Darkness in El Dorado - Archived Document
Anthropological Niche of Douglas W. Hume
Home | Darkness in El Dorado | Contact

Internet Source: Kin Selection and Conflict: an Analysis of a Yanomamö Ax Fight, November 7, 2010
Source URL: http://www.anth.ucsb.edu/projects/axfight/updates/chagnonandbugos.html


KIN SELECTION AND CONFLICT: AN ANALYSIS OF A YANOMAMÖ AX FIGHT

Napoleon A. Chagnon and Paul E. Bugos, Jr.

"The social behavior of a species evolves in such a way that in each distinct behavior-evoking situation the individual will seem to value his neighbors' fitness against his own according to the coefficients of relationship appropriate to that situation."
W.D. Hamilton, 1964
"The Genetical Evolution of Social Behavior: II"

"The span of social distance between those who exchange conditions the mode of exchange. Kinship distance, as has already been suggested, is especially relevant to the form of reciprocity. Reciprocity is inclined toward the generalized pole by close kinship, toward the negative extreme in proportion to kinship distance . . . close kin tend to share, to enter into generalized exchanges, and distant and non-kin to deal in equivalents or in guile."
Marshall D. Sahlins, 1968
"On the Sociology of Primitive Exchange"

From the days of Lewis Henry Morgan on, students of primitive culture have generally come to the same kind of conclusion so eloquently summarized in Sahlins' characterization of the nature of primitive exchange: closeness of kinship is an important mediator of interpersonal relationships of many kinds in human societies. *1

The implications of Hamilton's arguments about inclusive fitness (1964) and the nature of behavior involving related individuals raise the possibility that we can develop a general scientific explanation for kinship behavior in terms of natural selection. This is a powerful and, to some, uncomfortable possibility, for anthropologists have long considered kinship their special domain and an invasion from biological theory is, to some of us, inappropriate and perhaps even threatening.

The "nature" of kinship in human societies, particularly those kinds of societies traditionally studied by field working anthropologists, is a question for which many different kinds of answers exist. It is, to many, mostly symbolic in content; to others, the vehicle through which other kinds of relationships are expressed-economic, political, jural, domestic, etc. With few exceptions, the general and widely held anthropological views on human seem to play down-or even deny-that the facts of relatedness have much to do with human "kinship." In many societies, especially those whose members have been decimated by exotic epidemics, there may be a relatively poor relationship between kinship classifications and putative biological relationship; in some, there may be so few survivors that kinship systems are known primarily through one or just a handful of individuals, In others, where it might be possible to elucidate the relationship between genealogical relationship and kinship classifications, the investigations seem not to have been conducted at alland what we know about the "kinship systems" has been gotten by the observer from just a few informants. Or, if the investigations have been conducted, they tend not to be made available in the literature. There are, in fact, relatively few ethnographic descriptions that tell us how the observer collected his or her information about the "kinship system" found in any particular society, and for that reason it would be difficult to make a general summary of what it is anthropologists really do know about human kinship systems, especially the associated "kinship" behavior.

All of this is not to say that kinship in human societies is non-symbolic or does not serve as a vehicle for other kinds of relationships. An analysis of kinship behavior from the perspective of inclusive fitness theory does not obviate other, more traditional concerns and approaches to the same subject; it supplements them and increases our understanding of kinship. More important, we believe that such an approach can answer the larger question of why kinship is such an important category of human concern and why, as Robin Fox argues (chapter 5, this volume), it is one of the kinds of preoccupations with which humans seem to universally concern themselves . . . a preoccupation they apparently learn very easily.

Hamilton specifically phrased his argument in terms of "valuing" or "comparing" fitness according to closeness or remoteness of relationships in particular situations. Those of us who have spent considerable periods of time studying kinship behavior in the field will appreciate the significance of context in interpersonal kinship behavior. People manipulate relative, exaggerating closeness at times, denying it at others. In some cases, one can even get the impression that kinship is almost irrelevant to the humdrum workaday activities in which individuals engage, while at prerogative dominate what otherwise seems like a trivial matter to the uninformed outsider.

There is one context or situation in which, we believe, some of the most essential and rock-bottom characteristics of kinship come to the fore and reveal themselves clearly: conflict or crisis situations, Here, in Fortesian terms, the axiomatic qualities of human kinship as prescriptive altruism take on form and substance (Fortes 1969). Ambiguity and metaphorical aspects of kinship, it would seem, should be minimized as the actors elect to follow particular courses of action exercising choices that have obvious costs and benefits. If we are interested in examining individual human behavior with an eye toward understanding the extent to which that behavior is "tracking" biologically relevant dimensions of kinship relationships, it seems that crisis or conflict situations involving potential hazard to the actors are a reasonable place to begin looking.

This paper analyzes one such crisis event, an ax fight, that occurred in a Yanomamö village in southern Venezuela in 1971. The Yanomamö are characterized by a wide range of formalized and unformalized fights, ranging in seriousness from chest-pounding at the lowest end to outright raiding with intent to kill at the upper extreme (Chagnon 1966, 1967, 1968a, 1968b). Ax fights fall somewhat closer to the more violent end of the spectrum, since injuries during such fights can be very serious-and even fatal. The ax, of course, is an exotic item of material culture and such fights must therefore be of relatively recent origin in their culture. *2 Ax fights are, however, similar to the more traditional club-fights and probably are just refinements and improvements on them as a result of the introduction of new technology. Indeed, during the particular fight in question, the initial skirmish began with clubs-and clubs continued to be used when the fight escalated to more serious dimensions after a few men took up axes.

We have chosen to analyze this particular fight primarily because the event was filmed with 16mm motion-picture, synchronous-sound equipment and it has become the subject of a motion picture publication. The published film, however, deals with social organization, particularly marriage and alliance practices, and how the anthropologist, through his analytical techniques, renders apparently chaotic events intelligible. *3

In addition to that filming, the senior author of the present paper photographed the fight with 35mm still equipment from a slightly different and somewhat closer angle. This combination of photographic coverage of the event, which makes it possible to identify all the participants, and genealogical and demographic data collected by the senior author during the several years prior to and after the fight, make it possible to do a very detailed analysis of the event from the perspective of kin selection theory. While the senior author has witnessed many such fights during his twelve-year study of the Yanomamö, none have been so meticulously documented by photography and none have been described in such a way that all participants are identifiable. Such fights erupt with explosiveness and usually last such a brief time that without some kind of photographic documenting techniques, detailed analysis of the kind presented here would be very difficult except for just the smallest of fights. An important point to make, however, is that this fight is in no way unique or even unusual. It may be taken as a specific example of a much larger universe of similar kinds of Yanomamö conflicts characterizing the geographical area of the senior author's major field of research effort (see Chagnon 1974).

Another point is in order. The data required to make the following analysis were collected long before "sociobiology" entered the theoretical repertoire of social anthropologists and, therefore, could not have been systematically gathered with kin selection or reciprocal altruism arguments in mind. Two consequences of this fact should be obvious. First, it is not likely that the information systematically favors or disfavors the outcome of tests of kin selection theory. Second, had the data been collected with tests of kin selection in mind, it is likely that the definitiveness of our conclusions would be enhanced, for no one is more painfully aware of the kinds of supplementary detail necessary to make convincing statements about the applicability of kin selection theory to human kinship behavior than we. One consequence of our attempts in this paper is the strengthening of our conviction that extremely meticulous, quantitative studies of kinship behavior are absolutely essential if broad generalizations about the "nature" of human kinship behavior are to be developed.

The Fight

The fight, in its broad features and significance, was largely the extension of earlier disputes and of antagonisms that plague all Yanomamö villages as they grow in size and begin to fall apart. The mechanisms of kinship, marriage exchange obligations, and the authority of the village leaders become increasingly ineffective in organizing the villagers into a cohesive, cooperative whole (Chagnon 1974, 1975). The village in question, Mishimishimaböwei-teri, has fissioned several years earlier when its numbers had reached approximately 400 individuals, a size that is staggeringly large by Yanomamö standards. After the fission, one group, still calling itself Mishimishimaböwei-teri, vacillated in size from about 230 to about 275 people, depending on the numbers of families that temporarily rejoined it, the main group, from the splinter village. This pattern families leaving and rejoining characterizes most Yanomamö village fissions in this area of the tribe, since the "core" of the splinter group often consists of a few authoritative men and their families who have more antagonisms against the core leadership in the original village than do particular heads of households in either group. Thus, individual families often return to the original village in an attempt to patch up grievances and continue their social and economic pursuits in somewhat less strained circumstances. Purely in terms of energy, it is simply more efficient to return to an already-producing garden in the original village than to begin clearing a new garden at a great distance away (Chagnon 1968a, 1968c, 1974).

As the splinter group formed by a fission takes on its own political identity and characteristics, subsequent visits to the original village become, in themselves, political events. The relationships between visitors and hosts during such visits are no longer exclusively dominated by the kinship and marriage ties between them, but are increasingly subject to residential realities and political principles. Thus, visitors are received with considerable formality and tendered particular forms of hospitality and generosity that are assumed to be short-term prerequisites. Proper decorum in such cases requires that the visitors depart for home after a politely long visit, their departure usually being met with some relief by the hosts who, if they had fulfilled their obligations as proper hosts, can return to the workaday tasks of gardening and other domestic duties. [See Figure.8.2, which was created for the CD version of this paper ]

The fight that we are about to analyze erupted between a group of recently fissioned ex-residents of the village and a few of the members of the original village. The fission that separated them had occurred too recently in the past to permit the splinter group to assume its own political identity, yet they had returned as visiting dignitaries, expecting to be fed and feted. Moreover, they were deliberately finessing their kinship ties to the local group, expecting to be received not only as a delegation from an independent village but as kinsmen as well-they were trying to have it two ways. In addition, marriage exchanges over a number of generations linked one particular group of the hosts to some of the politically prominent men among the visitors in very complex and intimate ways. Indeed, the host men thus tied by kinship and marriage to the visitors supported them in the fight, and were sincerely interested in bringing these visitors back into the village as permanent residents. One of the more prominent local men had gone so far to demonstrate his sincerity that he cleared a moderately large garden, planted it, and offered to present it to the headman of the visiting faction-his brother-in-law. There was, however, much less enthusiasm among the rest of the members of the village (mostly the leaders from Lineage 1222) for having this group return as permanent residents, so the longer they stayed as visitors the more thinly veiled the local antagonisms became. The overall situation, in terms of the politics of fissioning, was extremely ambiguous. Many of the local people were growing weary of entertaining and feeding the visitors, who appeared to them to be taking advantage of a good thing: they stayed far too long for proper decorum, ate ravenously, complained about not being treated as was their just due as both visitors and kin, and loafed around all day while the hosts worked to support them.

Matters came to a head when one of the men from among the visitors (Mohesiwä 1246) ran into a party of women in the garden and demanded that one of the women (Sinabimi 1744) give him a share of the plantains she was carrying back to the village for her own family. She refused to give him plantains, punctuating her refusal with an insult. Mohesiwä was incensed, and beat Sinabimi with a piece of wood. She fled, screaming and crying, into the village, revealing her story to the others. This angered her kinsman Uuwä (1897) in particular. Some of the relevant social dimensions of the reaction by Uuwä are:

1 Uuwä was a member of the most prominent descent group (Lineage 1222) in the village, the group from which the undisputed local headman came. Thus Uuwä was a brother to the local headman, and in a sense, represented local authority. [ Figure.8.2 ]

2 Uuwä was also a half-brother to the beaten woman, and as is frequent in Yanomamö domestic group organization, brothers stand up for and defend sisters against cruelty inflicted on them by other men.

3 The affront-beating of a woman-was inflicted by a visitor during a situation of mounting tensions, tensions that were the result of visitors wearing out their welcome in the host village. In an important sense, overstaying the visit was a kind of political coercion, reflecting the capacity of the visitors to compel the hosts to treat them with deference beyond a socially appropriate point.

Uuwä took up a large club and rushed to the center of the village hurling insults at Mohesiwä, who met the challenge by coming forth with an unstrung palmwood bowstave. The two men flailed wildly at each other with their weapons, but since Uuwä's club was longer than Mohesiwä's bowstave, he managed to deliver a painful blow to the latter's forearm. By this time, supporters of both men were quickly arming themselves with long clubs, but Mohesiwä's supporters were clearly more committed to supporting him than were Uuwä's supporters, probably being more concerned about their own disadvantage as visitors and overall numerical inferiority. As soon as Mohesiwä's younger brother came to his aid with a long club and held Uuwä at bay, the conflict stabilized and the male contestants-just a handful of men-glared menacingly at each other. Their female supporters, especially Mohesiwä's sister and mother, hurled vicious insults at Uuwä. Still, the fight appeared to be over at this point and the main antagonists turned, stalked haughtily away, and returned to their respective hammocks.

Meanwhile, the husband of the beaten woman Yoinakuwä (2248), and his brother, Keböwä (0910), were scurrying about in their own houses looking for more dangerous weapons. No sooner had the club-fighters reached their hammocks than these two men emerged from their houses brandishing a machete and an ax, respectively. They were coming to the aid of their brother-in-law, Uuwä, and simultaneously protecting a "wife." They ran across the village clearing toward Mohesiwä's house, weapons raised conspicuously. They plunged into the house to attack Mohesiwä, but were met with a sea of arms and bodies and were partially immobilized and unable to strike Mohesiwä with their weapons. A struggle ensued within the house, and Keböwä wrested his ax from the resisters. He then grabbed Mohesiwä by the arm and began beating him on the legs and back with the blunt side of his ax, managing to deliver several crunching blows.

The fight had clearly escalated, and large numbers of men began arming themselves with clubs and other weapons. Mohesiwä's younger brother, Tourawä [Törawä] (1837) , again came to his rescue, discarding his club, taking up first a machete and then an ax. He attacked Keböwä from the blind side and managed to deliver a series of equally crunching blows to Keböwä's legs, arms, and back with the blunt side of his ax. Stunned and distracted-and in pain from the blows Keböwä stopped beating Mohesiwä and turned to identify his new adversary. Tourawä backed away a few steps and menacingly turned his ax head up, as if to strike Keböwä on the head with the sharp edge. As he stood there, poised to strike, someone reached up and grabbed his ax-handle from behind him, twisted it so as to turn the sharp edge back down, and dragged him out of the fight. The youth turned to struggle for control of his ax, but as soon as his back was turned, Keböwä rushed him from behind and delivered a powerful overhead blow with his ax, blunt side exposed, striking him squarely in the middle of the back between his shoulder blades, just missing his spine. The sound of Keböwä's ax thudding into Tourawä's back was sickening, and the youth collapsed instantly. *4 At this point, several older men stepped into the fight, enraged. A rapid series of blows were exchanged with clubs and the two groups alternately charged and withdrew.

With the entry of the most prominent men, the fight again stabilized. It gradually de-escalated to a series of insults, hostile stares, and ended. Tourawä, the youth felled by Keböwä's ax, painfully regained consciousness and was led off to his hammock. At this point, all the partisans of both sides returned to their homes and tempers cooled down, save for a period of time when the women of both groups screamed insults at each other before withdrawing. The next day some of the visitors packed their possessions and left for home.

The Analysis

Hamilton's formulation of the inclusive fitness hypothesis (1964) accounts for the preservation of altruistic tendencies in a population through natural selection operating at the level of individuals (cf. Williams 1966). Individuals, in maximizing their inclusive fitness, should be expected to favor other individuals who are genetically related to them, i.e., who are ultimately in a position to contribute the altruist's genes to succeeding generations. If the potential costs (in fitness units) to the altruist are the same should he/she favor (1) a close relative, (2) a distant relative, or (3) a non-relative, then one should expect the altruist to "invest" in that individual who, by receiving the benefit, is most likely to yield an equivalent or greater fitness benefit to the original altruist.

An analogy with banking and interest rates might be useful to illustrate the principle. Assume that you have three friends who each own a bank, and all "need" investors, or they will collapse. Your $100,000 would benefit all of them equally. But Bank A will give you 5 percent interest, Bank B 6 percent interest, and Bank C 10 percent. The cost to you in each case will be risking your $100,000, but the potential return in greater if you place your investment in Bank C. One would predict that you would invest in Bank C, on the assumptions that you would attempt to maximize your return and that the chances of each bank's folding are equal.

General predictions

Kin selection theory predicts that if you sustain the same "costs" to your potential fitness in helping a (1) close relative, (2) a distant relative, or (3) a non-relative, your inclusive fitness would be better served by aiding the relative that is genetically most related to you. In addition, the theory also predicts that as the costs to your own potential fitness increase, you would increasingly favor more closely related kin over more distantly related or unrelated individuals, providing that the benefits to the recipients of your aid remain approximately the same. Conversely, if the risks to your own fitness are very small, the theory would predict that individuals would help close relatives in a high proportion of cases, more distantly related relatives in a lesser proportion of cases, the proportion ultimately being predicted by the coefficient of relatedness of the distantly related kinsman to the altruist. If the risks are very small indeed, then it would be expected that an individual would aid even a non-relative on the expectation that such a person would ultimately, at some future date, reciprocate the aid in a similar circumstance (Trivers 1971; Alexander 1974).

There are, of course, intervening variables in the many sociocultural circumstances in which humans normally operate. One might expect, for example, an individual to aid a relative with high reproductive potential over an equally closely related individual with lower reproductive potential, i.e., a child over a parent. Since the child has a greater probability of further enhancing the altruist's inclusive fitness than the parents of the altruist, the altruist would potentially gain a higher inclusive fitness by aiding the younger relative (Alexander 1977b). In other circumstances, overlapping fitness interests of pairs of individuals must be considered. Thus, it might be more beneficial to take risks to aid an affine, say a brother-in-law, than to aid a lineage mate who might be more closely related. This is especially true in human societies characterized by systematic marriage exchanges involving cross-cousins. Since a brother-in-law (say, a second cross-cousin) ultimately provides you with your spouse, and your lineage mate (say, a first parallel cousin) of the same sex is your competitor for that spouse, your inclusive fitness interests might be better served by taking a higher risk for a less closely related individual (brother-in-law) whose inclination to favor you in return by continuing to provide you with mates might be a function of your willingness to help him. Other examples of asymmetry in costs and benefits for specific kinship and other social dyads are discussed elsewhere in this volume (see Irons, chapter 7).

For the example under discussion here, the ax fight, the major variable appears to be closeness of genetic relatedness: individuals seemed primarily to "decide" to aid others on the basis of the degree of relatedness obtaining between themselves and other participants in the fight. That is to say, the "risks" to the potential fitness of each individual seemed to be high and the potential fitness benefits each would drive from joining the fight would therefore appear to depend on closeness of relatedness.

One point of clarification is in order. It is not necessary, in testing kin selection hypotheses, to assume that the individual actors in any social situation are cognizant of "fitness" or are capable of "calculating" coefficients of genealogical relatedness. Natural selection favors organisms that behave adaptively, whether or not they are aware of the evolutionary consequences of their behavior. This aspect of the general theory of evolutionary biology constitutes a major stumbling block to social scientists in general and to social anthropologists in particular, many of whom conceive of "kinship" behavior in explicitly or implicitly mentalistic or ideological terms. For example, one prominent anthropological critic of kin selection theory, Marshall D. Sahlins, misunderstands this central aspect of natural selection when he claims that the "failure" of "sociobiologists" to address the problem of an organism's lack of cognitive appreciation of the evolutionary effects of its behavior ". . . introduces a considerable mysticism in their theory" (Sahlins 1976a:44-45). One could argue, with the same logic, that physicists are behaving "mysteriously" when they suggest that planetary orbits conform to Newtonian models, since it is "mysterious" to suggest that the planets describing these orbits are aware of Newtonian principles.

Specific predictions and relevant data

In the analysis of the ax fight that follows, then, we shall look for evidence that individuals help closer relatives over more distant ones: we expect that if someone comes to the aid of another person in the fight, the helper will be more closely related to the individual he or she is helping than he is to the village at large or to the opponent in the fight. One would predict, from kin selection theory, that the supporters of Mohesiwä will be more closely related to him and to each other than to Mohesiwä's opponent, Uuwä, and his supporters. Conversely, we also expect that Uuwä's supporters would be more closely related to him and among themselves than they are to Mohesiwä or his supporters.

By relatedness, of course, we mean actual genealogical relatedness, not merely terminological salutations. This is not to say that all men who are called "brother" by, for example, Mohesiwä actually participated in the fight as his supporters or that we are so naive as to assume that all kinship terms invariably reflect the minimal English biological equivalent when translated. We are, in short, examining the genealogical dimensions of kinship in this section of the paper. The extent to which these genealogies are accurate reflections of biological relationships can be judged by examining an earlier work of the senior author, where a detailed discussion of the field methods used in collecting the data is given (Chagnon 1974). We do no assert, however, that the genealogies are absolutely accurate and that they are entirely free of any errors; we only argue that a scrupulous attempt was made during the twelve-year period of the study to collect genealogical data that was as close to biological accuracy as possible, and that paternity-exclusion tests performed on blood samples suggest that there is good reason to believe that the genealogies are reasonable approximations to biological reality (Chagnon, chapter 4:98). We know of no reason to assume that the errors in the data are large or that they contain any bias that could make the statistics cited below specious.

The method we use to express kinship relatedness is essentially Wright's Inbreeding Coefficient (as described in chapter 4, this volume p. 110). See also Chagnon 1974, 1975), converted, by doubling the value, to the coefficient of relatedness, i.e., the statistic used by Hamilton in his formulation of the kin selection theory. We call this statistic the "coefficient of genealogical relatedness," Fg, to remind the reader that it is based on informants' statements about their genealogical relatedness.

A word of caution should be added here: we, as observers, might have more genealogical information at our disposal than the Yanomamö individuals might be utilizing as they "track" their social environment and make decisions about aiding various relatives. When we express the closeness of relationship between any pair of individuals as a coefficient of relationship, we are utilizing nearly all of the genealogical information available to us. *5 A major theoretical and methodological question is: do the Yanomamö appear to utilize the same amount of information?

An example is in order to clarify the problem. Because of the systematic exchange of marriage partners over many generations, individuals are, in many cases, related to each other in multiple ways. Some of these relationship "loops" are close, some are distant; i.e., an individual might be related to another as a first cousin by one genealogical loop but also as third cousin by a different loop. The coefficient of relationship would sum the values of all loops and express the relatedness between both individuals as a simple statistic, whether or not the more "remote" loops are even known to the actors. On the other hand, the multiplicity of relatedness loops does reflect the degree to which systematic marriage exchanges in previous generations have "bound" kinship groups to each other and obligated them to continue exchanges in the current or future generations. In general, the closest genealogical loops account for the major fraction of the value of the coefficient of relatedness; i.e., a multiplicity of relationship loops does not necessarily change the coefficient of relationship by a large increment, for in most cases the additional multiple loops reflect remote degrees of relationship. At this stage in the application of kin selection theory to predictions about Yanomamö behavior, we are utilizing all of the genealogical information in arriving at coefficients of relationships.

In a future publication, we shall refine the methodology in an attempt to determine the utility of segregating complex patterns of relationship into components that might enable us to predict social behavior more accurately. The issue to be addressed there has to do with the possibility that the closest genealogical connections are properly considered from the perspective of kin selection whereas the more remote connections might be taken to express marriage obligations, i.e., reciprocity expectations.

Table.8.1 [a mouseclick on bolded words will call up graphics; a click on the graphics will cause them to disappear] gives the matrix summarizing the degree to which each member of Mohesiwä's group (the visitors) is related to all other members. *6 The individuals are listed in the table according to their identification numbers, and occur in the table because they participated in the fight in some active way. Their participation is known from (1) notes taken in the field when the fight occurred, (2) their appearance in the 16mm motion-picture film and (3) their appearance in the 35mm still photographs taken during the fight.

Table.8.2 gives the matrix summarizing the degree to which each member of Uuwä's group (the hosts) is related to all other members of that group that supported him in the fight.

Table.8.3 summarizes the coefficients of relationship between members of the visiting group to the host group, and vice versa.

It is clear from the three tables that the members of each team are more closely related among themselves than they are to their opponents, i.e., that mutual supporters are more closely related genetically among themselves than they are to the individuals they are opposing. It is also clear that the visitors are, as a group, much more highly related among themselves than are the members of the fighting group that opposed them in the fight. Part of this is due to the fact that the host fighters came primarily from one of the smaller descent groups in the village (Yoinakuwä and his brother, Keböwä, are the chief spokesmen for that group), one that has not yet become intimately integrated into the genealogical structure of the village through multigenerational marriage ties. An examination of Table.8.4a and Table.8.4b , however, reveals that the members of the fighting team from the host group are more closely related among themselves than they are to the rest of the members of their own village. That is, the members of the host fighting group are somewhat more related among themselves than we would expect to occur by chance alone (they are approximately 8 percent more related among themselves than they are to the village at large), but, although the difference in relatedness in the right "direction," it is not statistically significant. The fighters from the visiting group and their local supporters, on the other hand, are related among themselves 234 percent more than they are to the village as a whole and 335 percent more among themselves than they are to their opponents.

Table.8.4a and Table.8.4b also summarize how the supporters of the major principals in the fight-Mohesiwä (1246) and Keböwä (0910) - are related to their respective champions and to the opponent of their champion. Thus, Mohesiwä's supporters are related to him 780 percent more closely than they are to his opponent, and Keböwä's supporters are related to him 210 percent more closely than they are to his opponent.

Finally, Figure.8.1 gives the distribution of average relationship coefficients for every individual in the village, including the visitors, for comparison. The mean value for egocentric relationship, using the Fg statistic, is 0.0790. That is, if any individual were taken at random from the village and systematically compared to all 267 other members of the village, we would predict that the average relationship would be Fg = 0.0790. Clearly the members of the fighting teams do not represent a random selection of the village; indeed, the average relationship among Mohesiwä's group falls entirely outside the range of the village distribution of average Fg values with its value of 0.2123.

Number of relationship loops: alliance versus descent

The number of different ways any pair of individuals are related-relationship loops-can be taken as a crude measure of the extent to which their immediate ancestors engaged in reciprocal marriage exchanges over several generations. This statistic, however, cannot be used by itself to express "closeness" of kinship, for many of the remoter relationships are associated with very low values of the Fg statistic. That is to say, individuals with two grandparents in common will be more closely related than individuals with two great-great grandparents in common, even though in both cases the number of "loops" may be just two.

Table.8.4a and Table.8.4b summarize the number of ways (or "loops") members of each fighting group are related to both their champion and their opponent, i.e., how may ways they are related to the person they supporting versus the person they are opposing in the fight. Thus the supporters of Mohesiwä are related to him, on the average, 4.75 ways, suggesting that they are in part obligated to support him because of previous marriage exchanges that tie their families to his. Note, however, that the magnitude of the value of Fg does not correspond perfectly with the number of relationship loops in all cases: individual 2505 is related to Mohesiwä (1246) four different ways, but the value of Fg derived by summing the four separate fractions of that value comes to only 0.0390. That is, the four different relationship loops involve remote ancestors. By comparison, individuals 0714 and 0723 are only related to him (Mohesiwä) by one relationship loop each, but the associated value of the Fg statistic in both cases is 0.1250, i.e., the common ancestor is a "close" ancestor (Mohesiwä is their sister's son).

Mohesiwä's supporters are much less related to Keböwä in terms of loops: they are, on the average, related to him only 1.00 ways each. In one case, however, the relationship is relatively close: individual 2505 is related to Keböwä two different ways with a summed Fg value of 0.1562 and, interestingly enough, 2505 is related to Mohesiwä four ways-but more remote ways (Fg = 0.0390).

Turning to the patterns of relationship Keböwä's supporters have to him, we find that while the Fg average values show they are more closely related to him than they are to his opponent, the number of relationship loops between themselves and his opponent is actually much larger: 4.09 compared to 1.50 for number of loops to Keböwä and, as in the case of Mohesiwä's supporters mentioned above, some of the supporters of Keböwä are more closely related to his opponent than they are to their own champion. However, it is clear that despite the relatively large numbers of loops connecting Keböwä's supporters to his opponent, they are in fact more closely related to their champion-even though the number of relationship loops is small. In brief, the larger number of relationship loops, because of their remoteness, does not necessarily result in closer relationships as measured by the Fg statistic.

The factors behind the composition of the team of supporters that formed around and defended Keböwä and his brother, Yoinakuwä, appear to be largely affinal in overall quality. Table.8.2 shows the descent and marriage patterns that are relevant. Three of Keböwä's thirteen supporters are men who have married his brother's daughters, men who would, by the extension of Yanomamö kinship logic, be his "sons-in-law" as well (Egos 1109, 0789, and 2134). Three other men (Egos 1062, 0390, and 1827) are his brothers-in-law, having either married his sister or having given a sister to him in marriage. This point was made in the more sociological analysis in the 16mm film-that alliance ties figured prominently in the structure and organization of the fight.

Viewing the affinal ties in a temporal and developmental perspective, it seems reasonable to argue that the incorporation of small lineage segments such as that represented by Keböwä and his siblings is initiated predominantly through marriage ties. The "solidarity" that characterizes the relationships of the small group to the larger co-residential group, in this view, rests primarily on affinal obligations and the expectations that future marriage exchanges will occur, cementing the inter-group ties more effectively. As these exchanges develop over several generations, filiative ties ramify, adding a new and perhaps even more enduring nexus, transforming the relationships into those of kinship or, to borrow Fortes' felicitous term, into relationships of "prescriptive altruism."

One might argue, further, that affinal relationships are essentially the initiators of social cohesion and yield only transient amity, and that they must be bolstered and supported by filiative links realized through additional marriage exchanges involving already related individuals. In terms of the important debate between Fortes and Leach (see Fortes 1959, 1969), alliance initiates the relationships of solidarity that bind lineal descent groups to each other. Descent and filiation add, over time, the more enduring and more binding cohesion, a cohesion that can be continually reinforced and renewed through subsequent marriages of consanguineally related individuals, but a cohesion that can likewise wane if such marriages do not occur. The relatively large number of consanguineal loops that relate Keböwä's supporters to his rival in this fight can be viewed as representing a set of earlier affinal exchanges that have been allowed to lapse because they were not consolidated through additional and more recent marriage exchanges. Conversely, both the number of relationship loops and the high values of Fg tying Mohesiwä's supporters closely to him represent a historical development of affinal exchanges that were continuously bolstered and reinforced through additional marriage exchanges in each generation, marriages that involved already-related individuals.

Interestingly enough, both affinal ties and consanguineal relationships are clearly involved as bases of recruitment-Keböwä/Yoinakuwä supporters seeming to rely on the former mode and Mohesiwä's supporters on the latter. One logical conclusion from this set of facts is that alliance and descent are not opposite or mutually exclusive alternatives to the development of social cohesion, but complementary and, in a temporal sense, additive. It might very well be that some societies exploit and utilize one or the other of these mechanisms in ordering the calculus of social life (Barnes 1962), but it seems apparent that the Yanomamö take advantage of both, and do so within the same village.

One feature of the fight that puzzled the senior author at the time the fight occurred was the intensity with which Keböwä reacted to the beating of Sinabimi-a woman who was married to his brother. To be sure, kin selection predicts that brothers would aid each other in this situation, but Keböwä, in effect, took the initiative to escalate the fight to axes. A "benefit" that ultimately accrued to Keböwä, unpredictable at the time of the fight, was the untimely death, two years later, of Yoinakuwä, the brother that he aided: Keböwä married Yoinakuwä's widow, Sinabimi, and assumed the responsibilities of rearing their family. That family consisted of six children, three of whom were nubile young girls whose future marriage and reproductive careers are now largely under Keböwä's control. Since marrying Sinabimi, Keböwä has sired one child, a son born in 1973, as of the last census update made by the senior author in 1974-75.

Conclusion

It is clear that closeness of relationship measured genealogically serves as a mediator of interpersonal behavior in the conflict situation under discussion. Members of the group that supported Mohesiwä in the fight are not a random set of individuals from the village and, to the extent that the genealogical data are an accurate reflection of biological relatedness among them, their participation in the fight can legitimately be seen as having considerable relevance to kin selection theory. Contrary to the extravagant claim made by Marshall Sahlins ". . . no system of human kinship relations is organized in accord with the genetic coefficients of relationship as known to sociobiologists" (1976a:57), there is good reason to believe that additional research will reveal similar patterns in other human societies. We believe that the evidence-and the theory-are of sufficient import at this point that new research projects should be developed to document the extent to which human kinship behavior is consistent with predictions based on inclusive fitness theory. The primitive world is, after all, on the wane and unless the research is done now, only questions will remain.

It is also clear that relatedness alone cannot account for all the bonds of attraction or tactics of recruitment in events such as the one analyzed here, and that affinity or alliance likewise operates to build coalitions. Apart from all considerations of theory in evolutionary biology, the collection of data of the kind described in this paper would, we believe, likewise contribute to a fuller understanding of systems of human marriage and mating, and by extension, to systems of social organization that can be appreciated in both quantitative as well as qualitative terms. We can make the invidious distinctions about the relative attractiveness of different styles of social science research later, but the empirical work must be done now.

Footnotes

*1 We wish to express our gratitude to the officers of the Harry Frank Guggenheim Foundation who generously supported the senior author's research efforts during 1974-7. We are also indebted to the National Science Foundation (Grant No. SOC75-14262) and the National Institute of Mental Health (Grant No. MH26008) for additional support during the same period of time.

*2 It may be that axes have replaced a particular kind of hard, palm-wood dueling weapon called himo, a type of double-edged, heavy, sword-like instrument. Himo clubs can still be seen in many villages, although the senior author had never seen a fight in which they were used.

*3 The film, The Ax Fight (Asch and Chagnon 1975), can be rented or purchased through: (1) Documentary Educational Resources, 24 Dane Street, Somerville, Massachusetts 02143 [sic. The address is currently: 101 Morse Street, Watertown, MA 02172; telephone numbers: (800) 569-6621 and (617) 926-0491; fax: (617) 926-9519; World Wide Web homepage address: http://der.org/docued ], or (2) Psychological Cinema Register, 17 Willard Building, The Pennsylvania State University, University Park, Pennsylvania 16802.

*4 The senior author of the paper was standing a few feet from Tourawä when Keböwä hit him with an ax. The filmmaker, Timothy Asch, was some thirty yards away filming the fight with a 16mm motion picture camera, unable to move closer with the bulky equipment without missing significant portions of the fight.

*5 We have not calculated the coefficients of relationship beyond common ancestors of the fourth ascending generation. Thus, increments of relatedness through more remote ancestors are not included in the Fg values.

*6 We are using the terms "visitors" and "hosts" to distinguish the two groups of fighters from each other in a simple way. It must be remembered, however, that some local people sided with and supported the visitors. See Table.8.4a and Table.8.4b for identities by village.

Bibliography

Alexander, R. D.
-------1974. The Evolution of Social Behavior. Annual Review of Ecology and Systematics 5:325-383.
-------1977b. Evolution, Human Behavior, and Determinism. Proceedings of the Biennial Meeting of the Philadelphia Science Association (1976). Vol. 2:3-21.
Chagnon, Napoleon A.
-------1966. Yanomamö Warfare, Social Organization and Marriage Alliances. Doctoral dissertation, University of Michigan. Ann Arbor: University Microfilms.
-------1967. Yanomamö: The Fierce People. Natural History 76:22-31.
-------1968a. Yanomamö: The Fierce People. New York: Holt, Rinehart & Winston
-------1968c. The Culture-ecology of Shifting (Pioneering) Cultivation Among the Yanomamö Indians. Proceedings, III International Congress of Anthropological and Ethnological Sciences, Tokyo, 1968. Vol. 3:249-255.
-------1974. Studying the Yanomamö. New York, Holt Rinehart and Winston.
-------1975. Genealogy, Solidarity, and Relatedness: Limits to Local Group Size and Patterns of Fissioning in an Expanding Population. In Yearbook of Physical Anthropology, 1975. Vol. 19:95-110. Washington: American Association of Physical Anthropologists.
-------1979. Chapter 4: Mate Competition, Favoring Close Kin and Village Fissioning Among the Yanomamö Indians. In Evolutionary Biology and Human Social Behavior: An Anthropological Perspective. N. A. Chagnon and William Irons, eds. Pp. 86-131. North Scituate, MA: Duxbury Press.
Fortes, M.
-------1959. Descent, Filiation and Affinity: A Rejoinder to Dr. Leach. (2 parts) Man 59 (309):193-197 and 59 (331):206-212.
-------1969. Kinship and the Social Order. Chicago: Aldine.
Fox, Robin.
-------1979. Kinship Categories as Natural Categories. In Evolutionary Biology and Human Social Behavior: An Anthropological Perspective. N. A. Chagnon and William Irons, eds. Pp. 132-144. North Scituate, MA: Duxbury Press.
Hamilton, W. D.
-------1964. The Genetical Evolution of Social Behaviour, Parts I and II. Journal of Theoretical Biology 7:1-52.
Sahlins, M. D.
-------1968. On the Sociology of Primitive Exchange. In The Relevance of Models for Social Anthropology. M. Banton, ed. Pp. 139-227. London: Tavistock. [Originally published in 1965.]
-------1976a. The Use and Abuse of Biology: An Anthropological Critique of Sociobiology. Ann Arbor: University of Michigan Press.
Trivers, R. L.
-------1974. Parent-Offspring Conflict. American Zoologist 14:249-264.