COMPETITION & WARFARE 

Intro

Most analyses of human conflict stress non-ecological causes:  psychological, social-structural, political, cultural, or ideological

However, there is a substantial literature in ecological anthropology on this subject, though relatively little agreement within it

In these lecture notes, I cover the following topics:

1) definition & prevalence of warfare
2) evolutionary theory of conflict
3) ecological theory of competition
4) ecological analyses of human warfare

Definition and Prevalence

Many possible definitions of warfare, but one common definition used by anthropologists is:  organized armed aggression between members of politically autonomous social groups

This avoids calling kin feuds, gang fights, etc "war" since such groups are not politically autonomous; it also avoids arbitrarily restricting term to societies with professional military & centralized political control

[Note that this definition also restricts warfare to human species, since other animals don't have what we would recognize as "politically autonomous groups" or military technology; in fact, organized inter-group aggression rare among other species (except social insects), but it has been described for a number of group-living primates (including chimpanzees) and carnivores (wolves, lions, hyenas)]

Using above definition, how common is warfare?

Data from a cross-cultural sample of 87 societies representative of various geographical regions and types of sociopolitical organization [see graph] reveals that:

• 61% of societies in this sample engaged in frequent warfare (at least yearly)
• 21% engaged in war less often (at least 1/generation)
• 18% of societies rarely/never went to war (only 8% of truly autonomous societies)
• nearly 40% had high mortality in war (>one third of combatants died)

Evolutionary Theory of Aggressive Contests

One commonly encounters the assertion that intentional killing of conspecifics (members of the same species) is an aberration of the human species, and in any case unlikely to be explainable in adaptive terms

This certainly seems logical if one believes that natural selection favors traits "for the good of the species" -- it would then seem that warfare (and other forms of intentional homicide) must represent an unfortunate consequence of human freedom from the dictates of evolution

But there are serious problems with this view

The first is empirical:  even if other species don't have "warfare" (since don't have culturally-transmitted technology or political organization), they do have organized inter-group aggression, and this is sometimes lethal:  

• one long-term study of wolves (Mech 1970) found that 25% of all deaths were due to intra-specific killing

• data on common chimpanzees indicates that organized group killings account for at least 18% of deaths in the Gombe study population (note that common chimps, Pan troglodytes, along with the bonobos or "pygmy chimps," Pan paniscus, are our closest living relatives)

Another problem is theoretical -- the naive assumption that natural selection wouldn't favor tendencies towards lethal conspecific violence since this harms the species

Instead, standard evolutionary theory predicts that conflict, including lethal aggression, will be favored by natural selection if it increases the average fitness of aggressors, regardless of its effect on average fitness in the population

To see why, consider a simple model of the "Hawk-Dove" contest (which will also serve to introduce you to evolutionary game theory models, useful for analyzing for all sorts of social interactions, not just conflict-based ones)

Suppose there are two strategies used in conflicts over resource:  

• Hawk always contests the resource, and if necessary fights for it, giving up only if seriously injured

• Dove never fights, instead either sharing (with other non-aggressors) or retreating (if threatened)

• these tactics are heritable in a very simple fashion:  a Hawk gives birth only to other Hawks, a Dove only to other Doves (this assumption can be relaxed, but we'll keep the exposition simple by adopting it for now)

Game theory models usually define the payoffs from various tactics in terms of mathematical variables, but to make things a little more concrete let's suppose the resource being contested is worth 100 "fitness points"

When two Doves face each other, they divide the resource equally, so payoff to each = 50 "fitness points"

In a population of Doves, all goes well -- until an aggressive Hawk mutant arises

Hawk wins every contest with Doves, gaining sole access to the resource (e.g., a patch of food) in every contest it engages in, and hence averaging 100 pts per encounter; since resources are converted into increased reproduction (by whatever conversion metric applies to our hypothetical 100 fitness pts), Hawks proliferate -- natural selection at work

You might suppose that Hawks will eventually replace Doves, since they win every resource contest, but this intuition is wrong

How so?  It has to do with the perils of success:  as Hawks become more numerous, they increasingly encounter other Hawks, and whenever they do serious fights ensue, with one Hawk gaining the resource and the other being seriously injured (see definition of the Hawk strategy, above)

Suppose that losing a fight costs 300 fitness points (due to injury); then the average payoff from Hawk vs. Hawk interactions = -100 (i.e., 100/2 - 300/2)

The result is that even though Doves can't win against Hawks, at least they never get hurt, so they walk away with zero points while Hawks who fight and lose to other Hawks (as we are assuming they do half the time they fight) get a negative score (-100)

So what will final outcome be:  All Hawks, all Doves, or some mixture?

General answer (which applies to a broad range of evolutionary games, regardless of the exact payoffs and tactics being modeled) is that selection will produce an equilibrium (termed an "evolutionarily stable strategy," or ESS for short) where the ratio of Hawks to Doves is such that average payoff of each is equal (see Table)

Of course, the specific equilibrium ratio depends on actual payoffs; these are hypothetical, and whole scenario is highly simplified -- but it illustrates a more general point:  where aggression has costs as well as benefits, it will be favored by selection only up the to point where it yields no further increase in fitness

But notice that the ESS does not maximize the average fitness of the whole population:  a "conspiracy of doves" would be best for everyone (avg. payoff = 50, vs. 33 for the ESS), but all-Dove is not an ESS, because any mutant (or immigrating) Hawk tactic will spread via natural selection (as will Dove in an all-Hawk population)

In conclusion, despite (or perhaps because of) its simplicity, the Hawk-Dove model helps us see that selection doesn't favor what's best for the population; if a new variant arises that has higher individual fitness, it will spread even if it thereby lowers average fitness of population. This is a very important point, with broad implications far beyond context of fighting (being relevant to population regulation, resource conservation, etc.)

By the way, I described this evolutionary game as if there are two distinct hereditary types, pure Hawk or pure Dove; but exact same logic (and ESS) holds if you assume a more complex animal that can play both Hawk and Dove -- then the ESS is a monomorphic (single-type) population whose members all play the same mixed strategy:  Hawk 1/3 of time and Dove 2/3 of the time

One can of course make the model more complex by considering other possible tactics; John Maynard Smith, the British evolutionary biologist who co-created the original Hawk-Dove model (and more or less single-handedly invented evolutionary game theory), analyzed a model with an additional strategy called "Bourgeois" (play Hawk if resident, Dove if intruder); in this model, all-Bourgeois is the ESS (since Bourgeois will never get injured by a Hawk, will always beat a Dove, and playing against itself will do as well as a Dove-Dove contest and better than a Hawk-Hawk one)

Theoreticians have also constructed models that allow for individual variation in fighting ability, assessment of your opponents fighting ability, bluffing, cognitive errors, randomized play, learning which tactic to play, and on and on; some of these variations lead to different results, but most end up at essentially the same ESS, and none result in an ESS that appears designed to maximize "the good of the species"

Ecological Models of Competition

Hawk-Dove model is useful for illustrating some fundamental points about how selection might favor strategic forms of conflict, but it doesn't say much about the ecological determinants of costs and benefits that shape resource contests in the real world (factors that will shape which resources are worth fighting over, the payoffs of winning, etc.)

One way ecologists have got into the act is by analyzing the population ecology of resource competition

Formal definition of competition: Active demand by 2 or more individuals or groups for a resource that is actually or potentially limiting

By limiting, we mean any increase in the per-capita amount of the resource leads to an increase in the population (or an increase in the reproductive success of individuals who gain access to the resource)

A more direct measure of competition between two groups (could be subgroups within a single population, or could be populations of two different species) is that an increase in N₁ (the size of one group) causes a decrease in dN₂/dt (the growth rate of the second group, as in exponential and logistic equations)

Note that this definition doesn't specify what form competition will take; even if competitors never physically encounter each other, if they both utilize same resource that is limiting for at least one of them then they are engaged in ecological competition

A resource need not be food; it could be shelter (from predators or weather), sunlight (as in competition between plants), access to medical care, or whatever limits members of a population

Two basic categories of ecological competition (see Boone 1992 reading):

1) scramble form = no direct interaction; each party tries to obtain resource as rapidly or efficiently as possible

2) interference form = active attempts to reduce competitor's access to resources (aggression, territoriality, etc.)

Can make some generalizations about when to expect one or other of these two forms of competition   [Logistic graph]

First, when density-independent controls predominate, or population frequently below K (e.g., because of frequent colonization of disturbed habitats), competition should be relatively rare, and when it occurs is more likely to be scramble form

In contrast, populations at K or under strong density-dependent controls should often encounter conditions favoring interference competition

Second, the characteristics of the limiting resources should affect the optimal competitive strategy:    [Graph]

1. When resources subject to competition are relatively scattered and fluctuate in abundance, scramble form is favored

2. Interference form favored when resources concentrated & temporally stable

These 2 generalizations (specifying relation to K and the resource distribution) allow us to predict the ecological conditions under which competition will shift from scramble to interference mode, including human forms such as enforcement of property rights and warfare (assuming warfare = an elaborate form of interference competition)

Alternative Ecological Explanations of Human Warfare

Let's now consider some alternative explanations for occurrence of warfare that have an ecological slant:

1. "Unnatural Crowding" explanation
2. Population Control explanation
3. Land scarcity/competition explanation

(1) "Unnatural Crowding" explanation:

This view (popular among the public but not among specialists) holds that warfare is a form of social pathology in response to overcrowding in urban civilization, in same way that rats in a cage or monkeys in a crowded zoo engage in excessive aggression and other unnatural forms of behavior not seen in wild populations of same species

It is ecological only in the very limited sense that it focuses on population density (in an absolute sense, not "population pressure") as a determinant; in some versions, it argues that warfare is a strictly post-agricultural or even post-urbanization phenomenon -- a result of our expulsion from the Eden of a "natural" foraging lifestyle

A direct test of this hypothesis would be to examine the occurrence of warfare archaeologically, but in very old sites it is difficult to distinguish warfare from other forms of homicide, or even homicide from mortuary rituals (e.g., were H. erectus fossils whose skulls had base broken out and cut marks on facial area victims of head-hunting, or did they die natural deaths and become objects of veneration by kin who wished to preserve their skulls by removing brains and flesh, or to engage in the kind of ritual consumption of deceased kin as found among some modern New Guinea peoples?)

Late in the prehistoric record it becomes quite easy to find archaeological evidence of warfare:  fortified villages, graveyards with victims of violence, weaponry and armor, etc.

The fact that such finds are almost always in context of agricultural, often urbanized, societies may support hypothesis the "crowding" hypothesis, or it may simply reflect the fact that the great increase in population density following development of agriculture makes such sites much more numerous (and easier to find), as well as the "hard" technological manifestations of specialized weaponry and defense that come with a sedentary way of life

An indirect but workable test of the crowding hypothesis is possible if we use subsistence mode (foraging vs. low-density horticulture vs. intensive agric.) as a proxy for population density, which is reasonable since foraging populations are 1 to 2 orders of magnitude less dense than agricultural populations

Returning to same 87-society sample we used earlier, we find that this version of the hypothesis can be rejected:  warfare is not statistically more common or more frequent among agriculturalists than foragers, nor among intensive agriculturalists (who include all the high-density, urbanized populations)    [see Graph]

These findings cast doubt on thesis that warfare is primarily a response to "unnatural" crowding or the stresses & frustrations of complex urbanized societies

(2) Population Control explanation:

This explanation takes various forms, starting with simple notion (going back to Malthus) that warfare directly controls population thru mortality

Data & logic argue against this, since unless folks practice strict monogamy, the effect of male mortality on population growth can be readily offset by female fertility

In fact, many warlike populations have moderate or large amount of polygyny, and female mortality from war is usually quite low; here are some examples (all swidden agricultural peoples):

[Note: there is even less demographic impact in modern state warfare:  annual growth rate in WW II = 5% Germany, 0.2% U.S.A.; among Vietnamese in 1960-1970 = 3%]

A more elaborate argument is offered by Marvin Harris, who holds that warfare among non-stratified societies is a cultural adaptation to prevent resource depletion

Harrus recognizes that warfare deaths themselves rarely have significant long-term impact on growth rates

Instead, Harris proposes a more complex, indirect link between warfare and population control:

• population growth ® protein scarcity ® need to regulate population

• female infanticide ® reduced population growth, but also ® shortage of women

• women shortage ® raiding other villages

• raiding complex ® selective advantage to warlike society ® preference for "fierce males" (more men also means more protein)

• preference for men ® more female infanticide (positive feedback loop)

• fierce warriors rewarded thru more wives ® polygyny ® more warfare (positive feedback again)

Thus Harris sees war in pre-state societies as adaptive in such cases (in preventing population from exceeding carrying capacity), but creating a sort of "ecological trap" from which it is hard to escape

Harris and others have applied this argument to large sample of societies (correlating warfare with male-biased sex ratios), but have focused particularly on well-studied case of Yanomamo Indians of NE Amazonia

•swidden horticulture (manioc) & hunting

•intensive warfare (ca. 30% adult male mortality)

•rapidly shifting village alliances & hostilities

•highly skewed child sex ratio (0-4 yrs) of 148:100 in areas with greatest warfare (periphery = 118:100); evens out by age 50 due to male war deaths

•high rate of polygyny (some men have 4-5 wives; 25% have at least 2)

•female infanticide + polygyny ® intense competition for wives

•most warfare related to this competition

Harris' explanation has been severely criticized on two grounds:

1) empirical critiques:  sufficiency of animal protein, population dynamics, infanticide

2) theoretical critique:  logic of population regulation

On protein issue, Chagnon & other Yanomamo specialists refute scarcity:

• animal protein averages 65gm/adult standard/day (NRC recommends 40 gm for Yanomamo body size)
• warfare more frequent between villages that don't share boundaries
• village with lower (33 gm) animal protein intake had same frequency of warfare as village with higher (77 gm) intake

Harris & colleagues have responded by saying high protein intake is sign of how effective warfare is (less warlike tribes have less protein), which seems like an ad-hoc argument meant to make the hypothesis unfalsifiable

More serious criticism is lack of evidence of population regulation among Yanomamo, who at time of Chagnon's main study had high rate of population growth (ca. 3% per annum)

On theoretical side, there is the problem of population regulation as a "public good" (see also "Levels of selection" section of "Concepts of Adaptation" lecture notes)

1) Naive group-selectionism (system is vulnerable to "free-riders" who do not restrain own reproduction while benefiting from restraint of others)

2) Internal contradiction: groups who control their population would necessarily have less warriors & be more vulnerable to military defeat, despite lower resource depletion

In sum, Harris proposes a very creative argument, but it does not hold up under scrutiny

(3) Land scarcity/competition explanation:

A third type of ecological explanation for warfare focuses on competition for land, often in context of (hypothesized) population pressure

In terms of small-scale, non-hierarchical societies, tropical forest swidden groups show highest rates of endemic warfare:  Amazon Basin, New Guinea, Borneo, New Zealand (Maori), W, Africa, Caribbean

One explanation for this is that game animals are limiting factor (a la Harris):

• most swidden crops = high carbohydrate but low protein (yams, manioc...)

• dispersed settlement argued to be means of maintaining "buffer zone" where game can flourish

• however, "buffer zone" theory remains controversial (smacks of mechanism-free functionalism, naive group-selectionism)

Another explanation is that warfare involves competition for land:

• Ember (1982) used data from New Guinea used to test this, responding to claim by Sillitoe (1977) that few instances of warfare in New Guinea can be accounted for via ecol. factors like land shortage

• test involves correlation of population density (proxy for land scarcity) with frequency with which defeat ® expulsion and loss of land to victors

Ember shows that there is a statistically significant correlation btwn density and land conquest in these populations:

Despite apparent support for land-scarcity explanation in New Guinea (though population density does not equal "population pressure"), doubtful that it applies to all tropical forest swidden societies

Specifically, among Amazonian Indians population density (and population pressure, in sense of resource supply/demand) is generally quite low (due most likely to depopulation from introduced diseases), many populations were expanding rapidly prior to sustained contact, and little or no evidence of land conquest associated with indigenous warfare

Scarcity of land (or other land-based resources, such as water) is also implicated in some instances of violent conflict and warfare within or between modern state societies (see Homer-Dixon et al. reading)

Have to be careful here to distinguish interests/goals of elites who make decisions about engaging in warfare from those of who carry out those decisions (and bear the heaviest price in mortality)

Since these elites often control mass media, they may manipulate populace into thinking ethnic violence or conflict is in their interest (i.e., exercise some degree of hegemony) and even use such conflict to deflect challenges to their rule (lots of contemporary examples of this, from U.S. racial conflict to Serbia/Bosnia to Rwanda to Iraq-Iran wars -- and some would suggest current Bush regime actions against Iraq)

Here again we see the importance of analyzing ecological interests and adaptation (including resource competition) at the level of individuals and interests groups, rather than looking for adaptive systems, societies, or populations per se

Finally, it's important to avoid simplistic environmental determinism in explaining resource-based conflicts, since such analyses often mask just the sort of interest-group conflicts I'm referring to

A classic example is the so-called "soccer war" between El Salvador & Honduras in 1970s, which was portrayed by some environmentalists as inevitable result of resource scarcity induced by population growth

As William Durham's (1979) analysis shows, this conflict was based on land/food scarcity, but a scarcity induced primarily by elites who dispossessed peasants of their land in order to grow export crops such as coffee and bananas [see graphs from Durham 1979]

In sum, while ecological factors (such as resource scarcity and competition) may be at root of many social conflicts and wars, these generally operate via a complex causal network of political, economic, and cultural (e.g., social construction of ethnic enmities) factors, and therefore should be analyzed in a socioculturally sophisticated manner (for a discussion of some of the current debates on this topic, see Kahl 2002 reading, and the book edited by Peluso & Watts which it reviews)

References

Boehm, Christopher (1992) Segmentary warfare and management of conflict:  a comparison of East African chimpanzees and patrilineal-patrilocal humans.  In Coalitions and alliances in humans and other animals, ed. A. Harcourt and F.B.M. de Waal, pp. 137-173.  Oxford: Oxford U. Press.

Boone, James L. (1992) Competition, conflict, and the development of hierarchies. In Evolutionary Ecology and Human Behavior, ed. E. A. Smith and B. Winterhalder, pp. 301-337. Hawthorne, NY: Aldine de Gruyter.

Chagnon, Napolean A. and Raymond Hames (1979) Protein deficiency and tribal warfare in Amazonia: new data. Science 203: 910-13.

Durham, William H. (1979) Scarcity and Survival in Central America: Ecological Causes of the Soccer War, pages 1-9. 21-62, & 169-173. Stanford, CA: Stanford University Press.

Ember, Melvin (1982) Statistical evidence for an ecological explanation of warfare. American Anthropologist 84: 645-49.

Haas, Jonathan (2001) Warfare and the evolution of culture.  In Archaeology at the millenium:  a sourcebook, ed. G. Feinman and T.D. Price, pp. 329-350.  NY: Kluwer/Plenum.

Harris, Marvin (1984) Animal capture and Yanamamo warfare: retrospect and new evidence. J. of Anthropological Research 40: 183-201.

Homer-Dixon, Thomas F., et al. (1993) Environmental change and violent conflict. Scientific American February: 38-45.

Homer-Dixon, Thomas F. (1999). Environment, scarcity, and violence. Princeton, NJ: Princeton University Press.

Kahl, Colin (2002) Review of Peluso & Watts, "Violent Environments." ECSP Report 8:135-143.

Manson, Joseph H. and Richard W. Wrangham (1991) Intergroup aggression in chimpanzees and humans. Current Anthropology 32: 360-390.

Maynard Smith, John (1974) The theory of games and the evolution of animal conflicts. J. of Theoretical Biology 47: 209-22.

Maynard Smith, John (1982) Evolution and the Theory of Games. Cambridge: Cambridge University Press.

Mech, L. David (1970) The Wolf:  The Ecology and Behavior of an Endangered Species.  Garden City, NY: Natural History Press.

Mesquida, Christian G. and Neil I. Wiener (1996) Human collective aggression: a behavioral ecology perspective. Ethology and Sociobiology 17(4):247-262.

Peluso, Nancy and Michael Watts, eds. (2001) Violent Environments. Ithaca, NY: Cornell U. Press.

Peoples, James G. (1982) Individual or group advantage?  A reinterpretation of the Maring ritual cycle. Current Anthropology 23:291-310.

Sillitoe, Paul (1977) Land shortage and war in New Guinea. Ethnology 16: 71-81.

Soltis, Joseph, Robert Boyd, and Peter J. Richerson (1995) Can group-functional behaviors evolve by cultural group selection?  An empirical test. Current Anthropology 36:473-483.

Vayda, Andrew P. (1974) Warfare in ecological perspective. Annual Review of Ecology and Systematics 5:183-93.

Vayda, Andrew P. (1989) Explaining why Marings fought. Journal of Anthropological Research 45:159-177.