Kinship, Culture, and an evolutionary perspective on the structure of cooperative big game hunting groups in Indonesia1

Work was conducted among traditional, subsistence whale hunters in Lamalera, Indonesia in order to test if kinship or lineage membership is more important for explaining the organization of cooperative hunting parties ranging in size from 8-14 men. Crew identifications were collected for all 853 hunts that occurred between May 3 and August 5, 1999. Lineage identity and genetic relatedness were determined for a sample of 189 hunters. Results of matrix regression show that kinship explains little of the hunters’ affiliations independent of lineage identity. Crews are much more related amongst themselves than expected by chance. This is due, however, to the correlation between lineage membership and kinship. Lineage members are much more likely to affiliate in crews, but beyond r = 0.5 kin are just as likely not to affiliate. The results are discussed vis-à-vis the evolution of cooperation and group identity. Alvard, M. Kinship, lineage, and cooperative hunting Page: 3 Introduction People frequently form cooperative groups in order to realize the benefits of collective action. While humans are not the only species that forms social groups, we are unique in the degree to which we regularly rely on the help of conspecifics to satisfy basic needs such as subsistence, defense, and offspring care (Hill 2002). In simple foraging groups, people help one another care for children (Ivey 2000), acquire difficult to obtain resources (Alvard and Nolin 2002), share food (Kaplan and Hill 1985), and fight other groups (Chagnon and Bugos 1979). In complex contemporary society instances are even easier to come by. Examples include unions, political parties, nation states, firms, college fraternities, sports teams, and universities. Kin selection (Maynard Smith 1964) and inclusive fitness theory (Hamilton 1964a, 1964b) offers good evolutionary explanations for why cooperation among relatives should be common. Kin selection is the process by which traits are favored because of their beneficial effects on the survival of relatives (Grafen 1984). Because kin share genes due to common descent, behaviors that increase the reproductive success of relatives can also increase the future representation of ego’s genes. Thus, kin selection theory predicts that, all other things equal, individuals will be more likely to favor kin than non kin, and close kin than more distant kin. Hamilton’s well-known rule predicts altruism can evolve if the following equation obtains rB C > 0, where r = the coefficient of relatedness between the actor and the recipient, B = the fitness benefit to the recipient, C = the fitness cost to the actor. The coefficient of relatedness is defined as the probability that two individuals share a copy of an allele through common descent (Wright 1922). For example, between sibs r = 0.5, between grandparent and grandchild r = 0.25, and between cousins r = 0.125. Thus, nepotism will evolve if the recipient of the favor is sufficiently related, the benefit is sufficiently great, or the cost sufficiently low. Kinship has been shown to be an important social organizing principle across a wide variety of taxa (Dugatkin 1997), and especially so in the social insects (Bourke 1997). For example, Sherman’s work on Belding ground squirrel alarms calls is well-known (Sherman 1977). Among humans, some of the best work shows that people are less likely to kill kin than non-kin (Daly and Wilson 1988; Johnson and Johnson 1991). Chagnon and Bugos’ (1979) analysis of a Yanomamo axe fight was among the earliest analysis to use kin selection theory to examine human social behavior. As predicted, combatants on each side were more related to one another than expected by chance. Within fields of study that take an evolutionary approach to human behavior, the role of kinship in explaining cooperative behavior within preindustrial societies is now taken for granted to some extent (Alexander 1987; Chagnon 1979, 1980; Hamilton 1975; see review by Voland 1998:363), although as I shall show, the extent of its importance is ambiguous (Brown 1991; Jones 2000; Richerson and Boyd 1998) Kinship has also long been argued by cultural anthropologists to be the primary organizing principle in tribal societies (Kuper 1982, 1996). In spite of what appears to a common ground, however, there has been little work over the last twenty years to integrate the two approaches. Indeed, kinship within the field has been “denaturalized” by many cultural anthropologists in the late twentieth century (Collier and Yanagisako 1987; Schneider 1984; for a review see Peletz 1995). A standard critique of kin selection theory applied to humans points to the incongruity between kin -genetically defined, and kin-culturally defined, to put it simply. This point was made the strongest by Sahlins (1976:58) when he stated “Kinship is a unique characteristic of human societies, distinguishable precisely by its freedom from natural relationships.” The hyperbole of this statement seems obvious. But while it is impossible to maintain the position that cultural kinship has nothing to do with genetic kinship, it is equally difficult to deny that people commonly organize themselves in ways that do not correspond to coefficients of relatedness. Sahlins (1976:26) wrote, “...local kinship networks...will comprise a determinate and biased proportion of any person’s genealogical universe.” 2 This is most apparent in Alvard, M. Kinship, lineage, and cooperative hunting Page: 4 systems of unilineal descent. Genetic kinship does not distinguish between individuals equivalently related. People who follow norms of unilineal decent, however, define as kin only those persons who share common decent through either the male or female parent. In such systems, two people who are each equally related to Ego genetically may be defined differently according to kinship norms -one as in-group member, the other as an outgroup member. In Sahlins’ words “...even so the son of a man’s brother may be one of the clan of the ancestor’s descendants while the son of his sister is an outsider and perhaps an enemy” (1976:12). In a patrilineal system the coefficient of relatedness for Ego and his mother’s brother’s son, and his father’s brother’s son (both called cousins in English) is 0.125. While the latter shares Ego’s lineage identity, the former does not. The difference is not simply a semantic one, as I will show below. The social relationships between ego and these individuals differ though the genetic relationship does not (see Figure 1). A kinship system based on kin selection, all other things being equal, predicts a bilateral descent or kindred system (implied by Murdock 1949:57). Kindreds are ego-based and consist of the group of all near relatives. In contrast to a unilineal system, in a kindred system no distinction is made between relatedness reckoned through one or other of the parents’ lineages. Cross-cultural work notes the relative rarity of such systems; only 36% of the 857 societies in Murdock’s ethnographic atlas have bilateral descent systems. Unilineal systems are much more common: 47% in Murdock’s sample practice patrilineal descent, 14% practice matrilineal descent, and 3% claim a double descent system (Murdock 1967). Structural-functional anthropologists of the mid 20 century offered a number of hypotheses to explain why groups might develop unilineal descent, though they did so without knowledge of kin selection theory, of course. The most compelling argument is that group members organized by unilineal kinship ties have the advantage of unambiguous group identity (Murdock 1949:60-61). This is because only full siblings share an identical network of kin since only sibs share parents. Other classes of relatives, like cousins for example, share only a portion of their kin in common. While cousins share a set of grandparents, they each also share a second set of grandparents with a different set of cousins (unless their parents were cousins themselves, of course). Figure 1. The relationships between ego and his two cousins marked A and B. The coefficients of relatedness beween Ego and both A and B are 0.125. In a patrilineal system B shares Ego’s lineage identity, A does not. Lineage 1