Scars for war: evaluating alternative signaling explanations for
cross-cultural variance in ritual costs
Richard Sosisa,b,4, Howard C. Kressb
, James S. Bosterb
a
Department of Sociology and Anthropology, The Hebrew University of Jerusalem, Mount Scopus, 91905, Jerusalem, Israel
b
Department of Anthropology, U-2176, University of Connecticut, Storrs, CT 06269-2176, USA
Initial receipt 29 October 2005; final revision received 2 February 2007
Abstract
While males in many societies endure traumatic and painful rites, in other societies male rites are mild or completely absent. To explain
these cross-cultural differences, we use data collected from the Human Relations Area Files electronic databases (eHRAF) to test two sets of
hypotheses derived from signaling theory. If costly male rites serve to signal mate quality, they would be expected to correlate with the
intensity of mating competition. If they serve to signal group commitments, they would be expected to be associated with the importance of
overcoming problems of collective action. Our results support the latter set of hypotheses: males in societies that engage in warfare endure the
costliest rites. Moreover, we show that whether wars are fought within cultural groups or against other cultural groups is an important
determinant of whether or not male rites result in permanent visible marks, such as ritual scars. We argue that costly male rites signal
commitment and promote solidarity among males who must organize for warfare.
D 2007 Elsevier Inc. All rights reserved.
Keywords: Cooperation; Costly signaling theory; Ritual; Religion; Warfare
1. Introduction
Evolutionary researchers have increasingly turned their
attention toward understanding the adaptive significance of
religious behaviors (for reviews, see Bulbulia, 2004a; Sosis
& Alcorta, 2003). One of the significant challenges these
scholars face is explaining the willingness of religious
adherents to perform acts that appear to entail significant
somatic and reproductive costs. These costs seem particularly
acute among male rites, which can include repeated
subincisions, skin hangings, exposure to fire, teeth pullings,
and other traumas. However, not all communities impose
such demands on their boys and men; some are totally
devoid of male rites and others may require nothing more
than the acquisition of clandestine teachings. Why do
societies vary so broadly in the ritual costs they demand
of their males?
Several researchers have previously tested evolutionary
hypotheses aimed at explaining the cross-cultural distribution
of various initiation and puberty rites, including
scarification and tattooing (Low, 1979; Ludvico &
Kurland, 1995; Singh & Bronstad, 1997).1
Motivated by
the logic of sexual selection and Zahavian signaling theory
(Hamilton & Zuk, 1982; Zahavi, 1975), these studies
examined whether males and females use scarification as a
form of mating competition in which performers signal their
resistance to pathogens, willingness to endure pain, or
overall genetic quality. Their results are largely unsupportive
or inconclusive, suggesting the need to explore
alternative explanations. With this goal in mind, here we
examine whether costly male rites serve to reliably signal
group commitments.
Anthropologists have long maintained that one of the
primary functions of ritual is the promotion of group
solidarity (e.g., Durkheim, 1995 [1912]; Rappaport, 1999).
Some researchers have argued that social bonding is not an
end in itself but a means to facilitate intra-group cooperation.
Irons (2001, 2004), for example, posits that the primary
1090-5138/$ – see front matter D 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.evolhumbehav.2007.02.007
4 Corresponding author. Department of Sociology and Anthropology,
The Hebrew University of Jerusalem, Mount Scopus, 91905, Jerusalem,
Israel.
E-mail address: richard.sosis@uconn.edu (R. Sosis).
1
Various researchers not working within a selectionist framework have
also conducted analyses aimed at explaining the cross-cultural distribution
of initiation and puberty rites (e.g., Paige & Paige, 1981; Schlegel & Barry,
1980; Whiting, Kluckholm, & Anthony, 1958; Young, 1965).
Evolution and Human Behavior 28 (2007) 234–247
adaptive benefit of religious behavior is its ability to foster
cooperation and overcome problems of collective action that
humans have faced throughout their evolutionary history.
He emphasizes that the costliness of ritual actions enables
them to serve as honest signals of commitment to the group
because only those who are committed to the group’s beliefs
and goals will be willing to incur the time, energetic, and
opportunity costs of ritual performance. In other words,
individuals pay the costs of ritual performance, but by doing
so they demonstrate their commitment and loyalty to the
group and can thus achieve a net benefit from successful
collective action.
Building upon Irons’s ideas and Cronk’s (1994) original
insights, several researchers have employed recent developments
in signaling theory to explain the emergence and
stability of ritual behavior under broad environmental
circumstances (Bulbulia, 2004b; Sosis, 2003, 2004). To
evaluate these arguments, Sosis and colleagues (Sosis,
2000a; Sosis & Bressler, 2003; Sosis & Ruffle, 2003,
2004; Ruffle & Sosis, 2007) examined the role of ritual
behavior in overcoming the inherent collective action
problems of cooperative labor that communal societies
confront. Their results showed that the frequency of
participation in costly ritual behavior was correlated with
individual levels of cooperation, and among religious
communes, commune longevity was correlated with the
costliness of ritual demands. While these studies focused on
how communities resolve the free-rider dilemmas surrounding
cooperative resource acquisition and consumption,
throughout our evolutionary history individuals have faced
an array of other collective action problems, most notably
warfare (Alexander, 1987; Keeley, 1996; Wrangham &
Peterson, 1996). Research has yet to untangle the relative
importance of these various collective action problems in
providing the ecological stimulus that favored the selection
of costly ritual behavior. Nor has research begun to evaluate
alternative signaling explanations of ritual behavior, namely,
whether rituals serve to advertise mate quality or intra-group
commitment. Here we take the first step toward addressing
both of these questions. In contrast to the pioneering studies
by Low and others (cited above), we focus solely on males
and examine a broad spectrum of rituals that they perform
across societies. We defer an analysis of female rites for
future research because we assume that the determinants of
ritual costs vary according to the ecological problems that
rituals emerge to solve. Distinct mating strategies and gender
roles in activities such as resource acquisition and intergroup
conflict suggest that the salient ecological problems
for males and females are likely to differ.
2. What do male rites signal?
2.1. Group commitment
Bliege Bird and Smith (2005) outline four necessary
conditions for the evolutionary stability of a costly signaling
system2
in a population: (1) there is within-group variance
in some unobservable attribute; (2) observers can benefit
from reliable information about this variance; (3) higherquality
signalers can benefit from accurately broadcasting
this information, but lower-quality signalers have the
potential to achieve benefits at the expense of recipients
through deception; and (4) the cost or benefit to the signaler
of sending the signal is correlated with the signaler’s quality.
Sosis (2005, 2006) has argued that religious behaviors meet
these conditions: (1) the intensity of religious beliefs varies
within communities and this variance is unobservable; (2)
individuals benefit from accurate information about this
variance because intensity of belief is related to one’s
commitment to the group and its goals, committed members
being more likely to be cooperators and thus preferred social
partners; (3) religious groups offer various benefits for
members that are mutually provided and are at risk of
exploitation by those not committed to group goals; and (4)
the cost or benefit of ritual performance is weighed against
opportunity costs that are expected to be higher for skeptics
than for believers since believers will have genuinely
forsaken many behaviors deemed unsuitable by religious
doctrine, while skeptics will not have abandoned these
behaviors.3
Thus, religious behavior can be understood as a
costly signal that reliably advertises the unobservable
condition of religious belief and group commitment. The
time, energetic, material, and opportunity costs of religious
activity serve to deter those who lack sufficient belief from
displaying the signal, consequently increasing trust and
solidarity among signaling group members. High-quality
signalers, in other words those committed to the group, are
assumed to realize reproductive gains via increased reputational
status and group-wide benefits achieved through
successful collective action.
As Irons (2001) notes, if religious behaviors signal
commitment to the group, individuals who have the most to
gain from signaling their group loyalties should be the ones
who exhibit the highest intensity of ritual performance.
These individuals may have the most to gain because they
either receive a higher fraction or are in greater need of the
2
To clarify a potential point of confusion, here we are discussing the
stability conditions for costly signals, not impossible to fake signals, i.e.,
indices. Unlike indices, such as vocal-signal frequency indicating physical
size, stable costly signals can be faked, but they are generally reliable
because it is not in the interest of deceivers to pay the costs of sending a
false signal.
3
There are alternative ways in which signaling theory could be
applicable to religious behavior. For example, if the benefits received are
delayed, signal costs could indicate long-term group commitment since
those faking commitment would generally not be willing to pay ritual costs
because they will not remain in the group long enough to receive the
delayed benefits of group membership (E.A. Smith, personal comment).
Sosis (2006) offers additional ways in which signaling theory could be
applicable to religious behavior, and future work will need to evaluate these
possibilities. Here we assume that one of these alternatives captures the
selective pressures operating on religious behavior, although we are not yet
in a position to determine which alternative is the most accurate.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 235
collective benefits, or they are at greater risk of being
mistaken for a free-rider, possibly as a result of personal
history (e.g., converts) or some phenotypic trait (e.g., skin
color). While there is often intra-group variance in ritual
performance, religious doctrines or group norms establish
minimum levels of ritual performance required for group
membership (Sosis, 2003). Groups that face the greatest risk
of exploitation by free-riders, owing to inherent challenges
in distinguishing committed group members from freeriders,
should on average require and exhibit higher levels of
ritual costs than groups that face lower risk of exploitation
by free-riders. Likewise, if ritual requirements safeguard the
benefits of group membership, required ritual costs should
be higher in groups that offer greater benefits to their
members. In other words, if facilitating cooperation is the
general ecological problem that rituals emerge to solve,
those groups that confront the most significant challenges of
collective action (owing to difficulties in distinguishing
members from free-riders or defending substantial group
benefits) will require the most demanding rituals of their
community members in order to deter free-riders from
exploiting the benefits attained through collective action.
If male rites serve to signal group commitment and
enhance male solidarity, we would expect group size and
genetic relatedness to independently impact the costliness of
ritual activities. Specifically, larger groups will face greater
free-rider problems (Olson, 1965) and they will therefore
experience increased selective pressures favoring costlier
rites that can deter free-riders. Ceteris paribus, we expect
groups composed of related male kin, such as patrilocal
societies, to have less costly male rites than communities
composed of unrelated males because closely related males
have expectations of solidarity on the basis of their kin ties,
and should therefore be less dependent on independent
costly signals of commitment. The impact of group size and
relatedness on the costliness of male rites should be
considerable regardless of the types of cooperative activities
that males must pursue.
Irons (1996) mentions two primary categories of
collective action problems that religious behaviors may
have evolved to solve: resource acquisitions and warfare.
Cooperative hunting and concomitant band-wide food
sharing pose widely recognized free-rider problems and
most anthropological research on these topics has been
aimed at explaining how groups manage these problems
(e.g., Alvard & Nolin, 2002; Gurven, 2004; Sosis, 2000b).
This literature, however, has not examined the possibility
that collective rituals might facilitate these activities by
enhancing intra-group trust and commitment. If costly ritual
actions promote cooperation, we may expect the costliness
of male rites across societies to be correlated with the
importance among males of cooperative resource acquisition
and consumption.
Warfare and group-defense likely pose even greater freerider
problems than these collective activities, owing to high
mortality risks. As Pinker (1997) observes, bA war party
faces the problem of altruism par excellence. Every member
has an incentive to cheat by keeping himself out of harm’s
way and exposing others to greater riskQ (p. 626). The
ethnographic literature on warfare is replete with examples
of men who defect en route to an attack or raid (e.g.,
Chagnon, 1997).4
When defending against an attack,
shirking may be harder to detect or punish, and therefore
also common. Whether attacking or defending, each
individual who defects places the remaining group members
at greater risk of injury or death. Thus, when warfare is
frequent within a society, reliable signals of intra-group
commitment that increase solidarity, such as ritual performance,
should be highly favored by selective mechanisms.5
The types of rituals that will be favored as signals of
commitment will depend upon the nature of warfare
prevalent within a society.6
In societies in which internal
warfare (fought within a cultural grouping) is common,
intermarriage is frequent and communities continually
fission and fuse; thus an enemy one day may be an ally
the next (Chagnon, 1997; Keeley, 1996; Otterbein, 1968).7
Because of the mobility of individuals across groups and
consequent shifting of alliances, individuals within communities
that engage in frequent internal warfare should
not be willing to submit to rituals that leave permanent
markers, such as tattoos or scars, which can signal group
identity. Such markers might hinder their ability to create
or join new groups, or at least minimize their credibility
amongst new group members. Warfare fought against other
cultural groups, referred to as external warfare, poses an
alternative problem. Groups engaged in external warfare
are concerned about uniting unrelated males and fielding
as large a combat unit as possible. However, when
imbalances of power occur within a region, smaller groups
are at risk of their members defecting to larger and more
powerful groups. For these communities, permanent
markers would serve to minimize the ability of men to
abscond to another group.
5
There are conditions, such as sustaining many casualties or difficulty
in recruiting warriors, under which warfare frequency might be negatively
correlated with signal costs (A. Cimino, personal comment), especially if
the costs of defection (to others) are low. We assume that the costs of
defection are generally high, hence our expectation that warfare frequency
is positively correlated with signal costs.
6
Otterbein (1994, p. 37) defines war as barmed combat between two
political communities. . .when political communities within the same
cultural unit engage in warfare, this is considered to be internal war. When
warfare occurs between political communities, which are not culturally
similar, this is referred to as external war.Q
7
As Thorpe (2003, p. 149) remarks concerning the Yanomamo, bthose
they kill are often relatives in closely related villages with whom they have
good relations at other timesQ (also see Ferguson, 2001).
4
Chagnon (1988) also demonstrates that reproductive gains can be
achieved through warfare, which may minimize the collective action
problem of warfare for some men. All potential warriors confront their own
expected payoffs (some probabilistic distribution) based on their fighting
skills. Skilled warriors may face payoffs that resemble coordination games,
whereas others may perceive high incentives for defection.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247236
Although all warfare is dangerous, we believe that external
warfare ultimately poses greater risks than internal warfare.
Internal warfare threatens the lives of individuals; external
warfare threatens the continuation of one’s entire kin group
and community (e.g., Harner, 1972). Because of the greater
inclusive fitness costs associated with defection in external
warfare, as well as the costliness of permanent markers, we
expect external warfare to elicit costlier signals of commitment
than internal warfare. Societies that engage in both
internal and external warfare will face tradeoffs in the costliness
of their commitment signals, especially their willingness
to submit to permanent markers, and we therefore expect
them to maintain rites at an intermediate level of costliness
between societies with only external or internal warfare.
2.2. Mate quality
As mentioned above, rather than signaling group commitment,
male rites may serve to advertise mate quality.
Returning to Bliege Bird and Smith’s conditions for the
evolutionary stability of a costly signal, it is clear that if ritual
behaviors signal mate quality, or characteristics associated
with mate quality, they also meet these necessary conditions.
For example, consider the argument that scarification or
tattoos signal pathogen resistance: (1) males vary in their
resistance to pathogens; (2) since this variance is likely to
correlate with the benefits males can provide directly to
mates (e.g., resource acquisition, protection), as well as the
quality of jointly produced offspring (to the extent this
resistance is heritable), females benefit from accurate
knowledge about this variance; (3) males who have high
pathogen resistance can increase their mating opportunities
by accurately broadcasting this information, but men who are
less resistant can benefit if they are able to deceive females
about their level of resistance; and (4) the cost to males of
exposing themselves to pathogens through ritual tattoos and
lacerations is negatively correlated with their resistance to
pathogens. Alternatively, male rites may also be signaling
ability to endure pain or other aspects of mate quality.
If male rites signal mate quality we expect the intensity
of these rites to increase as the competition for mates
increases. Societies that permit polygyny offer males greater
potential reproductive success than monogamous societies
and consequently there is greater competition for female
parental investment, as well as greater variance in male
reproductive success. In stratified polygynous societies, this
competition will be manifested in resource displays, and
wealthy males are likely to out-compete poorer males (e.g.,
Gaulin & Boster, 1990; Irons, 1979). However, in nonstratified
polygynous societies, where males do not compete
through resource competition, ritual signaling is likely to be
a more important mechanism to communicate mate quality.
3. Hypotheses
In sum, the above arguments suggest the following
hypotheses.
3.1. Demography and kinship
1. The costliness of a society’s male rites will be
positively correlated with community size.
2. Patrilocal societies will have less costly male
rites than societies in which males reside among
unrelated men.
3.2. Resource acquisition and sharing
3. The costliness of a society’s male rites will be
positively correlated with the importance of male
cooperative resource acquisition.
4. The costliness of a society’s male rites will
be positively correlated with the intensity of
food sharing.
3.3. Warfare
5. The costliness of a society’s male rites will be
positively correlated with the frequency of warfare.
6. The type of warfare a society engages in will be a
significant predictor of the costliness of male rites.
Societies with external warfare only will have the
most costly male rites, followed by societies with
both external and internal warfare, then societies
with internal warfare only, and societies without
warfare will maintain the least costly male rites.
7. Societies with external warfare will require their
males to bear more permanent markers than societies
without external warfare.
8. Societies with internal warfare only will require their
males to endure less permanent markers than
societies without internal warfare only.
3.4. Mating
9. Males in polygynous societies will perform costlier
male rites than males in nonpolygynous societies.
10. Males in nonstratified polygynous societies will
perform costlier male rites than males in polygynous
stratified societies.
4. Methods
To examine variation in the costs of male rites, we used
the Probability Sample File (PSF) found in eHRAF, the
web-based version of the Human Relations Area Files. The
PSF includes 60 of the 90 societies available in eHRAF and
is designed to offer a broad temporal and geographic sample
of world cultures that controls for cultural contact between
societies (Ember & Ember, 1998).
Eighteen undergraduate anthropology majors at the
University of Connecticut were recruited to collect data in
eHRAF on male ritual behavior. They were randomly
assigned cultures from the PSF and instructed to read the
appropriate ethnographic literature from eHRAF on each
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 237
society to determine the presence or absence of 19 male
rites, including tattooing, scarification, piercing, circumcision,
subincision, teeth pulling, body painting, and learning
secret knowledge.8
We did not limit data collection to
puberty or initiation rites; students documented all rituals
that were exclusively performed by males.9
All students
were ignorant of the hypotheses being tested.
From these data we wrote brief descriptions of the male
rites performed in each society. We offer two examples of
society descriptions representing the extreme ends of the
costliness distribution10
:
Eastern Toraja: In this society, boys are subincised (a cut
along the bottom of the penis) yearly beginning at 6 years
old and ending at 15. Around age 12, boys are
circumcised in a public ritual. Young boys have their
ears pierced by their mothers. There is a public initiation
rite into manhood where boys are cut on their arms,
hands, and legs, as well as, burned on the torso and arms.
Boys cannot show any pain during the public ceremonies.
Men paint their bodies daily.
Pawnee: Men paint their faces in this society.
A panel of four graduate students rated each society
description, assigning a score of 1 to the societies with the
least costly rites and 4 to those societies whose males
engage in the costliest rites. Inter-rater agreement was fairly
high (average Spearman r =0.72 of all six combinations,
range=0.66–0.81). Graduate students were also ignorant of
the hypotheses, as well as ignorant of the names of the
societies they were rating. A composite score was calculated
by summing the ratings of the graduate students and
subtracting by 3 to produce a score ranging between 1
and 13 (mean=7.38, S.D.=4.86). For example, the Eastern
Toraja received the maximum composite score of 13 (all
students rated as 4) and the Pawnee received the minimum
composite score of 1 (all students rated as 1). Fig. 1 presents
the distribution of costliness rankings for the 60 societies in
our sample.
Students coded the frequencies of external and internal
warfare for five levels following Ember and Ember (1992).
Owing to our small sample size and our lack of confidence
about the ability of HRAF materials to accurately offer fine
distinctions (e.g., whether war occurred once every year or
once every 2 years), data were collapsed to create variables
for the presence and absence of internal and external
warfare, respectively. Data were also recorded on the time
period during which the original author collected information
on warfare and whether or not the rituals described in
the files occurred during the same observation period. Our
descriptions of male ritual activity within each society
describe only the time period for which we have warfare
data and thus do not reflect the full history of each culture.
Students also coded for five levels of male cooperative
resource production and consumption, and collected data on
socialization for cooperation following the coding schema
of Poggie (1995). For reasons discussed above, these codes
were all collapsed to create dichotomous variables.
Data on all other independent variables were obtained
from the World Cultures CD (Gray, 1999). All independent
variables and coding schemes are presented in Table 1.
To test our hypotheses, we conducted ANOVAs with the
8
We included in our analyses only those rites that are expected to be
performed by all men. Furthermore, we did not exclude rites performed on
boys too young to refuse to participate. In these cases, it is not the child but
rather the father or other male kin who may be signaling their quality, or the
quality of their genetic line.
9
We chose to focus on all male rites because we expect that it is the full
suite of required rituals that offer a signal. For example, one who
successfully endures an initiation rite only to ignore subsequent ritual
requirements is not effectively signaling high group commitment or
mate quality.
10
All society descriptions are available from the authors upon request.
Fig. 1. Frequency distribution of cost ratings (N =60).
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247238
Table 1
Coding schema for independent variables
Independent variable Coding schema Analyses
Overall frequency of warfare 1=rare or absent 0=if category 1 or 2
2=warfare seems to occur once every 3 to 10 years 1=if category 3, 4, or 5
3=warfare seems to occur once every 2 years
4=occurs at least once a year
5=constant warfare
Frequency of external warfare 1=rare or absent 0=warfare absent
2=warfare seems to occur once every 3 to 10 years 1=warfare present
3=warfare seems to occur once every 2 years
4=occurs at least once a year
5=constant external warfare
Frequency of internal warfare 1=rare or absent 0=warfare absent
2=warfare seems to occur once every 3 to 10 years 1=warfare present
3=warfare seems to occur once every 2 years
4=occurs at least once a year
5=constant internal warfare
Socialization for cooperation 1=socialization mentioned but cooperation not mentioned 0=if category 1 or 2
2=cooperation mentioned as part of a suite of socialization values 1=if category 3 or 4
3=cooperation is discussed as the most important of values
4=cooperation is clearly identified as the primary social value
Subsistence type 1=foraging 0=nonforaging
2=horticulture 1=foraging
3=pastoral
4=agriculture
Cooperative production 1=male labor done individually 0=if category 1, 2, or 3
2=male cooperative production limited to the family 1=if category 4 or 5
3=some male cooperative production occurs outside the family
4=medium levels of nonfamilial male cooperative production
5=high levels of nonfamilial male cooperative production
Food sharing 1=food acquired by males is never shared 0=if category 1, 2, or 3
2=food acquired by males is only shared during festivals/special occasions 1=if category 4 or 5
3=some food acquired by males is regularly shared outside the family
4=medium levels of food sharing
5=most food acquired by males is regularly shared outside the family
Polygyny 0=polygyny absent
1=polygyny present
Social stratification 0=social stratification absent
1=social stratification present
Residence pattern 1=patrilocal 0=if categories 2–7
2=matrilocal 1=if category 1 or 8
3=ambilocal
4=avunculocal
5=neolocal
6=matrilocal–neolocal
7=ambilocal or neolocal
8=patrilocal–neolocal
Descent pattern 1=patrilineal 0=if categories 2–4
2=matrilineal 1=if category 1
3=ambilineal
4=bilateral
Community size 1V50 1V200
2V100 2V1000
3V200 3V5000
4V500 4N5000
5V1000
6V5000
7V50,000
8 N 50,000
Region 1=Africa 1=Africa
2=Asia 2=Asia
3=Europe 3=Americas and Caribbean
4=Middle East 4=Oceania
5=North America
6=Oceania
7=South America
8=Central America and Caribbean
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 239
Table 2
Data for primary variables in analyses4
Society Region
Community
size
Residence
pattern
Socialization
for
cooperation
Subsistence
type
Cooperative
production
Food
sharing Polygyny
Social
stratification
Overall
warfare
frequency
Cost
rating
Akan 1 2 0 0 1 1 . 1 1 1 7
Amhara 1 2 1 0 0 0 0 0 1 1 7
Andamans 2 1 0 0 1 0 0 0 0 1 11
Aranda 4 1 0 0 1 1 1 1 0 1 12
Aymara 3 3 1 0 0 0 0 1 0 1 5
Azande 1 2 . . 0 0 0 1 1 1 9
Bahia Brazil 3 4 1 0 0 1 0 0 1 0 1
Bemba 1 2 0 0 0 0 0 1 0 0 7
Blackfoot 3 1 0 1 1 1 1 1 0 1 13
Bororo 3 1 0 0 1 1 1 1 0 1 11
Central Thai 2 4 0 1 0 1 0 1 1 0 1
Chukchee 2 1 1 0 1 0 1 1 0 0 9
Chuuk 4 2 0 1 0 0 1 1 0 0 7
Copper Inuit 3 1 0 1 1 1 1 1 0 0 1
Dogon 1 2 1 0 0 0 0 1 1 1 8
Eastern Toraja 2 2 1 1 0 1 0 1 1 1 13
Ganda 1 2 1 0 0 1 0 1 1 1 1
Garo 2 2 0 0 0 0 0 0 0 0 9
Guarani 3 1 0 0 0 1 0 1 0 1 9
Hausa 1 4 1 0 0 1 0 1 1 1 11
Highland
Scotts
. 4 0 0 0 0 0 0 1 0 1
Hopi 3 3 0 1 0 0 0 1 0 1 8
Iban 2 4 0 0 0 0 0 1 0 1 9
Ifugao 2 3 0 1 0 0 0 1 0 1 9
Iroqouis 3 3 0 1 0 1 1 1 1 1 5
Kanuri 1 4 0 0 0 1 1 1 1 1 12
Kapauku 2 2 1 0 0 0 1 1 0 1 9
Khasi 2 4 0 1 0 . . 0 1 0 9
Klamath 3 1 1 1 0 1 1 1 0 1 12
Kogi 3 1 1 1 0 0 1 1 0 0 5
Korea 2 2 1 1 0 1 1 1 1 1 2
Kuna 3 3 0 1 0 1 1 0 0 0 1
Kurds . 4 1 0 0 1 0 1 1 0 6
Lau Fijians 4 4 1 0 0 0 1 1 0 1 10
Lozi 1 3 0 1 0 1 0 1 1 1 8
Lybian
Bedouins
1 1 1 0 0 0 0 1 0 1 6
Maasai 1 2 1 1 0 0 1 1 0 1 13
Mataco 3 1 0 1 1 1 1 0 0 0 9
Mbuti 1 1 1 1 1 1 1 1 0 1 10
Ojibwa 3 2 0 0 0 1 1 1 0 1 5
Ona 3 1 0 0 1 1 1 1 0 1 13
Pawnee 3 2 0 0 0 1 1 1 1 1 1
Saami . 1 0 1 0 1 1 0 0 0 1
Santal 2 2 1 0 0 . 1 1 0 0 10
Saramka 3 2 1 0 0 1 1 1 0 0 9
Serbs . 4 1 0 0 0 0 0 1 0 5
Shluh 1 3 1 1 0 0 0 1 1 1 5
Singhalese 2 3 0 1 0 1 1 1 1 0 1
Somali 1 3 0 1 0 0 0 1 0 1 8
Taiwan
Hokki
2 4 1 1 0 0 0 0 1 1 4
Tarahumarans 3 2 0 0 0 1 0 1 0 0 3
Tikopia 4 3 1 0 0 1 1 1 1 1 9
Tiv 1 3 1 1 0 1 0 1 0 1 13
Tlingit 3 1 0 1 1 1 0 1 1 1 12
Trobriands 4 2 0 0 0 1 1 0 1 0 1
Tukano 3 1 1 0 0 1 1 1 0 1 8
Tzeltal 3 4 0 1 0 0 0 0 1 1 8
Wolof 1 2 1 1 0 1 1 1 1 1 8
Yakut 2 2 0 0 0 1 1 1 0 0 13
Yanoama 3 2 0 0 1 0 1 1 0 1 10
4 See Table 1 for variable definitions.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247240
costliness rating of each society as the dependent variable.
The primary data used in our analyses are presented in
Table 2, and the bars on all figures indicate standard errors.
5. Results
5.1. Regional effects
The 60 societies which comprise the PSF are categorized
into eight regions: Africa (n =16), Asia (n=14), South
America (n=10), North America (n=8), Oceania (n=5),
Central America and Caribbean (n =3), Europe (n =3),
Middle East (n=1). To assess regional effects, we collapsed
North, South, Central America, and Caribbean into one
category and ignored societies from Europe and the Middle
East because of their low representation in the dataset. We
found no relationship between region and our costliness
ratings ( F3,56 =0.31, p =.818). Nor did we find any
relationship between region and any of our independent
variables (polygyny, cooperative labor, food sharing, foraging,
patrilocality, community size, social stratification, or
warfare; analyses not shown). One concern was that certain
ritual practices are associated with specific regions. Indeed,
we found genital mutilations to be more common in Africa
(81.3%) and Oceania (60.0%) than in Asia (21.4%) and the
Americas (0%). We also found piercing to be more common
in Oceania (40.0%) and the Americas (33.3%) than in Asia
(7.1%) and Africa (0%). However, there were no significant
regional effects associated with any other rituals in our
dataset (scarification, tattooing, ingestion of toxic substances,
isolation, and painting).
5.2. Patrilocality and community size
Contrary to our expectation, nonpatrilocal societies do
not have costlier male rites than patrilocal societies
( F1,59 =0.27, p=.61). As a related hypothesis, we examined
whether patrilineal descent was associated with the costliness
of male rituals, but also found no relationship
( F1,60 =0.40, p=.53). Costliness of rituals did not vary
either with community size. There was a nonsignificant
negative trend (opposite of our prediction) of the natural log
of community size estimates in our sample and the
costliness of male rites ( F3,60=1.25, p =.30).
5.3. Cooperative production, food sharing, and
socialization for cooperation
None of our measures of cooperative resource acquisition
and consumption are predictors of the costliness of
male rites: cooperative production ( F1,58=0.20, p =.66),
food sharing ( F1,58=0.37, p=.55), and socialization for
cooperation ( F1,59 =0.12, p =.73). Reliance on cooperative
production and consumption strategies is likely to vary by
subsistence type. We assume foragers maintain higher
levels of cooperation than agriculturalists, pastoralists, and
horticulturalists in order to overcome the considerable
Table 3
ANOVA of cost ratings by foraging and measures of cooperation
Independent variable F-ratio p value
Foraging 5.44 .02
Cooperative production 0.23 .64
ForagingÂCooperative production 0.10 .75
Foraging 6.16 .02
Food sharing 0.28 .60
ForagingÂFood sharing 0.32 .57
Foraging 5.91 .02
Socialization for cooperation 0.35 .56
ForagingÂSocialization for cooperation 0.33 .57
Fig. 2. Cost ratings by overall warfare frequency ( F =12.45, df =1,
p b.001).
Fig. 3. Cost ratings by type of warfare.
Table 4
ANOVA of cost ratings by external and internal warfare (N =60)
Independent variable F-ratio p value
External 7.86 .01
Internal 0.89 .35
ExternalÂinternal 5.90 .02
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 241
variance they often face in daily meat returns. Our data
indicate that foragers do indeed share food more
extensively than nonforagers ( F1,58=5.20, p=.03), yet
they do not significantly differ from other subsistence
groups in their reliance on cooperative production
( F1,58 =1.66, p =.20) or socialization for cooperation
( F1,59=0.03, p=.85). Foragers do exhibit costlier ritual
signals than nonforagers ( F1,60=6.63, p =.01), and foraging
remains a significant predictor of the costliness of
male rites in independent ANOVA models with cooperative
production, food sharing, and socialization for
cooperation, while these other measures of cooperation
remain insignificant (see Table 3).
5.4. Warfare
Overall, warfare frequency is the most significant
predictor of ritual costs in our dataset ( F1,60 =12.47,
p b.001; Fig. 2). The costliness of male rites for societies
in which warfare is present is significantly higher than in
societies without warfare ( F1,60 =7.33, p =.009); this
relationship holds for both external warfare ( F1,43=11.49,
p =.002) and internal warfare ( F1,39=5.95, p =.02). Fig. 3
indicates a trend in the predicted direction: males in
societies that engage in external warfare only perform the
costliest rites, followed by societies that engage in external
and internal warfare, internal warfare only, and, lastly, no
warfare ( F3,60=5.36, p=.003). However, the differences
between internal warfare only, internal and external
warfare, and external warfare only are not significant.
Table 4 shows that there is an interaction effect between
internal and external warfare; when external warfare is
present the presence of internal warfare does not impact the
costliness of rituals.
We predicted that societies with external warfare would
have costlier rites than societies with internal warfare
because we expect permanent markers (which are assumed
to be very costly) to be more prevalent among societies with
external warfare and nonpermanent markers to be more
prevalent among societies with internal warfare only. Our
data offer four categories of permanent markers: genital
mutilations (circumcisions and subinscisions), tattoos, scarifications,
and piercings.11
Fig. 4 shows that all of these
permanent markers are more prevalent when external
warfare is present in a society than when it is not; only
tattooing is not significant. Our data offer three categories of
nonpermanent markers, or specifically, ritual activity that
does not leave a permanent visible result: body and face
painting, ingestion of toxic substances, and periods of
isolation. Fig. 5 shows that all of these nonpermanent
markers are more prevalent in societies that engage in
internal warfare only than in those that do not.
5.5. Polygyny
Polygynous societies have costlier male rites than nonpolygynous
societies ( F1,60 =6.01, p =.017) and nonstratified
societies have costlier male rites than stratified societies
( F1,60 =6.85, p =.011). There is no interaction effect
between these variables ( F3,60 =.06, p=.81). Polygyny,
however, does not remain a significant predictor of ritual
11
We did not include teeth pulling and filing in the analysis because,
although damaging teeth is permanent, those who did not witness the
ritual would be unable to distinguish natural teeth loss and decay from the
ritual act. In addition, only five societies ritually pulled or filed teeth.
Consistent with our predictions though, four of these societies engaged in
external warfare.
Fig. 4. Frequencies of permanent markers by presence and absence of external warfare.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247242
costs when controlling for subsistence type ( F2,60=1.03,
p=.31) or overall warfare frequency ( F2,60=1.60, p=.21).
Social stratification remains a significant negative predictor
when controlling for warfare ( F2,60 =10.26, p=.002) but not
subsistence type ( F2,60=0.76, p=.39).
6. Discussion
Our results suggest that nonstratified foraging societies
that engage in warfare maintain the costliest male rites.
Moreover, the types of warfare in which communities
engage influence the costs and variety of rituals performed.
External warfare is associated with permanent markers,
whereas internal warfare is associated with ritual activities
that do not leave permanent visible signs. While polygynous
groups exhibit costlier rites than nonpolygynous groups, this
relationship is not significant when subsistence type or
warfare frequency is included in the model, suggesting that
male rites do not vary as a function of mating competition.
Ember and Ember (1992) have shown that warfare
frequency is a predictor of polygyny. Nonetheless, among
polygynous societies in our sample, warfare frequency
remains a significant predictor of the costliness of male
rites ( F1,47=5.24, p =.027). Among nonpolygynous societies,
high levels of warfare are also associated with costlier
rites, but this difference is not significant (average costliness
rating 7.5 vs. 4.1; n=13, t =1.75, p=.12).
Contrary to our prediction, patrilocality is not related to
ritual costs, even controlling for internal warfare frequency,
which has been shown to be correlated with patrilocality in
other cross-cultural research (Ember & Ember, 1971).12
Community size is not a significant predictor either of
variance in ritual costs. Small communities may actually
have costlier rites than larger ones ( F1,60=3.39, p=.07),
although when controlling for overall warfare frequency this
relationship is not significant ( F2,60=1.52, p=.22). In our
analyses, community size served as a proxy for the size of
the group that could potentially engage in collective action.
Future work will require more accurate measures and will
need to examine the relationship between ritual costs, group
size, and political systems. Societies such as nation-states
and chiefdoms, which tend to have large community sizes,
can coerce participation in warfare or other collective
activities through legitimate punishment threats (e.g., fines,
imprisonment). Thus, contrary to our original prediction,
these communities may be less dependent on ritual signs of
commitment than small egalitarian societies because they
can effectively punish defectors.
Cooperative production, food sharing, and socialization
for cooperation were not significant predictors of variance in
ritual costs either, although foragers exhibit costlier rites
than nonforagers, which is marginally significant when
controlling for overall warfare frequency ( F2,60 =3.27;
p=.07). We suspect that cooperative production and food
sharing are unrelated to ritual costs because in these pursuits
reciprocity can be employed to prevent free-riders from
accruing long-term benefits. In contrast, the life and death
stakes men face during each act of warfare or defense
demand a commitment mechanism that does not solely rely
on expectations of future cooperation, since men may not be
alive to reciprocate. Costly rituals offer such a mechanism,
Fig. 5. Frequencies of nonpermanent rites by presence and absence of internal warfare.
12
In our data, a 2Â2 Pearson chi-square model did not reveal a
significant relationship between the presence and absence of internal
warfare and patrilocal residence (v2
=2.13, df =1, p =.145). However,
analyzing the originally coded internal warfare data (five-point scale, see
Table 1) indicates a strong relationship between internal warfare frequency
and patrilocal residence ( F1,59 =8.95, p =.004).
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 243
which may explain why they are associated with warfare but
not resource acquisition and consumption.
Our analyses further demonstrated that permanent
markers are more common when external warfare is
prevalent, and rites that do not leave permanent signs are
more common when internal warfare is prevalent. These
findings should be interpreted cautiously, however, since
we were unable to control for overall warfare frequency.
All societies that engaged in external warfare also
experienced high frequencies of overall warfare, whereas
only 40% of those societies that engaged in internal warfare
only were reported to sustain high levels of overall warfare.
Since overall warfare frequency is positively correlated
with ritual costs (see above) and permanent markers are
assumed to be costlier than rites that do not leave visible
signs, our results may be spurious. We believe the
consistent pattern found across every ritual category we
examined provides evidence for our predictions concerning
warfare type and marker permanence; however, a dataset
with greater variance in the relevant variables is needed to
further explore these predictions.
It is also important to clarify that our results do not
demonstrate that costly male rites genuinely signal the
group commitments of individuals. As Maynard Smith and
Harper (2003) explain, for a signal to classify as a Zahavian
handicap the net benefits for displaying the signal must be
higher for a high-quality individual (in our case, someone
strongly committed to the group) than a low-quality
individual (someone weakly committed to the group). This
could mean that the costs are higher for low-quality
individuals, that the benefits are higher for high-quality
individuals, or both. Obviously, this would be extremely
difficult if not impossible to rigorously assess in an HRAF
study. Here we offer no precise measures of benefits (we
assume benefits accrue from status and successful collective
action) and costs were measured by outsiders on a relative
scale, and not in a currency that is directly related to fitness
(discussed further below). Indeed, given the rigorous
standards of evidence needed, with the exception of several
foraging studies (Bliege Bird, Smith, & Bird, 2001; Smith,
Bliege Bird, & Bird, 2003), few costly signals have been
convincingly demonstrated in the human evolutionary
literature. Nonetheless, our goal here was to explain
variance in the costliness of male rites and signaling theory
offered clear and testable predictions, as well as a
compelling interpretation of our findings.
We believe that costly male rites are an adaptive response
that emerges among warring communities to reduce
free-riding and promote cooperation during warfare
(cf. MacNeill, 2004). However, our analyses cannot entirely
eliminate alternative hypotheses, and indeed, it would be
imprudent to assume that one causal variable is responsible
for a set of behaviors as complex as male rituals. For
example, scars and paint may serve to make their bearers
look fierce or, alternatively, they may help distinguish
friends from foes on the battlefield. Even if warfare is the
primary selective force for the evolution of costly male rites,
the solidarity achieved through ritual actions may also
impact other collective endeavors. Likewise, although our
results did not support the hypothesis that costly male rites
serve to signal mate quality, male rites may be one of the
many arenas in which females evaluate males even if this is
not their main function.13
The ethnographic literature on bfailedQ rites is instructive
and suggests that costly male rites do indeed serve multiple
functions. For example, while Crocker and Crocker (1994)
interpret Canela ear piercing as symbolizing the importance
of obedience in a quasi-military society, Canela men claim
that pierced ears attract women. During the ear piercing
ceremony boys must remain motionless and show no pain,
or they will be publicly shamed, thus reducing their mating
prospects. In his cross-cultural survey on manhood, Gilmore
(1990) describes how women regard males who failed to
correctly perform rigorous initiation rites as unsuitable
mates, whereas men consider such individuals to be
unreliable and untrustworthy. Aside from illustrating that
rites may serve multiple functions, these examples suggest
that coerced initiations might have different signaling value
than those which are willingly performed, or at least
correctly performed, even if begrudgingly. Coercion and
improper expression of suffering, however, do not always
result in a failed signal of commitment (e.g., Bulbulia, in
press; Rappaport, 1999). In many boyhood rituals, such as
subincision or circumcision, since the father (or male kin)
rather than the boy is the signaler, expressions of pain are
often tolerated and even expected (Paige & Paige, 1981).
Moreover, once a marker exists, opportunity costs for outgroup
activities increase, thus enhancing the signal value of
the marker (Iannaccone, 1992).
We interpret the relationship between ritual costs and
warfare in our data as support for the claim that free-riding
during warfare is the foremost collective action problem that
costly male rites are designed to cope with. Alternatively, it
could be argued that male ritual violence is actually a cause,
rather than a consequence, of warfare. In other words, rather
than a mechanism to create male solidarity, ritual violence
during childhood produces violent men, and one manifestation
of this violence is warfare. In several cross-cultural
studies, Chick and colleagues (Chick & Loy, 2001; Chick,
Loy, & Miracle, 1997) found an association between
socialization for aggressiveness and warfare frequency,
although no causal relationship could be determined in
these studies either (also see Ember & Ember, 1994). One
limitation of this argument, however, is that it cannot
account for the relationships we predicted and found
between external warfare and permanent ritual markers
and internal warfare and rituals that do not leave permanent
13
Johnstone and Norris (1993) similarly argue that badges that
honestly signal status and serve to settle contests may also be used by
females when choosing a mate.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247244
visible marks. Future work should aim to distinguish
between these causal interpretations, possibly exploring
whether ritual costs are correlated with the frequency of
male intra-group violence against women or other men, as
would be expected if ritual violence causes boys to become
aggressive men.
This discussion raises an additional question: If ritual
violence is not a cause of warfare (but the converse, as we
believe), why are male rites so violent and traumatic?
Signaling theory may explain variation in ritual costs, but
it cannot explain how the specific form of ritual costs
varies across environments. Many rituals entail high
energetic and material costs (dancing, charity, etc.), but
most of the male rites considered here are physically
dangerous and painful. To explain why, we need to
consider the proximate mechanisms that underpin these
behaviors. In contrast to the positive affect induced by
ecstatic religious ritual, such as dance and chanting, the
male rites described in our data evoke intense negative
affective responses, including fear, pain, and awe. While
both positive and negative affect rituals appear to promote
solidarity, the neuropsychological impacts of these different
ritual forms offer some insight as to why male initiation,
puberty, and other rites are emotionally startling rather than
pleasing (Alcorta, 2006; Alcorta & Sosis, 2005). Increased
activation of the amygdala through fear, pain, and
alterations in body state, can result in the conditioned
association of arbitrary stimuli with heightened emotional
significance (Damasio, 1994). Through frightening and
painful rites, religious symbols can acquire deep emotional
significance that subsequently unites individuals who
shared the experience. The emotional impact of these rites,
many of which occur during critical developmental stages,
has profound long-term effects on memory and is
motivationally powerful (Alcorta & Sosis, 2005; McCauley
& Lawson, 2002; Whitehouse, 2000). The superior ability
of traumatic rituals to create enduring emotional bonds, in
comparison to positive-affect rituals, may explain why
traumatic rituals are associated with warfare, which even
when frequent is unpredictable and can occur many years
after the ritual experience. Our results suggest that these
rituals generate solidarity between men and serve as
reliable indicators of group commitment, thus reducing
the likelihood that men will defect when there is war.
7. Conclusion
This study is a modest step toward understanding crosscultural
variance in the costs of male ritual behavior. We
have made little effort to understand the psychology
underlying these behaviors, but this will need to be
addressed in future research. Future work must also extend
these analyses to female ritual behavior, which we suspect
often signals commitments to mates rather than to the larger
group (e.g., Knight, Power, & Watts, 1995; Strassmann,
1996). In addition, evolutionary research on religion must
further examine how individuals weigh various ritual costs
across currencies, a problem that has also arisen in other
tests of the costly signaling theory of ritual (e.g., Sosis &
Bressler, 2003). Estimating fitness costs of ritual activities is
challenging for both researchers and ritual participants. It is
unclear how individuals determine, for example, how much
costlier a ritual back scar is than a week of isolation or a day
of fasting. Here we used judgments of graduate students
who had not performed the rituals they were asked to rate,
but ideally we need to evaluate how ritual performers and
nonperformers differ in their judgments about ritual costs,
which will offer insights into the proximate psychology of
ritual behavior.
Religious behavior has often been cited as a cause of
warfare; here we have explored the converse, that warfare
can be a cause of extreme religious behavior. We
recommend caution in applying our results to understand
current geopolitical trends and the complex relationship
between religious fundamentalism and warfare. We do
believe that signaling theory can provide insights into the
rise of religious fundamentalism (Sosis, 2003) and its
association with warfare, terrorism, and militia movements,
but we are not claiming that warfare is a determinant of
contemporary fundamentalism. We would argue, however,
that the cooperation and intra-group trust achieved through
costly ritual behavior enhances the ability of religious
groups to organize for acts of terror and war (Atran, 2002;
Johnson, in press; Sosis & Alcorta, in press). This is a vital
area of research that is ripe for evolutionary investigation.
Acknowledgments
We thank Candace Alcorta, Rebecca Bliege Bird, Joseph
Bulbulia, Aldo Cimino, Dominic Johnson, Frank Marlowe,
Eric Alden Smith, and two anonymous reviewers for very
helpful comments on earlier drafts of this manuscript. We
also thank the following students for their invaluable help
on this project: Erin Anderson, Leon BenRimon, Casey
Callahan, Jackie Chan, Erica Colls, Israel Cordero, Susie
Divietro, Joseph Filush, Rebecca Floor, Amanda Garibaldi,
Heather Law, Miriam Lee, Anna Liamzon, Sarah Mallory,
Colleen Morris, Jaclyn Nadeau, Vanessa Nielan, Christopher
Nyberg, Reshma Oka, James Seeth, John Shaver, and
Asha Shipman.
References
Alcorta, C. (2006). Religion and the life course. In P. McNamara (Ed.),
Where God and man meet: The neurology of religious experience
(pp. 55–79). Westport, CT7 Praeger.
Alcorta, C., & Sosis, R. (2005). Religion, emotion, and symbolic ritual: The
evolution of an adaptive complex. Human Nature, 16, 323–359.
Alexander, R. (1987). The biology of moral systems. Hawthorne, NY7
Aldine de Gruyter.
Alvard, M., & Nolin, D. (2002). Rousseau’s whale hunt? Coordination
among big-game hunters. Current Anthropology, 12, 136–149.
Atran, S. (2002). In Gods we trust: The evolutionary landscape of religion.
Oxford7 Oxford University Press.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 245
Bliege Bird, R., & Smith, E. (2005). Signaling theory, strategic interaction,
and symbolic capital. Current Anthropology, 46, 221–248.
Bliege Bird, R., Smith, E., & Bird, D. (2001). The hunting handicap: Costly
signaling in human foraging strategies. Behavioral Ecology and
Sociobiology, 50, 9–19.
Bulbulia, J. (2004a). Religious costs as adaptations that signal altruistic
intention. Evolution and Cognition, 10, 19–38.
Bulbulia, J. (2004b). Area review: The cognitive and evolutionary
psychology of religion. Biology and Philosophy, 18, 655–686.
Bulbulia, J. (in press). Free love: Religious solidarity on the cheap. In J.
Bulbulia, et al. (Eds.), The evolution of religion: Studies, theories, and
critiques. Collins Family Foundation.
Chagnon, N. (1988). Life histories, blood revenge, and warfare in a tribal
population. Science, 239, 985–992.
Chagnon, N. (1997). Yanomamo. Fortworth, TX7 Harcourt Brace College
Publishers.
Chick, G., & Loy, J. (2001). Making men of them: Male socialization for
warfare and combative sports. World Cultures, 12, 2–17.
Chick, G., Loy, J., & Miracle, A. (1997). Combative sport and warfare: A
reappraisal of the spillover and catharsis hypotheses. Cross-Cultural
Research, 31, 249–276.
Crocker, W., & Crocker, J. (1994). The Canela: Bonding through kinship,
ritual, and sex. New York7 Harcourt Brace.
Cronk, L. (1994). Evolutionary theories of morality and the manipulative
use of signals. Zygon: Journal of Religion and Science, 29, 81–101.
Damasio, A. R. (1994). Descartes’ error: Emotion, reason, and the human
brain. New York7 Avon Books.
Durkheim, E. (1995 [1912]). The elementary forms of religious life.
New York7 Free Press.
Ember, C. R., & Ember, M. (1992). Resource unpredictability, mistrust, and
war. Journal of Conflict Research, 36, 242–262.
Ember, C. R., & Ember, M. (1994). War, socialization, and interpersonal
violence. Journal of Conflict Research, 38, 620–646.
Ember, C. R., & Ember, M. (1998). Cross-cultural research. In H.
R. Bernard (Ed.), Handbook of methods in cultural anthropology
(pp. 647–687). Walnut Creek, CA7 AltaMira.
Ember, M., & Ember, C. R. (1971). The conditions favoring matrilocal
versus patrilocal residence. American Anthropologist, 73, 571–594.
Ferguson, R. (2001). Materialist, cultural and biological theories on why
Yanomami make war. Anthropological Theory, 1, 99–116.
Gaulin, S. J. C., & Boster, J. S. (1990). Dowry as female competition.
American Anthropologist, 92, 994–1005.
Gilmore, D. (1990). Manhood in the making. New Haven, CT7 Yale
University Press.
Gray, J. P. (1999). A corrected ethnographic atlas. World Cultures, 10,
24–136.
Gurven, M. (2004). To give and to give not: The behavioral ecology of
human food transfers. Behavioral and Brain Sciences, 27, 243–283.
Hamilton, W. D., & Zuk, M. (1982). Heritable true fitness and bright birds:
A role for parasites. Science, 218, 384–387.
Harner, M. (1972). The Jivaro: People of the sacred waterfalls. Garden
City, NY7 Natural History Press.
Iannaccone, L. (1992). Sacrifice and stigma: Reducing free-riding in cults,
communes, and other collectives. Journal of Political Economy, 100,
271–291.
Irons, W. (1979). Cultural and biological success. In N. Chagnon &
W. Irons (Eds.), Evolutionary biology and human social behavior:
An anthropological perspective (pp. 257–272). North Scituate,
MA7 Duxbury.
Irons, W. (1996). Morality, religion, and evolution. In W. M. Richardson &
W. Wildman (Eds.), Religion and science: History, method, and
dialogue (pp. 375–399). New York7 Routledge.
Irons, W. (2001). Religion as a hard-to-fake sign of commitment.
In R. Nesse (Ed.), Evolution and the capacity for commitment
(pp. 292–309). New York7 Russell Sage Foundation.
Irons, W. (2004). An evolutionary critique of the created co-creator concept.
Zygon: Journal of Religion and Science, 39, 773–790.
Johnson, D. (in press). Gods of war: The adaptive logic of religious
conflict. In J. Bulbulia, et al. (Eds.), The evolution of religion: Studies,
theories, and critiques. Collins Family Foundation.
Johnstone, R., & Norris, K. (1993). Badges of status and the cost of
aggression. Behavioral Ecology and Sociobiology, 32, 127–134.
Keeley, L. (1996). War before civilization: The myth of the peaceful savage.
Oxford7 Oxford University Press.
Knight, C., Power, C., & Watts, I. (1995). The human symbolic
revolution: A Darwinian account. Cambridge Archaeological Journal,
5, 75–114.
Low, B. S. (1979). Sexual selection and human ornamentation. In N. A.
Chagnon & W. Irons (Eds.), Evolutionary biology and human social
behavior (pp. 462–487). Boston7 Duxbury Press.
Ludvico, L. R., & Kurland, J. A. (1995). Symbolic or not-so-symbolic
wounds: The behavioral ecology of human scarification. Ethology and
Sociobiology, 16, 155–172.
MacNeill, A. (2004). The capacity for religious experience is
an evolutionary adaptation to warfare. Evolution and Cognition, 10,
43–60.
Maynard Smith, J., & Harper, D. (2003). Animal signals. Oxford7 Oxford
University Press.
McCauley, R., & Lawson, E. T. (2002). Bringing ritual to mind:
Psychological foundations of cultural forms. Cambridge7 Cambridge
University Press.
Olson, M. (1965). The logic of collective action: Public goods and the
theory of groups. Cambridge7 Harvard University Press.
Otterbein, K. (1968). Internal war: A cross-cultural study. American
Anthropologist, 70, 277–289.
Otterbein, K. (1994). Feuding and warfare: Selected works of Keith F
Otterbein. New York7 Gordon and Breach.
Paige, K. E., & Paige, J. M. (1981). The politics of reproductive ritual.
Berkeley7 University of California Press.
Pinker, S. (1997). How the mind works. New York7 W.W. Norton
and Company.
Poggie, J. (1995). Food resource periodicity and cooperation values: A crosscultural
consideration. Cross-Cultural Research, 29, 276–296.
Rappaport, R. (1999). Ritual and religion in the making of humanity.
Cambridge7 Cambridge University Press.
Ruffle, B., & Sosis, R. (2007). Does it pay to pray? Costly ritual and
cooperation. The BE Journal of Economic Analysis and Policy, 7,
1–35.
Schlegel, A., & Barry, H. (1980). The evolutionary significance of
adolescent initiation ceremonies. American Ethnologist, 7, 696–715.
Singh, D., & Bronstad, P. M. (1997). The anatomical locations of human
body scarification and tattooing as a function of pathogen prevalence.
Evolution and Human Behavior, 18, 403–416.
Smith, E., Bliege Bird, R., & Bird, D. (2003). The benefits of
costly signaling: Meriam turtle hunters. Behavioral Ecology, 14,
116–126.
Sosis, R. (2000a). Religion and intra-group cooperation: Preliminary results
of a comparative analysis of utopian communities. Cross-Cultural
Research, 34, 70–87.
Sosis, R. (2000b). The emergence and stability of cooperative fishing
on Ifaluk Atoll. In L. Cronk, N. Chagnon, & W. Irons (Eds.),
Human behavior and adaptation: An anthropological perspective
(pp. 237–272). Chicago7 Aldine de Gruyter.
Sosis, R. (2003). Why aren’t we all Hutterites? Costly signaling theory and
religious behavior. Human Nature, 14, 91–127.
Sosis, R. (2004). The adaptive value of religious ritual. American Scientist,
92, 166–172.
Sosis, R. (2005). Does religion promote trust? The role of signaling,
reputation, and punishment. Interdisciplinary Journal of Research on
Religion, 1, 1–30.
Sosis, R. (2006). Religious behaviors, badges, and bans: Signaling theory and
the evolution of religion. In P. McNamara (Ed.), Where God and man
meet: Evolution, genes, and the religious brain (pp. 61–86). Westport,
CT7 Praeger Publishers.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247246
Sosis, R., & Alcorta, C. (2003). Signaling, solidarity, and the sacred:
The evolution of religious behavior. Evolutionary Anthropology, 12,
264–274.
Sosis, R., & Alcorta, C. (in press). Militants and martyrs: Evolutionary
perspectives on religion and terrorism. In R. Sagarin & T. Taylor (Eds.),
Natural security: A Darwinian approach to a dangerous world.
University of California Press.
Sosis, R., & Bressler, E. (2003). Cooperation and commune longevity: A
test of the costly signaling theory of religion. Cross-Cultural Research,
37, 211–239.
Sosis, R., & Ruffle, B. (2003). Religious ritual and cooperation: Testing for
a relationship on Israeli religious and secular kibbutzim. Current
Anthropology, 44, 713–722.
Sosis, R., & Ruffle, B. (2004). Ideology, religion, and the evolution of
cooperation: Field experiments on Israeli kibbutzim. Research in
Economic Anthropology, 23, 89–117.
Strassmann, B. (1996). Menstrual hut visits by Dogon women: A hormonal
test distinguishes deceit from honest signaling. Behavioral Ecology, 7,
304–315.
Thorpe, I. (2003). Anthropology, archaeology, and the origin of warfare.
World Archaeology, 35, 145–165.
Whitehouse, H. (2000). Arguments and icons. Oxford7 Oxford University
Press.
Whiting, J., Kluckholm, R., & Anthony, A. (1958). The function of male
initiation ceremonies at puberty. In E. Maccoby, T. Newcomb, &
E. Hartley (Eds.), Readings in social psychology (pp. 359–370). New
York7 Holt, Rinehart, and Winston.
Wrangham, R., & Peterson, D. (1996). Demonic males: Apes and the
origins of human violence. New York7 Houghton Mifflin Co.
Young, F. (1965). Initiation ceremonies. New York7 Bobbs-Merrill.
Zahavi, A. (1975). Mate selection: A selection for a handicap. Journal of
Theoretical Biology, 53, 205–214.
R. Sosis et al. / Evolution and Human Behavior 28 (2007) 234–247 247