Violence and Warfare in Precontact Melanesia

Levels of interpersonal violence and warfare for 30 Melanesian societies at the time of contact with Europeans are estimated based on ethnographic and historical records. While violence was common in indigenous Melanesia, it was not ubiquitous and some societies experienced extended periods of internal and external peace. Interpersonal violence and warfare were correlated-when one occurred there was a high probability of finding the other. Violence was not dependent on total population. It was, however, higher for population density greater than 50 persons per square kilometer. Violence inMelanesia may have been stimulated by the large number of relatively small polities, many of which competed with one another for prestige and, in some cases, land.


Introduction
Cross cultural studies of violence and warfare in indigenous cultures have revealed a number of general trends that appear to be independent of specific social contexts. Keeley [1] has examined a large number of small societies and has concluded that violence was more the rule than the exception in preindustrial societies. However, Loftin [2] found that the nature of violence changed with population, from a fight between individuals associated with distinct groups who knew one another to an impersonal collision of groups spurred into action by strong leaders. Ember et al. [3,4] found that the degree of participation in governance was an important factor in the frequency of warfare among a wide range of societies. Indeed, leadership appears to have played a role in both interpersonal violence and warfare, occasionally suppressing the former to retain the social cohesion necessary to carry out the latter.
Precontact Oceania offers a unique laboratory for the study of human behaviors, including leadership, violence, and warfare. Island populations varied from a few hundredjust sufficient to maintain a sustainable gene pool-to many tens of thousands. Social structures ranged from egalitarian bands to hierarchical systems with many of the aspects of nation states. In previous studies I examined the influence of leadership upon violence and warfare in Polynesia [5] and Micronesia [6]. In this paper I extend the same type of analysis to the third major cultural region in Oceania: Melanesia.
Violence and warfare in Melanesia have been the subject of a number of studies, including Camilla Wedgwood's [7] early paper "Some Aspects of Warfare in Melanesia. " More recently, Knauft [8] has surveyed violence in the context of evolving sociological and anthropological theory, with particular attention given to Papua New Guinea. He writes that ( [8]: 225) "the dominant sense one gets from available accounts, primary ones as well as contemporary reconstructions, is that warfare was indeed quite prominent if not endemic in most coastal and insular areas of Melanesia at the time these regions were first regularly contacted by Europeans. " However, Douglas [9] questioned the view that Melanesians were "constantly at war" and Chowning [10] suggested that they may have been no more violent than other Oceanians. This paper focuses on lethal physical violence (rather than sorcery, wife beating, fighting for sport, etc.) in 30 Melanesian societies prior to significant European contact. Most of the analysis is dedicated to islands, although four cultures from Papua New Guinea are described both to complement the island cultures and because of the excellent documentation available for these societies. The method used is one of controlled comparison [11], in which a similar analytical technique is applied to a set of cultures within a defined area.
I divide lethal violence into two categories: interpersonal violence and warfare. Interpersonal violence is defined as 2 Journal of Anthropology occurring in a dyadic relationship between two or more individuals where the target of the violence is a specific person. Assassinations, revenge attacks, and murders committed in the act of theft fall into this category. I use Tefft and Reinhardt's ( [12]: 154) definition of warfare as "an armed aggression between political communities or alliances of political communities. " Such communities can be kin groups, clans, villages, or coalitions of regional polities.
Raiding was a common form of warfare in much of Melanesia. In most cases a raid would consist of a few individuals who would enter an opponent's territory by stealth with the intention of killing one or a few people. More substantial raids, such as when a group would infiltrate a village before dawn to attack its sleeping residents, were also common. Here the attacking party had to be careful not to stay too long in the enemy camp for fear of being overwhelmed. Raids were often associated with feuds in which one individual or a well-defined group was the target of a revenge attack. When the target was a specific person, raids had some aspects of interpersonal revenge. When a substitute victim would suffice, they resembled a form of intergroup warfare.
Malinowski Casualties in formal wars ranged from a few warriors in a single battle to a significant part of the population in a protracted conflict. Feil [14] tabulated deaths due to warfare in Papua New Guinea highland societies and found that, in some cases, 25% of all deaths and over 30% of male deaths were attributable to war.
However, war was not uniformly destructive. Among the Maring, "nothing fights" were conducted where relatively little effort was taken to inflict injury on the other side [15]. Rather, these conflicts appear to have been designed to demonstrate resolve, a willingness to fight and defend the group's interests. Conquest, or even humiliation or domination of the other side, was not a driving factor in these battles. Similar behavior occurred on Tanna, where battles occurred as part of a formalized exchange process in which the disputants were represented by allies and did not themselves fight. Raids, which deviated from this formalism, were referred to as "stolen wars" [16].
Cross cultural studies of violence in indigenous societies (e.g. [17,18]) found that interpersonal violence and warfare are correlated with the degree of social stratification and/or political hierarchy. Melanesia offers particular advantages for testing such theories in that its societies ranged from largely egalitarian to those involving competitive leadership and hereditary chieftainships.
Sahlins [19] cites the Big Man concept of competitive leadership as characteristic of many Melanesia societies, though later analyses [20,21] suggest that chiefdoms existed in at least several parts of the region, notably New Caledonia. Within Big Man societies, care must be exercised in assessing the degree of actual "leadership" exercised by any single individual. First, many societies were fluid in their membership. If a person or family was dissatisfied with the actions of a Big Man, they could move to another group. Second, adherence to the wishes of the Big Man was sometimes optional even if one elected to remain in the group. Members of a group might decline to follow the suggestion of a leader. Third, the competition to be Big Man was not a level playing field in that sons of previous leaders had advantages, including wealth (which, however, was not always passed down to future generations), secret knowledge, familiarity with neighboring leaders. While nonhereditary leadership was common in Melanesia, its nature was complex and varied. I use the term Big Man for simplicity rather than to imply a common form of governance across different societies.

Data
This paper focuses on indigenous violence in societies in the area between Papua New Guinea and New Caledonia prior to significant contact with Europeans. Several criteria were used in the selection of cultures. First, data on population, interpersonal violence, and warfare had to be available. Second, I included societies ranging from small, relatively isolated islands such as Vokeo to groups in the large landmass of Papua New Guinea.
A particular complication in assessing violence and warfare in insular Melanesia compared to other parts of Oceania is the relatively large size of islands. Whereas the average area of inhabited islands in the Caroline Islands is 25 km 2 , in the sample of Melanesian societies considered here it is 3100 km 2 , exclusive of Papua New Guinea. The dense vegetation and mountainous topography of Melanesia, combined with a face-to-face leadership style in many of its societies, may have inhibited large-scale cultural and political unification on the larger islands. Sometimes there were significant cultural differences between groups living just a few kilometers from one another. For example, Hogbin [22] noted 18 languages spoken on Malaita. Ethnographic studies tend to focus on specific groups so that island-wide summaries can be challenging to construct.
Land areas are taken from the United Nations (n.d.) tables of islands. For Papua New Guinea societies, approximate domains are taken from ethnographic sources. Journal of Anthropology 3 Population estimates are uncertain for several reasons. Populations varied with time, affected by available nutrition, natural disasters, and, in the case of the smallest societies, variations in birth rate and sex ratio.
In some societies, violence was a significant factor in mortality. Knauft [23] estimates that nearly one third of all deaths among the Gebusi of lowland Papua New Guinea were attributable to homicide. In the Solomon Islands, some smaller islands were nearly depopulated by headhunting raids. The accuracy of population estimates is given below via a letter grade, with (A) corresponding to a detailed census count or other reliable method, (B) to a less precise but still generally reliable estimate, and (C) to a guess based on cursory observations or estimates of carrying capacity.
A numerical scale was used to represent the level of interpersonal violence per thousand people: (0) none: no recorded instances or very few per century; (1) rare: violence very rare, perhaps one death per decade; (2) occasional: one or a few deaths per year; (3) frequent: several deaths per year, but lethal violence socially disdained; (4) chronic: violence being a major part of the culture.
A similar scale, analogous to that of C. R. Ember and M. Ember [18] and Ember et al. [4], was used for warfare: (0) none: no recorded instances or only a few over several centuries; (1) rare: war very rare, perhaps several times per century; (2) occasional: wars every few years; (3) frequent: wars every year but not continuous; (4) chronic: warfare essentially continuous.
Lacking quantitative data over long periods of time, absolute measures of nearly any social quantity are difficult or impossible to assess. With regard to violence, Knauft  Recognizing these challenges, a letter grade-A, B, or C-assessed the quality of the information on interpersonal violence and warfare. Table 1 provides data on thirty societies in Melanesia. Ross [24] used data assembled in the Human Relations Area Files to estimate violence in 90 societies, including two in Melanesia. While his coding scheme was different than mine, a translation to my scale yields levels of interpersonal violence and warfare of 2 for the Trobriands, compared to my value of 2 for Kiriwina. However, for Manus he gives values of 2 for both interpersonal violence and warfare, whereas I estimated a much higher level of 4 for each, based on the later analysis of Bonnemaison [16]. Several trends are apparent in the data and can be compared with observations of other cultures. C. R. Ember and M. Ember [18] found that interpersonal violence and warfare are correlated in a broad range of societies. Studies of a sample of Polynesian islands [5] and the Caroline Islands [6] show a similar correlation. Figure 1 shows that this correlation also applies to my sample of Melanesian societies. However, as Figure 1 shows, the largest cluster of points is at the maximum value of both interpersonal violence and warfare.
It is not possible to establish a casual direction in the data, that is, whether war stimulates interpersonal violence (as proposed in [25]) or whether interpersonal violence leads to war. Rather, one can only surmise that internally violent societies are prone to war, and warlike societies are prone to high levels of internal violence.
Previous studies of violence and warfare in a selection of Polynesian islands [5] and the Caroline Islands [6] indicate that both interpersonal violence and warfare are likely to increase with total population, consistent with the broader cross-cultural studies of Ember [26], Leavitt [27], and Rosenfeld and Messner [28]. Figure 2 shows that this is not the case with the Melanesian societies sampled here. The degree of interpersonal violence and warfare ranged from rare to chronic across a wide range of population sizes, although higher levels of violence were more common for larger populations. Part of the reason for this may be the high degree of factionalism in Melanesia compared to Polynesia and Micronesia. The hierarchical leadership common in the latter two cultures permitted larger polities to be under the control of central leadership. Fewer individuals had   the authority to instigate intergroup violence. Internal violence was sometimes suppressed to strengthen social cohesion. Conversely, the weaker central control found in most of Melanesia resulted in smaller polities and a consequent higher probability of competition, including violent competition, between those smaller polities. Population density has been cited as a stimulus to violence, the thesis being that more frequent interaction will lead to more opportunities for violent conflict. Figure 3 shows that, with the exception of the Dani, all cases with population density greater than 50 persons per square kilometer had high rates of warfare and interpersonal violence. Whereas there is a spectrum of interpersonal violence and warfare among societies with lower population densities, both forms of violence are common in higher density cases.
The data for violence and warfare versus population density in Melanesia differ from those of Polynesia [5] and the Caroline Islands [6] where violence decreased with increasing population density. They also differ from computer simulations of violence and warfare in small societies [29] that highlight the beneficial effects of cooperation and the negative effects of violence on social performance. This difference will be discussed later in the paper. However, it is worth noting here that, in his global study of violence and warfare in preindustrial societies, including several in Micronesia, Keeley ([1]: 118) found that "absolutely no correlation exists between the frequency of warfare and the density of human population. " Insular Polynesia and Micronesia may be more the exceptions than the rule among small societies.
Cross cultural studies [2,17,30,31] have revealed a degree of correlation between social hierarchy and the frequency of warfare, supporting the notion that hierarchical societies are more prone to violence than less structured ones. Figure 4 shows that this is not the case in Melanesia. Recognizing that only six of the thirty cases studied here are chiefdoms, the levels of both interpersonal violence and warfare are higher for the Big Man societies than for those governed by chiefs. Two thirds of the Big Man societies have the maximum level of interpersonal violence and war.
Knauft [23] found that the egalitarian Gebusi had a remarkably high rate of homicide even though there was no well-defined leader. Conversely, Vokeo, which had headmen as leaders, had a lower level of violence, as did Kiriwina with its chiefs and Malaita where a Big Man style prevailed. This variation in violence versus political complexity suggests that one must look deeper for causes of interpersonal violence and warfare, beyond arguments based on leadership styles or the competition of leaders.

Discussion of Specific Islands and Cultures
To provide greater insight into the nature of violence in Melanesia, I give further details on several specific societies. I chose Vokeo and New Caledonia as small and large islands with chiefs, Malaita as a large island with Big Men, and the Maring as a Papua New Guinea culture with extensive ethnographic data.  Despite the high population density (146/km 2 ) of Vokeo, the natives believe that there is plenty of land for everyone. Although contact with Europeans occurred during Dutch explorations in 1616, significant interactions did not occur until the late nineteenth century. Many aspects of Vokeo culture were studied by Hogbin [32][33][34][35][36][37] during his visits in the period 1934-1948 and more recently by Anderson [38,39].
The island is divided into five districts, each of which contains a number of villages with populations ranging between 75 and 100. Houses are arranged in two or three clusters within a village, a cluster representing an exogamous patrilineal clan. Each clan-unit within the village is controlled by a kokwal, a position described by Hogbin ([32]: 318) as "more than a headman, though he is less than a chief for he has absolute control over only the members of his clan-50 or 60 individuals at a maximum. " The kokwal is chosen from among the eldest sons of the previous kokwal's wives, leading to not infrequent competition for the position. Vokeos feared being without a kokwal-he coordinated group activities and resolved disputes.
A complex set of social relationships linked all of the residents of Vokeo. People viewed themselves as part of a nested series of groups, starting with the family and extending outward to village clan unit, village, and district. Cooperation was considered essential to survival and every effort was made to avoid conflict. Sharing was routine, with baskets of small food items carried on routine journeys should another person be encountered. More formal exchanges occurred, including huge transfers of food initiated by the kokwal that were intended to demonstrate his prestige.
Hogbin ([37]: 179) estimates a relatively low rate of homicide on Vokeo, citing a murder rate of about one per three years during the period 1933-1949: "Petty raiding occurred from time to time, but the massing of hosts of armed warriors and wholesale carnage are unknown. " Ordinary rules of behavior were suspended at the start of a major feast. Individuals insulted and physically attacked one another (without the use of weapons) to resolve past grudges. These brief altercations-which might last only ten minutes or so-reset the social relationships and prevented more serious conflicts from developing.
While the rates of interpersonal violence and warfare on Vokeo were quite low, this prohibition did not extend to other societies.

violated has a small population and is isolated and self contained. Each resident is acquainted with all the rest, and if they carried on warfare against one another the effects would be devastating. In no time everyone would be dead. But on the mainland some groups are always outside the limits of regular peaceful relations, with the members regarded as foreigners. What does it matter if such people are wiped out? No kinship link is severed, no moral precept violated ([37]: 57).
Thus, even when geographical separation is small (Koil Island is only 19 km away), social isolation can still exist.

3.2.
Maring. The 7,000 or so Maring, divided into twentyone territorial clan groups of 100-2000 people each, occupied the Simbai and Jimi Valleys of the western highlands of Papua New Guinea. Principal pursuits of the Maring included swidden agriculture and pig raising. The overall population density appears to have been well below the carrying capacity of the land [15].
The Maring are of particular interest to the anthropologist since significant contact with western culture did not occur until 1954. Hence Rappaport [15] and Vayda [40,41], among others, had the opportunity to study a nearly pristine indigenous culture.
Maring society was strongly egalitarian. The clan cluster was the largest political unit and there were no formally recognized offices, ascribed or achieved. Rappaport [15] has questioned even the use of the term "Big Man" to describe a leadership style, citing the fact that the Maring did not engage in competitive feasting. All men could participate equally in discussions and decisions affecting the group. Nevertheless, there were certain men who, having proven themselves in the past, were respected more than others. They were typically outgoing personalities able to articulate a path of action, often by recognizing an emerging consensus within the group. Rappaport ([15]: 31) notes the tenuous hold that an individual had on power: "it frequently happens that he who attempts to instigate group action has misjudged the consensus and he is not followed. " In contrast to other Papua New Guinea peoples, such as the Mae Enga [42], the Maring were not constantly at war. Conflict was discouraged within the clan. Disputes were often resolved by compensation rather than by violence to avoid upsetting the complex web of mutual obligation that bound individuals within the social group. The frequency of external wars, which sometimes involved shifting alliances of clans, was about one or two per generation [41] and may have been tied to a ritual cycle involving the breeding and exchange of pigs. Long periods of peace separated most conflicts.
Wars were divided into two types: "nothing" fights that were highly formalized and "true" fights that were more lethal. Nothing fights, which involved the exchange of inaccurate arrows across the battle zone, produced few casualties. Rappaport [15] suggested that the Maring purposefully kept the killing efficiency low to avoid needless deaths. Fights were terminated to care for wounded or tend garden plots, or even on account of rain. If a rout were to occur, the land of the vanquished party was not immediately occupied since it was thought to be under the control of the ancestral spirits of the previous occupants. Still, casualties in a rout could be large, sometimes amounting to several percent of the defeated population [40]. In contrast to "nothing" fights, "true" fights involved more deadly weapons such as axes or spears. Wars ended when the opposing sides agreed that enough men had been killed or when one side was decisively routed by the other. The Maring also conducted informal raids, a particularly lethal form of conflict since they focused on killing.
Prolonged periods of peace occurred, broken by the kaiko feast ceremony at which all truces were nullified and the accumulated grievances of the years might lead to violence. Murder was a significant cause of intergroup fighting [41], leading at times to revenge homicide, raids, or even war. However, escalation was not inevitable and compensation could be offered in apology for an offense. Honor and reputation were important to the Maring, and fighting was one way to demonstrate ability and to attract one or more wives. Population pressure does not appear to have been a significant cause of war [41].

Malaita. Malaita is a large volcanic island in the
Solomons with a contact-era population estimated at 65,000. Mountain ridges divide the island into numerous small habitable areas, some containing less than 100 people. Agriculture, particularly taro farming, and pig raising were the principal economic activities.
Maliata follows a classic Big Man style of social organization and leadership. As Hogbin ([43]: 62) observed: qualifications is no recognized supreme ruler over even a small territorial group, and the individuals who command the respect of their fellows have not permanent legal claims to obedience but rather obtain by the distribution of their wealth the co-operation necessary for the enterprises they initiate. No one ever holds sway over more than at a maximum 200 followers, and although the heir to an old leader has an initial advantage over possible rivals, any ambitious young man can supplant him if he works hard, distributes sufficient wealth, and wins the respect and approval of his relatives through superior personal qualifications.
Sharing of food between the Big Man and the people was expected, as was sharing among his followers. Through his conspicuous generosity, the Big Man dominated the flow of gifts within the group and hence the network of mutual obligation [44].
Violence was a major part of the culture of Maliata and pacification was not complete until 1927. Hopkins ([45]: 168) found that "peace, except for short, uncertain intervals, was practically unknown. Everywhere any individual or tribe might be attacked anywhere at any time. " Boys were trained for war in mock battles and joined raiding expeditions as soon as they were old enough to wield a club. They were told that no woman would marry them until they had shed blood. Still, there was a recognized need for cooperation 8 Journal of Anthropology at the local level and residents within the same district lived relatively peaceably. Compensation could be offered for some offenses. Feasts provided a venue to release pentup grievances and a Big Man might intervene to prevent a fight among his followers, although certain crimes within the group, including murder and adultery, were considered sufficiently serious to warrant the death of the perpetrator. A Big Man's authority was limited to his local group and he had little ability to adjudicate quarrels involving people from another group. Retribution might be carried out by a hired killer or by offering a reward for the death of an individual [43]. Murders were a common occurrence: Coombe [46] reports 14 homicides in a period of 6 weeks in the relatively small area that she studied.
Raids involving 200-300 men were a common form of warfare on Malaita and could occasionally result in the total elimination of the target group [46]. If there were survivors, the attack might trigger a chain of revenge killings that could last for years. Kessing ([47]: 69) notes that "In a society where blood-feuding was endemic and there was no institutional political structure for imposing law and order or maintaining the peace, one may well wonder how some sort of stable life was possible at all. "

New Caledonia.
New Caledonia is a large island of over 16,000 km 2 in the southeasternmost portion of Melanesia. Estimates made during the mid-19th century suggest a population of about 70,000.
The primary social unit was the localized clan, headed by a hereditary patrilineal chief. There appears to have been a strong sense of "senior" and "junior" in relationships, one that extended to a hierarchical system of chiefs. Toward the end of the nineteenth century the island may have been headed toward greater centralization of authority [48].
If for no other reason, "the demands of subsistence provided a compelling incentive to limit and contain fighting ([49]: 57). " Wedgwood [7] estimated that wars occurred about every five years between two large groups in the north of the island. Douglas [9,50] assessed that while fighting was common, it was not continuous. Twenty-four fights were recorded between 1845 and 1853 within less than one quarter of the land area of the island, but it is not clear whether these were of sufficient scale to be called wars. Escalation from personal redress to group involvement could occur as obligations were called in from kin and friends [9,50]. Grievances could be nurtured for generations and could involve complex alliances that crossed clan boundaries. Withdrawal was a common form of conflict avoidance for individuals and groups and compensation was an alternative to violence. Internal fights were more numerous, but less lethal, than external fights. While exact figures are unavailable, less than 8% of the deaths but more than 85% of recorded injuries were due to internal (intragroup) violence. This implies normative limits on the types of violence within the group, limits that did not extend to external fights. Most intergroup fighting was in the form of raids and the Melanesian principle of equivalence, a form of reciprocity, demanded equal numbers of casualties on each side [9,50], although the complete elimination of a group sometimes occurred in a particularly destructive raid.
Bensa and Goromido [51] note that many wars occurred after the death of a chief as contenders sought to prove their leadership and fighting abilities. Concern for the destructiveness of chiefly succession sometimes led to the appointment of an outsider, a neutral party acceptable to all sides. After assuming power, violence served to validate a chief 's power while reducing that of his rivals. Cannibalism occurred, but more from a political than an economic need. Eating the bodies of one's enemies was a way to prevent the deceased's spirits providing future support for the enemy camp.

Discussion
The motivating factors leading to violence and warfare may be divided into proximate causes, those that drive immediate decisions, and ultimate causes, those that create the context for those decisions. Among proximate causes, economic factors were among the most important in that they affected the survival of both individuals and groups.
Land rights were a cause of war in Melanesia, but fights were not always for the purpose of acquisition. Land was sometimes considered as under the oversight of spiritual forces aligned to the original owners, so taking land without the consent of, or worse, with the antagonism, of its spiritual protectors, was considered unprofitable. Also, the small size of most Melanesian polities meant that additional land, once acquired, would lead to a population increase in the conquering group to the point where it could no longer be controlled by a single individual. Group fission would then occur, with a return to independent polities [52] and the potential for renewed conflict.
Women were frequently cited as a cause of interpersonal violence and of raids that could, in some cases, lead to largescale warfare. Adultery was more than an insult to the honor of the husband since women contributed to the economic well-being of groups by tending gardens, raising children, and performing other domestic tasks.
Intangibles, most especially individual and group prestige, were potent motivators of violence. This is evident in the very destructive headhunting practices conducted in the Solomon Islands, where the value of the objects sought (human heads) lay in prestige and ceremony rather than physical utility. However, there was also an economic impact in that the collection of heads demonstrated the power of the individual and hence enhanced his control of vital resources.
To begin to understand ultimate causes, violence must be viewed in a broader social context. Melanesians saw their identity in terms of the groups in which they lived. From a practical perspective, Hogbin ([37]: 47) observed on Vokeo that "No one can exist in isolation; he needs help not merely for major undertakings but also in lesser tasks. For him to fall out with his kinsmen or affines is therefore tantamount to doing himself a personal injury. " On Malaita, Hopkins ([45]: 32) found that membership in the tribe, "is the only way of securing any measure of protection of life, of property, or home. A tribe defends, shelters, feeds its members, and avenges for their death or any damage done to them or the tribe concerned. " More starkly, Hogbin ([53]: 51) found that on Guadalcanal "nobody can afford to forego the help of his fellows. " Survival outside of the group might have been possible, but it was tenuous and highly undesirable.
Individuals depended on group membership for more than physical needs. To an extent, individual and group identities were intermingled. As Ross ([54]: 191) writes of the Baegu of Maliata, "prestige and power are group attributes rather than individual ones. Men acquire renown through the demonstrated power and wealth of their groups. . ." Identification with the group transferred ambition from the individual to the group. Conversely, the focus on the group may have, in some cases, reduced individual ambition. Hopkins ([45]: 31) writes: "it is rather against a man's interests to make the most of the piece of land he works. It belongs to the tribe really, and if he gets a surplus off it they will enjoy it. " Thus, "ambition is. . . tribal. " The notion of collective identify, in which an individual saw his or her personal identity as intrinsically related to membership in a group, made the defense of the group and its reputation a priority. A threat to the group was a de facto threat to the individual.
Individual and group identification could go in the other direction, wherein an individual represented the group. On Malaita, eating a single enemy body absorbed the mana of the entire tribe, making it unsustainable. Its members dispersed to other groups.
Group identification does not equate to group rigidity. Competitive leadership in Melanesia recognized that a perceived abuse of power could cause people to leave one group and join another. Presumably, they would then be willing to defend their new group with the same vigor as they did their original one. It was the need to be part of group, rather than a particular group, that was most important. However, the very notion of belonging to one group implies that one is not a member of another group. This negative component of group identity is an important factor in competition between groups.
The strong association of individual and group suggests that an ultimate cause of conflict in Melanesia was associated with the concept of identity, the need to show that the group of which one was a member was important enough to risk injury or death in a fight. As Scheffler ([55]: 399) found for Choiseul Island, "Intergroup conflict was an essential element in the maintenance of the individual identities of those groups and, consequently, of the continuation of the larger system. . ." Fighting was a form of costly signaling in which the identity and continuity of a group was maintained. There was a distinct economic component of violence in that it validated the individual's and group's rights to land and other resources essential to survival.
The absence of territorial ambition in much of Melanesia, combined with the inability of a single individual to exert authority over a wide area, suggests that a second ultimate cause of conflict was the notion of stability, the desire to maintain the status quo or, if it was disturbed, to return things to their previous state. On Kiriwina,Irwin ([56]: 52) writes "one might wonder whether the normal outcome of war was the eventual reestablishment of the former political situation. " Powell ([57]: 142) came to the same conclusion regarding conflict on Kiriwina, namely that the aim of war "was not to obtain a decisive conclusion. . . but rather to restore the balance of power. " Writing of the small islands off the coast of Malekula, Layard ([58]: 588) noted that wars were fought "almost entirely on questions involving the prestige of one group against another, in order to maintain the existing order of society by wreaking vengeance on any who seek to disturb it. " Here, as in other places, the number of dead needed to be the same on both sides to prevent a resurgence of violence. Among the Maring, Rappaport ([15]: 113) found that "a principle of absolute reciprocity is supposedly in force; every death at the hands of an enemy group requires the killing of one of that group's members, and peace should not be made until both antagonists have revenged all of their losses. " A truce might last ten years or more before fighting resumed to even the score.
The concept of stability can be applied to other elements of violence. A revenge attack was more than emotionally motivated vengeance. From an economic perspective, it reduced the recipient of revenge by the same measure of productivity as was suffered in the original offense. Revenge could be nonviolent; compensation was often offered as an alternative to killing. It reduced the economic state of the compensating party relative to the aggrieved party. Compensation had an element of justice and as such "never centers on misdeeds, rather on the transfer of pigs, money, and women needed to return the victim to its previous levels of reproductive power ([59]: 73). " Revenge had a deterrent value. Sillitoe [60,61] suggests that the need to compensate the other side for losses suffered in battle discouraged groups from fighting. It was not worth the cost. Douglas ([50]: 27) writes "the basic premise of Melanesian political interaction and fighting [is] that maintenance of relative equivalences demanded revenge for insults and injuring, that kinspeople, friends and allies might be held responsible for the actions of individuals. " There were limits to violence. While leaders sometimes encouraged violence to increase their influence and prestige, they suppressed it when it threatened their interests or the interests of the group. Incessant internal violence reduced social solidarity and hence the ability of a group to withstand external attack or conduct attacks itself. Hogbin ([43]: 75) writes that on Maliata, "the ramifications of mutual dependence. . . are so complex that people cannot afford to pursue their quarrels very far. " Conversely, the lack of kinship or other connections with people in different groups led to a greater acceptance of violence against them, as in the case of Vokeo cited above.
The relationships existing between individuals and groups, while dynamic, need not be violent. The pacification of Melanesian societies following the arrival of western authorities shifted competition from violent to peaceful processes. Feasts and other competitions served to establish and maintain status differentials. Such events were a type of costly signal, demonstrating that reputation was important enough to justify the expenditure of significant economic resources. Jennes and Balantyne ( [62]: 256) suggest that the end of fighting not only encouraged but "necessitated" the elaboration of food exchanges. (Note that gift exchange was a vital part of Melanesian society even during violent times.) Gift exchange served to create a network of mutual obligation between the members of a society or set of societies. Such obligations could extend to reluctance to use violence against one to whom one is indebted. Peoples ([63]: 300) proposed that "natural selection between the groups favors the survival of groups with many reciprocators. " In practice, Melanesian society employed both gifts and violence, often with one closely following the other, as mechanisms of social interaction. It is of interest to compare the level of violence in Melanesia with that of other Pacific culture areas. Table 2 gives the average values of interpersonal violence and warfare for Melanesia compared to a sample of societies from Polynesia [5] and the Caroline Islands [6].
Chowning [10] argues that violence in Melanesia was probably no greater than in other parts of Oceania, but Table 2 suggests that Melanesia had substantially larger rates of both forms of violence than did the other culture samples. However, while the averages dispute Chowning's broad conclusion, it is certainly true that high levels of violence-as well as some cases of relatively peaceful societies-occurred in each of the three culture areas. For example, the Polynesian island of Niue (population 5,000) and the Micronesian island of Chuuk (population 11,000) each had fragmented societies similar to those of the larger islands in Melanesia. In each case the degree of political integration was low and a constantly shifting set of alliances engaged in nearly continuous warfare.
It is significant that Niue and Chuuk had traditions of ascribed leadership, although the acceptance of inherited authority was weak in both cases. This suggests that the size and number of autonomous polities may have been as important or more important in determining the frequency and severity of warfare than the style of leadership. This is the situation found in many parts of Melanesia. In terms of the concepts of identity and stability, many small groups competed for their right to exist, for prestige and, in some cases, for land.
Ascribed leadership in Polynesia, Micronesia, and parts of Melanesia assisted the development of multilevel distributed governance that could span large areas and populations, although complete unification of a large island or island group was seldom achieved. In parts of Melanesia where the span of control of a Big Man was limited to the range of his personal interactions, integration was limited to alliances rather than hierarchies and, as on Chuuk and Niue, violent competition was frequent. Without defined roles and responsibilities for leaders, anyone could attempt to take a group to war, increasing the risk of intergroup conflict.
Thus Chowning [10] is correct in saying that the nature of violence in Melanesia may not have been different from that in other parts of the Pacific. However, the combination of competitive achieved leadership and the larger land areas of Melanesia may have exacerbated the frequency and severity of violence compared to Polynesia and Micronesia.