Landscape and warfare in Anglo-Saxon England and the Viking campaign of 1006

This paper outlines the state of research into early medieval conﬂict landscapes in England and sets out a theoretical and methodological basis for the sustained and systematic investigation of battleﬁeld toponymy and topography. The hypothesis is advanced that certain types of place were considered particularly appropriate for the performance of violent conﬂict throughout the period and that the social ideas that determined the choice of locale are, to some degree, recoverable through in-depth, interdisciplinary analysis of landscapes, place names and texts. The events of 1006 and the landscape of the upper Kennet are introduced as a case study that reveals the complex interplay of royal ideology, superstition and place that were invoked in the practice of violence in late Anglo-Saxon England. In the course of the discussion, this paper seeks to demonstrate the value of applying a similar approach to the full range of evidence for conﬂict landscapes in early medieval England and beyond. this ), possible battleﬁeld locations. In ‘Searching for Brunanburh’ (n. 3 above), Michael Wood presents an exemplary study that highlights the limitations of a mono-disciplinary approach to battleﬁeld studies. Wood also makes a persuasive case for identifying the site of Brunanburh with a location on the river Went, at a site with long continuity of use as an assembly and muster point in a highly militarized landscape.

addressed the subject directly and in general terms. 1 It is true enough that individual early medieval battlefields have been a frequent subject of enquiry for historians. This, however, has generally been in the context of a desire to fix a location for an iconic event or to reconstruct the probable movement of troops on the battlefield. Studies of the former sort -often resting heavily on the evidence of place names -abound in local history and place-name publications. Indeed, the number and specificity of such studies militates against a systematic review; 2 some iconic battles have spawned minor publishing industries in their own right. 3 Nevertheless, attempts to establish the location of most early medieval battlefields with precision have generally been unsuccessful and, as John Carman has recently pointed out, the nature of enquiry has long been dictated by the preoccupations of military historians, arguably arresting the development of battlefield archaeology as an independent sub-discipline. 4 The last twenty-five years have seen huge advances made in the way that battlefields can be recorded and understood through archaeological techniques, but these methods have only recently been accepted as a useful complement to traditional military history. Only since 2012 has English Heritage formally included archaeological approaches in the selection guidance for designating sites for the English Heritage Register of Historic Battlefields. 5 The seminal work combining historical and archaeological approaches to battlefields was undertaken in the late 1980s at the Little Bighorn, 6 but these techniques have not so far been applied successfully to any early medieval battlefield -a situation that arises in part from the difficulties in securely identifying battlefields of the period to the degree of precision that the techniques of the archaeological method demand. 7 Of the forty-three battlefields that English Heritage has included on its register, only three date from before 1100. Two of these -Hastings and Stamford Bridge -are from the same year (1066). The other is the battle of Maldon (991). The Battlefields Trust is slightly less pessimistic, and its register includes an additional two early medieval battlefields. 8 One of these is another from 1066 (the controversial site of the Battle of Fulford), 9 the other is the endlessly contested location of the Battle of Brunanburh. 10 Other battles of the period are represented as dots on a map, without names, dates or any other information (or even an indication of whether the map is a comprehensive representation of the number of early medieval battlefields). 11 This picture has recently been clarified to a great extent by Glenn Foard and Richard Morris's review of English battlefields, which enu- merates conflict at the supra-regional level and offers a fairly pessimistic assessment of the potential for future research in the field. 12 Moreover, degradation of the archaeology of such places through development, failures in the protection legislation and unscrupulous metal-detectoring, means that the potential to fruitfully apply the approaches of battlefield archaeology is ever diminishing. 13 This means not only that the unrecorded evidence from the handful of locatable sites is under constant threat, but also that the likelihood of identifying battlefields described in documentary sources from their archaeological signature is also diminishing; a gloomy prognosis for future study. So far, only the late medieval battlefields at Towton and Bosworth have shown themselves amenable to the sort of detailed investigation pioneered in the USA and applied by Glenn Foard to the civil war battlefields at Naseby and elsewhere. 14 Various archaeological remains can be used to suggest traces of past conflict: in particular, bones showing weapon trauma, defensive structures, and material remains of battlefield technology. Mags McCartney has also suggested that these categories can be added to with a subtle reading of site morphology -looking for what he calls 'traces of fear' in the social organization of past societies that might reveal the presence or threat of aggression. 15 The latter raises some interesting issues, especially for the changing layout of Anglo-Saxon burhs, settlements and manorial enclosures. 16 The other three categories of archaeology have all left their traces in early medieval Britain, but rarely in ways that help to locate landscapes of battle. As Andrew Reynolds succinctly puts it, 'Such remains, both monumental and artefactual, represent direct evidence for militarized society, yet they do not (necessarily) explicitly indicate actual conflict.'   Whilst the archaeology that could be used to locate and identify conflict is largely absent, warfare looms large in the chronicles and histories that survive from the early medieval period. However, whilst written accounts of the period often provide information about the protagonists and associate the battle with a place name, information about the nature and conduct of warfare is rare and a relatively small amount of it before the ninth century seems to be contemporary with the events described. 18 Even when accounts are elaborated at greater length (take for example the poetic treatment of the battle of Maldon or Asser's description of the battle of Ashdown), they are made problematic by the use of literary devices, traditional motifs and archaic references -often drawing on a Roman military vocabulary that had long passed out of everyday use. 19 Equally sparse is the sort of topographical, strategic and logistical detail that might make early medieval battlefields more readily identifiable. Moreover, problems with the sources mean that even when a place might be identifiable through a place name, it is not always clear that the event described did in fact occur at the place recorded. A famous example is the battle of Maldon (991) which, were it not for the independently surviving poem describing the battle, might be assumed to have taken place at the burh mentioned in the Anglo-Saxon Chronicle rather than on the shores of the Blackwater estuary where it is now generally agreed to have occurred -at some remove from the fortified town. 20

A way forward?
Given these conditions, one might be forgiven for abandoning the study of early medieval conflict landscapes as a lost cause. However, by considering sites of conflict in relation to their wider landscape settings and the symbolic associations of topographical and man-made features, it may be possible to find a way forward. Typically, approaches that have considered the relationship of warfare to landscape have done so in the shadow of the military historical/geographical tradition that has its roots in post-Enlightenment efforts to establish a theory of warfare in line with the principles of scientific rationalism. 21 The concept of landscape as 'terrain' -a passive object to be used or overcome in response to pragmatic strategic, tactical and logistical concerns -is central to this way of thinking, and has been enormously influential. 22 The ahistorical bias implicit in this approach has been justified (and compounded) by frequent recourse to the concept of 'Inherent Military Probability' (IMP), a principle popularized by the military historian Alfred Burne, 23 which attempts to get around the problems of applying modern military precepts to the past by postulating that commanders behave according to universal laws of rational military behaviour. The questionable basis on which this theory rests can be summarized as follows: 1. IMP assumes that all commanders everywhere are equally competent, have received similar or equivalent training, and that effective command and control structures have been present or effective in all armies at all times. 2. It denies the influence of culturally specific rituals and taboos around the appropriate conduct of warfare and/or perceptions of the landscape. 3. It requires no corroborating archaeological or historical evidence.
In many cases, IMP has served as a means by which historians who have received military training (a sizeable constituency throughout the twentieth century) have sought to legitimize assumptions made on the basis of their own experience rather than on a foundation of primary evidence. Although rarely applied explicitly, its influence can be felt in a number of works dealing with early medieval conflict landscapes. Kerry Cathers, for example, makes a good number of salient points in a chapter that seeks to explain the monumental correlates for battlefields identified by Guy Halsall, 24 especially in recognizing that battles had to be fought by agreement and in places known to the combatants (the image of the frustrated warlord leading his tired men aimlessly around the landscape in search of a fight is quickly and rightly dismissed). However, when she asserts that 'it is unlikely that military commanders would prefer a site with religious or supernatural connections if it were not a suitable battle- 21 Most influential has been the early nineteenth-century view of Carl von Clausewitz, written in the aftermath of the Napoleonic wars and published posthumously in 1832. C. von Clausewitz, On War (1832), ed. and trans. M. Howard and P. Paret, rev. edn (Princeton, 1984), pp. 348-9: 'Geography and ground can influence military operations in three ways: as an obstacle to the approach, as an impediment to visibility, and as cover from fire.'  site, over another location which was more suited to battle', or that to suggest that roads, fords and monuments could be included as part of a battlefield is 'arguing contrary to common sense', a fatal epistemological bias is introduced. 25 The precise factors that may have made a battle site 'suitable' cannot be assumed from a modern rationalistic perspectivethat they might include religious or supernatural considerations is, moreover, highly probable given what else is now believed about perceptions of the landscape in early medieval society. 26 Likewise, what may have constituted 'common sense' between the fifth and eleventh centuries was based on a radically different (and changing, contradictory and heterogeneous) stock of cultural assumptions: this was, after all, a society that believed in the efficacy of judicial ordeal as a means of establishing criminal guilt. 27 It would be wrong to deny any element of shared humanity with the members of past societies: early medieval warriors got hungry, preferred to sleep indoors when they could and -all other things being equalscorned the cross-country hike through moor, fen and fastness when dry and well-trodden paths were available. To suggest otherwise would be absurd, and it is for this reason that it remains imperative to consider the location of conflict sites in relation to wider military networks: muster sites, communication networks, signalling systems, fortified places. However, it is also important to recognize with John Carman that 'battlefields are carefully chosen and necessarily reflect attitudes to the appropriateness of the use of space for particular functions' -not merely a rehearsal of prosaic assumptions about military probability, but a complex relationship of social, political and cosmological ideas. 28 This is essentially to argue that the study of early medieval battlefields should sit 25 alongside other modern approaches to medieval landscapes and monuments, not least the reframing of debates relating to the later medieval castle where theoretical approaches to the symbolic use of space are significantly advanced in comparison to contemporary battlefield studies. 29 Such a method has the potential to draw on comparable approaches to warfare in other disciplines, most notably in anthropology, prehistoric archaeology and geography. Drawing on the writings of Hegel, the cultural geographer Michael J. Shapiro has framed an important distinction between what he describes as the two 'faces' of warfare. The first of these is the visible face, the face that is presented by politicians and historians to show warfare as 'an instrument of state policy': 'Instrumental and rationalistic talk links the features of war with enduring projects of the state: maintaining security, clearing spaces for effective and vital functioning, meeting obligations to friends, and so on.' 30 It is this face, with some notable exceptions, that has received most attention from those writing on the subject of early medieval warfare, in particular the ways in which it can be linked to the development of social and civic institutions and the implementation of organized systems of civil defence. 31 The second face is ontological, concerned with the identity of the war-making group. It uses violence as a communicative strategy to build meaning within the self-group, to express and delimit values and ideas, and to share these ideas with others. In other words, warfare functionson one level -as semiotic exchange, and a number of other writers have similarly emphasized the communicative, performative qualities of 29 See D. Stocker, 'The Shadow of the General's Armchair', Archaeological Journal 149 (1992), pp.
415-20, for an important salvo in the 'revisionist' approach to castles. This review paper helped to firmly establish the position, long advanced by historians such as Charles Coulson, that the castle should not be seen as exclusively, or even primarily, shaped by military functionalism, but rather that in form, location and landscape context these ostensibly militarized places were contingent on a variety of cross-cutting social, religious, economic and ideological realities.  warfare and violence. 32 In order to apprehend this 'second face', Shapiro suggests that the semiotic function of violence can be sought in so-called 'structures of expression' -the ways in which a society conducts and presents itself in war. 33 The places that were considered appropriate to warfare -as glimpsed through the evidence of place names and the circumstantial details of battle -can be thought of precisely as a 'structure of expression' with the potential to reveal wider social attitudes to the practice of violence. Indeed, this has been the underlying theoretical basis for Carman's diachronic study of conflict landscapes. 34 Seen from these perspectives, early medieval traditions that describe a disproportionate number of battles fought at places that -for example -imply the existence of ancient monuments or ancestral memorials, take on a profound significance for understanding how early medieval societies in Britain thought about the conceptual relationship between landscape and warfare. In the study alluded to above, Guy Halsall has pointed out that, of twenty-eight battlefields named in written sources and dated between 600 and 850 (excluding Viking raids and civil wars), almost all can be placed at ancient monuments and/or water crossings on the evidence of the documentary record. 35 More recently, Ryan Lavelle has demonstrated the relationship of battle sites with the political geography of later Anglo-Saxon Wessex. 36 Whilst these tendencies can be understood in the light of strategic and organizational considerations, the specific choice of places associated with antiquity, boundaries and royal identity points towards an engagement with other symbolic associations. 37 The success of these studies in associating conflict with clearly differentiated landscape settings and their semiotic content provide a partial model for a more comprehensive analysis of the available evidence.

Landscape and warfare in Anglo-Saxon England
However, as Aliki Pantos incisively demonstrates in relation to the investigation of assembly sites, 38 problems arise in relying exclusively on toponymic and documentary evidence. A place name may, in fact, conceal more than it illuminates concerning the characteristics of a location. The example of Bradford-upon-Avon in this regard is instructive. The site of a battle recorded in the Anglo-Saxon Chronicle under the year 652, Bradford might be included in Halsall's list of conflicts at 'water-crossings'. 39 The Chronicle entry is exceptionally tight-lipped, even for this famously laconic document: 'Cenwalh fought' is all that the scribe saw fit to relate -even the object of Cenwalh's belligerence goes unidentified. The written evidence alone does not permit of any further analysis. However, investigation of the wider landscape reveals that Bradford-on-Avon was the site of a Roman settlement, an iron-age hill fort and an eighth-century monastic foundation. 40 Some of these may have coloured the choice of the location as an appropriate theatre of war -whether by Cenwalh and his foes in 652, or by an unknown West Saxon scribe in the ninth century -and the investigation of wider battlefield environments can therefore furnish a great deal of additional relevant data.
As Foard and Morris recognize, a substantial number of early medieval conflicts can be located on a broad, sub-regional basis (generally too broad, in fact, to be useful in the context of the current research agendas and methodologies of battlefield archaeology). 41 However, as John Baker and Stuart Brookes have recently demonstrated, an approach that combines the evidence of place names with the wider traces of militarism in the landscape can go a long way towards reconstructing historical systems of civil defence. 42 Battlefields naturally sit within or in relation to those systems, 43 and are amenable to investigation through similar data sets. It is proposed that through combining an understanding of the militarized landscape with a close reading of the wider historical context and the   evidence of place names, a number of desirable research outcomes can be generated -not least the provision of a sophisticated toolkit for identifying battle sites where etymological analysis alone has failed to convince or has thrown up multiple competing sites. 44 In addition, research carried out along these lines across a large number of conflict sites has the potential to reveal why certain types of place figure more frequently than others in the documentary record and to identify the place of warfare in the organization of the early medieval landscape. On an individual level, as the following case study is intended to illustrate, the close study of individual conflict sites can reveal a great deal -not just about the landscape of civil defence or strategies of inter-polity aggression, but also about the ideas, attitudes and ideologies that underpinned the performance of violence in the early Middle Ages.

The battle at the Kennet in 1006
Historical context The events of 1006 are recorded in the CDE manuscripts of the Anglo-Saxon Chronicle and are given here in translation: Then when it came near to winter, the army travelled home, and after Martinmas the raiding-army came to its secure base in the Isle of Wight, and there provided themselves everywhere whatever they needed. And then towards midwinter they took themselves to their prepared depots, out through Hampshire, into Berkshire at Reading; and they did, in their custom, ignite their beacons as they travelled; and travelled then to Wallingford and scorched it all up; and then turned along Ashdown to Cwichelm's Barrow and there awaited the boasted threats, because it had often been said that if they sought out Cwichelm's Barrow they would never get to the sea. Then they turned homewards by another route. Then the army were assembled there at Kennet [aet Cynete], and there they joined battle; and the Danes soon brought that troop to flight, and afterwards carried their war-booty to the sea. There the people of Winchester could see the raiding-army, 44 E.g. Brunanburh, Ashdown, Assandun, etc. See Foard and Morris, The Archaeology of English Battlefields, p. 45 for a tabulation of the multiplicity of possible battlefield locations. In 'Searching for Brunanburh' (n. 3 above), Michael Wood presents an exemplary study that highlights the limitations of a mono-disciplinary approach to battlefield studies. Wood also makes a persuasive case for identifying the site of Brunanburh with a location on the river Went, at a site with long continuity of use as an assembly and muster point in a highly militarized landscape.
proud and not timid, when they went by their gates to the sea and fetched themselves provisions and treasures from over 50 miles from the sea. 45 This campaign came in the context of major raids undertaken from the 990s onward that exploited the apparent inability of the late Anglo-Saxon state to adequately defend itself from a resurgent threat from Scandinavian aggression across the North Sea, and which would ultimately result in the conquest of England by Swein Forkbeard in 1013 and then again in 1016 by his son, Cnut. On this occasion, the ability of the Viking army to range (relatively) unopposed for scores of miles inland may have been intended to hasten the payment of tribute by demonstrating the inadequacy of English civil defence. In the following year, the Viking army would receive a payment of 36,000 pounds from the English crown. However, the Chronicle account of 1006 is remarkable for several reasons. Firstly, it explicitly describes resistance offered by an Anglo-Saxon army during the course of the Viking march, which implies the existence of an effective system of mobilization. Secondly, it offers a unique record of attitudes relating to a named landscape feature that suggest an association between ideas about conflict and places of communal memory. Finally, the Viking route through Wessex suggests a deliberate manipulation of the associations of place that would have had maximum psychological impact for the harassed English population. Close investigation of the landscape -particularly of the probable location of the battle at the Kennet -can shed further light on the significance of all of these points, and suggest ways in which the conflict landscape of late Anglo-Saxon England can be further interpreted.

Topography and archaeology
On the face of it, the battle at the Kennet is one of ten conflict sites in Wessex whose primary locational characteristic is the proximity to water -in this case the vicinity of the river Kennet, which has its origins in the Avebury region of Wiltshire. There is, however, sufficient documentary and circumstantial evidence to provide the battle with an unusually secure and investigable topography. Nick Baxter, in an unpublished article on the location, significance and historical context of the battle, has surveyed the evidence of tenth-century charters, and the associated chronicle account provided by John of Worcester, to locate the battle at the crossing of the river Kennet near the village of East Kennet. 46 In summary, this rests on the description of the battle as 'at' or 'next to' Kennet ('aet Cynete'; 47 'juxta Kenetan' 48 ), with the implication that the references imply a specific place, rather than the river in general. Kennet (Cynetan) occurs as a place name in a 972 charter of Athelstan granting land within Overton hundred, 49 indicating that an area within the hundred was known by that name as early as the tenth century. The modern village names of East and West Kennet preserve this local toponym. It seems highly probable on the basis of this charter that, to a West Saxon audience, the place name Kennet had quite specific associations, a suggestion made more plausible by the precision with which other places in the chronicle entry are identified (including a rare allusion to a named assembly site). 50 The most likely place, on this reading of the evidence, at which the Anglo-Saxon fyrd would have intercepted the raiding army, is just north of the point at which the Ridgeway crosses the river Kennet just south of Overton Hill near the aforementioned village of East Kennet (Fig. 1). For the Viking army to reach the river from Cwichelm's barrow (Cwichelmeshlaew, the shire meeting place of Berkshire but now within the post-1972 boundaries of Oxfordshire), the most straightforward route would have been to follow the line of the Ridgeway. 51 Additional reasons to support this thesis arise in the course of the ensuing discussion.
The crossing of the Kennet lies in the river valley at the foot of Overton Hill, over which the Ridgeway climbs -eventually reaching a height of over 250 metres. Immediately to the north of the crossing, after a steep climb from the valley bottom, the gradient becomes markedly shallow as the Ridgeway continues to climb northwards and, at the change in gradient (approx 350 metres from the crossing), is the site of the 'Sanctuary' -a stone, and earlier wooden, circle that was still largely intact as late as 1725 (Fig. 2)  sanctuary . . . and the veneration for it has been handed down thro' all succession of times and people'. 52 How much this assertion reflects reality is of course unknowable, but one suspects that Stukely was probably correct in his assessment: it is self-evident that a monument like the Sanctuary would have impressed itself forcibly on the imaginations of anyone who came in contact with it.
In 1685, the apothecary Robert Toope wrote to John Aubrey with description of a cemetery he had discovered by chance adjacent to the still-standing stone circle. He describes how the bodies lay 'So close one by another that skull toucheth skull . . . At the feet of the first order I saw lay the heads of the next, their feet intending the temple: I really believe the whole plaine, on that even ground is full of dead bodies.' 53 Sadly, Toope's habit of pulverizing human bones for use in medicinal remedies seems to have taken its toll, 54 and no trace of Toope's cemetery has been found in subsequent excavations. It is entirely likely that any such cemetery was associated with periods other than the early medieval. Nevertheless, the presence of a mass grave of late tenth-/early eleventh-century date in a similar landscape context on the Dorset Ridgeway at least  provides a possible analogue for such a burial in the late Saxon period, although the deposition of remains at the Ridgeway cemetery does not correspond to the sardine-like scenario that Toope's description implies. 55 Several modern excavations have taken place at the Sanctuary. 56 None has found evidence of Toope's cemetery. These excavations have understandably focussed on the substantial Neolithic and bronze age activity at the site and are not -for that reason -summarized here. Early medieval remains in the vicinity are not substantial and certainly cannot be used to prove or disprove the location of the battle. 57 Only a late Anglo-Saxon stirrup strap mount, found in the river Kennet near Silbury Hill, provides material evidence for a military presence of the right date in the immediate vicinity of the likely battle site. 58 This is not surprising, however, given the total absence of remains from any first-millennium English battlefield. 59 Three hundred and sixty degree views from the Sanctuary demonstrate lines of sight with the top of Silbury Hill to the west north-west, the numerous bronze-age round barrows that dominate the level ground to the north-east, and the rising hillside north north-west (as well as single outlying barrows to the north and immediately south of the stone circle site on either side of the Ridgeway - Fig. 3). It is this funerary landscape that is presumably referred to as seofan beorgas (seven barrows) in the boundary clause of the 972 charter. 60 West and East Kennet long barrows are also visible from the Sanctuary, forming part of the wider Neolithic ritual landscape. Directly to the north-west lies the line of the stone avenue that proceeds from the henge   it is important to place the battle site in this context. Most strikingly, the battle site is situated in close relationship to the postulated beacon relay linking Yatesbury and Avebury (via Silbury Hill), to Totterdown and thence to Marlborough. 64 It seems likely that a station at Totterdown would have seen any approaching army on the Ridgeway from a considerable distance and been able to relay that knowledge to the minor burhs at Avebury and (probably) Marlborough almost instantly, providing perhaps as much as a day's warning of an approaching enemy. Certainly the Chronicle entry for 1006 implies a preparedness on the part of the West Saxon levy that faced the Viking raiding army at the Kennet -even though they were ineffective in dealing with the threat. The presence of this system implies that this crossroads was recognized to be an important strategic junction and thus formed part of a militarized landscape. 65 Archaeological excavations at Silbury have revealed evidence of a structure, including a palisade, dating from c.1010. 66 It is possible that, in reaction to the failure of the local fyrd to counter the Viking threat in 1006, this area was considered to require investment to reinforce its defensive capability.
The fact that an army assembled near the Sanctuary would have been able to see signals from Silbury and Totterdown adds to the likelihood that the battle was fought at Overton Hill; sophisticated messagesutilizing different woods for variation in colour or intensity of flame, some perhaps carried by runners 67 -could potentially have been communicated right up until the enemy were in view of the assembled fyrd. Reynolds and Brookes also emphasize the role of minor burghal garrisons in contributing to local policing and defensive duties, and it seems probable that the forces assembled at the Kennet were drawn from these settlements. 68 Certainly the communication routes make this plausibleas mentioned above, the battle site is probably near the junction of the Ridgeway with the Roman road from Bath to London, which leads west to Silbury Hill and east to Marlborough. This route runs to the south of a circular network that incorporates a more northerly road system connecting Avebury and Marlborough, segments of which are described as herepaeþ on the first edition Ordnance Survey map and in a boundary clause from a charter of 939 (S 449). A further possibility is that West Civil Defence', pp. 584-6. 67 Reynolds and Brookes, 'Anglo-Saxon Civil Defence', pp. 590-1. 68 Reynolds and Brookes, 'Anglo-Saxon Civil Defence', p. 596. Saxon forces could have approached the Ridgeway using the stone avenue that leads directly from the fortified enclosure at Avebury to the Sanctuary, which may well have been a more clearly delineated route in the eleventh century than it now appears. The fact that, in this case, a battle was fought near to -but not within -a robustly defensible site is unremarkable. 69 The object of initiating the conflict from a West Saxon perspective was to prevent the progress of the Viking raiding army towards the West Saxon heartlands and ultimately to the sea -the futility of remaining within a stronghold in these circumstances is neatly illustrated by the 'proud and not timid' progress of the Viking army past the walls of Winchester. 70 The events of 1006 at the Kennet thus seem to suggest an efficient system of civil defence put into action; even if the end result was disastrous, the network seems to have functioned remarkably well. The Danish passing of Cwichelm's Barrow can thus be interpreted as a trigger that ultimately mobilized an Anglo-Saxon force to block passage from the Ridgeway south into the Pewsey Vale. If so, this might provide a partial explanation for the prophetic passage contained in the Chronicle entry.

The symbolic and mythological landscape
The practical function of a defensive network such as that described by Reynolds and Brookes was to enable lethal force to be brought efficiently to bear on groups perceived to be the enemies of the established system of social and territorial organization. 71 Considered in these terms, it is entirely reasonable to view warfare against an invading army as a form of sanctioned, legitimate, quasi-judicial violence. 72 The laws of King Ine (688-94) -whence the oft-cited definition of a here (the normal way in which Viking armies are described in the Chronicle) as a group of more than thirty-five armed men is derived -makes explicit the association of the term with criminality. A here could thus very well be defined as simply meaning a large group of thieves, as opposed to a small one (hloð) or fewer than seven (simply, ðeofas 'thieves'). 73 Other semantic links -words for wolf (such as wearg), for example, are applied to monsters, Vikings and other social deviants (particular those implicated in illicit violence) 74make the 'outlaw' status of illegitimate armed groups very clear. 75 Seen from this perspective, a battle site can be expected to take on some of the characteristics of execution sites: these are places where lethal violence was meted out to transgressors of social and legal norms. Firm evidential links between warfare and judicial killing in the early medieval period are uncommon, but the recent Weymouth relief road (A354) excavation on the Dorset Ridgeway provides a possible exception. Around fifty headless skeletons were excavated, along with a slightly smaller number of skulls. Isotope analysis demonstrates that the individuals possessed diverse, but predominantly Scandinavian-centred origins and migration histories, 76 and carbon dating places the burial between 970 and 1025. 77 It is highly probable that this is an execution cemetery for Viking prisoners -perhaps a captured raiding party -and the site exhibits the traits that Andrew Reynolds has identified as characteristic of late Saxon execution sites: high elevation and proximity to major routeways (both ensuring high visibility) and association with prehistoric monuments. 78 The Dorset Ridgeway mass grave thus provides a West Saxon example of an association between execution and mass violence -possibly even contemporary with the events of 1006. Given all of this, it is striking that the battlefield at the Kennet exhibits many of the same topographical and geological characteristics. 79 Halsall, in providing some interpretation of his locational findings, raised the intriguing possibility that 'certain kinds of [battle] site were set 73 aside as "places of fear" ', suggesting that the practice of Anglo-Saxon warfare engaged on some level with the supernatural danger associated with specific landscape types. 80 The apparent conceptual association between execution sites and battlefields, alongside aspects of the morphology of the Kennet battlefield, support this suggestion. Most dramatically, the upstanding stone circle may have been thought of as a particularly unwholesome place. There is certainly some suggestion that the similar monument at Stonehenge was regarded as peculiarly dreadful -not only does it appear to have been used as a place of execution and deviant burial in the early Saxon period, but its absence from any documentary sources of pre-eleventh century date is sufficiently remarkable to imply a deliberate silence. It may have represented something especially horrible to the Anglo-Saxon imagination. 81 The similarly surprising omission of the upstanding Sanctuary from all contemporary documentsincluding charter bounds, where it might be expected to figure as a prominent landmark given that the neighbouring barrows do feature in the charter of 972 (S 784; see above) -might be considered in the same light, and the possible presence of a mass burial draws parallels with both the Stonehenge execution burial and the Dorset Ridgeway mass grave. Moreover, the presence of a major crossroads in a landscape dominated by barrows brings into combination features with sinister associations; AElfric was, as late as the 990s, warning of the recourse by witches to both barrows and crossroads for the purpose of summoning the dead. 82 The fact that the location is situated on a territorial boundary (in this case an estate boundary defined by the 972 charter, S 784) also links the battle site with other places associated with judicial violence. 83 The seofan beorgas (seven barrows) that define the location in the charter bounds referred to above are -stone circle aside -the defining monumental characteristic of the Kennet battlefield landscape. There are, in fact, more than seven, and they dominate the field of view in almost Guthlac's demons are similarly bellicose, boasting that 'the throng will come trampling in with troops of horses and with armies. Then they will be enraged; then they will knock you down and tread on you and harass you and wreak their anger upon you and scatter you in bloody remnants.' 87 In both cases, however, neither dragon nor demon seems to have any long-term designs on the wider human realm -even the dragon, having wreaked vengeance for a theft from his hoarded treasure, returns to his dwelling to await the coming of Beowulf and 'trusted now to the barrow's walls'. 88 In every description of the dragon, his guardianship of the barrow is emphasized. Guthlac's demons, too, express themselves in sympathetic terms as the injured party, claiming that Guthlac 'had perpetrated the greatest affliction upon them when, for the sake of bravado in the wilderness, he violated the hills where they, wretched antagonists, had formerly been allowed at times a lodging-place'. 89 It is in defence of their possessions and barrow-homes that these supernatural forces are mobilized; the terrifying behaviour exhibited or threatened by dragons and demons tends to obscure the fundamentally defensive nature of the violence directed towards their respective human adversaries. In both cases the human intervention represents a usurpation of hoard and home, the very things that it was a king's responsibility to protect on behalf of his people (as the poetic entry for the Anglo-Saxon Chronicle in 937 iterates with great clarity). 90 There is thus an ambiguity about the motivations of these mound-dwellers in the literature that complicates their monstrosity and in fact aligns these supernatural forces with traditional aristocratic values. It is thus insufficient to regard the mounds occurring in battlefield contexts as exclusively defined by their negative associations; we need to recognize the ambiguity of violence and monstrosity implied by the battlefield topography. 91 Richard Bradley has perceived in the reuse of ancient monuments an active process whereby social memories are created to legitimize claims to power and land: 'In such cases, links with a remote past, which could not be evaluated before the development of archaeology, may have been used to legitimize the social order', 92 and this idea has been taken up widely in regard to the reuse of pre-historic (and Roman) structures in Britain and Scandinavia during the early medieval period. 93 Both the reuse of ancient sites and the rehearsal of mythic narratives have been interpreted as behaviours that act to collapse temporality, merging contemporary events into what has variously been described as 'myth-time' or 'ritual-time', 94 whereby the protagonist becomes indissoluble from the mythic archetype. It can be argued that the wider locational evidence for conflict presented here points towards battles functioning as rituals that take place in highly charged symbolic loci and which position actors within this myth-ancestral time -thus emphasizing legitimacy, tradition, right to rulership and king-worthiness on a deeply rooted psychosocial level. raises the possibility of some sort of local communal memory or folklore now lost. 96 The battlefield does, however, sit in relation to two other remarkable myth-ancestral conflict sites, both centred on mounds, where mythological archetypes were arguably present and actively deployed: Cwichelm's Barrow and Woden's Barrow. Both are worth discussing briefly here as they relate closely related to the Kennet battle and the choice of battleground.

Invoking Cwichelm
As Guy Halsall has pointed out, there are perfectly sound strategic reasons why prominent monumental features might be chosen as battlefields. In support of that view, he used the example of the 1006 Chronicle entry and reference to the 'boasted threats' made by the English in relation to Cwichelm's Barrow. 97 The function of the barrow as the shire meeting place for Berkshire means that the mound would have occupied a central and defining position within the local area, and was possibly conceived of as being at the heart of regional identity by the local community. It was, in other words, exactly the sort of place where a community might choose to arrange a punch-up; a highly visible and meaningful manifestation of 'home turf '. And yet, whilst this goes some way to explaining this extraordinary Chronicle entry, detailed consideration of the site -now known as Skutchmer Knob (or Cuckhamsley Hill, or variants of both of these names) -supports a deeper reading. Although the site is now severely damaged by earlier and illicit excavations, Sarah Semple and Alexandra Sanmark's excavations at the site have demonstrated that the mound was originally constructed as a prehistoric barrow, probably in the bronze age. 98 The combination of a funerary monument with a name linking it to an early Anglo-Saxon figure -probably the seventh-century West Saxon leader Cwichelm 99 -makes it probable, as Howard Williams has discussed in some detail, that the mound functioned as a memorial to a royal ancestor (possibly even a belief that it contained his grave). 100 The reuse of Cwichelm's Barrow for administrative and legal functions recalls suggestions that funerary landscapes were used in the early Anglo-Saxon period as meeting places where the whole community -living and dead -could engage in identity-shaping ritual. 101 The supposed resting place of a legendary king may have played precisely such a role for the people of Berkshire; a powerful link to the past and a potent symbol of cultural identity and shared heritage that conferred ancient authority to a place associated with the exercise of power. 102 That ancestral traces could be considered manifest in times of assembly leads naturally to the suggestion that they could also be present when the same community was under threat -particularly when conflict centred on the myth-ancestral site itself, and there is growing evidence not only of conceptual and practical associations between battle sites and military assembly (mustering) but also between battle sites and funerary monuments. 103 The recorded traditions surrounding Cwichelm identify him as a notable warrior-king who fought against Penda of Mercia and against the British. Of particular interest is his reported slaying, with his brother King Cynegils, of 2,065 Britons 'on Bea's hill'. 104 A communal memory of this sort -as Williams suggests -is highly likely to have contributed to ideas about the place being particularly appropriate for aggression directed towards the enemies of Wessex; 105 one can perhaps see in the wording of the Chronicle entry a folkloric echo of the power of the mythic archetype as it was deployed in a military context. 106 Neil Price has made a compelling case for a Viking-age Scandinavian belief in the active agency of supernatural forces on the battlefield. 107 Leaving aside the possibility of cross-cultural influence (in whichever direction) in the later Saxon period, there is some reason to believe that similar ideas may have held some force in the Anglo-Saxon world-vieweven in a late and overwhelmingly Christian environment. The belief that ancestral forces could be accessed at prehistoric sites has been discussed above, and was apparently current as late as the 990s, when AElfric warned against the recourse of witches to heathen burials for the purpose of summoning the dead. 108 The increasing prevalence of military visions and expectations of divine intervention in military affairs that characterize the later Middle Ages can be traced from notably early expressions in Anglo-Saxon texts. 109 It is also notable that royal military saints come into view in England at a very early date -unparalleled elsewhere in early Christian Europe. 110 It may be that these precocious developments should be understood in the light of pre-Christian frameworks of understanding that encompassed a belief in battlefield magic and the role of supernatural antagonists in war. Taken in this light, what Williams has called 'the curse of Cwichelm's mound' 111 takes on an even more direct significance: the complex relationship of fear, power, ancestral memory and warfare may have contributed to a belief that supernatural agents -variously associated with the winning of territory, the establishment of folk groups and royal genealogies, and the aggressive defence of hord and hamas -could be accessed and employed in certain prominent locations, possibly influencing the perceived appropriateness and choice of certain places for conflict.
The links between Cwichelm's Barrow and the Kennet battlefield are perhaps to be understood in this light. Although the threat of Cwichelm appears to have gone unfulfilled, it is surely no accident that the conflict -when it came -should have come in an environment that, through its funerary topography, invoked not only Cwichelm's mound, but Bea's Hill as well. In more than one sense, the Danish visit to Cwichelm's Barrow may have acted as a trigger: the challenge activated a network of conceptual, spatial and narrative associations that may have brought, alongside the practical military deployment, an expectation of supernatural assistance that could be assumed to act with juridical authority and royal legitimacy. The doomed Saxon fyrd, nervously waiting amongst the ruins of prehistory for the approaching Viking army, were perhaps able to stand a little taller, grip their spear shafts more firmly, in the belief that the spirits of mighty ancestors were close at hand.

On Woden's ground
Following the path of the Ridgeway three miles south of the Kennet, the walker reaches the line of Wansdyke cutting east-west across the chalk downs (Fig. 4). Crossing it, one enters a landscape that contains one of the densest concentrations of theophoric place names in England. Here can be found Woden's Dyke (Wansdyke), Woden's Gate (Woddes geat), Woden's Valley (Wodnesdene) and Woden's Barrow (Woddes beorg, Wodnesbeorg). 112 It is a location where the memory of conflict and of royal mythology and genealogy intersect, and was almost certainly an important symbolic locale for West Saxon kings who claimed descent from Woden and derived their legitimacy from military conquest -both of which were expressed in the landscape through association with monumental prehistoric remains. 113 I have articulated the argument elsewhere that this location may have been of particular importance in the early to mid-Anglo-Saxon period as a ritual complex associated with a cult of martial kingship. 114 By the later Anglo-Saxon period, it is likely that many of these functions and associations had been transplanted to other places similarly associated with kingship and spiritual authority, much as the locations of royal burial were also to be translated. 115 Nevertheless, it is unlikely that these associations -particularly at a local level -would have been entirely forgotten, certainly while place names remained intelligible markers in common currency (the Woden names of this particular landscape are known from ninth-and tenth-century charters, as well as ninth-century Chronicle entries). 116 It seems likely that an intrusion into Fig. 4 Map of the route from the probable battle site at Overton Hill (a) to the area around Woden's Barrow (b), marked on the first edition Ordnance Survey map (1889, 1:10,560 scale). © Crown Copyright and Landmark Information Group Limited (2014). All rights reserved this symbolically rich landscape would have been profoundly disturbing to the local West Saxon population, and perhaps part of the deliberate psychological campaign that the Viking here seems to have been waging in 1006.
The topography of the site is also worth considering. Approaching from the Ridgeway, the crossing of Wansdyke into Woden's Valley is nothing short of breathtaking (Fig. 5). On either side of the path at the edge of the chalk ridge stand the peaks of Knap Hill and Walker's Hill, the former crowned with iron-age and Neolithic enclosures, 117 the latter with the Neolithic long barrow now known as Adam's Grave, but formerly -and in a charter of 825 (S 272) -as Woden's Barrow. 118 Between these peaks the path plunges to the edge of the ridge that separates the Wiltshire uplands from the Vale of Pewsey. Passing through that symbolic gateway meant unfettered access to the agricultural heartlands of Wessex

Conclusion
A landscape such as that at the Kennet can be seen to express a number of different social ideas that were open to manipulation by both the West Saxons and their Viking enemies. By the early eleventh century, and after two centuries of warfare and settlement, it can very plausibly be supposed that many Scandinavians had been exposed to the cultural norms and geographical vocabulary of English warfare. By their aggressive entrance into this landscape -first at Cwichelm's Barrow, then at the Kennet, and afterwards crossing Wansdyke and Woden's Field before ultimately passing the walls of the royal burh at Winchester -the Viking army was able to mount a profound challenge to the ideals and obligations of English royal authority. It was a challenge that operated on a number of levels: it challenged the military effectiveness of systems of civil defence, and made mockery of the king's role as law enforcer; it invited the defence of land, hoard and home; it defiled the burial places of ancestors and dared the wrath of heroes and euhemerized gods. Taken together, the campaign of 1006 can be seen as a blow struck against English, and particularly West Saxon, identity and claims to effective control over the landscape. Through the exploration of these events, this paper has attempted to show how the close study of individual conflict landscapes can make a substantial contribution to how warfare in early medieval England should be understood, particularly in its relationship to matters of royal and ancestral mythology and the intangible attributes of landscape that have been increasingly recognized as an important key to matters of ideology