War without warriors? The nature of interpersonal conflict before the emergence of formalised warrior élites - Rick J Schulting

War without warriors? The nature of interpersonal conflict before the emergence of formalised warrior élites
DRAFT PAPER FOR THE SESSION:
The experiential role of violence and combat in the creation of social identities
the Sixth World Archaeological Congress
29 June - 4 July 2008
University College Dublin
Session organisers: Barry Molloy and Angelos Papadopoulos
Rick J Schulting
School of Archaeology
University of Oxford
Not for citation without author’s written permission

Abstract
The appearance of ‘warrior élites’ marks a well-defined, or well-imagined, role that emerges at various times and places around the world. In prehistoric Europe, this image first appears most clearly in the Bronze Age, yet there is abundant evidence for earlier violence, from both the Mesolithic and Neolithic periods. Who was carrying these acts out, and how did they figure into the creation of social identities? ‘Warrior’ identity in these earlier periods was arguably indistinguishable from ‘adult male’ identity. Understanding the conditions preceding the emergence of more formal and specialised warriors should shed light on the nature of conflict and identity, and the place of material culture in the formation and maintenance of specialised social roles.

Introduction
Many of the papers in this session address the nature of élite warrior status in the context of chiefdoms and early states. But what about before this? Certainly there was no absence of interpersonal violence, as demonstrated by many sites exhibiting skeletal injuries. The earliest example with evidence for multiple injuries that can be attributed to large-scale, inter-group conflict comes from the oft-cited Epipalaeolithic cemetery 117 at Jebel Sahaba in Sudan, dating to ca. 13,000 years ago. A sufficient number of Mesolithic skeletons are found across Europe with embedded projectile points, and blows to the head, to strongly suggest more than occasional violence. For much of Europe, the frequency and scale of conflict may increase further with the
appearance of the Neolithic, though the comparison is not a straightforward one, and is greatly hampered by the relative paucity and uneven distribution of Mesolithic human remains. But in none of these cases is there unequivocal evidence for the emergence of a specialised warrior identity: there are few indications of formal weapons that could not equally serve other purposes, iconographic portrayals are rare, and there are few if any contenders for distinct warrior graves. This paper addresses the nature of ‘war without warriors’ from the perspective of Mesolithic and Neolithic Europe. Ethnographic cases of band and tribal conflict are also drawn upon to help elucidate the nature of warfare in these small-scale societies.
A discussion of this nature will inevitably rest on definitions. So to make these clear from the outset, ‘warfare’ here refers to armed conflict of lethal intent between two or more autonomous sociopolitical groups. This is distinct from feud, which occurs within a single group, or homicide, which can occur between individuals either within or between sociopolitical groups (though more often the former, through simple physical proximity; the latter situation, unless quickly resolved, almost invariably escalates into warfare). The notion of ‘warrior identity’ is perhaps slightly more difficult, though less so in its more extreme forms, where it can be seen as a specialised identity, often, though not always, associated with young/middle adult males, and carrying with it the expectation that these are the individuals primarily responsible for a group’s defence and offence directed against other groups (thus, to state the obvious, the existence of warriors pre-supposes the existence of war). This status will often be accompanied by a distinct material culture, including first and foremost weaponry, which may not be generally available to other members of the group; it may also include other insignia.
This definition points the way to my general thesis, that ‘warrior identity’ in Mesolithic and Neolithic Europe was largely indistinguishable from ‘adult male’ identity. I will return to this after a consideration of the range of evidence for conflict in the Mesolithic and Neolithic Europe.

Evidence for interpersonal conflict in the Mesolithic and Neolithic
Evidence for conflict in the European Mesolithic takes the form of embedded bone and stone projectiles, and both healed and unhealed cranial trauma. For cranial trauma, healed injuries are in the majority, which may very well indicate a different context for violence: an element of ritualised dispute resolution is often inferred when faced with similar cases in other parts of the world. The major exception to this is the site of Ofnet in Bavaria, with its two so-called ‘skull-nests’. Depending on the analysis, between 25 and 50 percent of the skulls in these two pits show perimortem injuries. The heads were removed while the bodies were still fleshed, with cutmarks visible on a number of vertebrae, and deposited in a formal manner, in some cases wearing animal tooth ornamentation, in two shallow pits in a cave. While the nature of this event, or even whether it is a single event, has been debated, inter-group hostilities would seem to offer a reasonable scenario.
Neolithic evidence for conflict takes two main forms, skeletal trauma and architecture, the latter in the form of enclosures. Though there is much debate concerning the roles of Neolithic enclosures, itself a varied category, there is good evidence that at least some were subjected to large-scale attacks, even if they were not necessarily intended as defensive structure when first conceived and built. Foremost among these for Britain are Hambledon Hill, Crickley Hill and Carn Brea. The large-scale conflicts implied by attacks on these sites must quality as warfare by any definition: the events are on a scale that is simply too large for them to represent in-group fighting. Some 400 projectiles were found concentrated at the entrance into the Crickley Hill promontory enclosure. Undoubtedly this represents only a fraction of the missiles fired at and from the enclosure, as many would be retrieved at the time, and many others would be lost in the intervening 5500 years since this event – and it does seem to be a single event, as a recent dating programme by Alasdair Whittle and Alex Bayliss has shown. The same can be said for the even more impressive figure of 800 arrowheads found by Roger Mercer at Carn Brea. Given the unknowns, trying to estimate the number of people involved in such conflicts on the basis of the number of recovered arrowheads is probably futile, but it does not seem unreasonable to suggest the involvement of hundreds of antagonists.
Skeletal evidence is on occasion equally dramatic, best exemplified by the paradigm-shifting late LBK mass grave at Talheim in SW Germany. Thirty-four men, women and children were found here, unceremoniously thrown into a large pit, with many showing perimortem cranial injuries. Evidence for even larger-scale violence comes from the late LBK enclosure ditches of Asparn-Schletz in Austria, where more than 120 individuals were found in the ditches, with many showing perimortem injuries. Talheim is perhaps slightly more ambiguous than the British enclosures as regards whether it can be said to relate to ‘warfare’; it could conceivably represent a feud between, for example, lineages within a larger social group, though this might imply a degree of supra-household or supra-village organisation beyond what is usually attributed to the LBK. This is harder to argue for the dead in the ditches of Asparn-Schletz, if these relate to a single episode, which is not inconsistent with the available 14C dates and stratigraphy. The relative dearth of young women here compared to an expected normal demographic profile has been used to suggest their removal as captives.
Then there are numerous other sites with a small number of individuals showing skeletal injuries attributable to interpersonal violence. In many cases these are yet more ambiguous, and it cannot be said with any certainty whether they reflect in-group or out-group conflict. The overall pattern that emerges, however, is intriguing. With very few exceptions, projectile injuries affect males, while both males and females are more often equally affected by healed as well as unhealed cranial injuries. This pattern demonstrably applies to the Mesolithic and the Neolithic, continuing into the Late Neolithic and Chalcolithic (and presumably beyond this, though I do not have specific data to hand for the Bronze Age). This implies at least two different contexts for violence, one involving predominantly males fighting at distance with bows, and the other involving close-quarters conflict that could equally affect males and females. While it is perhaps tempting to infer that this marks the distinction between out-group and in-group conflict, this is probably too simplistic a reading of what was a more complex and varied situation, particularly in the case of the contexts leading to cranial trauma.
However, the notion that projectile injuries do relate for the most part to organised conflict between primarily men from different groups does appear a reasonable one. What little iconographic evidence there is for conflict in the European Mesolithic/Neolithic comes from a number of well-known rock art sites in the Spanish Levant, and appears to support the role of bow conflict. The dating of these panels as either Mesolithic or Neolithic is controversial, but for our purposes here this distinction does not really matter. The panels are interesting for a number of reasons, one being that some seem to depict small groups of archers in battle line arrangements that are not dissimilar to those known in a number of ethnographic tribal contexts, such as seen for the Dani of New Guinea in the film Dead Birds.
From the above discussion, it is clear that the weapon of choice for long-range conflict in earlier prehistory was the bow and arrow, and probably most men, and young boys, would own at least one. Of course the bow can also be used to hunt. But for the British Neolithic at least, the large numbers of leaf-shaped arrowheads that have been found can be seen as somewhat puzzling, given the paucity of hunted game known from contemporary faunal assemblages. Indeed, there are no known examples of embedded projectile injuries in wild fauna from the earlier Neolithic (though, oddly, embedded point fragments are known from domestic cattle and pig at the Late Neolithic site of Durrington Walls). Yet four cases of directly embedded projectile fragments are known from Neolithic human remains in Britain (Ascott-under-Wychwood, Tulloch of Assery, Penywyrlod and, most recently, Wayland’s Smithy), with another case from Poulnabrone in Ireland. The existence of many other examples is suggested through the close association of projectile points, often with broken tips, with the skeleton (e.g., Hambledon Hill, Cat’s Water, Wayland’s Smithy, just to name a few; a recent study of a Beaker age example from Feizor Nick Cave presents a rare example of a female with a projectile injury). Even if wild game are underrepresented because of the ‘ceremonial’ nature of many – though by no means all – of the sites, it is unlikely that hunting played more than a minor economic role. What it might do, however, is serve to train young men in proficiency with the bow. This training of boys from a young age in both shooting and dodging arrows is a common feature of the ethnographic literature for tribal societies ranging from Papua New Guinea to the North American Pacific Northwest. It begins as a game, but is in deadly earnest in later life, though sometimes retaining some of the qualities of a ‘sport’ – was the shooting of domestic species at Durrington Walls perhaps part of a ceremonial hunt with an underlying purpose of maintaining, and advertising, the bow skills of the various participating groups?

Small-scale conflict and war without warriors
Another important consideration in the present discussion is the general nature of tribal conflict. Based largely on ethnographic accounts of tribal warfare, but informed where possible by the archaeological record of Mesolithic and Neolithic Europe, some suggestions in this direction can be made. One of the most common themes to emerge is that in band and tribal conflict the first choice of encounter is almost invariably ambush, preferably against an unarmed opponent. The night-time or early dawn raid features prominently, maximising the likelihood of killing the enemy while minimising the danger to one’s own group. Indeed, the enterprise would often be abandoned if the alarm were raised. Open battles would sometimes be fought by some groups ethnographically, and indeed might even be arranged beforehand, but ambush and deception are found more often, and were probably more deadly in the long run. Advantages, however they arose, were often relentlessly pursued, to the point of the extermination of the opposing group. This rarely had anything to do with territorial expansion, though this might follow on by default. One result of such climates of uncertainty seen ethnographically is that during times of unrest, people would sharply curtail their travel, and men would go about their days and nights fully armed, even within their own communities.
Overtly economic motivations are often more evident in expansionist chiefdoms and states. Nevertheless it is clear historically that some tribal groups are very expansionist, such as the segmentary house societies of the Iban of Borneo, who pursued an explicit policy of expanding into new territory by force. Many additional examples could be found among African pastoralists. Then there are also many different kinds of economic motives that can operate at a smaller scale. The question is who decides that they are worth pursuing in a society lacking strong central leadership, and how this ambition is realised. One of the features of bands and tribes is that the constituent social units are relatively small, though in the case of tribes they can amount to large numbers of individuals in aggregate. And, in both bands and tribes one of the main structuring principles is kinship. These two points taken together – small numbers of people, often closely related, whether fictitiously or not – mean that any economic advantage would be easily perceived as beneficial both to specific individuals and to the overall group. Carrying out the action, then, becomes a matter of convincing enough individuals to join an expedition.
In the case of the Mesolithic, one might mention the likelihood of stored food reserves, particularly in coastal situations that were at least partly sedentary. This, and conflict over access to particularly productive fishing locales may have been factors in accounting for the high incidence of violence seen in Mesolithic Scandinavia and along some of the large rivers of eastern Europe: the Iron Gates of the Danube, and the rapids of the Dnieper River in the Ukraine. Food stores would of course also be present in the Neolithic, in the form of stored grain, and perhaps more enticingly, in the more mobile form of herds of domestic animals, and in particular cattle. The economic and cultural value attributed to cattle in the Neolithic of northwest Europe emerges fairly clearly from number of lines of evidence. Under such conditions, as I and others have argued elsewhere, they make particularly attractive targets for raiding. And such raids can be carried out by very small numbers of individuals. Again, stealth and ambush are the ideals.
The preference for this kind of conflict itself may militate against elaborate and showy weaponry. It might also be argued to provide little incentive for the development of an élite warrior identity.
In addition to the preference for ambush, a critical factor is the combination of a high degree of both autonomy and responsibility placed on the individual in small-scale societies. The relevance of this here relates to the central importance of revenge, which emerges again and again as one of the strongest motivators for both within-group feuding and between-group warfare. The duty of vengeance falls most directly on close kin, and next, in the case of between-group conflict, on the aggrieved community and its allies as a whole. While help would certainly be sought, neither the means nor the responsibility for revenge are delegated to any authority. One of the consequences of this is that it is very difficult to cease hostilities once begun – none have the clear and binding authority to do so, though attempts might be made. In societies with more centralised authority, the personal/family responsibility for vengeance is less emphasised, with the decision for, and means of, retribution coming under firmer control by the central authority.
Thus, the reason that we do not see a specialised warrior identity in the Mesolithic or Neolithic is that every able-bodied male would be expected to perform this role alongside his other roles: hunter, farmer, herder, fisher, weaver, potter, etc. This is part of the autonomy and responsibility of small-scale, acephalous societies. This is not to say that all would be equally proficient in warfare, and no doubt some would display a greater aptitude and become recognised as leaders, though if the ethnographic accounts are anything to go by, such leadership roles would usually be restricted to the duration of the immediate conflict at hand, and would not extend into other fields of activity. Nevertheless, it is possible that success here, as with any important endeavour, might offer one avenue to elevated status, particularly when it presented the opportunity to add to the group’s valuable cattle herds. The ability to obtain the necessary support and organise a retaliation or raiding party against another group would, whatever the ideal, probably not be an option open to all.
The question of whether warfare proper appears only with the Neolithic is not one I propose to go into in any detail here, though it warrants a brief mention. Raymond Kelly’s argument to this effect hinges on the useful notion of social substitution – that is, warfare exists when group identity is such that one individual can be made to stand in for another. Thus, in the case of revenge for a killing, rather than seeking the killer him- or herself, any member of the killer’s group is seen as an appropriate target, satisfying the requirements of blood vengeance. Kelly argues that this only occurs with the appearance of agriculture in the Neolithic. It is this point that is contestable, as it adopts a rather stereotypical view emphasising one extreme of the gamut of hunter-gatherer societies, seeing them as egalitarian, highly mobile, and with shifting group membership such that no real notion of group identity emerges. Not only are there numerous ethnographic counterexamples, but the archaeological record of Mesolithic Europe strongly suggests that in many areas people had much more developed notions of territoriality, and that differences were marked by material culture on a relatively small spatial scale, strongly indicative of the emergence of group identities. But the thesis of the crucial role of social substitution itself seems sound. There is also a strong element of mutual reinforcement here, since inter-group conflict encourages, or even necessitates, the taking of sides, which further encourages the use of material culture to demarcate those sides.

A paradox
It is with the Chalcolithic (ca. 3200-2500 BC) that we first seem to see the image of a specialised male warrior, on carved stelae in the Alpine region, and, from ca. 2500 BC, the widespread appearance of Beaker ‘warrior graves’, with fine arrowheads, stone wristguards, copper or flint daggers and the Beaker drinking vessel itself (the stereotypical, hard fighting, hard drinking warrior). Of course, as has been often emphasised, this is an image, one constructed in death by those doing the burying, and is undoubtedly an idealised one. But, regardless of its ‘reality’ in individual cases, the simple fact that adult male identity is being constructed and portrayed in this way indicates the importance of the warrior image. This is precisely what is lacking in the earlier Neolithic. It could be said that the emphasis on commingled burial in mortuary monuments that dominates the ‘Early’ Neolithic of northwestern Europe (Britain, northwest France, southern Scandinavia) precludes any kind of particular identification in death other than an overall community membership. But, aside from the fact that this is in itself says something about the structure of this society, this would not apply to the single flat graves that predominate in the LBK of central Europe. Male graves here do tend to be associated with stone axes and projectile points, but because of the ambiguous nature of these implements these are not seen as warrior graves. Given the clear evidence for the use of stone axes to inflict horrific head injuries at Talheim and other sites, it may be time to re-consider this. Perhaps part of the male identity being marked in death here does relate to roles in warfare, as much as to carpentry, exchange and hunting.
All this begs the question, then, of whether the appearance of more formal and unambiguous weaponry, seen particularly with the advent of first copper but more particularly bronze metallurgy, is associated with an increase in actual levels of violence. As is so often the case, the question is not a straightforward one, partly because the skeletal evidence has not yet been systematically reviewed. Experience with the Neolithic evidence suggests that many early published accounts of injuries are unreliable in both directions; that is, claimed injuries are often not sustained when reviewed with modern forensic criteria, while many real injuries are missed. Yet, on present information, it seems that there may actually be less skeletal evidence for interpersonal violence in the Beaker period and Early Bronze Age, at least in Britain. This has been argued by Roger Mercer and Nick Thorpe. John Robb has drawn attention to a comparable phenomenon in Italian prehistory, where periods with greater material elaboration of conflict imagery actually show a lower prevalence of skeletal injuries than periods lacking such imagery.
What happens to shift the role of conflict from the many to the few? Does this even happen with the appearance of formalised weaponry, or does this only mark the emergence of élite leaders, with most adult males still participating in warfare, but using more mundane forms of weaponry – bows and arrows, wooden clubs, etc.
A traditional Marxist approach offers one approach to this question. Increasingly conspicuous differences in wealth within communities in the Beaker period and EBA may have increased tensions that were either largely absent or latent in the preceding Neolithic. At the same time, as has often been suggested, it may be that traditional kin-based networks were breaking down, and allegiances could be more easily shifted. Under such circumstances, it may have been prudent to control the means of violence, since the threat might come internally as much as externally. This is all rather hypothetical, as there is no clear evidence for such within-group conflict (though how would it be recognised?), but it may be worth exploring further.
Pacific Northwest Coast chiefs to a large extent controlled, or tried to control, specialised weaponry, though these were often just elaborated versions of weapons – clubs and daggers – more widely available. Firearms were an obvious exception when they became available through the Fur Trade, and a very concerted effort was made by chiefs to control these and distribute them to chosen followers.
The development of metallurgy and its increasing importance in the creation of bronze weapons provides obvious possibilities for control. But this implies that such weapons really were the state of the art, and the means by which warfare was conducted. Yet, unless the nature of warfare changed significantly from earlier period, projectile weapons would remain the most effective long-distance weapon, and it would matter less whether these were stone- or metal-tipped. For close-range fighting, the ethnographic preference is almost invariably for surprise attack rather than melee, and again wood, antler and stone clubs would arguably be as effective for this purpose as metal weapons. This could imply that the nature of warfare did indeed change, or that, as many have argued, the main purpose of bronze weaponry was for display
rather than real combat. This position, however, seems difficult to sustain in the light of research demonstrating that bronze swords often do exhibit edge damage consistent with use in combat, and experimental archaeology shows that objects such as halberds previously thought to be unwieldy and ineffective, actually work quite well (at least on stationary sheep’s heads).

Conclusion
Neither violence in general nor warfare specifically, are marked by the appearance of specialised, formal weaponry in the European Bronze Age. Abundant and compelling archaeological evidence demonstrates that warfare was a feature of human societies long before this. The nature of small-scale societies – their small size, the paucity of specialised roles, and, perhaps most importantly, the generally high degree of autonomy and personal responsibility – is such that all men were ‘warriors’ as the situation arose, in addition to all their other roles. This flexibility and lack of specialisation extends to the ‘weapons’ of this time, which are for the most part indistinguishable from tools used for other purposes. A most interesting question is whether the emergence of formalised weaponry sees an increase or decrease in the actual prevalence of violence, or changes in the forms it takes. Some intriguing preliminary suggestions are that the two are inversely correlated. But further work is required, both on use-traces on the weapons themselves, and on the skeletons that bear the brunt of human violence.