Causes, Consequences, and Kin Bias of Human Group Fissions
Fissions of human communities are monumental occasions with consequences for cultural and genetic variation and divergence through time by means of serial founder effects. An ethnographic review shows that most human group fissions are fueled primarily by internal political conflict and secondarily by resource scarcity. As found for other social animals, human fissions lead to subgroups that have higher levels of relatedness as compared with the original community because of kin-biased assortment known as the lineal effect. Fission processes that increase the average relatedness of subgroups are important because relatedness governs how strongly kin/group selection favors social behaviors such as warfare, peacekeeping, and other forms of collection action. However, random individual assortment is not an appropriate null model for evaluating lineage assortment because nuclear families and extended households are expected to remain together, which in and of itself forces higher relatedness in smaller subgroups. We develop a lineage assortment index where low values represent subgroups with coefficients of relatedness near those expected if nuclear and extended households had chosen to associate into random groupings. Two fissions of Ache villages (Paraguay) are examples of this type of fission with a low lineage assortment index not significantly different from zero as evaluated with controlled simulations. On the other extreme, a lineage assortment index near unity represents a lineal fission that maximizes the relatedness of subgroups such as the perfect split of a lineage into sublineages. A fission of Piaroa (Venezuela) fits this scenario. While previous discussions of fission have emphasized similarities among human studies and even other social mammals, we highlight the full range of potential kin bias in the formation of new communities.
KeywordsGroup splits Lineal fissions Human kinship Genealogical relatedness Cross-cultural comparison
It is difficult to overestimate the importance of group fissions for human history (Carneiro 1987). Fission events, including migration and various types of subgroup formation, have contributed to the colonization of humans around the globe and the phylogenetic “speciation” of the world’s more than 7,000 languages (Lewis et al. 2013) and cultures, analogous to the founder effect in biological speciation (Melnick 1987; Templeton 1980). Comparative ethnographic information on fission events may help shed light on social dynamics and micro-evolutionary forces that increase cultural, linguistic, and genetic variation through time (Cavalli-Sforza et al. 1988; Freedman 1984; Neel and Salzano 1966).
In social mammals, fissions often occur along divisions in matrilineal kinship, as in some rodents (Hoogland 1995; Waterman 2002), bats (Metheny et al. 2008), baboons (Altmann et al. 1985; Van Horn et al. 2007), macaques (Chepko-Sade and Olivier 1979; Melnick and Kidd 1983), and elephants (Archie et al. 2006). In general, when multimale/multifemale groups split, maternal kinship ties affect individual decisions on which subgroup to pertain. In species where paternal kin form social bonds and fathers are recognizable (e.g., gorillas: Nsubuga et al. 2008), patrilineal kinship may also influence group membership.
Humans are unusual, indeed unique, because we have large social groupings combined with pair-bonding, recognition of social fathers, and the institution of marriage (Alexander 1990; Chapais 2008; Rodseth et al. 1991). Human fissions have kin biases as in other social species, tending either toward patrilineality or matrilineality, but with the added complexity of embedded nuclear and extended family households that generally remain cohesive around marriage bonds even during fission events. In this study we found only one case of divorce that co-occurred with fission out of the several hundred marriages that could have disintegrated. The cohesion of familial units works toward kin-biased association and the “lineal effect” (Neel 1967; Neel and Salzano 1966) only to a point. There is also a corrosive force working against the lineal effect of having one of the spouses remain in a lineage that is generally not their natal one given that they have married into another family owing to inbreeding avoidance and exogamous marriage rules (i.e., marriages often mix lineages and dilute lineal effects). A perfect lineal fission, perhaps possible in other social species, is impossible for humans given that marriage ties generally cross-cut lineages.
Segmentary lineages (Evans-Pritchard 1940), if defined in a loose sense (contra Sahlins 1961) as human kinship systems in which descendants of close kin stand together in conflict against more distant kin, may be a generalizing principle for understanding human fissions. Evolutionary anthropologists have led research in this area from an inclusive fitness perspective (Hamilton 1964), finding kin biases in residential associations during and after conflict and fission events, perhaps most famously in the Yanomamö film The Ax Fight (Chagnon and Bugos 1979; also see Alvard 2009). Indeed, previous discussions of fission have emphasized similarities in lineal fissions among the Yanomamö (Chagnon 1975, 1976, 1979), Semai (Fix 1975), Hutterites (Olsen 1976, 1987), and a few other human studies both matrilineal and patrilineal (Chepko-Sade and Olivier 1979). Comparisons of humans and other social animals have noted an increase in coefficients of relatedness (Wright 1922) during fission from a larger mother community to smaller daughter communities (e.g., Chagnon 1976; Chepko-Sade and Olivier 1979). However, perfect congruence across species would be somewhat surprising given the unique human social structure of embedded family units within larger communities discussed above.
Summary of 11 fission studies (of 49 total) where information is available on the population size at fission and the fraction of the population that splits (split fraction)
Size at fission
Political conflict; adultery trigger
Early and Peters 2000
Competition over hunting
Disputes over women and leadership
Land and labor needs; factionalism
Disputes over women and leadership
Jealousy and in-fighting
Internal conflict between 2 leaders
Hostiles vs. Friendlies
Summary of five fission events with coefficients of relatedness (r) before and after fission, relatedness expected if households remain cohesive and randomly assort (rrandom family), and a maximum relatedness possible if a fission occurred perfectly along lineal divisions (rmax with households remaining cohesive and population that approximates the number of observed emigrants)
Friendlies vs. Hostiles
Splits into 3
0.186, 0.191, 0.257
The most famous, well-studied fission in all of anthropology is the 1906 split of the Hopi stronghold at Old Orayvi. Extensive research on this event includes work by Leslie White and his field school, Eggan (1950), Whiteley (1988, 2008) and Levy (1992). The actual breakup of Old Orayvi was a complex process that eventually created three new villages, enlarged a fourth, and mostly disintegrated Orayvi itself (Whiteley 2008). For simplicity, we analyzed the primary and roughly equal division between “Friendlies,” those who wanted to go along with the White Man’s way, and “Hostiles,” those who did not. We used the composite census compiled by Whiteley (2008) using government records, Titiev’s (1972) household census, and Leslie White’s genealogy to develop estimates of relatedness for the 1,048 residents by including 98 deceased ancestors (although most are females given that only matrilineal genealogies are available from White’s work). Living fathers were identifiable for younger generations using census information, but fathers were not known for many of the earlier ancestors.
Less well known are the Piaroa of Venezuela, who traditionally lived in circular multifamily huts (Kaplan 1975), and the Waimiri-Atroari of Brazil, who live in large circular villages (Silva 2009). Fairly complete genealogies (depth of around four generations), including deceased male and female ancestors, are available for these swidden horticulturalists.
We contribute to these data two fissions in a village of 400–500 Ache (Paraguay) in the years 1990 and 2000. The Ache were traditionally nomadic hunter-gatherers, but since contact they have conglomerated into sedentary horticultural communities. In the first few weeks of the 2000 fission, only a few families followed a strong political leader to a new community, but within a few months about a third of the original settlement had packed up and left. This was a power struggle between the resident leader, who was supported by the priest, and the emigrating leader, who was a church pastor with many followers. Interestingly, a number of adult sibling sets were split by the fission, with some leaving and others staying behind. Economic factors were also involved. The resident leader was involved in a corrupt timber harvest with the priest, ran the colony as a dictatorship, and reaped most of the benefits from collectively held tribal resources while excluding the emigrating leader and his followers. Some of the emigrants described the split as traditionalist conservationists versus greedy loggers. Some that stayed behind simply did not want to incur the cost of moving. A previous split in 1990 of the same community also led to a splinter colony of Ache. It was a longer process, with people leaving over a period of several years before the two communities stabilized. This event was less overtly political. A group of people simply did not feel comfortable in the situation and had the option to move out into the bush and live more as foragers. Many of them were low status and had little to lose by setting up their own community. The Ache genealogy is complete and extends back to the late nineteenth century (Hill and Hurtado 1996).
We used controlled simulations in Microsoft Excel to answer the following questions: (1) What range of relatedness is expected if subgroups form by random individual assortment? (2) What range of relatedness is expected if subgroups are comprised of cohesive households that randomly assort? and (3) What is the type of fission that maximizes the relatedness of subgroups (rmax)? The first two questions were answered straightforwardly by running 1,000 simulations that gave probabilities of each individual or family unit dispersing in fissions that split their group roughly in half (Hopi), thirds (Waimiri-Atroari, Piaroa, Ache in 2000), or fifths (Ache in 1990; Table 2). The third question, how to find rmax, required searching through large numbers of possible combinations of splits under the constraints that households remain cohesive and that the size of the subgroup approximates the observed number of emigrants. We discovered that combinations of large lineages always gave a higher rmax than combinations of smaller lineages, even when smaller lineages themselves were more closely related—via reciprocal marriage exchange, for example. For the 2000 Ache split, a combination of the three largest patrilineages (all larger than the largest matrilineage) yielded rmax, whereas the 1990 split took the largest patrilineage. For the Hopi, six matrilineages resulted in rmax (patrilineages are small because the genealogies are matrilineal). For the Piaroa, the observed split of a lineage into sublineages was the one that yielded rmax. Household designations were not available so we did not simulate random household movement. Finally, for the Waimiri-Atroari, rmax resulted from the single largest patrilineage splitting off. Calculations of relatedness, size of lineages, and genealogical error checking were conducted using Ed Hagen’s Descent software (http://itb.biologie.hu-berlin.de/~hagen/Descent).
Population sizes at fission (available from 11 different societies) vary widely given the range of socioecological contexts (Table 1). As expected, fission tends to occur in villages that are larger than average, as seen in societies that have a sample of different fissioning events (Hutterites, Kayapó, and Yanomamö). The percent of the original population that splinters off ranges in the overall sample from 26 to 49%, with an average of 35% (mean of means).
Village fissioning has long been attributed to stresses related to competition between ambitious leaders (e.g., Ndembu: Turner 1957), frequent quarrelling (e.g., Siuai: Oliver 1955), and the “roughly geometric increase in irritation” as populations become more dense (Tsembaga: Rappaport 1968). The counterbalances to fissive factors, and factors that promote fusion, probably include reciprocal marriage exchange, strong leaders, threat of warfare, and other economies of scale, but we do not address these forces here. Reasons for fissioning are difficult to discern, mostly because the same event can be said to have multiple motives, and different informants and anthropologists often give different reasons (Carneiro 1987). Our cross-cultural assessment shows that the most common purported reason, including multiple answers, is internal political conflict (29 of 37, or 81%), and secondarily, resource scarcity (8 of 37, or 19%), although the two are often conflated, and resource scarcity likely contributes to the internal strife. Conflicts over sex (e.g., Kwaio: Keesing 1978) or witchcraft (e.g., Apinayé: Matta 1970) are the more immediate triggers precipitating some fissions (Carneiro 1987).
Lineal fissions are purported to have occurred in 26 studies, with statements such as “fission along kin lines” or “patrilineages stick together”; in only one case is potential dissent mentioned: “fission of lineages does not always follow strict genealogical lines when the genealogical factor is in conflict with the corporate factor” (Chuuk: Goodenough 1951). The lineal effect is measured in terms of an increase in coefficients of relatedness for the Yanomamö (Chagnon 1975, 1979), Semai (Fix 1975), and Hutterites (Olsen 1976), to which we can now add Ache, Hopi, Piaroa, and Waimiri-Atroari (Table 2). However, as discussed above, comparisons against a null model of random family movement are necessary to test for lineal fissions.
On the other extreme is the Piaroa case mentioned above, with perfect sublineage formation in this small group where indices are unity. All same-sex adult siblings remained together and siblings-in-law separated (Kaplan 1975). The Hopi case of Old Orayvi represents a much larger village split with a statistically significant lineage assortment index of 0.50. Old Orayvi then is an intermediate example with lineage assortment in terms of relatedness that is exactly half the maximum possible relatedness.
Our comparative results suggest that most fissions are purportedly driven by internal political conflict. Comparative findings also appear to imply a consistent kin bias in residential assortment driven by lineal fissions. However, the small number of well-documented fissions represents the full spectrum of possible outcomes from limited lineal fission (Ache and Waimiri-Atroari) to perfect lineal fission (Piaroa), with the Hopi at Orayvi being intermediate. It seems reasonable to assume that individual residential decisions often involve considerations not just of consanguineal kin but also spouses and affines, as well as potential reproductive and economic opportunities.
Although previous studies purport to show similarities among different human groups and even other social species regarding kin biases in residence patterns following fission events, we raise several cautionary notes. First, human social structure is unique, with embedded human families within larger lineages and communities. This unique human social structure is expected to have an impact on fissioning because of the cohesion of human nuclear families and extended families living together in households. Second, although cohesion of immediate families is ubiquitous, the degree of lineage cohesion is expected to vary considerably across different human systems—weakest in nomadic hunter-gatherers and strongest in robust segmentary lineage systems. Third, previous studies generally report lineal fissions, either anecdotally or by measuring an increase in relatedness, but this assumes random assortment as a null model and does not address the cohesion of the lineage directly. For these reasons, we find little evidence for much congruence across fission studies.
Some have argued that genetic, cultural, and linguistic micro-evolution may have been accelerated by the lineal effect in humans that strongly affect FST values, potential for group selection, rates of divergence, and effective population sizes (Fix 1999; Freedman 1984; Neel and Salzano 1966; Smouse et al. 1981). This appears to be true, at least to a point, in that every fission studied thus far shows an increase in the coefficients of relatedness of the subgroups. In some cases, however, the lineal effect is tempered in that the cause is only immediate families sticking together and not the genealogical cohesion of wider or deeper lineages as indicated by small lineage assortment indices. The fluid fission-fusion dynamics of many contemporary hunter-gatherers (Hill et al. 2011; Kelly 2013; Marlowe 2005), arguably the best representatives of all but the past 10,000 years or so of human evolution, would argue against strong lineal fissioning, particularly in nomadic societies often lacking strong lineages in the first place (although there may be exceptions in the purportedly strong patriclans of Australian Aborigines and in contexts of intense warfare, for example). The emergence of larger, more sedentary farming and herding communities may have led to stronger segmentary lineage systems, but large, dense, sedentary populations are themselves expected to dampen the lineal effect in the same way that they weaken kin-structured migration and genetic drift (Fix 2004).
Until research provides more examples of substantial lineal fissions like the Piaroa fission (and less like the Ache or Waimiri-Atroari fissions), we cannot fully evaluate the potential of the lineal effect on human micro-evolutionary forces. More empirical results are needed to generate a frequency distribution of the kin and lineage assortment indices across human societies and even over time in the same society. Genetic studies also provide insight. For example, Hunley and colleagues (2008) found that Yanomamö splits do indeed appear to be lineal, to a point, with male and female relatives generally staying together, but village movement, fusions, and frequent migration between villages likely produced more Y-chromosome homogeneity as contrasted with more mtDNA heterogeneity, contrary to expectations.
Understanding social dynamics of fissions is greatly facilitated by the availability of complete genealogies combined with census information. It led to the discovery that our measure of the lineage assortment index on a continuous scale does not demonstrate congruence among different human groups but instead highlights the full range of variation from random assortment of families to perfect lineage splitting. The assortment indices are easily extendable to other social species, although the restriction of cohesive families would need to be removed, which would extend rmax considerably. Many other important questions and projects of this nature require primary genealogical and residence data. It is a tragedy of modern anthropology that such data are often not readily available in sharable databases.
This paper benefited from help and conversations with Peter Whiteley, Greg Blomquist, Rob Boyd, Craig Palmer, and Stephen Beckerman. Financial support was provided by a National Geographic Society Research and Exploration grant (#9165-12).
- Alexander, R. D. (1990). How humans evolved: Reflections on the uniquely unique species. Museum of Zoology (Special Publication No. 1). Ann Arbor: University of Michigan.Google Scholar
- Århem, K. (1981). Makuna social organization: A study in descent, alliance, and the formation of corporate groups in the north-western Amazon. Stockholm: Almqvist & Wiksell.Google Scholar
- Carneiro, R. (1987). Village splitting as a function of population size. In L. Donald (Ed.), Themes in ethnology and culture history. Meerut: Archana Publications.Google Scholar
- Chagnon, N. A. (1975). Genealogy, solidarity, and relatedness: limits to local group size and patterns of fissioning in an expanding population. Yearbook of Physical Anthropology, 19, 95–110.Google Scholar
- Chagnon, N. A. (1976). Fission in an Amazonian tribe. New York Academy of Sciences, 16(1), 14–18.Google Scholar
- Chagnon, N. A. (1979). Mate competition, favoring close kin, and village fissioning among the Yanomamö Indians. In N. A. Chagnon & W. A. Irons (Eds.), Evolutionary biology and human behavior. North Scituate: Duxbury Press.Google Scholar
- Chagnon, N. A., & Bugos, P. E. (1979). Kin selection and conflict: An analysis of a Yanomamo ax fight. In N. A. Chagnon & W. A. Irons (Eds.), Evolutionary biology and human behavior (pp. 213–288). North Scituate: Duxbury Press.Google Scholar
- Chapais, B. (2008). Primeval kinship: How pair bonding gave birth to human society. Cambridge: Harvard University Press.Google Scholar
- Dumont, J.-P. (1978). The headman and I: Ambiguity and ambivalence in the fieldworking experience. Austin: University of Texas Press.Google Scholar
- Early, J. D., & Peters, J. F. (2000). The Xilixana Yanomami of the Amazon: History, social structure, and population dynamics. Gainesville: University Press of Florida.Google Scholar
- Eggan, F. (1950). Social organization of the Western Pueblos. Chicago: University of Chicago Press.Google Scholar
- Evans-Pritchard, E. E. (1940). The Nuer: A description of the modes of livelihood and political institutions of a Nilotic people. Oxford: Clarendon.Google Scholar
- Fix, A. G. (1999). Migration and colonization in human microevolution. Cambridge: Cambridge University Press.Google Scholar
- Goodenough, W. H. (1951). Property, kin, and community on Truk. New Haven: Yale University Publications.Google Scholar
- Hill, K. R., & Hurtado, A. M. (1996). Ache life history: The ecology and demography of a foraging people. New York: Aldine de Gruyter.Google Scholar
- Hoogland, J. L. (1995). The black-tailed prairie dog: Social life of a burrowing mammal. Chicago: University of Chicago Press.Google Scholar
- Kaplan, J. O. (1975). The Piaroa: A people of the Orinoco Basin. Oxford: Clarendon.Google Scholar
- Keesing, R. M. (1978). Elota’s story: The life and times of a Solomon Islands big man. New York: Henry Holt & Company.Google Scholar
- Levy, J. (1992). Orayvi revisited: social stratification in an “egalitarian” society. Santa Fe: School of American Research Press.Google Scholar
- Lewis, M. P., Simons, G. F., & Fennig, C. D. (Eds) (2013). Ethnologue: Languages of the world, 17th edition. Dallas, Texas: SIL International. Online version: http://www.ethnologue.com.
- Matta, R. (1970). A divided world: Social structure of the Apinayé. PhD dissertation. Harvard University.Google Scholar
- Maybury-Lewis, D. H. P. (1967). Akwa-Shavante society. Oxford: Clarendon.Google Scholar
- Melatti, J. C. (1970). O sistema social Craô. PhD dissertation. Sao Paulo: Universidade de Sao Paulo.Google Scholar
- Melnick, D. J. (1987). The genetic consequences of primate social organization: a review of macaques, baboons and vervet monkeys. Genetics, 73, 117–135.Google Scholar
- Neel, J. V. (1967). The genetic structure of primitive human populations. Journal of Human Genetics, 12, 1–16.Google Scholar
- Neel, J. V., & Salzano, F. M. (1966). Further studies on the Xavante Indians: some hypotheses-generalizations resulting from these studies. American Journal of Human Genetics, 19(4), 554–574.Google Scholar
- O’Leary, T. J., Levinson, D., Hays, T. E., Hockings, P., Bennett, L. A., et al. (Eds.). (1994). Encyclopedia of world cultures. New York: Hall Publishing.Google Scholar
- Olsen, C. L. (1976) The demography of new colony formation in a human isolate: An analysis and history. PhD dissertation, University of Michigan.Google Scholar
- Rappaport, R. A. (1968). Pigs for the ancestors: Ritual in the ecology of a New Guinea people. New Haven: Yale University Press.Google Scholar
- Silva, M. F. (2009). Romance de primas e primos: Uma etnografia do parentesco Waimiri-Atroari. Manaus: Valer Editora.Google Scholar
- Smouse, P. E., Vitzthum, V. J., & Neel, J. V. (1981). The impact of random and lineal fission on the genetic divergence of small human groups: a case study among the Yanomama. Genetics, 98, 179–197.Google Scholar
- Templeton, A. R. (1980). The theory of speciation via the founder principle. Genetics, 94, 1011–1038.Google Scholar
- Titiev, M. (1972). The Hopi Indians of Old Oraibi: Change and continuity. Ann Arbor: University of Michigan Press.Google Scholar
- Turnbull, C. M. (1965) The Mbuti Pygmies: an ethnographic survey. Anthropological papers of the American Museum of Natural History, 50, part 3. New York: American Museum of Natural History.Google Scholar
- Turner, V. W. (1957). Schism and continuity in an African society. Manchester: Manchester University Press.Google Scholar
- Verswijver, G. (1992). The club-fighters of the Amazon: Warfare among the Kaiapo Indians of Central Brazil. Gent: Rijksuniversiteit Te Gent.Google Scholar
- Whiteley, P. M. (1988). Deliberate acts: Changing Hopi culture through the Oraibi split. Tucson: University of Arizona Press.Google Scholar
- Whiteley, P. M. (2008). The Orayvi split: A Hopi transformation. Anthropological Papers of the American Museum of Natural History. Number 87. New York: American Museum of Natural History.Google Scholar