. 4
( 13)


15 20 25 30 35 40 45 50 55 60 65 70

-2000 Total Net Production
Net Caloric Demand


age of male

Figure 3.3
Family demands and net family production.

term average hunting returns between the best and worst hunter in the
sample of Ache men (Hill et al. 1987). Similar discrepancies in hunting
ability across men have been found among the !Kung (Lee 1979),
Hadza (Hawkes, O™Connell, and Blurton Jones 2001), Hiwi (Gurven,
Hill et al. 2000), Gunwinggu (Altman 1987), Agta (Bion Grif¬n 1984),
and Machiguenga (Kaplan unpublished data).4 Therefore, even among
men of the same age, there must be net transfers over the long term
from families producing a surplus to families producing a de¬cit.
These food transfers provide great reproductive bene¬ts. The ability
to ˜˜borrow™™ and ˜˜lend™™ across the life course is necessary for subsid-
izing the juvenile learning period (see Kaplan and Robson 2002 and
Robson and Kaplan 2003 for theoretical models of the evolution of
such age transfers and their relationship to learning). If families had to
˜˜balance their budget™™ at every period, they would either have had to
lower their fertility or force their older children to fend for themselves.
This would most likely increase childhood and adolescent mortality
and lower rates of skills acquisition. Adolescent males could not afford
to hunt, because their returns are so low during the learning period.
Moreover, there would be no way to buffer the risks associated with
94 Kaplan and Gurven

the stochasticity of family size and child demands. If families needed
to support all of their individual food needs, regardless of whether
few or many children survived, they would be forced to lower fertility
or reduce child subsidies. Similarly, the ability of women to shift pro-
duction across time without changing consumption probably increases
infant survival and decreases the length of interbirth intervals, thereby
increasing the total reproductive success of women. When the oppor-
tunity costs of food acquisition are high due to the need to care for
infants, women may produce less when they have infants and then
work harder when those opportunity costs are low (i.e., when they
have no infant to nurse and protect).

3.3.3 The Problem with Dyadic Reciprocal Altruism
It is unlikely that such a system of sharing would be stable with
strictly dyadic reciprocal altruism. Reciprocal altruism will only emerge
among self-interested actors if there is repeated interaction that re-
wards cooperation and punishes defection. In terms of reciprocal altru-
ism, there is no incentive for young or older adults with small families
to support older adults with large families. Those older adults with
large families will never produce a surplus to ˜˜pay back™™ those subsi-
dies, because they are likely to die before the young adults reach the
age in which they need assistance to support their families. While it
might be argued that there is intergenerational reciprocity where the
children of the older adults, in turn, support the families of those who
helped them, the long time periods between changes in directional
¬‚ows would make such arrangements inherently risky. There is a
great deal of mobility between hunter-gatherer bands and residential
arrangements are not stable over long periods. There is no guarantee
that children who are helped when they are young will live in the
same band as those who helped them. The same argument applies to
sharing between non-nursing and nursing women. Additionally, time
discounting of bene¬ts received in the distant future (relative to the
present consumption payoffs from not sharing) also makes intergen-
erational reciprocity unstable (Hawkes 1992).
Similarly, reciprocal arrangements regarding the stochasticity of
family size are unlikely to emerge with dyadic relationships. If family
size variation is due primarily to random luck, it may be bene¬cial for
two individuals to agree at the start of their reproductive careers to an
arrangement in which the individual who ends up with fewer surviv-
ing offspring agrees to support the one with more surviving offspring.
The Natural History of Human Food Sharing and Cooperation 95

However, once the outcome is known, there is no incentive for the one
with fewer children to provide the support, since his family will never
need the payback and there is no way to enforce the bargain.
Our thesis is that humans have found ways to take advantage of the
gains from such trades, even though these gains would not emerge
through dyadic reciprocal altruism. We propose that multi-individual
negotiations result in the emergence of social norms that are collec-
tively enforced. We base this proposal on a result obtained by Boyd
and Richerson (1992), and treated more recently by Bowles and Gintis
(2000), in which cooperation is modeled with punishment. These four
researchers found that cooperation can be stable in large groups, if
noncooperators are punished and if those who do not punish noncoop-
erators are also punished. In fact, they found that any social norm
could be stable as long as both those who disobey and those who fail
to punish those who disobey are punished. However, we suggest that
self-interested actors also negotiate these norms, weighing the individ-
ual costs and bene¬ts of different social norms.

3.3.4 Two Thought Experiments
Imagine the following scenario. A woman returns from collecting ber-
ries and pounding palm ¬ber with her bawling infant. A wingless
wasp stung her baby while she had put him down to pound the ¬ber,
and the baby is in great pain. She is frustrated and says to the other
women in camp, ˜˜This is crazy for me to go out and pound ¬ber when
I have such a young baby. I would gladly work twice as hard when he
is a little older if I could concentrate on watching him now.™™ A few
days later when the baby™s wound is infected and the child has a fever,
another woman, remembering a similar incident she experienced a few
years ago, says, ˜˜You know, Singing Deer is right. We should work
hard when we have no baby on the breast and allow those with a
young one to care for it well.™™ Another woman, who has not had a
child in the last ten years, says, ˜˜Why should we work to feed other
people™s babies? If you have a baby, you must feed it.™™ Other men and
women consider their own situation (as well as the situation of their
children) and present their opinions. Eventually a consensus (or at
least, an agreement) is reached, with those in the minority either agree-
ing to go along with the new norm or leaving to live with ˜˜less foolish™™
people. However, one woman, who is not nursing, hardly pounds ¬ber
at all. Other women begin to gossip about her, remarking upon how
lazy she is, because she has no child to care for. She notices that the
96 Kaplan and Gurven

shares she receives in food distributions start to become less generous
and begins to suspect that others are talking about her behind her
back. She leaves and pounds a large quantity of ¬ber, which she gener-
ously shares with the rest of the group. She can feel the warmth of
others return and has learned her lesson.
We consider another similar scenario. A ¬fty-year old man turns to
another older man and exclaims, ˜˜Look at these lazy young men! They
come back to camp at midday and play around. Yet you and I have
lots of children to feed and no food to give them. What will those boys
do when they have big families to feed?™™ The other older man agrees,
adding, ˜˜How do I know if that lazy one is good enough for my
daughter? How do I know if he will get enough food to keep her chil-
dren healthy? He should come to my ¬re and bring me lots of meat”
then I will know.™™
The young men are not very enthusiastic, because they do not like
hunting all day long, but they are reluctant to anger the men whose
daughters they favor. One young man, who is a good hunter for his
age, realizes that he could take advantage of such a system and starts
to hunt longer hours, giving the older men generous shares from his
hunt. The other young men, afraid of being outdone, also begin to
hunt longer hours and share the fruits of their labor more generously.
While admittedly hackneyed, these scenarios are meant to re¬‚ect the
ongoing discussions and commentaries about sharing, work effort, and
laziness that are so pervasive in foraging societies. We do not mean to
suggest that all social norms are explicitly negotiated with words or
that norms solidify over a short period as a result of a few conversa-
tions. In some circumstances, lack of compliance and ˜˜voting with
one™s feet™™ are almost surely involved in those negotiations. In fact, we
know virtually nothing about how standards for appropriate behavior
emerge and change in small-scale societies without of¬cial means of
enforcement. It is likely that majority-rule voting arrangements are not
adhered to in a strict sense, since some individuals exercise undue
in¬‚uence (e.g., kombeti among Aka, kapita among Efe [Hewlett and
Walker 1990], and chiefs among Yuqui [Stearman 1989]). Nevertheless,
we propose that such multi-individual negotiations, partly verbal and
partly nonverbal, do result in social norms and that the weight of col-
lective opinion, based upon the individual costs and bene¬ts of norms
in given contexts, determines accepted patterns of behavior. In the next
section, we develop a preliminary framework for explaining variation
in norms regarding cooperation.
The Natural History of Human Food Sharing and Cooperation 97

3.4 Part III: A Preliminary Framework for Explaining Sharing

We propose that social norms of sharing re¬‚ect the relative strengths of
two opposing forces: gains from cooperation and possibilities for free-
riding. Socioecological variation in potential bene¬ts of cooperation
and possibilities to free-ride on cooperative behavior determine cul-
tural variability in norms of sharing and cooperative labor.
We also propose that in the course of our evolutionary history, natu-
ral selection has shaped our psychology to possess certain traits.

1) perceptual sensitivity to potential gains from cooperation
2) motivation to take advantage of those gains
3) perceptual sensitivity to opportunities for free-riding
4) motivation to avoid being a victim of free-riding
5) motivation to take advantage of opportunities for free-riding
6) perceptual sensitivity to the short- and long-term personal costs and
bene¬ts of social norms regarding cooperative behavior (from the per-
spectives of both the self and others)
7) motivation to negotiate social norms so that one™s own personal
bene¬ts from cooperation and free-riding are maximized
8) motivation to obey and enforce social norms so that punishment
is avoided, and those who disobey norms or fail to enforce them are

Our proposal is that this psychology, the complex analytical brain,
and the extended life history coevolved in the hominid line”all be-
cause of the dietary shift towards large, high-quality food packages
and hunting. It is this feeding adaptation that generates the gains from
group cooperation. The large size of the packages and the dif¬culty of
their acquisition through hunting
a) facilitate sharing (imagine sharing blades of grass back and forth)
b) increase short-term variation in acquisition luck, since large pack-
ages are not abundant
c) require signi¬cant learning and experience
d) increase the disparity between production and consumption at the
individual and family levels over the medium and long term
98 Kaplan and Gurven

e) increase the bene¬ts of collective action and cooperative pursuits,
especially in hunting
f ) generate economies of scale, since foods are often distributed in
larges patches distant from residential locations.

These qualities generate large gains from intertemporal substitution
in consumption and production over the short, medium, and long
term; gains from specialization by age, sex, and perhaps individual
qualities; gains from joint production and cooperative acquisition; and
gains from turn-taking in acquisition of patchily distributed foods. The
distribution and relative importance of each of those gains is likely to
vary with local ecology and the foods exploited.
Possibilities for, and gains from, free-riding act against cooperation.
Three factors are likely to in¬‚uence the threat of free-riding. First, a
larger number of individuals in cooperative networks is likely to in-
crease the threat, because the ability to detect and punish free-riders
probably diminishes with partner number. As group size increases,
the probability that more than one individual free-rides may also in-
crease (Boyd 1988). As the number of free-riders increases, costs of
punishment increase and the incentive to cooperate decreases. Second,
the quality of information about behavior is also likely to affect oppor-
tunities for free-riding. If work effort is dif¬cult to monitor and if it is
dif¬cult to determine whether variance in productivity is due to acqui-
sition luck or work effort, opportunities for free-riding may increase
(Cosmides and Tooby 1992). Third, gains from free-riding are also
likely to vary according to kinship relationships between participants.
As overall relatedness decreases, the differences among optimal alloca-
tions of work and distribution across individuals are likely to increase.
Those opposing forces may have led to the evolution of some gen-
eral moral sentiments”supported both by the motivational psychol-
ogy of individuals and common cultural norms. Variation in need and
production among individuals due to stochasticity should engender
generosity and cultural norms emphasizing the value of generosity”
perhaps mediated through costly signaling and reciprocal altruism.
Sharing sentiments and norms would favor those who were unfortu-
nate over the short or long run and require generosity from the more
fortunate. Virtually every investigator who seeks to establish friend-
ships with members of traditional subsistence populations, who are
much poorer, feels the pressures associated with those sentiments.
Similarly, temporary states affecting production or need”such as ill-
The Natural History of Human Food Sharing and Cooperation 99

ness, nursing, and high dependency ratios”would also promote gen-
erosity. As mentioned above, the rule that larger families deserve and
receive larger shares is very widespread. Conversely, variation due to
lack of effort or laziness would not generate generosity and perhaps
invoke ridicule or punishment. Indeed, laziness and stinginess are
constant themes for gossip in traditional societies. Other things being
equal, people should feel more generous towards (and trusting of)
close kin, because of the reduced scope of con¬‚icting interests.
At the same time that moral systems are likely to have such general
guiding principles, there is scope for considerable variation in the
norms of cooperation and sharing, depending upon the speci¬c con-
stellation of gains from cooperation and possibilities for free-riding. Of
critical importance is the relationship between the size and composi-
tion of residential groups and the optimal size of cooperating units. In
general, people will tend to organize residential groups so that they
can take maximal advantage of the gains from cooperation and re-
duce risks of free-riding. Thus, many forager-horticultirists in South
America”such as the Machiguenga, Piro, and Tsimane”settle in
extended family units, characterized by an older couple, their adult
sons and/or daughters, and the founding couple™s grandchildren. La-
bor is divided by age and sex, and food is eaten communally. This sys-
tem of communal production and consumption maximizes gains from
specialization and from spreading consumption and production needs
through the entire age-structure, while kinship and shared genetic
interests in the young children minimize con¬‚icts of interest.
Several factors may cause residential groups to differ in size and
composition from their optimal structure for cooperation. Aggregation
of larger groups is common, because of threats of violence (e.g., Yano-
mamo [Chagnon 1983]), lack of resources such as water or groves of
trees (e.g., Dobe !Kung [Lee 1979]), and now schools and delivery
points of social services (e.g., Chacobo [Prost 1980]). In these cases,
restricted sharing”where some or all foods are shared with only a
subset of the residential group”is the norm.
Restricted sharing systems appear to be particularly common when
the primary gains from sharing derive from variance reduction in
consumption and when gains from cooperative pursuits are small or
restricted to only some resources or times of the year. A common prin-
ciple evidenced in restricted sharing systems is that the breadth and
depth of resource sharing depends on the size of food packages avail-
able. When food packages are small, they are shared with a few special
100 Kaplan and Gurven


Percent Received

PACK < 4 kg
30 PACK >= 4 kg



0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
RANK of Recipient NF

Figure 3.4
Ache sharing by package size.

partners, with whom reciprocal sharing is very common. As package
size increases, the size of sharing networks grows (increased breadth)
and the percentage of the food kept by the acquirer™s family is reduced
(increased depth).
Figures 3.4 and 3.5 illustrate features of this system. Figure 3.4 shows
the percentage of sharing events by resource package size, in which
speci¬c Ache families receive shares at their permanent horticultural
settlement. For each individual, sharing partners were rank-ordered
from those who received most often to those who received least often.
The x-axis displays the rank order and the y-axis gives the average per-
centage of occasions in which partners of each rank received shares.
The data show that small packages are repeatedly shared with few
individuals and that the size of sharing networks expands with large
packages. Figure 3.5 (derived from data collected among the Hiwi and
adapted from Gurven, Hill et al. 2000) is a path analysis predicting the
total accumulation of food transferred between families over a six-
month sample period, giving additional information about how part-
ners are selected and about the size of shares given. Kinship predicts
the spatial proximity between givers and receivers, which, in turn, pre-
dicts both how much was received in the past and the amount given in
the present. In addition to kinship and proximity, the past history of
sharing also predicts the amount given, suggesting that giving is con-
tingent upon past receipts when controlling for these other factors.
The Natural History of Human Food Sharing and Cooperation
B >A
Age B
-0.29 Family Size Residential
0.19 Relatedness
of B Distance
of A to B
from A to B -0.42
0.04 0.14 -0.15

A >B


% A kept
in NF

Figure 3.5
What affects how much nuclear family A gives to nuclear family B among Hiwi foragers?

102 Kaplan and Gurven

Larger families also receive larger shares, as would be expected if need
is being taken into account. Qualitative and quantitative reports from
other societies suggest that similar patterns”kin bias, differential rules
for sharing different resources (with increased breadth and depth of
sharing with increased package size), contingency of sharing on the
basis of past receipts, and larger shares to larger families”are found
in other societies (Ifaluk [Sosis 1997], Eskimo [Damas 1972], Batak
[Cadelina 1982], Yora [Hill and Kaplan 1989]). It is not always the
case, however, that the residential group is larger than the optimal
sharing network for all resources acquired. In cases where very large
packages are sometimes acquired (e.g., giraffe among the !Kung), it is
sometimes necessary to inform members of neighboring groups about
kills because the optimal sharing group size is larger than the optimal
residential unit (Lee 1979).
Such systems tend to take advantage of the gains from cooperation
while minimizing risks of free-riding. Reducing daily variation in the
consumption of small packages requires fewer partners than in the
case of larger packages. Thus, a small circle of trusted partners, fre-
quently kin and neighbors, is most ef¬cient. As package size increases,
the bene¬ts of a greater number of partners increase, but so too do the
costs of free-riding.
Another important principle of restricted sharing systems is that
work effort in cooperative activities is rewarded. Thus, when coopera-
tive task groups form, food is often shared equally among the partici-
pants. When those task groups do not include members from all the
families in the residential group, a system of primary and secondary
sharing is very common. In the primary distribution, all participants in
the cooperative activity receive approximately equal shares of the total
catch (see part I of this chapter for a list of groups engaging in this
practice). In secondary distributions, each individual that received
shares redistributes his or her share to families that did not participate.
Those shares are smaller and tend to be shared according to the size of
the packages acquired in the manner discussed earlier in this section.
Figure 3.6 from the Yora illustrates this pattern (see Hill and Kaplan
1989). The ¬rst two bars show the primary distribution to members of
the foraging party and the second two bars show the secondary distri-
bution. This is an ˜˜incentive compatible™™ system in which work effort
is rewarded in the primary distribution and the other bene¬ts of shar-
ing (e.g., intertemporal substitution in consumption and production)
are handled in the secondary distribution. In cases where representa-
The Natural History of Human Food Sharing and Cooperation 103




Other Families


Foraging At Camp Overnight
Party Trips

Figure 3.6
Yora meat sharing.

tives from every family in the residential group become involved in co-
operative pursuits, such as the Ache when living in the forest and the
Yora on trek (the third set of bars), food tends to be eaten communally.
In addition to rewarding work effort, sharing systems also appear to
reward special capital contributed to cooperative efforts. For example,
cooperative ¬shing and whaling among some coastal groups (e.g., Ifa-
luk [Sosis 1997], Lamalera [Alvard and Nolin 2002], and Makah [Sin-
gleton 1998]) requires boats and large work parties. Again, there is a
primary distribution to all those who worked and secondary distribu-
tions for further sharing. However, in this case, boat owners receive
larger or preferential shares. This suggests that not all individuals are
weighted equally in the negotiation of sharing norms. While it is possi-
ble that those without boats could form a coalition to enforce equal
sharing (since they are greater in number), it appears that those with
special capital have more to offer in the market for cooperative part-
ners and use this leverage to their advantage. Similarly, among Mbuti
pygmies who hunt with large nets, net owners receive more food
(Turnbull 1965) and among Efe and Aka Pygmie hunters, food shares
depend upon the task performed in the cooperative hunt (Ichikawa
1983; Kitanishi 1998).
Finally, sharing systems undergoing transition also illustrate impor-
tant principles in the negotiation of sharing norms. For example, the
Ache have experienced several changes in food sharing and labor orga-
nization. Their economy transformed from full reliance on hunting and
gathering in small groups to a mixed economy of foraging, farming,
104 Kaplan and Gurven

and wage labor in larger settlements after their establishment of per-
manent peaceful contact with the larger Paraguayan society. For the
¬rst ¬ve or so years following settlement, agricultural ¬elds were
cleared and planted communally. All able-bodied men were expected
to contribute labor in large work parties. This pattern resembled the
cooperative economy of the past. However, within a few years, it be-
came apparent that some people were often absent from work parties
and resentments began to build. Some men tired of this system and
cleared their own personal gardens. Communal ¬elds became smaller
and a system of private ¬elds, with fewer friends helping each other,
came to predominate. Similarly, even with hunted and gathered foods,
the system changed from communal sharing of all game to a restricted
pattern resembling the Hiwi one shown in ¬gure 3.5. It is interesting to
note that the Ache still retain the traditional sharing pattern when trek-
king in the forest, even though they revert to the new pattern when
residing in the settlement. Similarly, the !Kung San appear to have
undergone major changes in their system of food distribution, since
becoming involved in a mixed economy and the larger state society.
Again, the trend seems to be from more communal distributions to-
wards more restricted sharing, with a great deal of bickering and strife
during the transition (see Shostak 1981 and the associated N/ai ¬lm
for qualitative accounts).
The transition to horticulture among the Ache and !Kung was very
rapid, and encouraged through missionary assistance. As mentioned
above, the establishment of private ¬elds was quickly advocated and
voted upon in local Ache meetings. This contrasts with the pattern
in other groups such as the Hadza (Woodburn 1982) and the Batek
(Myers 1988), where initial attempts at horticulture by a minority of
the population met with abrupt failure. The ¬rst harvests of the few
transitional farmers were exploited by incessant demands from those
who did not farm, ultimately making farming an unproductive activity
due to mutual adherence to more traditional norms of sharing.

3.5 Conclusion

We have proposed that in addition to individual reciprocal arrange-
ments, humans appear to be able to take advantage of gains from co-
operation in ways that are unexpected by pair-wise game models. We
suggested that people engage in multi-individual processes of norm
negotiation (both verbal and nonverbal) that allow gains from cooper-
The Natural History of Human Food Sharing and Cooperation 105

ation and minimize risks of free-riding. The framework we proposed,
however, is qualitative and far from fully speci¬ed. It clearly requires
formal models to evaluate its plausibility.
We suspect that given the absence of state controls, the systems of
exchange and cooperation found in traditional societies would not be
stable without the complex web of kinship connections characterizing
their residential groups. Those connections have two effects. First, as
discussed earlier in this chapter, they reduce con¬‚icts of interest be-
tween individuals and families. In fact, the marriage alliances between
families (observed and commented on since the earliest days of anthro-
pology) may be a way to minimize such con¬‚icts of interest through
the production of descendents sharing genes from both sets of families.
Second, kinship connections lower the variance in the payoffs associ-
ated with norms of sharing and cooperation. For example, norms that
allocate larger shares to families with more children to feed may be
disadvantageous for individuals in small families, but because, mem-
bers of small families are likely to have close kin (nieces, nephews,
brothers, sisters, and grandchildren) in large families, the total net
results of the norm for their genetic lineage may be positive. Since
most other species that have elaborate systems of resource sharing and
cooperation”such as social insects and group-hunting predators”
organize cooperation along kinship lines, it is likely that kinship
played an important role in the evolution of cooperation in humans.
Models of multi-individual norm negotiation with and without kinship
will be particularly useful in evaluating this intuition.
In part III of this chapter, we suggested that norms of sharing and
cooperation would re¬‚ect the ecology of subsistence, as well as the
associated variability in the gains from cooperation and possibilities
for free-riding. However, it is possible that similar ecologies may re-
sult in very different equilibria, depending upon historical conditions
and perhaps even essentially random perturbations. Formal models
would also be useful for evaluating this possibility. If multiple equilib-
ria are possible, then cultural or trait group selection may determine
which equilibria come to dominate over time. Given the kinship rela-
tions organizing the formation of groups in traditional societies, cul-
tural and genetic selection among groups and lineages may occur
Finally, informal observation (and the results of behavioral genetics
studies) suggest that there may be signi¬cant individual differences
within groups in terms of free-riding and obedience to group norms.
106 Kaplan and Gurven

The existence of varying degrees of free-riding by individual members
of social groups may be an inevitable outcome of cooperative norms
that can only be partially enforced. The optimal amount of effort allo-
cated to police free-riding may itself be subject to negotiation, as are
allocations to law enforcement in state societies.
This chapter represents a ¬rst step in a developing a multi-
individual approach to cooperation among traditional human societies
and to the psychology that underlies it. Our hope is that this paper will
help stimulate the development of formal analyses of those processes.


1. Chimpanzee mothers do share some dif¬cult-to-acquire solid foods with weaned
offspring (Silk 1979), but chimpanzee young are largely self-suf¬cient after they are
2. While computer simulations reveal that signi¬cant correlations between individuals in
amounts given and received are possible when tolerated theft is the sole cause of food
sharing, correlations greater than 0.2 were only found in groups of fewer individuals
than was common in the above groups.
3. The consumption and production of women is not included in this calculation since,
on average, women produce just enough to support their own consumption or a bit less.
4. Among the Machiguenga of Yomiwato, the best hunter produced more than half of
the meat for the whole village over a year period.


Alexander, R. 1987. The Biology of Moral Systems. New York: Aldine de Gruyter.
Allen-Arave, W., M. Gurven, and K. Hill. n.d. Is Food Sharing Among Close Kin Main-
tained by Kin Selection or Reciprocal Altruism? Evidence from Ache Food Transfers. Uni-
versity of New Mexico. Ms.
Altman, J. 1987. Hunter-Gatherers Today: An Aboriginal Economy of North Australia. Can-
berra: Australian Institute of Aboriginal Studies.
Altman, J. 1980. Baboon Mothers and Infants. Cambridge: Harvard University Press.
Alvard, M. 2002. Carcass ownership and meat distribution by big-game cooperative
hunters. Research in Economic Anthropology, 99“131.
Alvard, M., and D. Nolin. n.d. Rousseau™s whale hunt? Coordination among big game
hunters. Current Anthropology 43(4):533“559.
Aspelin, P. 1979. Food distribution and social bonding among the Mamainde of Mato
Grosso, Brazil. Journal of Anthropological Research 35:309“327.
Bahuchet, S. 1990. Food sharing among the pygmies of Central Africa. African Study
Monographs 11:27“53.
The Natural History of Human Food Sharing and Cooperation 107

Bailey, R. C. 1991. The Behavioral Ecology of Efe Pygmy Men in the Ituri Forest, Zaire. An-
thropological Papers, No. 86, Museum of Anthropology, University of Michigan.
Baksh, M., and A. Johnson. 1990. Insurance policies among the Machiguenga: An ethno-
graphic analysis of risk management in a non-Western society. In Risk and Uncertainty in
Tribal and Peasant Economics. Ed. Elizabeth Cashdan, Boulder, CO: Westview Press.
Balikci, A. 1970. The Netsilike Eskimo. New York: Natural History Press.
Barnard, A., and J. Woodburn. 1988. Property, power and ideology in hunting-gathering
societies: An introduction. In Hunter-Gatherers, Volume II: Property, Power and Ideology.
Eds. Tim Ingold, David Riches, and James Woodburn. 4“31. Oxford: Berg.
Bion Grif¬n, P. 1984. The acquisition and sharing of food among Agta foragers. The shar-
ing of food: From phylogeny to history conference. Homburg, Germany.
Bliege Bird, R. L., and D. W. Bird. 1997. Delayed reciprocity and tolerated theft: The be-
havioral ecology of food-sharing strategies. Current Anthropology 38:49“77.

Bliege Bird, R., E. A. Smith, and D. W. Bird. 2001. The hunting handicap: costly signaling
in male foraging strategies. Behavioral Ecology and Sociobiology 50:9“19.

Blurton Jones, N. 1984. A sel¬sh origin for human food sharing: Tolerated theft. Ethology
and Sociobiology 5:1“3.
Blurton Jones, N. 1987. Tolerated theft, suggestions about the ecology and evolution of
sharing, hoarding, and scrounging. Social Science Information 26:31“54.
Bock, J. 2002. Learning, Life History, and Productivity: Children™s lives in the Okavango
Delta of Botswana. Human Nature 13(2):161“198.
Boehm, C. 1993. Egalitarian society and reverse dominance hierarchy. Current Anthropol-
ogy 34:227“254.
Boyd, R. 1988. The evolution of reciprocity in sizable groups. Journal of Theoretical Biology
Boyd, R., and P. Richerson. 1989. The evolution of indirect reciprocity. Social Networks
Boyd, R., and P. J. Richerson. In press. Solving the puzzle of human cooperation, In Evo-
lution and Culture. Ed. S. Levinson. Cambridge MA: MIT Press.
Boyd, R., and P. Richerson. 1992. Punishment allows the evolution of cooperation (or
anything else) in sizable groups. Ethology and Sociobiology 13:171“195.
Cadelina, R. V. 1982. Batak Interhousehold Food Sharing: A Systematic Analysis of Food
Management of Marginal Agriculturalists in the Philippines. Ph.D. dissertation. Univer-
sity of Hawaii, Honolulu.
Carneiro, R. L. 1983. The cultivation of manioc among the Kuikuru of the Upper Xingu.
In Adaptive Responses of Native Amazonians. Eds. R. Hames and W. Vickers, 65“111. New
York: Academic Press.
Cashdan, E. 1985. Coping with risk: Reciprocity among the Basarwa of Northern Bo-
tswana. Man (N.S.) 20:454“474.
Chagnon, N. 1983 [1968]. Yanomamo: The Fierce People. New York: Holt, Rinehart and
108 Kaplan and Gurven

Chicchon, A. 1992. Chimane Resource Use and Market Involvement in the Beni Bio-
sphere Reserve, Bolivia. Ph.D. dissertation. University of Florida.
Clements, K. C., and D. W. Stephens. 1995. Testing models of animal cooperation: Feed-
ing bluejays cooperate mutualistically, but defect in a massively iterated Prisoner™s Di-
lemma. Animal Behaviour 50:527“535.
Clutton-Brock, T. H. 1991. The Evolution of Parental Care. Princeton, NJ: Princeton Univer-
sity Press.
Cosmides, L., and J. Tooby. 1992. Cognitive adaptations for social exchange. In The
Adapted Mind: Evolutionary Psychology and the Generation of Culture. Eds. J. H. Barkow, L.
Cosmides, and J. Tooby. New York: Oxford University Press, 163“228.

Damas, D. 1972. Central Eskimo systems of food sharing. Ethnology 11:220“240.
Dowling, J. 1968. Individual ownership and the sharing of game in hunting societies.
American Anthropologist 70:502“507.

Dugatkin, L. A. 1997. Cooperation among Animals: An Evolutionary Perspective. New York:
Oxford University Press.

Endicott, K. 1988. Property, sharing, and con¬‚ict among the Batek of Malaysia. In Hunter-
Gatherers, Volume II: Property, Power and Ideology. Eds. Tim Ingold, David Riches, and
James Woodburn. Oxford: Berg, 110“128.
Feinman, S. 1979. An evolutionary theory of food sharing. Social Science Information
Frank, R. 1988. Passions within Reason. New York: Norton.
Gintis, H. 2000. Strong Reciprocity and Human Sociality. Journal of Theoretical Biology
Gould, R. A. 1980. Comparative ecology of food-sharing in Australia and northwest
California. In Omnivorous Primates. Ed. R. Harding and G. Teleki. New York: Columbia
University Press, 422“454.
Grafen, A. 1990. Biological signals as handicaps. Journal of Theoretical Biology 144:517“
Gurven, M. 2004. Reciprocal altruism and food sharing decisions among Hiwi and Ache
hunter-gatherers. Behavioral Ecology and Sociobiology 56(4):366“380.

Gurven, M. 1999. Transitions in food sharing patterns from nomadic foragers to seden-
tary horticulturalists. Society for Cross-Cultural Research, Santa Fe, NM.
Gurven, M., and H. Kaplan. n.d. Determinants of Time Allocation to Production across
the Lifespan among the Machiguenga and Piro Indians of Peru. University of New Mex-
ico. Ms.
Gurven, M., K. Hill, and H. Kaplan. 2002. From forest to reservation: transitions in food
sharing behavior among the Ache of Paraguay. Journal of Anthropological Research
Gurven, M., W. Allen-Arave, K. Hill, and A. M. Hurtado. 2000. ˜˜It™s a Wonderful Life™™:
Signaling generosity among the Ache of Paraguay. Evolution and Human Behavior 21:263“
The Natural History of Human Food Sharing and Cooperation 109

Gurven, M., W. Allen-Arave, K. Hill, and A. M. Hurtado. 2001. Reservation sharing
among the Ache of Paraguay. Human Nature 12(4):273“298.
Gurven, M., K. Hill, H. Kaplan, A. M. Hurtado, and R. Lyles. 2000. Food transfers among
Hiwi foragers of Venezuela: Tests of reciprocity. Human Ecology 28(2):171“218.
Hames, R. 1987. Garden labor exchange among the Ye™kwana. Ethology and Sociobiology
Hames, R. 1989. Time, ef¬ciency, and ¬tness in the Amazonian protein quest. Research in
Economic Anthropology 11:43“85.
Hames, R. 1990. Sharing among the Yanomamo: Part I, The effects of risk. In Risk and Un-
certainty in Tribal and Peasant Economies. Ed. E. Cashdan. Boulder, CO: Westview Press.
Hames, R. 2000. Reciprocal altruism in Yanomamo food exchange. In Human Behavior and
Adaptation: An Anthropological Perspective. Eds. N. Chagnon, L. Cronk, and W. Irons. New
York: Aldine de Gruyter.

Hamilton, W. D. 1964. The genetical evolution of social behavior. Journal of Theoretical Bi-
ology 7:1“52.

Harako, R. 1976. The Mbuti as hunters. A study of ecological anthropology of the Mbuti
Pygmies (Ituri, Zaire). Kyoto University African Studies 10:37“99.
Harpending, H. 1998. Comment. Current Anthropology 39:88“89.

Hawkes, K. 1991. Showing off: Tests of an hypothesis about men™s foraging goals. Ethol-
ogy and Sociobiology 12:29“54.

Hawkes, K. 1992. Sharing and collective action. In Evolutionary Ecology and Human Behav-
ior. Eds. E. A. Smith and B. Winterhalder. Aldine de Gruyter, New York: 269“300.
Hawkes, K. 1993. Why hunter-gatherers work: An ancient version of the problem of
public goods. Current Anthropology 34:341“361.
Hawkes, K., J. F. O™Connell, and N. G. Blurton Jones. 1989. Hardworking Hadza grand-
mothers. In Comparative Socioecology of Mammals and Man. Eds. R. Foley and V. Standen.
London: Basil Blackwell, 341“366.
Hawkes, K., J. F. O™Connell, and N. G. Blurton Jones. 1991. Hunting income patterns
among the Hadza: Big game, common goods, foraging goals and the evolution of the
human diet. Phil. Trans. R. Soc. Lond. B 334:243“251.

Hawkes, K., J. F. O™Connell, and N. G. Blurton Jones. 1997. Hadza women™s time alloca-
tion, offspring provisioning, and the evolution of long post-menopausal life spans. Cur-
rent Anthropology 38:551“577.
Hawkes, K., J. F. O™Connell, and N. G. Blurton Jones. 1998. Hadza Hunting and the Evo-
lution of Nuclear Families. University of Utah. Ms.
Hawkes, K., J. F. O™Connell, and N. G. Blurton Jones. 2001. Hadza meat sharing. Evolu-
tion and Human Behavior 22:113“142.
Hawkes, K., H. Kaplan, K. Hill, and A. M. Hurtado. 1987. Ache at the settlement: Con-
trasts between farming and foraging. Human Ecology 15:133“161.
Henry, J. 1941. Jungle people: A Kaingang tribe of the highlands of Brazil. New York: J. J.
110 Kaplan and Gurven

Henry, J. 1951. The economics of Pilaga food distribution. American Anthropologist
Hewlett, B., and P. Walker. 1990. Dental health, diet and social status among Central
African Pygmies and Bantu. American Anthropologist 92:383“398.
Hill, K. 1988. Macronutrient modi¬cations of optimal foraging theory: An approach us-
ing indifference curves applied to some modern foragers. Human Ecology 16:157“197.
Hill, K., and K. Hawkes. 1983. Neotropical hunting among the Ache of eastern Paraguay.
In Adaptive Responses of Native Amazonians. Eds. R. Hames and W. Vickers. New York:
Academic Press, 139“188.
Hill, K., and A. M. Hurtado. 1996. Ache Life History: Ecology and Demography of a Foraging
People. New York: Aldine De Gruyter.
Hill, K., and H. Kaplan. 1988. Tradeoffs in male and female reproductive strategies
among the Ache: Part 1. In Human Reproductive Behavior. Eds. L. Betzig, P. Turke, and M.
Borgerhoff Mulder. 277“290. Cambridge: Cambridge University Press.
Hill, K., and H. Kaplan. 1989. Population description and dry season subsistence patterns
among the newly contacted Yora (Yaminahua) of Manu National Park, Peru. National
Geographic Research 3:317“324.
Hill, K., and H. Kaplan. 1993. On why male foragers hunt and share food. Current An-
thropology 34:701“710.
Hill, K., K. Hawkes, H. Kaplan, and A. M. Hurtado. 1987. Foraging decisions among
Ache hunter-gatherers in eastern Paraguay. Human Ecology 12:145“180.
Hoffman, E., K. McCabe, and V. Smith. 1998. Behavioral foundations of reciprocity: Ex-
perimental economics and evolutionary psychology. Economic Inquiry 36:335“352.
Holmberg, A. R. 1969 [1941]. Nomads of the Long Bow: The Siriono of Eastern Bolivia. New
York: Natural History Press.

Hurtado, A. M., K. Hawkes, K. Hill, and H. Kaplan. 1985. Female subsistence strategies
among Ache hunter-gatherers of eastern Paraguay. Human Ecology 13:1“28.

Ichikawa, M. 1983. An examination of the hunting-dependent life of the Mbuti pygmies,
eastern Za±re. African Study Monographs, 4:55“73.
Isaac, G. 1978. The food-sharing behavior of protohuman hominids. Scienti¬c American
Ivey, P. K. 2000. Cooperative reproduction in Ituri Forest hunter-gatherers: Who cares for
Efe infants? Current Anthropology 41:856“866.
Johnstone, R. 1997. The evolution of animal signals. In Behavioural Ecology: An Evolution-
ary Approach. Eds. J. R. Krebs and N. B. Davies. London: Blackwell Scienti¬c, 155“178.
Kaplan, H. 1994. Evolutionary and wealth ¬‚ows theories of fertility: Empirical tests and
new models. Population and Development Review 20:753“791.
Kaplan, H., and K. Hill. 1985. Food sharing among Ache foragers: Tests of explanatory
hypotheses. Current Anthropology 26:223“245.
Kaplan, H. and Robson, A. 2002. The co-evolution of intelligence and longevity and the
emergence of humans. Proceedings of the National Academy of Sciences 99:10221“10226.
The Natural History of Human Food Sharing and Cooperation 111

Kaplan, H., K. Hill, and A. M. Hurtado. 1990. Risk, foraging, and food sharing among the
Ache. In Risk and Uncertainty in Tribal and Peasant Economies. Ed. E. Cashdan. Boulder,
CO: Westview Press, 107“144.
Kaplan, H., K. Hill, K. Hawkes, and A. M. Hurtado. 1984. Food sharing among the Ache
hunter-gatherers of eastern Paraguay. Current Anthropology 25:113“115.
Kaplan, H., K. Hill, A. M. Hurtado, and J. B. Lancaster. 2001. The embodied capital
theory of human evolution. In Reproductive Ecology and Human Evolution. Ed. P. T. Ellison.
Hawthorne, NY: Aldine de Gruyter.
Kaplan, H., K. Hill, J. Lancaster, and A. M. Hurtado. 2000. A theory of human life history
evolution: Diet, intelligence, and longevity. Evolutionary Anthropology 9:156“185.

Kaplan, H. S., Lancaster, J. B., and Robson, A. 2003. Embodied capital and the evolution-
ary economics of the human lifespan. In: Lifespan: Evolutionary, Ecology, and Demographic
Perspectives. Eds. J. R. Carey and S. Tuljapakur. Population and Development Review 29,
supplement, 152“182.
Kent, S. 1993. Sharing in an egalitarian Kalahari community. Man (N.S.) 28:479“514.
Kitanishi, K. 1998. Food sharing among the Aka hunter-gatherers in northeastern Congo.
African Study Monographs 25:3“32.
Kramer, K. L. 1998. Variation in Children™s Work among Modern Maya Subsistence Agri-
culturalists. Ph.D. dissertation. University of New Mexico.
Krebs, D. 1987. The challenge of altruism in biology and psychology. In Sociobiology and
Psychology: Ideas, Issues, and Applications. Eds. C. Crawford, M. Smith, and D. Krebs. Hill-
sdale, NJ: Lawrence Erlbaum Associates, 81“118.

Lancaster, J. B., and C. S. Lancaster. 1983. Parental investment: The hominid adaptation.
In How Humans Adapt. Ed. D. Ortner. Washington, DC: Smithsonian Institute Press, 33“69.
Lancaster, J. B., H. Kaplan, K. Hill, and A. M. Hurtado. 2000. The evolution of the human
life course and investment in human capital. In Perspectives in Ethology: Evolution, Culture
and Behavior, vol. 13. Eds. F. Tonneau and N. S. Thompson. Plenum, NY.
Lee, R. B. 1979. The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge:
Cambridge University Press.
Marlowe, F. n.d. Sharing among Hadza Hunter-Gatherers. Harvard University Depart-
ment of Anthropology. Ms.
Marshall, L. 1976. Sharing, talking, and giving: Relief of social tensions among the !Kung.
In Kalahari Hunter-Gatherers. Eds. R. Lee and I. Devore, 350“371.
Myers, F. 1988. Burning the truck and holding the country: Property, time and the nego-
tiation of identity among Pintupi Aborigines. In Hunter-Gatherers, Volume II: Property,
Power and Ideology. Eds. Tim Ingold, David Riches, and James Woodburn. Oxford: Berg,
Peterson, J. T. 1978 [1974]. The Ecology of Social Boundaries: Agta Foragers of the Philip-
pines. Urbana: University of Illinois Press.
Peterson, N. 1993. Demand sharing: Reciprocity and the pressure for generosity among
foragers. American Anthropologist 95:860“874.
Price, J. A. 1975. Sharing: The integration of intimate economies. Anthropologica 17:3“27.
112 Kaplan and Gurven

Prost, G. 1980. Chacobo: Society of Equality. University of Florida. M.A. thesis.
Ritchie, M. A. 1996. Spirit of the Rainforest. Chicago: Island Lake Press.
Robson, A., and Kaplan, H. 2003. The evolution of human life expectancy and intelli-
gence in hunter-gatherer economies. American Economic Review 93:150“169.
Rogers, A. R. 1993. Why menopause? Evolutionary Ecology 7:406“420.
Rothstein, S. I., and R. Pierroti. 1988. Distinctions among reciprocal altruism, kin selec-
tion, and cooperation and a model for the initial evolution of bene¬cient behavior. Ethol-
ogy and Sociobiology 9:189“209.

Sahlins, M. 1972. Stone Age Economics. London: Tavistock.
Shostak, M. 1981. Nisa: The Life and Words of a !Kung Woman. New York: Random House.
Silberbauer, G. 1981. Hunter/gatherers of the central Kalahari. In Omnivorous Primates.
Ed. R. Harding and G. Teleki. New York: Columbia University Press, 455“498.
Silk, J. 1979. Feeding, foraging, and food sharing of immature chimpanzees. Folia Primato-
logica 31:123“142.
Simon, H. 1990. A mechanism for social selection and successful altruism. Science

Singleton, S. 1998. Constructing Cooperation: The Evolution of Institutions and Comanage-
ment. Ann Arbor: University of Michigan Press.

Smith, E. A. 1988. Risk and uncertainty in the ˜˜original af¬‚uent society™™: Evolutionary
ecology of resource-sharing and land tenure. In Hunter-Gatherers, Volume 1: History, Evo-
lution and Social Change. Eds. Tim Ingold, David Riches, and James Woodburn. New
York: Berg, 222“251.
Smith, E. A., and R. Bliege Bird. 2000. Turtle hunting and tombstone opening: Public gen-
erosity as costly signaling. Evolution and Human Behavior 21:245“261.
Sosis, R. H. 1997. The Collective Action Problem of Male Cooperative Labor on Ifaluk
Atoll. University of New Mexico. Ph.D. thesis.
Sosis, R., S. Feldstein, and K. Hill. 1998. Bargaining theory and cooperative ¬shing partic-
ipation on Ifaluk Atoll. Human Nature 9(2):163“203.

Speth, J. D. 1990. Seasonality, resource stress, and food sharing in so-called egalitarian
foraging societies. Journal of Anthropological Archaeology 9(2):148“188.
Stahl, I. 1972. Bargaining Theory. Economic Research Institute at Stockholm School of
Economics, Stockholm.
Stanford, C. B. 1995. Chimpanzee hunting behavior and human evolution. American Sci-
entist 83:256“261.
Stearman, A. M. 1989. Yuqu± foragers in the Bolivian Amazon: Subsistence strategies,
prestige, and leadership in an acculturating society. Journal of Anthropological Research
Tanaka, J. 1980. The San Hunter-Gatherers of the Kalahari. Tokyo: University of Tokyo
The Natural History of Human Food Sharing and Cooperation 113

Trivers, R. L. 1971. The evolution of reciprocal altruism. Quarterly Review of Biology
Turnbull, C. M. 1965. Wayward Servants: The Two Worlds of the African Pygmies. Westport,
CT: Greenwood Press.
Wilson, D. S. 1990. Weak altruism, strong group selection. Oikos 59:135“140.

Wilson, D. S. 1998. Hunting, sharing, and multilevel selection: The tolerated-theft model
revisited. Current Anthropology 39:73“97.
Winterhalder, B. 1986. Diet choice, risk, and food sharing in a stochastic environment.
Journal of Anthropological Archaeology 5:369“392.
Winterhalder, B. 1996. A marginal model of tolerated theft. Ethology and Sociobiology
Winterhalder, B. 1997. Social foraging and the behavioral ecology of intragroup resource
transfers. Evolutionary Anthropology 5:46“57.
Woodburn, J. 1982. Egalitarian societies. Man 17:431“451.
Zahavi, A., and A. Zahavi. 1997. The Handicap Principle: A missing Piece of Darwin™s Puzzle.
New York: Oxford University Press.
4 Costly Signaling and
Cooperative Behavior

Eric A. Smith and Rebecca
Bliege Bird

There is, deep down within all of us, an instinct. It™s a kind of drum major instinct”a
desire to be ¬rst . . . We all want to be important, to surpass others, to achieve distinc-
tion, to lead the parade . . . Don™t give it up. Keep feeling the need for being ¬rst. But I
want you to be ¬rst in love. I want you to be ¬rst in moral excellence. I want you to be
¬rst in generosity.
(From a sermon by Dr. Martin Luther King, Jr.)

4.1 Introduction

The last few decades have witnessed an increasing convergence and
interaction between economic and evolutionary approaches to human
behavior, a trend certainly exempli¬ed in the present volume. In this
chapter, we draw on a framework we will refer to as costly signaling
theory (CST) that has been elaborated more or less independently in
both economics (e.g., Veblen 1899; Spence 1973) and evolutionary biol-
ogy (e.g., Zahavi 1975; Grafen 1990). In keeping with the theme of the
present volume, we explore the ways in which CST might illuminate
strong reciprocity and other forms of cooperative behavior.1 In contrast
to most of the contributors to this volume, we argue that many of the
phenomena classed as strong reciprocity (as de¬ned in chapter 1)
might be individually optimal (i.e., produce a net ¬tness bene¬t) and
thus not require cultural or genetic group selection, at least not for
their evolutionary origins.
This chapter is organized as follows. Section 4.2 summarizes the fun-
damental features of costly signaling theory, and section 4.3 outlines a
game-theoretical model of cooperative behavior based on this theory.
Section 4.4 discusses the conditions under which we might expect
group-bene¬cial signaling to be favored over neutral or ˜˜sel¬sh™™
116 Smith and Bliege Bird

signaling. We then apply these arguments to a variety of contexts in
which cooperative behavior is commonly observed, and for which
standard models of conditional reciprocity seem inadequate. First, we
consider cases of unconditional generosity involving the provisioning
of collective goods, such as public feasts or ¬ghting on behalf of one™s
community. Section 4.5.5 considers the special, but crucial, case of
enforcement of group-bene¬cial norms and punishment of those who
defect from them. We then discuss ways in which CST may illuminate
situations involving trust and commitment (section 4.6). In each of
these sections, we present a variety of ethnographic and historical
examples that illustrate the application of CST to understanding coop-
erative behavior. Section 4.7 offers a brief set of conclusions that both
review the material presented in this chapter and suggest the areas
where major questions remain.

4.2 Costly Signaling Theory

Costly signaling theory proposes that expensive and often seemingly
arbitrary or ˜˜wasteful™™ behavioral or morphological traits are designed
to convey honest information bene¬ting both signalers and observers
(Zahavi 1975; Grafen 1990; Johnstone 1997). These signals reveal in-
formation about underlying qualities of the signaling individuals (or
groups). By ˜˜qualities,™™ we mean characteristics of the signaler that are
of importance to observers (i.e., elements that will affect their payoffs
from social interaction with the signaler), but that are directly observ-
able only with dif¬culty or not at all (e.g., disease resistance, competi-
tive ability, resource endowment, dedication to an ongoing social
relationship). Readers unfamiliar with CST should note that it is rele-
vant to a much wider range of behavioral and morphological features
than are considered here (see for example Johnstone 1995; Zahavi and
Zahavi 1997).
There are two key conditions required for evolutionary stability of
such signaling. First, both signalers and receivers must bene¬t from
sharing information about signaler variation in the underlying quality.
The second condition is that signals impose a cost on the signaler that
is linked to the quality being advertised. This link can take one of two
forms: either lower-quality signalers pay higher marginal costs for sig-
naling or they reap lower marginal bene¬ts. These two conditions are
related, since quality-dependent cost (the second condition) serves to
Costly Signaling and Cooperative Behavior 117

ensure that the signal honestly advertises the relevant underlying qual-
ities of the signaler (the ¬rst condition).
CST provides a powerful framework for explaining how honest
communication can be evolutionarily stable despite the pervasive con-
¬‚icts of interest generated by natural selection. When the conditions
outlined above are met, honest signals will be of bene¬t to both sig-
naler and observer, even when their interests overlap very little. The
payoff to the observer derives from the information inferred from the
signal”he or she should be able to evaluate the signaler™s qualities as
competitor, mate, or ally by attending to the signal rather than through
more costly means of assessing the signaler™s abilities, qualities, or
motivations. The payoff to the signaler results from the observer™s re-
sponse. Note that the mutuality of interest in information sharing
can exist even when in a broad sense signaler and observer have
strongly opposed interests and hence incentives to engage in deceit”
for example, interactions between predator and prey, or between en-
emy soldiers.
It bears emphasizing that the logic of CST is not based on standard
conditional reciprocity (see table 4.1). For example, when we say that
signal observers may use the information they have received to choose
someone as a (future) ally, we are not proposing that this is a favor
reciprocated to the signaler, any more than a peahen that chooses the
peacock with the showiest tail is ˜˜paying back™™ the cock for having
expended high signaling costs. Rather, CST explanations propose that
responding to signals in a way that bene¬ts the signaler is simply the
best move the responder can make given the available information.
The mere fact that a costly action (e.g., hosting an expensive feast)
results in a bene¬cial response (e.g., an increase in social status) does
not entail conditional reciprocity. It is important to keep this distinction

Table 4.1
Comparison of conditional reciprocity and costly signaling accounts of cooperation.
Are features below expected with: Conditional reciprocity? Costly signaling?
Donor obtains net gain in the long run Yes Yes
Donor is paid back by recipients Yes Not necessarily
Unilateral provisioning of public good No Possibly
Donors have higher status than recipients No Yes
Requires punishment of free riders Yes No
Stability less likely with larger group size Yes No
118 Smith and Bliege Bird

in mind when considering the special case of group-bene¬cial

4.3 Group-Bene¬cial Signaling

In most cases, CST is applied to contexts where the bene¬ts in question
are privately consumed (e.g., mating opportunities) and any wider so-
cial bene¬ts absent or incidental. In principle, signaler-observer rela-
tions can range from highly cooperative to blatantly antagonistic, as
in the case of prey signaling their vigor to predators (Caro 1994), or
individuals or social groups competing for social dominance (Neiman
1997). The situation that concerns us here is when costly signaling
ensures that competitors for various social goods (e.g., alliances, mat-
ing opportunities, leadership positions) advertise their relevant qual-
ities honestly, thus allowing observers to discriminate amongst the
signalers and make their best move (such as ally with, mate with, or
defer to those signaling more often or more intensely).
Several authors (Zahavi 1977, 1995; Boone 1998; Roberts 1998;
Wright 1999) have proposed that costly signaling could provide an ex-
planation for cooperation and group-bene¬cial behavior. In an earlier
set of papers (Bliege Bird, Smith, and Bird 2001; Smith and Bliege Bird
2000), we argued that unconditionally providing a collective good
when it was otherwise not in the provider™s best interest to do so could
be favored if such provisioning served as a reliable signal of the pro-
vider™s quality. Those who provide this group bene¬t, or who provide
more of it (i.e., signaling more intensively), assume costs greater than
their personal share of the collective good, but in doing so honestly ad-
vertise their quality as allies, mates, or competitors. This information
could then alter the behavior of other group members to act (for purely
sel¬sh motives) in ways that provide positive payoffs to signalers”for
example, preferring them as allies or mates, or deferring to them in
competitive situations (Smith and Bliege Bird 2000).
A formal model of this proposal, framed as an n-player public goods
game, has been developed by Gintis, Smith, and Bowles (2001); we will
refer to this as the GSB model.2 In this model, cooperation involves
providing a bene¬t to all members of the group regardless of any re-
ciprocation in kind. Given the public goods game payoff structure
and non-repeated interactions, the unique equilibrium of this game
involves universal defection as the dominant strategy, and hence in-
dividually costly cooperation could not evolve (unless there were
Costly Signaling and Cooperative Behavior 119

strong group selection in its favor). Even if interactions among group
members were repeated, cooperation among more than a few indi-
viduals would require implausible forms of coordination (Boyd and
Richerson 1988). The GSB model is meant to apply to such cases,
where conditional reciprocity is unlikely to emerge and is vulnerable
to free-riding.
It seems reasonable to suppose, however, that providing the group
bene¬t serves as an honest signal of the provider™s underlying quality
(as de¬ned in section 4.2). Speci¬cally, suppose that providing the
group bene¬t is differentially costly as a function of the provider™s
quality. For simplicity, GSB assume that members of the social group
come in two types, high quality and low quality. The model further
assumes that every individual knows his or her own quality (but not
that of others) and that any other group member has probability p of
being high quality (and probability q ¼ 1 À p of being low quality).
In the GSB game, each member plays two roles in any given period:
signaler and responder. The signaler role takes two forms: providing
the collective bene¬t (e.g., hosting a feast) or not providing it. The
responder role consists of observing signalers (including partaking in
any collective bene¬ts they may provide) and then making a decision
whether or not to interact with one of them. This interaction, like the
signal, is stated in the most general terms here, but could involve such
things as mate choice, coalition formation, partner choice, deference in
competitive situations, and so on.
With these options, in each role a player can use one of four strat-
egies, as listed in table 4.2. Speci¬cally, signalers can chose to signal
(provide the collective bene¬t) (1) always, regardless of their quality;

Table 4.2
Strategies in the n-person signaling game.
AS ¼ always signal, regardless of quality
SH ¼ signal only if one is high quality
SL ¼ signal only if one is low quality
NS ¼ never signal
AR ¼ always respond, whether or not signaler signals
RS ¼ respond by interacting only with a signaler who signals
RN ¼ respond by interacting only with those who do not signal
NR ¼ never respond
120 Smith and Bliege Bird

or they can make signaling conditional on their type”signaling (2)
only if high quality, or (3) only if low quality; or (4) decide to never
Similarly, responders can interact with an individual chosen at ran-
dom (1) from all the other n À 1 group members; (2) from the subset of
other members who provided the bene¬t; (3) from the subset of other
members who did not provide the bene¬t; or (4) the responder can
forgo interacting with any group member in this period.
A signaling equilibrium will occur if all players chose to a) signal
only if high quality, and b) respond by interacting only with those
who signal. Following the labels in table 4.2, this means that all play
˜˜SH™™ as signalers, and ˜˜RS™™ as responders. To determine if this signal-
ing equilibrium will be favored (i.e., if it will be a strict Nash equilib-
rium), we need to specify some assumptions about payoffs from the
various strategies. First, following the standard logic of CST, we as-
sume that high-quality individuals pay a lower cost to signal than
low-quality ones and that interacting with high-quality individuals
will yield a higher payoff to responders than if they interact with low-
quality individuals. We also assume that any signaler who interacts
with a responder will gain a positive bene¬t from this interaction; in
the GSB model, this bene¬t is the same irrespective of the signaler™s
type and regardless of whether or not the Signaler in fact signaled
(provided the collective good).3
These assumptions produce the payoff matrix outlined in table 4.3
(for a full explication, see Gintis, Smith, and Bowles 2001). The analyti-
cal results discussed in GSB reveal that three conditions are necessary
and suf¬cient for honest signaling (SH, RS) to be a strict Nash equilib-
rium. First, the bene¬ts of signaling must exceed its expected cost for
the high-quality type. Second, the opposite must hold for low-quality
types. Finally, responders must gain greater bene¬ts from interacting
with a high-quality type than with a low-quality type. Note that these
conditions are essentially the minimal assumptions needed to apply a
costly signaling framework.
In addition, as long as p (responder™s payoff from interacting with a
high-quality individual) þ q (responder™s payoff from interacting with
a low-quality individual) > 0, there is a non-signaling equilibrium
(NS, AR) in which no one signals and responders choose randomly
from all other group members. Similarly, if the above inequality is
reversed, there is a non-signaling equilibrium (NS, NR) in which no
Costly Signaling and Cooperative Behavior 121

Table 4.3
Payoff matrix for the n-person signaling game. Adapted from Gintis, Smith, and Bowles
2001. See table 4.2 for key to strategy abbreviations.
s À pc À qc 0 s/p À pc À qc 0 Àpc À qc 0 Àpc À qc 0
ph þ ql ph þ ql 0 0
s À pc s À pc s À pc Àpc
ph þ ql h l 0
s À qc 0 qs/p À qc 0 s À qc 0 Àqc 0
ph þ ql l h 0
s s
NS 0 0
ph þ ql ph þ ql
0 0
s ¼ signaler™s payoff from interacting with a responder
c ¼ signaling cost for a high-quality type
c 0 ¼ signaling cost for a low-quality type
h ¼ responder™s payoff from interacting with a high-quality type
l ¼ responder™s payoff from interacting with a low-quality type
p ¼ proportion of n group members who are high-quality types
q ¼ 1 À p ¼ proportion of n group members who are low-quality types

one signals and responders never choose interaction partners. The
GSB analysis indicates that the honest signaling equilibrium will have
higher payoffs than either non-signaling equilibria when, holding all
other parameters of the model ¬xed, (a) high quality types are suf¬-
ciently rare (p is small); (b) the responder™s bene¬t from consuming
the collective good provided by the Signaler is suf¬ciently large; (c)
the advantage of interacting with high quality types is suf¬ciently
large; and (d) the cost of signaling is suf¬ciently small (for high-quality
GSB also show that the form of signaling outlined in the previous
paragraphs will proliferate when rare and be evolutionarily stable, as
long as the cost of signaling is suf¬ciently greater for low-quality than
for high-quality players, and high-quality individuals are neither too
common nor too rare. The reason for the latter condition is that if
high-quality individuals become too common ( p is very high), res-
ponders have a very high probability of interacting with such indivi-
duals even if they choose randomly, and thus those who avoid the
costs of signaling will still have a high probability of being chosen for
bene¬cial interactions. GSB provide an analysis showing that p will at-
tain an equilibrium value under a range of plausible conditions.
122 Smith and Bliege Bird

In summary, the n-player costly signaling model developed by Gin-
tis, Smith and Bowles (2001) shows that cooperative acts can function
as ordinary costly signals and be favored by selection acting on either
cultural or genetic variation. Over a broad range of parameter values,
honest signaling of high quality by providing collective bene¬ts is a
strict Nash equilibrium, and a large basin of attraction grants it robust
evolutionary stability. The conditions for this equilibrium are simply
that (a) low-quality types pay greater marginal signal costs than do
high-quality types; (b) other group members bene¬t more from in-
teracting with high-quality than with low-quality types; and (c) this
interaction provides bene¬ts to high-quality signalers that exceed the
signaling cost.
All of these results, however, apply equally to ordinary noncoopera-
tive signals, and thus the GSB model in itself speci¬es only necessary,
but not suf¬cient, conditions for understanding why cooperative sig-
naling might be favored over other forms with equivalent individual
costs and bene¬ts. The remainder of this chapter examines this last is-
sue, both theoretically and empirically.

4.4 Why Group-Bene¬cial Signaling?

Honest signaling of quality need not be bene¬cial to the signaler™s
group. Indeed, the GSB model applies equally well to socially neutral
or harmful forms of costly signaling. This raises the question of why
costly signaling should ever take the form of providing collective
goods. After all, in other species such signaling generally involves
displays such as peacock™s tails, roaring contests between red deer, or
ritualized struggles between male elephant seals, which provide no
overall group bene¬ts. Furthermore, there appear to be numerous hu-
man examples of such socially wasteful displays: foot-binding, head-
hunting, various forms of conspicuous consumption, duels, violent
brawling, and even the conspicuous ¬‚outing of social norms.
We can think of three possible answers to this question. One”
invoking group selection among alternative evolutionarily stable equi-
libria (Boyd and Richerson 1990)”will be discussed brie¬‚y in a later
section of this chapter. First, we explore two other explanations in
greater detail, one involving the superiority of collective goods in
attracting a larger audience and the other proposing that such provi-
sioning is a superior signal of group-bene¬cial qualities (i.e., that coop-
eration is an intrinsic element of the qualities being signaled).
Costly Signaling and Cooperative Behavior 123

4.4.1 Broadcast Ef¬ciency and Signal Escalation
Because signals evolve not only to convey honest information, but also
to attract the attention of observers, advertising levels can escalate as a
result of competition among signalers over such attention (Arak and
Enquist 1995; Guilford and Dawkins 1993). Signal design may thus be
directly related to competition over observer attention. This process
could transform a socially neutral signal, such as an individual show-
ing off his skill by spearing a few small ¬sh, into a socially bene¬cial
one, such as investing in construction of a stone ¬sh weir allowing
hundreds of kilograms of ¬sh to be caught and shared throughout the
community. Providing larger amounts of food than a competitor for
˜˜no-strings-attached™™ public consumption will tend to attract more
attention from more observers (Hawkes 1993). This argument could
easily be generalized to a wide range of public goods and correspond-
ing appetites.
Put another way, one of the advantages to the signaler of providing a
collective good over some more ˜˜wasteful™™ display of handicap may lie
in the broadcast ef¬ciency of the signal (Smith and Bliege Bird 2000).
By ˜˜broadcast ef¬ciency,™™ we mean the number of observers attracted
per unit of signaling effort. A man who expends a given amount of en-
ergy and risk in ¬ghting with his neighbor might broadcast his abilities
to far fewer people than one who assumes the same costs in publicly
defending his village against an attack. We would expect individuals
to take advantage of any means for increasing broadcast ef¬ciency
when they can bene¬t from increasing the number of observers and
thus to signal by providing collective goods if doing so has such an
effect. Furthermore, we expect that competition among such signalers
will often result in increasing quantities of collective goods being pro-
vided to attract larger audiences (up to some equilibrium level, of
Grafen (1990) has modeled the role of differential quality in setting
levels of competition in costly signaling games. His analysis indicates
that as differences among competitors become more acute (e.g., as
the differences in quality between the best and worst males increases),
the level of advertising effort among all competitors increases corre-
spondingly. Individuals near the low end invest heavily in advertising
to distinguish themselves from slightly worse males; those of higher
quality have no choice but to increase their effort to outdo those below
them. This effect will be strongest when there are many competitors,
especially if quality is continuously varying, rather than discrete. For
124 Smith and Bliege Bird

example, male frogs will call more frequently and produce calls of
longer duration when the number or density of competitive callers
increases (Wells 1988). Levels of advertising tend to spiral upward
under these conditions, in an arms race to outdo one™s competitors.
Note that this broadcast ef¬ciency argument does not reduce to say-
ing that signalers gain bene¬ts by providing goods that attract an audi-
ence. For CST to apply, there must also be a relation between the signal
(in this case, the goods used to attract an audience) and variation in the
underlying qualities being signaled. If signaling were simply a matter
of attracting audiences by supplying collective goods, we would expect
that quantity of goods supplied would be the only relevant dimension.
Yet this fails to account for observations that only certain resources”
and often relatively scarce ones at that”are provided for public con-
sumption. The argument we are making here is that certain types of
collective goods yield greater signal value per unit produced because
they reveal relevant underlying qualities of the signaler. Resources that
are more sensitive to marginal differences in levels of skill, strength,
knowledge, or leadership will allow observers to discriminate amongst
competing signalers in terms of these qualities more effectively.
For example, in a foraging economy, large game (e.g., marine turtles)
is usually harder to locate and capture than smaller, more abundant
game (e.g., sardines) or most plant resources. When the amount of a re-
source harvested does not re¬‚ect differences in underlying quality, the
marginal payoff to the signaler of harvesting enough to provide a pub-
lic good might not be high enough to justify the increased labor costs.
In addition, if observers are interested in qualities such as skill or dedi-
cation, harvests of gathered resources or low-variance game should
generally attract a smaller audience than an equivalent amount of a
more challenging resource that does facilitate such discrimination.
This would further increase the difference in payoff to the signaler
of producing and providing quality-correlated collective goods versus
other resources. This may explain why the marine foragers we have
studied (see section 4.5.3) voluntarily assume higher labor costs and
failure rates to provision a feast with 50 kilograms of turtle meat,
rather than providing 50 kilograms of sardines with greater reliability
and at lower labor costs.

4.4.2 Signaling Cooperative Qualities
Ordinarily, CST views signals as ˜˜indicator traits™™ of underlying qual-
ities, with simply a contingent connection between signal and quality.
Costly Signaling and Cooperative Behavior 125

Thus, a signal such as a peacock™s tail is an indicator of male vigor and
hence (on average) genetic quality; only those cocks who are vigorous,
disease-resistant, and excellent foragers can afford the cost of produc-
ing, maintaining, and dragging around a heavy and showy tail (Petrie
1994). But any trait that reliably indicated genetic quality would serve
as well; there is no inherent reason that peahens should favor showy
tails over some other equally reliable indicator. However, signal ob-
servers may value cooperative traits in themselves. Consequently, such
traits may have intrinsic value to observers that extends beyond their
role as indicator traits.
We can expect that responders will prefer signals that provide a col-
lective good worth G over some equally informative signal that pro-
vides no collective good because, in addition to the gains from the
information transferred in the signal, each of the n responders™ payoffs
will also be increased by G/n. Note, however, that this responder pref-
erence will not be enough to favor group-bene¬cial signaling if (as
assumed earlier in this chapter) the interaction is a one-shot game, and
all group members receive a share of the collective good whether they
ally with the signaler producing it or not. In a more realistic model,
however, group-bene¬cial signals may enhance the signaler™s value to
a potential ally because they strongly predict the signaler™s ability to
produce such signals in the future. For instance an individual who
punishes wrongdoers within the group has honestly signaled his abil-
ity to also punish enemies of the political alliances of which he is part.
Similarly, one who harvests surplus resources and generously shares
them with others rather than conspicuously consuming them person-
ally has honestly signaled his ability to do the same with an ally or
mate. In both cases, we are proposing that a high-quality individual is
more likely to provide the social bene¬t because the cost of doing so is
lower than the cost for a low-quality one. The quality being signaled
might be anything that lowers the cost of behaving in a cooperative
manner, such as superior strength or greater foraging skills.


. 4
( 13)