eternal Victory: Triumphal Rulership in Late Antiquity, Byzantium and the Early Medieval West (paperback edition; Cambridge, 1990).  J.M. Wallac-Hadrill, Early Germanic Kingship in England and on the Continent (Oxford, 1971) is dated but has useful insights.  On the Carolingians, see the works of J.L. Nelson collected in ead. Politics and Ritual in Early Medieval Europe (London, 1986).

[51] Y. Hen, ‘The uses of the Bible and the perception of kingship in Merovingian Gaul’ Early Medieval Europe 7.3 (1998), pp.277-89, and the literature cited there.

[52] Halsall, Warfare and Society in the Barbarian West, pp.29-30, and references.

[53] Halsall, Warfare and Society in the Barbarian West, pp.145-47, for discussion of the timing of campaigns.

[54] Childebert II’s Decrees are to be found in The Laws of the Salian Franks trans. K.F. Drew, (Philadelphia, 1991), pp.156-59.

[55] See above, n.47.

[56] See, for example, the charter issued by King Ecgberht of Wessex on 19 August 825, when the king ‘moved against the Britons’ (whom he defeated at the battle of Galford): Cartularium Saxonicum ed. W. de G. Birch (3 vols.; 1885-93), no.389.  See Anglo-Saxon Chronicle, s.a. 825: English Historical Documents, vol. 1, ed. & trans. Whitelock, doc. No.1.

[57] G. Halsall, ‘Gender and the end of Empire’, Journal of Medieval and Early Modern Studies 34.1 (Winter, 2004), pp.17-39.

[58] There has been much debate on the significance of weapons in graves.  For Anglo-Saxon England, classic studies are H. Härke, ‘Early Saxon weapon burials: frequencies, distributions and weapon combinations’ in Weapons and Warfare in Anglo-Saxon England, ed. S. Chadwick-Hawkes (Oxford, OUCA Monograph 21, 1989), pp.49-61; id., ‘“Weapon graves”? The background of the Anglo-Saxon weapon burial rite.’ Past & Present 126 (1990), pp.22-43.  For analogous issues among the Franks, see G. Halsall, ‘Material culture, gender, sexuality and transgression in sixth-century Gaul’, ‘Merovingian masculinities’, and ‘Growing up in Merovingian Gaul’, in id., Cemeteries and Society in Merovingian Gaul: Selected Studies in History and Archaeology, 1992-2009 (Leiden 2010), pp.323-55, 357-81, & 383-412 respectively.

[59] Annals of St-Bertin, s.a. 859: The Annals of St-Bertin. Ninth-Century Histories, Volume 1, trans. J.L. Nelson (Manchester, 1991).

[60] Edict of Pîtres (25 June, 864), ch.26: Monumenta Germaniae Historica: Legum Sectio 2: Capitularia Regum Francorum 1-2 ed. A. Boretius & V. Krause (2 volumes; Hanover, 1883, 1895-7), no.273.

[61] The classic statement is H.M. Chadwick, The Heroic Age (Cambridge, 1912).

[62] The Gododdin: Text and Context from Dark-Age North Britain, ed. & trans. J.T. Koch, (Cardiff, 1997).  The Taliesin Poems: New Translations ed. & trans. M. Pennar (Llanerch, 1988).

[63] The Gododdin, ed. & trans. Koch.

[64] Walthari: A Verse Translation of the Medieval Latin Waltharius trans. B. Murdoch (Glasgow, 1989). The Tain, trans. T. Kinsella (Oxford, 1969).

[65] W. Goffart, ‘Conspicuously absent: martial heroism the Histories of Gregory of Tours and its likes’, in The World of Gregory of Tours, ed. K. Mitchell & I.N. Wood (Leiden 2002), pp.365-393.

[66] For the frequency of Frankish Warfare in the later eighth century, one should read The Royal Frankish Annals and The Revised Annals (above, n.48).  For a sample from an earlier period, see Halsall, Warfare and Society in the Barbarian West , Appendix, pp.231-33. For peace treaties, consult the series of treaties between the Frankish rulers of Italy and the Doges of Venice, contained in Capitularia Regum Francorum 1-2 ed. Boretius & Krause, nos. 233-4 (840), 235 (856), 237 (883), 238 (888) and 239 (891).

[67] Halsall, Warfare and Society in the Barbarian West, pp.2-3, 142.

[68] Halsall, Warfare and Society in the Barbarian West, pp.137-38.

[69] Halsall, Warfare and Society in the Barbarian West, pp.137, 140-41.

[70] Halsall, Warfare and Society, pp.119-33 for  survey of the topic, with references.

[71] Halsall, Warfare and Society in the Barbarian West, pp.156-62.

[72] This, for example, was the fate of the Slavic dux Liudewit.  Royal Frankish Annals, s.a. 822.

[73] H. Clarke & B. Ambrosiani, Towns in the Viking Age (revised edition, Leicester, 1995), pp.156-58.

[74] Halsall, Warfare and Society in the Barbarian West, pp.223-27.

[75] Halsall, Warfare and Society in the Barbarian West, pp.90-91, 135-37.

[76] Classically, T. Reuter. T. Reuter, ‘Plunder and tribute in the Carolingian empire.’ Transactions of the Royal Historical Society, 5th series, 35 (1985), pp.75-94; id. ‘The end of Carolingian military expansion’, in Charlemagne’s Heir. New perspectives on the Reign of Louis the Pious, ed. P. Godman & R. Collins  (Oxford, 1990), pp.391-405

[77] Halsall, Warfare and Society in the Barbarian West, pp.177-80.

[78] See. E.g. Waltharius’ account of a battle between two mounted forces, and note the similarity with Anglo-Saxon poetic accounts of the conflicts between shield-walls.  It ought to be noted, however, that Anglo-Saxon armies are very rarely, even in these sources, actually specified as being entirely dismounted.

[79] Halsall, Warfare and Society in the Barbarian West, pp.89-101.

[80] See the works of Tim Reuter (above, n.77)

[81] For the debate on the ‘difference’ of Viking warfare, see, e.g., S. Coupland, ‘The rod of God’s wrath or the people of God’s wrath? The Carolingians’ theology of the Viking invasions’, Journal of Ecclesiastical History 42.4 (1991), pp.535-54; G. Halsall, ‘Playing by whose rules? A further look at Viking atrocity in the ninth century.’ Medieval History 2.2 (1992), pp.3-12; N. Lund, ‘Allies of God or man?  The Viking expansion in a European perspective’, Viator 20 (1989), pp.45-59; P.H. Sawyer, The Age of the Vikings (2nd edition; London, 1971); C.P. Wormald, ‘Viking Studies: whence and whither?’ in R.T. Farrell (ed.), The Vikings(Chichester, 1982), pp.128-53.

[82] Annals of St-Bertin, s.a. 858.

[83] Annals of St-Bertin, s.a. 833; Thegan, Life of Louis the Pious, ch.42: Carolingian Civilization: A Reader, trans. P.E. Dutton (Peterborough, Ontario, 1996), pp.141-55; The ‘Astronomer’, Life of Louis, ch.48: Son of Charlemagne: A Contemporary Life of Louis the Pious, trans. A. Cabaniss (Syracuse NY, 1961)

[84] Annals of St-Bertin, s.a. 866.

[85] Halsall, Warfare and Society in the Barbarian West, pp.96-101.

[86] For a good, solid study, see Renard, ‘La politique militaire de Charlemagne at la paysannerie franque’ (above, n.17).

[87] Italian Capitulary of 801, ch.2, 3; Capitulary on Mobilisation for the use of the missi (808), ch.2; Aachen Capitulary (810), ch.12, 13: Capitularia Regum Francorum, ed. Boretius and Krause, nos. 98, 50, 64.

[88] Recent excellent discussions of later Carolingian rulers with useful discussion of their military policies are to be found in E.J. Goldberg, Struggle for Empire: Kingship and Conflict under Louis the German, 817-876 (Ithaca, 2006), & S. MacLean, Kingship and Politics in the Late Ninth Century: Charles the Fat and the End of the Carolingian Empire (Cambridge, 2003).

[89] Capitularia Regum Francorum, ed. Boretius and Krause, no.218.

[90] Ludovici II Diplomata, ed. K.Wanner (Rome, 1994), nos.20, 47 and 35.

[91] Annals of St-Bertin, s.a., 869.

[92] Carolingian Civilisation, trans. Dutton, pp.485-99.

[93] For later Anglo-Saxon military service see Abels, Lordship and Military Organisation.  The standard view prior to Abels’ work is represented by C. Warren Hollister, Anglo-Saxon Military Institutions on the Eve of the Norman Conquest (Oxford, 1962).

[94] On Viking armies, see: N. Lund, ‘Allies of God or man?  The Viking expansion in a European perspective’, Viator 20 (1989), pp.45-59, at pp.52-54; Halsall, Warfare and Society in the Barbarian West, pp.106-108.

[95] For Ashdown, see Anglo-Saxon Chronicle, s.a., 871; Asser, Life of King Alfred, ch.: Alfred the Great: Asser’s Life and other Contemporary Sources, trans. S. Keynes & M. Lapidge (Harmondsworth, 1982), pp.65-110; for Saucourt see Annals of St-Vaast, s.a., 881: Monumenta Germaniae Historica Scriptores Rerum Germanicarum 12, ed. B. de Simson (Hanover, 1909), pp.40-82; Ludwigslied: Carolingian Civilisation, trans. Dutton, pp.482-3; for the Dyle, see Annals of Fulda, s.a. 891: The Annals of Fulda. Ninth-Century Histories Volume II, trans. T. Reuter (Manchester, 1992); Regino of Prüm, Chronicle, s.a. 891: History and politics in late Carolingian and Ottonian Europe. The chronicle of Regino of Prüm and Adalbert of Magdeburg, trans. S. MacLean(Manchester, 2009).

[96] For attacks late in the year in Britain see the Anglo-Saxon Chronicle (which at this date began its year in September), s.a.. 866, 867, 868, 870, 872, 873, 874, 875, 878.  In Francia, an attck was made on Poitiers around Christmas: Annals of St-Bertin, s.a.863; one attack on Tours was almost certainly aimed at arriving on 11 November, the feast day of Tours’ patron Saint, St Martin, though it arrived a few days early on the 8th: Annals of St-Bertin, s.a. 853.

[97] The kings of Northumbria: Anglo-Saxon Chronicle s.a. 867; Edmund of East Anglia: Anglo-Saxon Chronicle, s.a. 870; Æthelræd of Wessex: Anglo-Saxon Chronicle, s.a. 871.

[98] Eg. The battles of Brisarthe (866); Reading (870); and Saucourt (882): Regino of Prüm, Chronicle, s.a. 866; Annals of St-Bertin, s.a., 866; Anglo-Saxon Chronicle, s.a.871.  For Saucourt, see above, n.94.

[99] Anglo-Saxon Chronicle, s.a., 880.

[100] For the discussion of Viking atrocity, see above, n.82.

[101] Halsall, Warfare and Society in the Barbarian West, pp.111-116.

[102] S. Coupland, ‘‘The fortified bridges of Charles the Bald.’ Journal of Medieval History 17 (1991), pp.1-12; B. Dearden, ‘Charles the Bald’s fortified bridge at Pîtres (Seine): recent archaeological investigations’ Anglo-Norman Studies 11 (1989), pp.107-12; J. Hassall & D. Hill, ‘Pont-de-l’Arche: Frankish influence on the West Saxon burh?’ Archaeological Journal 197 (1970), pp.188-95; Halsall, Warfare and Society in the Barbarian West, pp.99-100, 219-20.

[103] On the Burghal Hidage, see above all D. Hill, ‘The Burghal Hidage: The establishment of a text.’ Medieval Archaeology 13 (1969), 84-92; D. Hill & A. Rumble (ed.), The Defence of Wessex: The Burghal Hidage and Anglo-Saxon Fortifications (Manchester, 1996).  R. Abels, ‘English logistics and military administration, 871-1066: The impact of the Viking wars’, in Military Aspects of Scandinavian Society in a European Perspective AD 1-1300, ed. A. Nørgård Jørgensen & B.L. Claussen (National Museum Studies in Archaeology and History 2; Copenhagen, 1997), pp. 256-65

[104] Halsall, Warfare and Society in the Barbarian West, pp.102-106. Abels, Lordship and Military Obligation, pp.97-115 is excellent.  See also N.P. Brooks, ‘England in the ninth century: the crucible of defeat’, TRHS 5th ser. 29 (1979), pp.1-20, reprinted with postscript in his Communities and Warfare 700-1400, pp.48-68.

[105] In The Edict of Pîtres (above, n.61).

[106] See, e.g., Annals of St-Bertin, s.a., 862.

[107] Weland’s death: Annals of St-Bertin, s.a.863; Edict of Pîtres (above, n.61).

[108] On tenth-century warfare, the best works remain those of Karl Leyser and his student, Tim Reuter: K. Leyser, ‘The battle at the Lech, 955’, in his Medieval Germany and its Neighbours (London, 1982), pp.43-67; id., Communications and Power in Medieval Europe: The Carolingian and Ottonian Centuries (ed. T. Reuter; London, 1994); T. Reuter, ‘Carolingian and Ottonian warfare’, in Medieval Warfare: A History, ed. M. Keen (Oxford, 2000), pp.13-35; id. ‘The recruitment of armies in the Early Middle Ages: what can we know?’ in Military Aspects of Scandinavian Society in a European Perspective AD 1-1300, ed. Nørgård Jørgensen & Claussen, pp.32-37.

[109] See above, n.102.

[110] Reuter, ‘Carolingian and Ottonian warfare’, pp.24, 28; A classic is K.F. Werner, ‘Heeresorganisation und Kriegführung im deutschen Königreich des 10. und 11. Jahrhunderts’, Italia’, Ordinamenti Militari in Occidente Nell’Alto Medioevo (Settimane di Studio 15) (Spoleto, 1968), pp.791-843.

[111] Leyser, ‘The battle at the Lech, 955’, p.58.

[112] T. Reuter, Germany in the Early Middle Ages, 800-1056 (London, 1991), p.143.

[113] Reuter, Germany in the Early Middle Ages, pp.70-180.

[114] N. Lund, ‘If the Vikings knew a Leding – What was it like?’, in Developments around the Baltic and the North Sea in the Viking Age, ed. B. Ambrosiani & H. Clarke (Birka Studies 3; Stockholm, 1994), pp.98-105; id., ‘The Danish Empire and the End of the Viking Age’, in The Oxford Illustrated History of the Vikings, ed.P.H. Sawyer (Oxford, 1997), pp.156-81; G.A.E. Morris, ‘Violence and late Viking Age Scandinavian social order’, in Violence and Society in the Early Medieval West, ed. G. Halsall (Woodbridge, 1998), pp.141-56.  For the circular fortresses, see E. Roesdahl, Viking Age Denmark (London, 1982), pp.147-55.

[115] C. Sánchez-Albornoz, ‘El ejército y la guerra en el reino Asturleones 718-1037’, Ordinamenti Militari in Occidente Nell’Alto Medioevo (Settimane di Studio 15) (Spoleto, 1968), pp.293-428; more recently, J. Escalona Monge, ‘Las prestaciones de servicios militares y la organización de la sociedad feudal castellana: los infanzones de Espeja’, Castillos de España 94 (1987), pp.55-60.

[116] J. Dunbabin, France in the Making, 843-1180 (Oxford) remains a very useful introduction to this period of French history.

[117] An excellent recent overview is S. MacLean. ‘Review article: apocalypse and revolution: Europe around the year 1000.’ Early Medieval Europe 15 (2007), pp.86-106.  A debate on the issue in the journal Past & Present contained the following contributions: T.N. Bisson, 'The Feudal Revolution', Past & Present 142 (1994), pp.6-42; D. Barthélemy, ‘Debate: The Feudal Revolution. I.’ Past & Present 2 (Aug., 1996), pp.196-205; S.D. White, ‘Debate: The Feudal Revolution. II.’ Past & Present 152 (Aug., 1996), pp.205-23; T. Reuter, 'Debate: The Feudal Revolution. III.' Past & Present155 (May 1997), pp.177-95; C.J. Wickham, ‘Debate: The Feudal Revolution. IV.’ Past & Present 155 (May 1997), pp.196-208; T.N. Bisson, ‘Debate: The Feudal Revolution. Reply.’ Past & Present 155 (May 1997), pp.208-25.  Other important contributions include: D. Barthélemy, ‘The year 1000 without abrupt or radical transformation’, in Debating the Middle Ages. Issues and Readings, ed. L.K. Little, &  B.H. Rosenwein, (Oxford, 1998), pp.134-47; G. Bois, The Transformation of the Year 1000. The Village of Lournand from Antiquity to Feudalism (Manchester, 1992); The Apocalyptic Year 1000. Religious Expectation and Social Change, 950-1050 ed. R. Landes, A., Gow & D.C. Van Meter (Oxford, 2003); J.-P. Poly & E. Bournazel, The Feudal Transformation (New York, 1991).

[118] For late Carolingian ‘castles’, see: C.L.H. Coulson, ‘Fortresses and social responsibility in late Carolingian France’, Zeitschrift für Archäologie des Mittelalters 4 (1976), pp.29-36.  A. Debord, ‘Castrum et castellum chez Ademar de Chabannes’, Archéologie Médiévale 9 (1979), pp.97-113; Deyre, ‘Les châteaux de Foulque Nerra’, Bulletin Monumentale 32 (1974), pp.7-28; M. Fixot, Les fortifications en terre et la naissance de la féodalité dans le Cinglais(Caen, 1968); J. Le Maho, ‘De la Curtisau château: l’exemple du Pays de Caux’, Château Gaillard 8 (1977), pp.171-83; Les Fortifications de terre en Europe occidentale du Xe au XIIesiècles [Colloque de Caen, 2-5 Octobre 1980] (= Archéologie Médiévale 11 (1981); G. Noyé, ‘Les fortifications de terre dans la seigneurie de Toucy du Xe au XIIIesiècle.  Essai de typologie’, Archéologie Médiévale 6 (1976), pp.149-218; A. Renoux, ‘Châteaux normands au Xe siècle dans le de moribus et actis primorum Normanniae ducum de Dudon de Saint-Quentin’, Mélanges d’archéologie et d’histoire médiévales en l’honneur du doyen Michel de Boüard(Geneva, 1982), pp.327-46.