ArticlePDF Available

Power in Cultural Evolution and the Spread of Prosocial Norms

Authors:

Abstract

According to cultural evolutionary theory in the tradition of Boyd and Richerson, cultural evolution is driven by individuals’ learning biases, natural selection, and random forces. Learning biases lead people to preferentially acquire cultural variants with certain contents or in certain contexts. Natural selection favors individuals or groups with fitness-promoting variants. Durham (1991) argued that Boyd and Richerson’s approach is based on a “radical individualism” that fails to recognize that cultural variants are often “imposed” on people regardless of their individual decisions. Fracchia and Lewontin (2005) raised a similar challenge, suggesting that the success of a variant is often determined by the degree of power backing it. With power, a ruler can impose beliefs or practices on a whole population by diktat, rendering all of the forces represented in cultural evolutionary models irrelevant. It is argued here, based on work by Boehm (1999, 2012), that, from at least the time of the early Middle Paleolithic, human bands were controlled by powerful coalitions of the majority that deliberately guided the development of moral norms to promote the common good. Cultural evolutionary models of the evolution of morality have been based on false premises. However, Durham (1991) and Fracchia and Lewontin’s (2005) challenge does not undermine cultural evolutionary modeling in nonmoral domains.
THE QUARTERLY REVIEW
of Biology
POWER IN CULTURAL EVOLUTION AND
THE SPREAD OF PROSOCIAL NORMS
Nathan Cofnas
Balliol College, University of Oxford
Oxford OX1 3BJ United Kingdom
e-mail: nathan.cofnas@balliol.ox.ac.uk
keywords
cultural evolution, gene-culture coevolution, power, prosocial norms, evolution
of morality
abstract
According to cultural evolutionary theory in the tradition of Boyd and Richerson, cultural evolution is
driven by individualslearning biases, natural selection, and random forces. Learning biases lead people
to preferentially acquire cultural variants with certain contents or in certain contexts. Natural selection
favors individuals or groups with tness-promoting variants. Durham (1991) argued that Boyd and
Richersons approach is based on a radical individualismthat fails to recognize that cultural variants
are often imposedon people regardless of their individual decisions. Fracchia and Lewontin (2005)
raised a similar challenge, suggesting that the success of a variant is often determined by the degree of
power backing it. With power, a ruler can impose beliefs or practices on a whole population by diktat,
rendering all of the forces represented in cultural evolutionary models irrelevant. It is argued here, based
on work by Boehm (1999, 2012), that, from at least the time of the early Middle Paleolithic, human
bands were controlled by powerful coalitions of the majority that deliberately guided the development
of moral norms to promote the common good. Cultural evolutionary models of the evolution of morality
have been based on false premises. However, Durham (1991) and Fracchia and Lewontins (2005) chal-
lenge does not undermine cultural evolutionary modeling in nonmoral domains.
The Quarterly Review of Biology, December 2018, Vol. 93, No. 4
Copyright © 2018 by The University of Chicago Press. All rights reserved. This work is licensed under
a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0),
which permits non-commercial reuse of the work with attribution. For commercial use,
contact journalpermissions@press.uchicago.edu.
0033-5770/2018/9304-0001$15.00
Volume 93, No. 4December 2018
297
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Introduction
ACCORDING to mainstream cultural
evolutionary theory, individuals pos-
sess cultural variants”—“idea[s], skill[s], be-
lief[s], attitude[s], and value[s](Richerson
and Boyd 2005:63) mostly stored in hu-
man brains(Richerson and Boyd 2005:61).
Through interacting with each other, people
may adopt and transmit these variants. Cul-
tural evolutionists seek to identify the disposi-
tions that underlie the transmission process,
and explain why those dispositions were fa-
vored by natural selection under ancestral
conditions (Henrich and McElreath 2003;
Lewens2015:17).Having(purportedly)iden-
tied how people acquire and transmit cul-
ture, mathematical modelers can, given the
tness values of cultural variants and their
distribution in a population, predict the fu-
ture evolution of a culture, or explain retro-
actively why some variants proliferated and
others disappeared.
Fracchia and Lewontin (2005:2122)raise
a fundamental objection to thisresearch pro-
gram. They note that the success or failure of
a cultural variant can depend crucially on
the amount of power behind it(Fracchia
and Lewontin 2005:22). Cultural evolution-
ary models assume that people more readily
acquire beliefs and values with certain intrin-
sic properties, or preferentially learn from
certain kinds of people in certain contexts.
The models do not seem to account for the
possibility that cultural variants can be im-
posed on people coercively, regardless of
theirlearningrulesortransmissionbiases
(Richerson and Boyd 2005). Indeed, in a list
of cultural evolutionary forces discussedin
Not by Genes Alone, Richerson and Boyd (2005:
Table 3.1) say nothing about power or co-
ercion. All of the forces mentioned involve
voluntary imitation (see also Henrich and Mc-
Elreath 2003:Box 3). This potentially leaves
out a big factor that operates in real life.
Whether punishment,which Richerson and
Boyd do not include on their list of cultural
evolutionary forces but which they do discuss
in other parts of the book and in other works,
or prestige biascaptures the phenomenon
of power as it is exercised in real life will be
discussed presently.
Some years before Fracchia and Lewontin
(2005), Durham (1991) criticized Boyd and
Richerson for what he described as their
radical individualismthat ignores struc-
tured asymmetries or power relations, and
coercion(Durham 1991:182). When Boyd
and Richerson do mention power or coer-
cion at all (e.g., Boyd and Richerson 1985:
229230), they treat it as individually deliv-
ered punishment of noncooperators’” (Dur-
ham 1991:182).
Mainstreamculturalevolutionistshave not
changed their approach in response to Dur-
ham or Fracchia and Lewontin. To illustrate,
consider a classic empirical study in the Boyd
and Richerson tradition in which Gürerk
et al. sought to investigate institutional se-
lection(Gürerk et al. 2006:108). In their
experiment, anonymous subjects played a 30-
round public goods game, with each round
consisting of three stages: an institution choice,
avoluntary contribution,and a sanction-
ingstage. In the rst stage, subjects simul-
taneously and independentlychose to adopt
either a sanctioning institution(SI) or a
sanction-free institution(SFI; Gürerk et al.
2006:108). In the second stage, subjects (anon-
ymously) played a public goods game with all
of the others who had chosen the same insti-
tution, and they were informed about how
much each of the other players in both groups
had contributed to the public good (thus
the SI group members knew how well they
were doing compared with SFI group mem-
bers and vice versa). In the third stage, sub-
jects in the SI group were allowed to pay to
punish or reward each other.
In the rst couple rounds, most subjects
chose the SFI. However, freeriding in the SFI
group led contributions to fall precipitously.
Punishment in the SI group quickly led con-
tributions to rise to the near-maximum level.
The vast majority of subjects who initially chose
the SFI ended up switching to the SI and
adopting a prosocial strategy. Those who par-
ticipated in the SI received a signicantly
higher payoff than those who stuck with the
SFI. Henrich describes this study as an ex-
perimental demonstration of cultural group
selection in action.A key lesson, he says, is
that the playersdegree of rationality did
298 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
not permit them to foresee the nal outcome
and select the higher payoff institution on
the rst interaction(Henrich 2006:61).
Can we really draw conclusions about cul-
tural group selectionfrom Gürerk et al.s
(2006) experiment? Or were there power
relationsand coercionthat were decisive
in the real-life cultural evolution of cooper-
ation and punishment but which were not
captured in the experimental setup? It is true
that subjects in the SI group were able to
punisheach other, and in that sense they
wielded a kind of power. But there are some
essential differences between power, pun-
ishment, and coercion in real life and that
among Gürerk et al.s (2006) subjects in the
SI group. First, subjects were able to choose
beforehand whether they wished to be sub-
ject to punishment or not (i.e., they chose
whether to participate in the SI or the SFI).
Henrich (2006) imagines that this is analo-
gous to hunter-gatherers in the ancestral en-
vironment choosing which band to migrate
to. However, the option to avoid being sub-
jected to punishment by preemptively mi-
grating was probably only an option for our
hunter-gatherer ancestors in very limited cir-
cumstances(becausemigrationopportunities
are limited, hunter-gatherers often receive
the threat of expulsion as a death threat. See
the section below titled The Reverse Domi-
nanceHierarchyandDeliberateGuidanceof
Cultural Evolution for a real-life illustration.
When hunter-gatherers do have the chance
to migrate, it is typically to culturally similar
bands). Second, power in the experiment was
symmetrical not, as it typically is in real life,
asymmetrical (see Singh et al. 2017:461
462)in the experiment anyone could pun-
ish anyone for any reason (that is to say, in
real life, those with less power cannot punish
individualsor groupswith more power.
And a noncooperative individual who vio-
lates group rules cannot go around punish-
ing cooperative group members). Third, and
perhaps most important, the experimental
setup totally eliminated the possibility of col-
lectivedecision-making. Again,anonymous sub-
jects simultaneously and independentlychose
whether to participate in the SI or the SFI.
Henrich points out that most subjects failed
to foreseethat higher payoffs would ulti-
mately be reaped by participants in the SI.
He argues that this shows that successful in-
stitutions are not generally designed with ra-
tional foresightpeople just cannot predict
the consequences of different institutional ar-
rangements. But if instead of simultaneously
and independently choosing their institu-
tionsubjects had been allowed to meet
and coordinate a strategy, it is quite likely
that most everyone would have been able
to foreseethat the best strategy would be
toadopttheSIandtocollectivelypunishfree-
loaders. The experimental setup articially
prevented this from happening because it
was predicated on radical individualism that
sees cultural evolution as driven by the deci-
sions of uncoordinated, independently act-
ing individuals.
The results of one of Ostrom et al.s (1992:
412413) experiments seem to support the
assumption that, if people are allowed to
communicate andmake a collective decision,
they can anticipate the benets of punish-
ment. Groups of eight subjects played a com-
mon pool resource game for20 rounds. After
the 10th round, subjects met face-to-face for
10 minutes and were allowed todecide (by a
majority vote) whether to institute a sanc-
tioning mechanismwith a fee-to-ne ratio of
1/2 for their future interactions. Out of four
groups in this experimental condition, two
instituted a sanctioning mechanism (with
fee-to-ne ratios of 10¢/20¢ and 20¢/40¢,
respectively). However, Ostrom et al. (1992)
note that the subjects in this experiment had
previously participated in another experi-
ment where they played a common pool re-
source game with a sanctioning mechanism but
no chance for communication. Out of the 14
subjects who voted against the sanctioning
mechanism, 11 had previously participated
in a design with a relatively high fee-to-ne
ratio of 20¢/80¢. Out of the 18 who voted in
favor of the mechanism, just three had been
in the 20¢/80¢ design. They infer from this
result that the high level of sanctioning ac-
tivity in the 20¢/80¢ design, the lack of over-
all efciency gains and the presence of blind
revenge combined to impede the willingness
of participants to choose a sanctioning mech-
POWER IN CULTURAL EVOLUTIONDecember 2018 299
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
anism(Ostrom et al. 1992:413). They also sug-
gest that the experience of the rst 10 rounds
of the . . . game had an effect on mechanism
choice(Ostrom et al. 1992:413). To test this
possibility, they conducted two more experi-
ments where subjects could communicate
and adopt a sanctioning mechanism before
playing any rounds.Theresult:In both of
these experiments, the subjects quickly agreed
to an investment strategy and a sanctioning
mechanism to punish defectors. Across the
two experiments, net yields averaged 95%
94% with fees and nes included(Ostrom
et al. 1992:413).
What Is Power?
In order to properly articulate Durham
and Fracchia and Lewontins(DFL)challenge,
we should have a clear idea of what power
is. Fracchia and Lewontin (2005; Lewontin
2005) refer simply to power,which they il-
lustrate with examples of rulers imposing
cultural variants by decree. They do not ex-
plicitly describe the mechanisms that make
suchdecrees effective. Durham usuallyrefers
just to power,and at one point to struc-
tured asymmetries or power relations, and
coercion(Durham 1991:182)phraseology
that seems to imply that power relationsand
coercionare different things. Poweris
clearly a multifarious concept. Which type(s)
of power might pose a challenge to tradi-
tional cultural evolutionary modeling?
In his inuential analysis, Dahl (1957) sug-
gests that the intuitive idea of power ...is
something likethe following: Ahas power
over Bto the extent that he can get Bto
do something that Bwould not otherwise
do(Dahl 1957:202203; cf. Lukes 2005:30).
Whether or not this denition describes our
intuitive idea of power, it is clearly too broad
for the present purposes. On Dahlsdeni-
tion, virtually all cultural transmission would
have to be conceived as a manifestation of
power. Any time Bcopies A,Aexercises power
over B, even if Bis guided by one of the learn-
ing biases represented in standard cultural
evolutionary models.
Bachrach and Baratz (1970) offer a useful
taxonomy of power (see discussion in Lukes
2005:2122), which may allow us to pick out
those forms that present a challenge for cul-
tural evolutionary modeling. Their taxon-
omy includes inuence, authority, force, and
manipulation. Ainuences Bwhen A,with-
out resorting to either a tacit or an overt
threat of severe deprivations, causes [B]...
to change his course of action(Bachrach
and Baratz 1970:30). Aexercises authority
when Bcomplies because he recognizes that
the command is reasonable in terms of his
own values(Bachrach and Baratz 1970:34).
Auses force when Aobtains compliance by
depriving Bof choice. Amanipulates Bwhen
Bscompliance is forthcoming in the ab-
sence of recognition on the compliers part
either of the source or the exact nature of the
demand upon him(Bachrach and Baratz
1970:28).
According to Bachrach and Baratzs (1970)
jargon, powershould refer only to cases
where Bcomplies with Abecause Bfears
that Awill deprive him of a value or val-
ues which he regards more highly than those
which would have been achieved by non-
compliance(Bachrach and Baratz 1970:24).
Force,they say, should refer only to cases
of physical manipulation. However, follow-
ing Lukes (2005)and common usagewe
can regard inuence, authority, force, and
manipulation as species of power. When A
deprives Bof the option of following a course
of action, this is an exercise of force.
Inuence, as Bachrach and Baratz dene
it, is a form of power that can, in general, be
easily accommodated by cultural evolution-
ary models. When people freely copy each
other, this is a way of being inuenced. Stan-
dard cultural evolutionary models simply
reect the rules, or the learning biases,thatde-
termine how people freely copy each other.
Force is the form of power that poses the
mostobvious(apparent)challengetothemod-
els. If cultural variants are imposed on peo-
ple because the option to reject them is taken
away by force, then it seems that standard
cultural evolutionary models cannot explain
what happened.
Authority is established when individuals
freely recognize the right of a leader to make
certain demands on them. If cultural variants
spread on a large scale because of the exer-
cise of authority, this would seem to under-
300 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
mine cultural evolutionary modeling, at least
to some extent. If people accept the authority
of a leader to tell them what cultural variants
to adopt, to the extent that the leader exer-
cises their authority, the learning biases of
cultural evolutionary models will cease to
play a decisive role in (further) cultural evo-
lution. Learning biases might explain why
people choose to accept the authority of a
particular leader in the rst place, but they
would not determine which variants subse-
quently spread. When religious leaders pro-
mulgate opinions about values or doctrine,
their followers may adopt those opinions be-
cause they accept the authority of the leader.
To reiterate, cultural evolutionary models
might help explain how authority itself comes
to be established (e.g., how people become
Catholic and thereby accept the Popes au-
thority). But, just as when cultural variants
are spread through force, when they are
spread through the exercise of authority the
learning biases of cultural evolutionary mod-
els play no role. When Durham (1991) dis-
tinguished between power relationsand
coercion,perhaps he had in mind the dis-
tinction between authorityand force.
Authority and force can interact in inter-
esting ways. For example, one reason that we
accept authority might be that it is backed
by force (would Prime Minister Theresa May
retain her authority if the military and the
police disbanded? Probably not for long).
Conversely, the power to employ force is of-
ten made possible by widespread acceptance
of the legitimacy of the authority (people tol-
erate Theresa May giving orders to the mili-
tary and the police because they accept her
right to be Prime Minister). Sorting out how
this works in detail is not important for the
present discussion. The point is that when
cultural variants are spread through the ex-
ercise of either force or authority, the learn-
ing biases featuring in cultural evolutionary
models are to that extent irrelevant.
Manipulationsurreptitiously generating
desires in people to make certain choices
generally works by taking advantage of learn-
ing biases. For example, advertisers use ce-
lebrity endorsements to take advantage of
prestige bias. The Sara Lee Corporation paid
former basketball star Michael Jordan a great
deal of money to endorse Hanes underwear
because many people are motivated to copy
successful athletes. Richerson and Boyd (2005:
124) note that some men also started shaving
their heads because this was Michael Jordans
practice. They cite imitating Jordans choice
of underwear and imitating his hairstyle as
illustrations of prestige bias.But these two
cases of imitation illustrate different kinds of
power. When men shave their heads because
of Jordan, they are being inuenced. When
they buy Hanes because of him, they are
(most likely) being manipulated.
In principle, cultural evolution driven by
manipulation can be modeled in the same
way as that driven by inuence. Of course, if
manipulation is a decisive factor, an explana-
tion of cultural evolution that leaves this fact
out and appeals only to the learning biases of
the people who were manipulated would be
impoverished. But there is no reason why
standard cultural evolutionary models can-
not be used to track how cultural variants
spread in such circumstances. Prestige, how-
ever, has a supercial resemblance to author-
ity. The following section considers whether
the latter can be reduced to the former.
The fact that which cultural variants peo-
pleadopt maybedetermined bytheir position
in networks of powerspecically networks
of force and authoritycannot be captured,
at least in any obvious way, by models that
are predicated on a strong form of methodo-
logical individualism. Models that are com-
mitted to a highly individualistic picture of
decision-making and action will also miss the
fact that, when it comes to force and author-
ity, the locus of power is very often a group of
collectively acting individuals.
Can Power Be Reduced to
Coordinated Punishment
or Prestige Bias?
Coordinated punishment and prestige bias
are two phenomena that already feature in
mainstream cultural evolutionary models.
If either one captures the phenomenon of
powernamely, the exercise of force or au-
thorityin a sufciently realistic way, then
cultural evolutionary modeling would not be
threatened by DFLs challenge. This section
POWER IN CULTURAL EVOLUTIONDecember 2018 301
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
considers coordinated punishment and pres-
tige bias in turn.
Boyd et al. acknowledge that previous mod-
els of the evolution of punishment have been
unrealistic in their assumption that punish-
ment is an unconditional and uncoordinated
individual action automatically triggered by
defection(Boyd et al. 2010:617). To rectify
this shortcoming, they propose a model (also
defended in Bowles and Gintis 2011:Chap-
ter 9) of what they call coordinated punish-
ment of defectors.This sounds like it may
be a step toward incorporating collective so-
cial phenomena (including collectively exer-
cised force) into cultural evolutionary models.
However, closer inspection suggests that the
kind of coordination captured by Boyd et al.s
(2010) model is still unrealistically individu-
alistic. In their model, coordinated punish-
mentmeans that, when individuals decide
whether to punish defectors, they take into
account whether other individuals have sig-
naled their intention to punish defectors. But,
as shall be argued, the decision to punishas
it is represented in this modelis made
without the sort of collective plan that sup-
ports cooperative behavior in real life, and
that, according to anthropological evidence,
played an essential role in the evolution of
prosocial norms.
Boyd et al. say that, in their model, pun-
ishment is coordinated among group mem-
bers so that it is contingent on the number of
others predisposed to participate in the pun-
ishment(Boyd et al. 2010:617). But there is
much more to the coordinated punishment
described by anthropologists, or even by Boyd
et al. in their verbal description of anthro-
pological observations. As Boyd et al. them-
selves say:
ethnographicevidenceindicatesthatpunish-
ment is coordinated by means of gossip and
other communication among punishers, is
contingent on the expected effectiveness of
punishment in inducing cooperation, and it
is not undertaken unless it is judged as legit-
imate by most group members (Boyd et al.
2010:617).
The collective agreement and action that
emerges from this kind of cooperation can-
not be reduced to punishing in response to
signs that others will punish.
In Boyd et al.s (2010) model, individuals
are drawn from a population of punishers
and nonpunishersto form groups. Mem-
bers of each group engage in an initial three-
stage interaction. In the rst stage, punishers
exhibit a signal.In the second stage, if a
threshold number of other individuals in the
group gave the signal, punishers cooperate
(with probability 1 e)otherwise they de-
fect. Nonpunishers always defect at this stage.
In the third stage, if the threshold number
of punishers revealed themselves in the rst
stage, the punishers impose a cost on those
who did not cooperate. Group members then
engage in further interactions consisting only
of cooperation/defection and punishment
stages.
In administering a punishment, each pun-
isher incurs a cost. The more punishers there
are in a group, the less the total cost of in-
icting punishment on a defector, since the
defectors ability to retaliate at all can be as-
sumed to diminish exponentially with increas-
ing numbers of punishers.
Because punishment imposes a cost on de-
fectors that is, on average, greater than the
cost of cooperating, in general, cooperation
maximizes expected payoff if defectors face
punishment. In the model, defectors who
are punished in the rst round subsequently
behave like punishers, cooperating with prob-
ability 1 e. They defect with probability e
either because of error or because, due to
random differences in circumstances, coop-
eration may be costlier than being punished.
In the model, after engaging in a (ran-
domly varying) number of interactions, in-
dividuals produce offspring according to the
payoff they received, and new groups are
formed from a mix of the offspring. Offspring
inherit the strategy of punishingor non-
punishingfrom their parent.
Boyd et al.s basic conclusion is that the
initial proliferation of punishment occurs
underplausiblelevelsof group geneticdiffer-
ences and results in persistent and high levels
of cooperation(Boyd et al. 2010:620). Pun-
ishers are able to proliferate in their model
becauseunlike in other modelsthey re-
frain from inicting costly punishment un-
302 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
less there are enough other punishers in their
group to reduce the cost, and they are able
to do this because punishers coordinateby
signaling their intentions. This paper will
argue that prosocial-norm-enforcing punish-
ment did not evolve in this way, with individ-
ual contingent, signaling punishers gradually
inltrating groups of nonpunishers in the
ancestral environment. Rather, individuals
within groups of nonpunishers formed coali-
tions to collectively impose prosocial norms
on all group members. And, through a pro-
cess of gene-culture coevolution, we became
genetically adapted to be receptive to such
norms, including norms to punish noncoop-
erators. The evidence will show that Boyd
et al.s (2010) model gives an incorrect his-
torical account of how punishment arose in
human evolution, and of how we evolved a
disposition to punish norm violators.
The discussion above suggests that power
cannot be reduced to coordinated punish-
ment. Can it be reduced to prestige bias?
Prestige bias refers to our (well-established)
tendency to preferentially copy individuals
who are successful, or to whom others are pay-
ing attention or deferring (see, e.g., Chudek
et al. 2012). This is, of course, one of the cen-
tral learning biases taken into account by cul-
tural evolutionary models.
It is true that prestigious and prominent
peoplehave(ceterisparibus)moreinuence
overculturethan the lessprestigious,andthis
inuence is a form of power.When celeb-
rity Kylie Jenner tweeted that she did not use
the app Snapchat anymore, the companys
stock lost 1.3 billion dollars in value. That is
because when Jenner says she uses or does
not use something, many people will be in-
clined to imitate her.
But the powerof Kylie Jenner is funda-
mentally different from the power of a king,
a hunter-gatherer coalition enforcing rules,
or even the Pope. Jennersinuence derives
from the fact that many people preferentially
imitate her. When she exhibits behavior, or
expresses an opinion, we can explain why it
spreads by appealing to the learning biases
(namely, the prestige bias) of her audience.
The kind of power that threatens cultural evo-
lutionary modeling is that which is backed by
force or authority, and which thereby neutral-
izes all of the learning biases featuring in
mathematical models. Neither force nor au-
thority can be reduced to prestige bias, since
force and authority neutralize all learning
biases. When, in 1492, Isabella I and Ferdi-
nand II issued the Alhambra Decree ordering
the Jews of Spain to convert to Catholicism,
leave, or be executed (as discussed in more
detail below), it is theoretically possible that
a small number of Jews were impressed by
the religious passion of the prestigious queen
and king and converted for that reason. But
most of the Jews who converted probably
did so because their freedom to choose had
been coercively restricted. Living in Spain
as a practicing Jew was literally no longer a
possible choice. In hunter-gatherer groups, in-
dividuals did not have the choice to refrain
from sharing meat or contributing to group
defenseif they did not do these things, and
were recalcitrant enough in the face of re-
proach, they would be subject to execution
or (what was often effectively the same thing)
expulsion.
Approaches to Accommodating Power
Lewontin (2005) illustrates the role of
power in cultural evolution with an exam-
ple. Bavarians are mostly Catholic, Westpha-
lians Protestant. This state of affairs was not
brought about (according to Lewontin) by
any mechanism recognized by cultural evo-
lutionistsnot by biased transmission, the
relative appeal of the content of the two reli-
gions, drift, Darwinian selection, or any other
(see the list in Richerson and Boyd 2005:
Table 3.1). Rather, Lewontin says, in 1555 the
German princes and the Holy Roman Em-
peror adopted the rule of cuius regio, eius re-
ligio, which allowed rulers to enforce their
own religion in their own dominions and to
expel those who were recalcitrant(Lewon-
tin 2005). If cultural change is usually driven
in this wayby sweeping diktats issued by
powerful individuals or small groupsthen
Boyd and Richerson-type modeling, which
tracks the aggregated effects of small-scale
events(Lewens 2015:17), will have a very lim-
ited range of application. Most of history
and the Paleolithic culture to which our minds
are adaptedwould better be explained by
POWER IN CULTURAL EVOLUTIONDecember 2018 303
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
a study of how individuals take power and
impose their favored cultural variants on
the masses, and cultural evolutionary models
would not shed much light on the selective
forces that shaped our social learning dispo-
sitions.
As Lewens notes, it would surely be a huge
overestimation of the power of the typical
ruler to suppose that they can change the
religious beliefs of a population by mere de-
cree. In regard to the specic example of
cuius regio, eius religio, it is not clear that the
German princes even tried to change peo-
ples religious beliefs on a large scale. They
tended to enforceCatholicism or Luther-
anism on their subjects according to what
was already the prevailing religion, and even
subjects in the minority did not necessarily
convert in response to the ofcial order (Le-
wens 2015:133).
But that is just one example. It seems clear
that sometimes cultural variants, even reli-
gious beliefs, can spread as a result of being
backed by power. As discussed, the forms of
power that seem to undermine cultural evo-
lutionary modeling are force and authority.
Sometimes power makes all of the difference
in which cultural variants prevail in a group.
The Alhambra Decree, which expelled prac-
ticing Jews from Spain in 1492, caused more
than two-thirds of the not already-converted
Jewish population to become Catholic, at
least nominally. The decree, initiated by one
person (Isabella I) and signed by two (the
aforementioned and Ferdinand II) radically
changed the distribution (and attractiveness)
of certain cultural variants in Spain, and its
consequences reverberated for centuries.
In the 4th century, Christianity surely got a
boost from being supported by Constantine,
although whether this was responsible for its
ultimate triumph is difcult to say.
Lewens (2015:138139) proposes three
ways in which cultural evolutionary research
programs could accommodate the phenom-
enon of power. First, restriction: we could
simply restrict the application of models to
times and places where cultural evolution is
not driven by Lewontin scenarios—“where
human societies have been free of interest
groups, governments, unions, and perhaps
gross asymmetries of individual power(Le-
wens 2015:138). Second, presumption: we
could take systems of power as xed back-
ground conditions against which individuals
interact(Lewens 2015:138), using cultural
evolutionary theory to track cultural change
resulting from those interactions. And, -
nally, reduction: we could try to reduce sys-
tems of power—“governmental structures,
judicial procedures, and so forth(Lewens
2015:139)to small-scale interactions dealt
with by standard cultural evolutionary theory.
By restricting cultural evolutionary mod-
eling to times and places free of interest
groups, governments, unions, and . . . gross
asymmetries of individual power(Lewens
2015:138), restriction undermines any mod-
eling applied to modern societies. Lewens
suggests that we could still use cultural evo-
lutionary models to investigate early hominin
history, since hunter-gatherers were largely
egalitarian with respect to individualsaccess
to economic resources(Lewens 2015:138,
citing Boehm 1999 and Knauft 1991). But
even this is not certain. Despite their lack of
formal leaders, great differences in individ-
ual power among adult males, or great dif-
ferences in individual access to economic
resources, hunter-gatherers do have systems
of rules, coercion, and punishment. In fact,
as shall be argued below, the egalitarianism
of hunter-gatherers is itself maintained by
the exercise of power wielded by the rank
and le over their potential dominators (or
alphas), creating what Boehm (1993, 1997,
1999) calls a reverse dominance hierarchy.
This means that even hunter-gatherers pos-
sess a power apparatus that could theoreti-
cally be used to impose cultural variants on
individuals willy-nilly and force deviants into
line. Whether hunter-gatherers actually do
use this power to inuence the spread of cul-
tural variants in a way that undermines Boyd
and Richerson-type modeling is the central
question addressed in the present paper.
The problem with presumptiontaking
power relations as xed background condi-
tions against which individuals interact(Le-
wens 2015:138)is that it seems to ignore
DFLs challenge. Suppose we want to explain
the spread of different sects of Christianity in
mid-16th-century Germany. An explanation
emphasizing the forces listed in Richerson
and Boyd (2005:Table 3.1) and relegating
power relations to xed background condi-
304 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
tionswould clearly be impoverished and
misleading, in that it would be taking for
granted the factor that ought to be a primary
target of a scientic explanation (even if, as
Lewens argues, it was not the whole story).
How could we construct an explanation of
the spread of Catholicism and Lutheranism
taking the power of the princes as xed
background conditions? Perhaps we could
treat the positive and negative incentives
offered for adopting different beliefs in dif-
ferent places as being features of the environ-
ment like the weather. The fact that people
prefer, ceteris paribus, denominations that
do not cause them to be exiled from their
home could be considered a content bias.
Although it may be possible to model the
decision-making process of German peasants
alongthese lines, thisapproachwould ignore
the fact that a major driving force of cultural
evolutionis the asymmetricalpowerrelations
of the princes and the peasants. Fracchia and
Lewontins (2005) point is that explaining
historical trendstrends about how cultural
variants spreadrequires us to recognize
power as a major factor. In Durhams words,
to ignore [power] . . . is to ignore what may
often be the leading cause of transforma-
tion(Durham 1991:182). The fact that we
can treat the inuence of power as a given
and focus on individual decision-making
does nothing to counter the claim that power
often plays a decisive role in the fate of cul-
tural variants. Lewens raises a similar problem
for presumption, saying that as a strategy it
simply omits to explain many features that
are admitted as important in determining cul-
tural change(Lewens 2015:138), although
he says presumption can be useful for un-
derstanding . . . individual interactions and
their long-term effects(Lewens 2015:138).
This leaves reductionreducing the inu-
ence of power to individual-level interactions
that can be handled by standard cultural evo-
lutionary models. In Lewenss view, the jus-
tication for this approach is that power
relations are all grounded in individual inter-
actions, which should in principle be ame-
nable to populational modeling. This does
seem to be the approach advocated by cul-
tural evolutionary modelers themselves (e.g.,
Boyd and Richerson 2002; Richerson and
Henrich 2012), although in practice their
models do not generally reect power rela-
tions at all.
But redescribing the exercise of power in
terms of individual interactions does not ne-
gate the fact that power is a factor, and it
does not neutralize DFLs challenge. Con-
sider again the Alhambra Decree. Isabella
and Ferdinand issued an order (Jews had to
convert, leave, or be executed). Since those
with the ability to enforce the orderthe sol-
diersaccepted the authority of the monar-
chy, they conveyed the threat to the Jews, who
chose whether to convert or leave (no one
volunteered for execution). If we want to
explain why Spanish soldiers accepted the
monarchy, we would probably need to refer
to all of the learning biases that gure in cul-
tural evolutionary models. If we want to ex-
plain how Jews responded to the threat, we
could say that the cost of being overtly com-
mitted to Judaism had increased, and since
people have a content bias to acquire cultural
variants with lower costs, Judaism became less
attractive. So rather than saying that the Jews
converted in response to the exercise of power
by Isabella and Ferdinand, we can say that
they decided which practices to adopt in
light of the incentives. But we still cannot ex-
plain what happened without referring to
the fact that, because the social structure gave
power to the monarchs, two individuals drove
cultural evolution in a specic direction. To
explain why the powerful individuals acted
as they did, we would need to engage in tra-
ditional historical, sociological, or psycholog-
ical analysis.
Can Cultural Evolutionary
Modeling Be Restricted to the
Ancestral Environment?
Modern societies are founded on exten-
sive, complex systems of power. The informa-
tion we are able to obtain, the opinions we
are exposed to, and the options that are pre-
sented to us are constrained in all sorts of
ways by the (often hidden) exercise of force
andauthority. DFLschallengeto culturalevo-
lutionary modeling appears to have great
force when it comes to understanding how
cultural variants spread in extremely hierar-
chical societies where powerful individuals
and groups exercise such far-reaching con-
POWER IN CULTURAL EVOLUTIONDecember 2018 305
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
trol.This is nottodeny that themodelscanbe
used to explain some aspects of cultural evo-
lution even in modern societies. When peo-
ples choices in a particular domain are
largely unconstrained by power, the models
can sometimes explain how and why culture
develops as it does. For example, cultural
evolutionary models can explain the demo-
graphic transition in the West (Richerson
and Boyd 2005:Chapter 5), since people in
Western countries in recent history have been
largely free to choose how many children to
have. Power is rarely the only factor inuenc-
ing what people believe and how they act, so
models might be used in conjunction with
traditionalhistorical/anthropological/socio-
logical methods. But to explain phenomena
such as the spread of communism and the
concomitant values in the 20th century, a
study of the inuence of power will often be
more fruitful than a study of the aggregate ef-
fect of the decisions of many individuals.
Cataloging all of the ways in which power
can or does inuence the spread of cultural
variants,andin whichit might underminecul-
tural evolutionary modeling, is not the aim of
the present paper. This paper has a narrower
focus, which is to consider whether or to what
extent power undermines cultural evolution-
ary modeling applied to nomadic, foraging
societies. It seeks to test the possibility raised
by Lewens (2015) that we can restrict the ap-
plication of Boyd and Richerson-style models
to conditions that prevailed during the long
period before the advent of sedentarism, food
storage, and agriculture, when human soci-
ety was reputedly egalitarian.
There are two reasons why it is worthwhile
to investigate whether Boyd and Richerson-
style models can be applied unproblemati-
cally to nomadic foragers. First, the models
have been used to give accounts of the histor-
ical origins of specic cultural variants, such
as certain ethical norms. If the models can-
not be applied to nomadic foraging societies
(where the variants in question originated),
thenwe will havetoreconsider theseaccounts.
Second, the modelspredicated on the idea
that cultural variants spread in the ancestral
environment due largely to content and con-
text learning biaseshave been used to make
predictions about what psychological adap-
tations we might have. Given that variants
spread due to content and context learn-
ing biases, cultural evolutionary theorists ask
what psychological dispositions we might
have evolved in order to extract adaptive
information from the cultural environment.
But if power was a major factor inuencing
how variants spread in the ancestral environ-
ment, then we might make different predic-
tions about how our psychology evolved. We
might predict, for example, that we evolved
to have different attitudes toward variants
imposed on us by power and those that we
adopted due to learning biases.
Chimpanzees (Pan troglodytes)haveastrong
dominance hierarchy based on physical force.
The alpha individual rules because he (some-
times with the help of his coalition partner)
can beat up every other chimp. Beta can beat
up gamma, and so on down the line. Under
naturalconditions, alladult males,beingphys-
ically stronger, are dominant to all females
(Goodall 1986).
Dominant chimps use their social position
primarily to secure access to food and mating
opportunities. Dominance relations have lit-
tle direct inuence on cultural transmission.
Chimps have different, culturally transmitted
ways of shing for termites, building nests,
and performing a few other tasks. Which cul-
tural variant an individual acquires depends
on the practice of the most accessible models
and features of the local environment (Mc-
Grew 2004:Chapter 7). Under experimental
conditions, chimps preferentially attend to
dominant individuals when learning how to
solve foraging problems, and are more likely
to copy their methods (Horner et al. 2010).
But dominant individuals have no interest
inandmayevenlackawarenessofthecul-
tural practices of their subordinates, hence
they make no attempt to forcibly spread their
favored practices. The common ancestor of
chimps and humans probably had a similar
hierarchical social structure (Boehm 1999).
At some point, hominins developed much
greater awareness of cultural possibilities, and
a much greater capacity for deliberately trans-
mitting learned practices. It became feasible
for dominant individuals to force subordi-
natesto adopt theirpreferredcultural variants.
Richerson and Henrich note that [p]reda-
tory elites and other self-interested subgroups
with some form of coercive powercould
306 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
establish norms and institutions that dis-
proportionately benet them. Ideologically
motivated groups with coercive power may
sustain equilibria at mad extremes, at least for
brief periods of time(Richerson and Hen-
rich 2012:50; they do not comment on the
implications this might have for cultural evo-
lutionary modeling). But elites who impose
cultural variants on others do not necessarily
have to be motivated by narrow self-interest
they could be prompted by altruism, concern
for the commonweal, or the belief that a well-
functioning group will ultimately benet
themselves (i.e., enlightened self-interest).
In any case, if dominant individuals or coali-
tions in the Paleolithic did exercise power to
impose cultural variants on others, it could
create a problem for cultural evolutionary
modeling for the reasons discussed above.
There are two crucial empirical questions
to answer. First, was there any locus of power
in Paleolithic communities capable of impos-
ing cultural variants on subordinates? Sec-
ond, if the answer to the rst question is yes,
was this power routinely exercised in a way
thatsupportsDFLs critiqueof culturalevolu-
tionary modeling? These questions are dealt
with in turn in the following two sections.
Based on work by Boehm, the rst will be an-
swered in the afrmative. The second will
also be answered in the afrmative, but with
a qualication. It will be argued that, in the
ancestral environment, power was used pri-
marily as a means of enforcing moral norms.
Therefore, DFLs critique undermines stan-
dard cultural evolutionary modeling applied
to moral norms that arose in the past 250,000
years, but it does not necessarily undermine
modeling applied to nonmoral norms in the
ancestral environment. It also does not nec-
essarily undermine modeling applied to the
evolution of moral norms and intuitions that
might have occurred before 250,000 bp (the
specic example of Baumard et al.s 2013
model will be discussed).
The Existence of Power
in the Paleolithic
Knauft notes that male status differentia-
tion in human evolution is U-shaped(Knauft
1991:397). The social organization of the
chimp-human common ancestor is presumed
to have been as hierarchical as that of mod-
ern chimps. After hunter-gatherers settled
down and developed the means to stockpile
food, human societies became even more hi-
erarchical than those of chimps. For some
millions of years after the chimp-human lin-
eages split and before sedentary, hierarchi-
cal societies were established, humans lived
as nomadic hunter-gatherers. All nomadic
hunter-gatherers that have ever been de-
scribed, without exception, live by an ideology
of extreme egalitarianism (among men). No
nomadic hunter-gatherer society tolerates adult
males issuing direct orders to other adult
males. Some groups have a titular leader,
but his role never goes beyond leading by ex-
ample and helping to organize group discus-
sionsaimedatreachinga consensuson group
actions (whether to go to war, where to mi-
grate, and the like; Boehm 1999).
Asnotedearlier,Boehm(1993,1997,1999)
argues that nomadic foraging societies are
characterized by a reverse dominance hier-
archy: men who would otherwise be sub-
ordinate to an alpha band together to keep
would-be alphas from assuming positions of
dominance. In contrast, Erdal and Whiten
(1994:177) argue that there is an absence,not
areversal, of hierarchy. According to them,
would-be subordinates simply refuse to obey
would-be leaders. A hunter-gatherer group
does not (in their view) start with a hierarchy
and [then] reverse it(Erdal and Whitten
1994:177). Rather, counterdominantbe-
havior prevents leaders from arising in the
rst place.
The debate between Boehm and Erdal
and Whiten is not just scholastic. If Erdal and
Whiten are right that forager egalitarianism
is based on a widespread refusal to obey or-
ders, then there would have been essentially
no locus of power among Paleolithic adult
male foragers that could have potentially di-
rectedculturalevolution.Ifnooneeverobeys
authority,no individual orcoalitioncanexer-
cise power over anyone else. But if Boehm is
right that egalitarianism is maintained by the
asserted efforts of the rank and le to keep
down ambitious, aggressive, and accomplished
individuals, Paleolithic bands wouldhave had
groupscoalitions of the majoritywith the
power to potentially impose cultural variants
on those whom they dominated.
POWER IN CULTURAL EVOLUTIONDecember 2018 307
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
There are compelling reasons to side with
Boehm in this debate. Perhaps most signi-
cantisthe factthat nomadic hunter-gatherers
deliberately and actively enforce egalitarian-
ism, employing a variety of coercive methods
ranging from gossip to ridicule to ostracism
to expulsion (often a de facto death sentence)
to outright execution. They do not simply re-
fuse to obey commands and rebuff would-be
dominators. They espouse an explicit egali-
tarian ideology (Cashdan 1980) and take ac-
tive measures to head off possible power grabs
by individuals. As Boehm (1999) argues, our
primate heritage makes us readily suscepti-
ble to developing orthodox hierarchies (i.e.,
with alpha-types dominating everyone else).
To prevent orthodox hierarchies from emerg-
ing requires the group to continually keep
down those who would take control if given
the opportunity.
What is more, the group actively controls
the behavior of potential alphas, compelling
them to procure meat for others and allow-
ing them to act as informal leaders in hunt-
ing and warfare (so that the group benets
from their expertise without submitting to
theirdomination;Boehm1997:S104).Again,
this goes beyond refusal to obey orders and
reects bona de domination and control
of potential alphas.
Everyone, not only potentially dominant
individuals, is subject to group rules. People
who are not suspected of political ambitions
but who horde food or engage in deception
canbetargetedby someof thesame sanctions
as those who try to gain individual power.
Boehm (1999) quotes Services (1975:48
49) observation that in foraging societies
social life is intensely regulated by codes,
rules, expectations, habits, and customs that
are related to etiquette, ethic, and role. And
because these are not [normally] explicit,
nor revealed by frequent breaches, the soci-
ety might give the impression of freedom
and lack of conict(Boehm 1999:84). But
the threat of collectively imposed sanctions
of varying degrees of severity hangs over ev-
eryones heads.
The foregoing suggests that, in the sort
of hunter-gatherer bands that existed since
at least the early Middle Paleolithic and to
which our social psychology is presumed to
be adapted, there did exist a locus of power
that was theoretically capable of driving cul-
tural evolution in the way DFL suggested
(i.e., by at issued by a consensus of groups).
Whether this threatens cultural evolutionary
modeling turns on the answer to an empiri-
cal question: Did hunter-gatherer coalitions
of the majority, as a matter of fact, exercise
the power described above to promote their
favored cultural variants, or did they use it
only to suppress power grabs by would-be
alphas? It will be argued that the empirical
evidence supplied by Boehm (1999, 2012)
strongly suggests that hunter-gatherers did use
their collective power to impose cultural var-
iantsthat theythoughtwould promote group
success and well-beingvariants broadly re-
lated to morality (note that a great variety of
behavior can be moralized, from food pref-
erences to factual beliefs). They did not use
power to impose variants related to practices
that were not thought to directly concern the
group,likeindividualsspecicspear-making
techniques. Consequently, cultural evolution-
ary models can be applied more or less un-
problematically to the nonmoral domain of
culture. It is cultural evolutionary model-
ing of the transmission of group-concerning
moral practices such as reciprocity and pun-
ishment (e.g., Henrich and Boyd 2001) that
may be problematic.
The Reverse Dominance Hierarchy
and Deliberate Guidance of
Cultural Evolution
ThissectionbrieyoutlinesBoehms (1999,
2012) account of how subordinate humans
circa 250,000 bp overthrew their alphas and,
around the same time, used their newly ac-
quired collective power to impose a deliber-
ately engineered moral code. In the light of
Boehms theory, the following section will
critically analyze some well-known modeling
work on the coevolution of prosocial norms
and norm psychology.
From our primate ancestors we inherited
three key dispositions: an enjoyment of domi-
nating, a capacity for submission, and a dislike
of being dominated. Among chimpanzees,
these same dispositions lead to the develop-
ment of a dominance hierarchy. Subor-
308 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
dinatesparticularly maleswait for the
opportunity to challenge higher-ups and as-
cend the hierarchy, if possible. Subordinate
chimpanzees (P. troglodytes) often form coali-
tions to challenge the power of alphas (Boehm
2012:9596) but, because of limits on their
coordination abilities and the physical su-
premacy of the dominant individuals, they
never succeed in doing away with the prac-
tice of alpha rule itself.
Chimps, bonobos, and gorillas have the
ability to make generalizations about what
sort of positive or negative behavior will elicit
the wrath of their superior, and purposefully
comply with the superiorsdemands(Boehm
2012:106; the behavior demanded by domi-
nant apes mostly concerns feeding priority
and mating; Boehm 2012:107108). This was
an important preadaptation for our ancestors
to develop the ability to appreciate group-
imposed rules. Taking advantage of our abil-
ity to understand and adopt regular patterns
of behavior, coalitions of the majority in hu-
man groups instituted rules to enforce egal-
itarianism and promote the good of the
collective. For hunter-gatherers, the coalition
of the majority is the object of fear, rather
than, as for chimps or gorillas, powerful dom-
inant individuals. Boehm expounds: People
like Bushmen or Pygmies gossip incessantly
and are highly judgmental, and group opin-
ion is something to be feared because moral
outragecanleadtoostracism,expulsionfrom
the group, or even execution. This is true of
all hunter-gatherers(Boehm 2012:107).
Darwin speculated that, when language
developed, the wishes of the members of
the same community could be distinctly ex-
pressed,thecommon opinionhow eachmem-
ber ought to act for the public good, would
naturally become to a large extent the guide
to action(Darwin 1871:72). The powerof
public opinion derived, he said, not from co-
ercion but from our natural love of approba-
tion and horror of scorn and infamy,which
itself is rooted in our [i]nstinctive sympathy
inherited from our distant ancestors (Dar-
win 1871:86; see Lewens 2007:162167 for an
overview of Darwins account; Darwin 1871
does not explain precisely how [i]instinctual
sympathygives rise to a desire for moral ap-
proval). Although Darwin was right about the
new possibilities that language created for so-
cial organization, the ethnographic evidence
discussed below suggests that hunter-gath-
erers employ coercion to enforce behavior
approved by public opinion. There is good
reason to think that coercive methods were
employed by ancient humans as well. We do
value the moral approval our fellows, and
this is also an important source of motivation
to conform to group rules (Mameli 2013:
911). But our desire for moral approval is
most likely an evolved response to the wide-
spreadsocial practiceofcollectively punishing
norm violators, and is not simply an outgrowth
of sympathy.
Language allows members of coalitions in
a hunter-gatherer band to collectively agree
on what Boehm terms a blueprintupon
which to model their society. Echoing Dar-
win, he says that because humans are able
to communicate in great detail, . . . groups
can develop precise notions about the kind
of society in which they wish to live(Boehm
1999:193). Whether, as a matter of historical
fact, humans began collectively suppressing
alphas and then started enforcing other moral
behavior or vice versa (see Boehm 1999:
194), ultimately hunter-gatherers came to
implement blueprints of egalitarian, moral
societies. Boehm (2012:Table 1) reports the
number of cases recorded by ethnologists
where capital punishment was meted out for
various offenses in 50 mobile foraging bands.
There are records of capital punishment in
24 of the 50 bands, for a total of 45 instances.
Boehm notes that this is likely a signicant
underestimate of the prevalence of capital
punishment, because foragers have learned
to conceal such practices from outside au-
thorities (the count here also does not in-
clude expulsion, which, as noted, is often a
defacto deathsentence). The mostcommon
crimes eliciting a lethal response were vari-
ous forms of intimidating the group with
aggression, violence, or malicious sorcery.
But people were also executed for such mo-
ralviolations as theft,failingtoshare meat,in-
cest, adultery, and failing to respect taboos.
To reiterate, execution is just the most ex-
tremeform ofpunishment. The vastmajority
of norm violators in hunter-gatherer bands
reform in response to less serious sanctions
POWER IN CULTURAL EVOLUTIONDecember 2018 309
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
such as ridicule. People use the collective
power of the group to make society conform
to a desired blueprint.
Humans were hunting with wooden weap-
ons at least 400,000 bp, and were taking down
big game as a regular source of food 250,000
bp (Boehm 2012:146). Boehm speculates that
the key impetus for reversing the dominance
hierarchy came when meat from large ani-
mals became a crucial part of our diet. Hunt-
ing bands can only be effective if all hunters
are reasonably well nourished and motivated,
and this requires that the meat be shared
rather than hoarded by an alpha and his
coalition partners (Boehm 2012:151152) or
taken by greedy thieves and cheaters(Bo-
ehm 2012:155).
When coalitions of the majority started to
collectively enforce group rules, this created
a strong selection pressure for the develop-
ment of a conscience, and concomitant emo-
tions such as shame and moral pride. To
have a conscience is to personally [identify]
with community values, which means internal-
izing your groups rulesso that you are emo-
tionally invested in following them (Boehm
2012:113). The human capacity for language
allows for an extreme degree of effective sur-
veillance. No matter who witnesses a rule vio-
lation,word canspreadthroughoutthewhole
group. People who failed to control their
predatory tendencies(Boehm 2012:67)
who bullied or cheated or otherwise violated
group normswould have been at a great
disadvantage. Those who were able to inter-
nalize norms as goals in themselves would
have avoided trouble much better than those
who, like chimps, observed the rules only
when they thought an authority was watch-
ing (Mameli 2013). Given the prospect of
group punishment, developing a good repu-
tation became an important way to increase
individual tness.
Consider an illustrative example of group
enforcement of norms, reported by Turnbull
(1961:94108) and cited by Boehm (2012:
3743). The Mbuti have a hunting practice
where the men set up long nets in the shape
of a semicircle while the women and chil-
dren make noise to drive animals (small- and
medium-sized game) toward them. Accord-
ing to their rules, a man can kill any animal
that falls into his own net and take the meat
for his own family. Although there may be
chance variation in how well a man fares on
a given hunting expedition relative to other
men, on average this is a fair way of distribut-
ing the spoils.
In the course of one such hunt, a man
named Cephu felt he was getting short-
changed, so he surreptitiously moved his net
ahead of the others. The strategy worked as
far as increasing his catch but, unfortunately
for him, he was spotted by another hunter.
Word quickly spread of his misdeed. Most of
the other families arrived back at the camp
before him, and a man named Kenge an-
nounced: Cephu is an impotent old fool. No,
he isnt, he is an impotent old animalwe
have treated him like a man for long enough,
now we should treat him like an animal. Ani-
mal!(Boehm 2012:3839). This triggered an
outpouring of gossip and condemnation of
Cephu as, in Boehms words, a group con-
sensus materialized(Boehm 2012:39).
Cephu soon returned, and Kenge shouted
at him that he was an animal. Everyone else
was silent and for a short while just ignored
him. Then, after the whole crowd confronted
Cephu with vague accusations that he was
selsh, someone nally made a direct accu-
sation against him that he had stolen meat.
One man said that he hoped Cephu would
fall on his spear and kill himself like the an-
imal he was. Who but an animal would steal
meat from others? There were cries of rage
from everyone, and Cephu burst into tears
(Turnbull 1961; quoted in Boehm 2012:39
40).
Atrst Cephu argued that he had made an
honest mistake. When no one accepted this
excuse, he claimed that he deserved to place
his net in a better position, considering that
he was an important man, a chief, in fact,
of his own band(referring to his extended
family;Boehm 2012:40). Someoneresponded
that the Mbuti do not have chiefs. If Cephu
was a chief of his own band, let him go with
it and hunt elsewhere and be a chief else-
where(Boehm2012:40)asuggestionthat,
were it followed, would have meant starva-
tion in the forest for Cephu and his family.
310 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Faced with the threat of expulsion, Cephu
began to apologize profusely. He agreed to
turn over all of his meat from the days catch
to his accusers, and the other band members
then took everything from him and his wife.
Despite all the drama, a few hours later he
participated with everyone else in the evening
singing and all was set right againand pre-
sumably he never repeated his misdeed.
Notice how this real-life example of en-
forcing a cooperative norm is fundamentally
different from what was possible in the ex-
perimentalsetupofGürerketal.(2006;dis-
cussed in the introduction to this paper). All
of the members of the group explicitly coor-
dinated in advance, agreeing on what they
foresaw would be a set of practices that would
lead the group to collectively acquire as much
meat as possible and to distribute it fairly.
Members ofthe society did not havea realistic
option to opt out of the systemthey could
not choose to participate in a sanction free
institution.The decision to adopt a coopera-
tive institution and punish defectors/deviants
was made collectively. In Cephus case, we
see that punishment was preceded by collec-
tive coordination. Kenge led the group by
asserting that they should change the way
they treat him: for his misdeed they should
treat him as an animal. Before Cephu ar-
rived, all of the members of the group (be-
sides his immediate family members) had
shared theirsense of outrage with each other,
and established that they were on the same
page.Wheneveryone confrontedCephu, they
were not doing so as individuals expressing
their personal anger, but as members of a
group that had already come to a collective
understanding.
The collective rebuke and punishment of
Cephu illustrates how our moral sense plays
an important role in undermining the radi-
cal individualistic approach of cultural evo-
lutionary models. Chimp society is despotic
and without morality. In humans, collectively
imposed and accepted moral rules allow co-
ordinated group action among all those who
accept the rules. Hunter-gatherers use mo-
rality to enforce egalitarianism, whereas sed-
entary, resource-hording people use morality to
enforce despotism. Whether morality is used
to promote egalitarianism or despotism, it
binds people into a collective decision-making
body that cannot be legitimately atomized, as
in cultural evolutionary models.
Cultural Evolutionary Models of the
Development of Prosocial Norms
In cultural evolutionary models, cultural
variants are distributed in a population, and
individuals with content and context learn-
ing biases choose whom to imitate. Modeling
work suggests that, in populations composed
of social learners interacting with each other,
stable behavioral patterns will emerge as a
by-productof our learning biases (Chudek
et al. 2013:442). These behavioral equilibria
are more likely to be group detrimental than
benecial (Boyd and Richerson 1992; Hen-
rich and Boyd 2001:86; Chudek and Henrich
2011:222). But, according to other models,
cultural group selection favors the spread of
those equilibria that happen to be benecial
(e.g., Boyd and Richerson 2002). Then, in a
process of gene-culture coevolution, we be-
come genetically disposed to be receptive to
prosocial behavioral patterns, or norms(Hen-
rich and Boyd 2001; Chudek and Henrich
2011; Chudek et al. 2013).
The aforementioned models assume that
there is no organized enforcement of norms
such as the collective enforcement described
in the previous section. Boyd and Richerson
say that [f]or most of human history, states
were weak or non-existent, and norms”—
such as rules against murder”—“were not
enforced by external sanctions(Boyd and
Richerson 2002:287). Each individual makes
a decision whether or not to imitate the be-
havioral patterns represented in the group.
Among these behavioral patterns are ten-
denciesto punish those whofailto exhibitcer-
tain other behaviors. Punishing norm violators
is just another cultural variant that can spread
among individuals. A norm (whether it is ben-
ecial or detrimental to the group) can be sta-
bilized (in part) by punishment. Punishment
is presumed to involve a cost, but it can be re-
inforced by the punishment of nonpunishers.
Punishment of nonpunishers does not have
POWER IN CULTURAL EVOLUTIONDecember 2018 311
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
to go on forever (punishing those who fail to
punish nonpunishers) because a combina-
tion of conformism and fear of punishment
will drive cooperation at the rst level to
near xation, which in turn lowers the cost
of rst-order punishment (Boyd and Richer-
son 1992).
As noted, Boyd and Richersons models
are based on the assumption that during the
period of gene-culture coevolution in which
norms rst evolved and we became adapted
to live in a norm-governed environment,
norms were not enforced by external sanc-
tions(Boyd and Richerson 2002:287). The
ethnographic evidence discussed in connec-
tion with Boehms theory suggests that this is
false. Although there were no statesor gov-
ernments in the modern sense, there was a
decision-making bodya coalition of the ma-
joritythat exercised coercive power, which
amounts to what is effectively the same thing.
The cultural variants to refrain from murder,
to share meat, and so on were collectively
adopted and formally enforced, not just pun-
ished by isolated individuals who possessed
a variant to punish those who did not exhibit
the relevant behaviors (or to punish in re-
sponse to signals that others will join in their
punishment, as in Boyd et al.s 2010 model,
discussed earlier).
Chudek et al. (2013) list three possibl e ways
that different behavioral equilibria could be
selected. The rst is that rational, forward-
looking individualsperceive the ultimate ben-
ets of being in a cooperative equilibrium,
assume others are similarly sensible, and
choose the prosocial state(Chudek et al.
2013:439). They give three reasons why they
think this was not a signicant factor in real
life. First, Chudek et al. say, people are not
actually good at making the calculations re-
quired to determine what practices would
be benecial. Second, group decisions are
often heavily inuenced by leaders and coali-
tions whose interests diverge from the over-
all group(Chudek et al. 2013:439). Third,
we see many examples of patently nonproso-
cial institutions in societies throughout the
world. Chudek et al. seem to recognize the
possibility of group decisions(in the sec-
ond reason), but they do not explore the im-
plications of this phenomenon for cultural
evolutionary theory. Insofar as group deci-
sions do play a role in cultural evolution,
these authors think that they tend to reect
the narrow self-interest of powerful political
factions. They do not consider the possibil-
ity that group decisions can be made by a
coalition of the majority that is more or less
interested in the success of the group as a
wholebut if Boehm was right then this is
in fact a big part of how groups in the ancestral
environment selectedamong different possi-
ble behavioral equilibria to collectively en-
force (cf. Singh et al. 2017:470471).
The second possible way Chudek et al.
(2013) mention to select among behavioral
equilibria is stochasticity.By chance, groups
move from one equilibrium to another, and
groups are likely to spend more time in equi-
libria with larger basins of attraction.
The third way is cultural group selection,
whichacts on thevariationprovided (in part)
by stochasticity. Cultural evolutionists iden-
tify cultural group selection as by far the most
important factor in the spread of group-
benecialnorms. Groupsrandomly adopted
more or less group-benecial or detrimental
norms, and those with benecial norms sur-
vived, enjoyed more immigration, and their
individuals were preferentially copied by in-
dividuals in other groups.
Boyd and Richerson (2002) provide a
model showing that, in a population consist-
ing of small groups with behavioral norms
at different equilibria, group-benecial norms
can spread if people tend to copy more suc-
cessful individuals in either their own or in
neighboring groups. Individuals in groups
with more group-benecial norms will tend
to be more successful and, therefore, their
norms will be copied by the members of other
groups. But this model fails to account for
the empirical fact that moral norms are for-
mally enforced (by collective action) within
foraging bands. Individuals from one band
cannot copy the moral norms of individuals
from another (imagine if Cephu responded
to the accusations against him by saying that
he subscribed to the different meat-sharing
practices of a neighboring people). However,
agroupcancollectively decide to copy another,
more successful group, and begin enforcing
a new suite of norms.
312 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Henrich describes prestige-biased group
transmission(Henrich 2016:168) as a mat-
ter of individualsin one group copying in-
dividualsin more successful groups. But
he also describes some striking cases where
the leadership of one group decides to adopt
the practices of another group. In the case
of the Irakia Awa of New Guinea, senior
men(Henrich 2016:173; emphasis added)
decided to copy the pig-rearing practices of
their more economically successful neighbors,
the Fore. The process of transmission did not
involve individuals copying individuals. In
Henrichs words: [T]his transmission be-
tween groups occurred rapidly because the
Irakia already had a political institution in
the village, which involved a council of the
senior members of each clan, who were em-
powered by tradition (social norms) to make
community-level decisions(Henrich 2016:
174).
Power played a complicated role in the
adoption of new communal practices among
the Irakians. A few years after they copied the
Fore, some young men did not want to con-
tinue raising pigs anymore. When they com-
municated their preference to the village
elders, the elders would not even discuss it
[and] . . . disparaged the idea and criticized
the younger people for being lazy and un-
willing to lead proper lives. . . . The young
men . . . admitted that it would be impossible
to make such a change with the elders rmly
against it(D. J. Boyd 2001:270). However,
another group of young adults converted to
Seventh-Day Adventisma religion that pro-
hibits the consumption of pork. They outright
refused to raise pigs, and no serious sanctions
were imposed on them. Apparently the power
(that probably took the form of authority)of
the village elders was substantial but not ab-
solute.
Although the Irakians were not foragers,
this story illustrates how the assumptions of
cultural evolutionary models (that transmis-
sion is from individual to individual and that
variants are not imposed on groups of peo-
ple all at once by means of power) do not
apply to many real-life scenarios that have
been studied by cultural evolutionists. Hen-
richsverbal description of how the Irakians
adopted the Fores practices refers to forces
that cultural evolutionary models do not ac-
commodate.
Gene-Culture Coevolution: The
Development of Norm Psychology
As discussed, cultural evolutionary theo-
rists argue that the learning biases that lead
us to acquire adaptive cultural information
tend to lead, when many individuals inter-
act with each other, to the development of
norms—“stable group-wide patterns of be-
havior(Chudek et al. 2013:442). Cultural
evolutionary models of the development of
transmitted behavioral patterns show that a
variety of norms can be stable (Chudek et al.
2013:438439). Some equilibria are group
benecial, such as those involving widespread
cooperation and punishment of defection.
Most equilibria are group detrimental, such
as those involving widespread noncoopera-
tion, or widespread, enforced performance
of costly rituals (Chudek and Henrich 2011:
222 give several examples, including dis-
ease-spreading endocannibalism). Although
group-detrimental norms are more likely than
benecial norms to develop out of the in-
teraction among individuals within a group
(according to the modelers), cultural group
selection will nevertheless tend to favor the
proliferation of the latter.
Groups that happen to have cooperative
norms tend to outsurvive other groups, re-
ceive more immigrants, and (as discussed
above) their individuals tend to be selected
as models to imitate by individuals in other
groups. On Chudek et al.s (2013) account,
because norms inevitably arise as a by-prod-
uctof cultural learning, natural selection
would have favored innate dispositions for
handling normsanorm psychology.They
suggest that our innate norm psychology pre-
paresustorecognize,andmotivatesustoob-
serve, the social norms in our environment.
Since cultural group selection caused coop-
erativenorms(includingnormsto punish de-
fectors and punish nonpunishers) to prevail
in the ancestral environment, our norm psy-
chology should make us especially disposed
to acquire prosocial norms and punish vio-
lators (see also Henrich and Boyd 2001:87).
Indeed, in accordance with Chudek et al.s
POWER IN CULTURAL EVOLUTIONDecember 2018 313
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
predictions,evidence suggeststhatyoung chil-
dren automatically infer behavioral rules by
observing people (particularly adults), and
that they internalize rule adherence as a per-
sonal goal and even enforce observance in
other children (Chudek and Henrich 2011;
Chudek et al. 2013).
The theories of both Boehm and Chudek
et al. predict the existence of a norm psy-
chology.Both claim that Paleolithic hunter-
gatherers lived in norm-governed societies
and that individual success was tied to the
ability to recognize and follow local norms,
which tended to be prosocial. Laboratory ex-
periments revealing our disposition to infer
and follow norms will not adjudicate between
the theories.
The theories, although making virtually
the same predictions about how recent hom-
inin evolution shaped our psychology, give
different explanatory accounts of how pro-
social norms initially spread, and they paint
radically different pictures of the human ca-
pacity to deliberately guide our own evolution
(both cultural and genetic). The empirical
evidence reviewed in this paper seems to fa-
vor Boehm. Among nomadic hunter-gather-
ers around the worldall that have been
studied by anthropologistspowerful coali-
tions enforce prosocial norms with a great
deal of explicit awareness of the social bene-
ts that observance of these norms will pro-
duce. These coalitions enforce the norms
because individuals have an explicit blue-
print in their minds for what kind of society
they want to live in, and norm enforcement
is a deliberate strategy for bringing that soci-
ety about. In many cases humans are blind to
the consequences of their socially transmitted
practicesthe evidence for that is undeni-
able (Henrich 2016). But our hunter-gatherer
ancestors were not blind to the consequences
of all of their practices. When coalitions of
the majority seized power from alphas, they
gained the ability to impose practices that
had consequences that they favored.
an alternative explanation for the
development of prosocial behavior:
the mutualistic theoryof morality
Baumard et al. (2013) defend an alterna-
tive explanation for the evolution of moral
judgment and prosocial behavior. They sug-
gest that, in the ancestral environment, peo-
ple could benet by cooperating with each
other in a variety of ways: they could hunt to-
gether, share food with the expectation of fu-
ture reciprocation, and so on. Each person
would bring a certain amount of resources,
effort, and talent to cooperative ventures.
This led people to compete to be chosen as
partners for cooperation. If Awill get a lower
return by cooperating with Bthan Awould
get on average from cooperating with some-
one else, Awill be better off not choosing B
as a partner, and Bwill lose out. Our ances-
tors faced the adaptive challenge of seeking
out good exchange partners while making
themselves attractive exchange partners to
others. Baumard et al. argue that, as a conse-
quence, we evolved a moral disposition to
value fairness”—the notion that people are
entitled to share in the product of coopera-
tion in proportion to their contribution to
creating that product. In hunter-gatherer
tribes the moral order was not sustained by
collective agreement about rules and collec-
tively administered punishment, but by indi-
viduals exercising partner choice.
Baumard et al.s (2013) theory seems to
explain a range of ndings in experimental
games. For example, in dictator games, where
one person (the dictator) decides how to di-
vide a pot of money with an anonymous part-
ner, dictators tend to keep most of the pot
for themselves. But suppose the dictator is
asked to distribute money that was earned by
the anonymous partner by performing well
in a quiz contest or on an exam. In that case,
dictators tend to become very generous, and
give their partners more or less what they
earned (Rufe 1998; Oxoby and Spraggon
2008). We seem to intuitively feel that people
are entitled to the product of their efforts
and, according to Baumard et al. (2013), we
object to inequality only when it is the conse-
quence of unfairness.
The theories of Baumard et al. and Boehm
can, to some extent, be reconciled. It could
be that groups began collectively enforcing
moral codes around 250,000 bp (as Boehm
argues), but voluntary mutualistic coopera-
tion contributed to the evolution of our
moral sense (in the way described by Baumard
et al.) before that time. Even if early human
314 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
groups were dominated by alphas who bul-
lied group members and sometimes appro-
priated resources for themselves by force, not
all interactions had to be based on bullying:
group members could still have cooperated
with each other in some contexts, and com-
peted to be chosen as cooperation partners.
Furthermore, even after hunter-gatherers es-
tablished a reverse dominance hierarchy and
egalitarian political norms, people were still
allowed to obtain unequal rewards due to
greater ability in some contexts. Although the
rules among nomadic foragers for sharing
meat from big game generally demand a
more or less equal distribution, the rules for
sharing other kinds of food are often much
looser, and people who contribute more to
obtaining such food may have a degree of
freedom to distribute it as they wish (see Gur-
ven 2004). Even if, since 250,000 bp, hunter-
gatherer bands were politically egalitarian
(amongmales)and biggame hunting/distri-
bution was a largely socialist enterprise, peo-
ple still formed cooperative relationships with
each other on a smaller scale, and it would
have been advantageous for them to follow
norms of fairness even if these norms were
not always enforced by the collective effort
of the group.
Although some of our behavior and moral
intuitions can be explained by Baumard
et al.s (2013) account, we cannot discount
the evidence that antisocial behavior is (and
presumably was) often collectively punished
in egalitarian bands, and that our moral psy-
chologyistosomeextenttailoredtolivingin
such egalitarian bands. As noted, Baumard
et al. suggest that antisocial hunter-gatherers
are, in general, not formally punished,but
rather they suffer when other members of
their group refuse to cooperate with them.
But the anthropological evidence reviewed
above suggests that hunter-gatherers do not
merely withdraw cooperation from antiso-
cial individuals. Groups collectively mete out
positive punishmentsthe death penalty is
widely administered for a range of offenses.
It is true that, as Baumard et al. say, [p]un-
ishment . . . is uncommon in societies of for-
agers(Baumard et al. 2013:66). But this can
be explained by the fact that the threat of
punishment preemptively stops most serious
offenses (cf. Service 1975:4849). The death
penalty rarely needs to be administered be-
cause it is preceded by many escalating warn-
ings of what is to come should offenders fail
to reform.
The hypothesis that our moral sense is cal-
ibrated for living under conditions of some-
what enforced egalitarianism also explains
some results from experimental games, which
Baumard et al. (2013) cannot so easily ex-
plain. They discuss the following experiment
conducted by Dawes et al. (2007). Subjects
were divided into groups containing four
anonymous members for a one-shot interac-
tion.Eachsubjectwasgivenarandomamount
of money by a computer, and was informed
how much money had been granted to the
other three members. The players then had
the chance to give positiveor negativeto-
kens to each other. Giving a positive or a neg-
ative token cost the giver one monetary unit,
and increased or decreased (respectively) the
recipients payoff by three monetary units.
Groups were broken apart and formed with
new anonymous members to play again for
a total of ve rounds.
The results of Dawes et al.s study suggest
that people are willing to pay a penalty to
equalize outcomes. In the course of ve
rounds, 68% of subjects reduced another
players income at least once, 28% did so ve
times or more, and 6% did so ten times or
more. ...74%...increased another players
income at least once(Dawes et al. 2007:
794). Those who received a very high initial
endowment tended to receive many nega-
tive tokens, while those who received a very
low initial endowment tended to receive
many positive ones. Subjectstoken-distrib-
uting behavior cannot be seen as a rational
strategy because, again, the interactions were
one shot. It cannot be interpreted as retali-
ation or punishment, since subjects knew
that all of the payoffs had been determined
randomly by a computer. Rather, subjects
seemed to be bothered by inequality per
se. Baumard et al. offer the following ex-
planation: Overall, the distribution of [to-
kens] . . . displays the logic of fairness: The
moreaparticipant received money,themore
others would taxher. Conversely, the less
she received, the more she would get com-
pensated’” (Baumard et al. 2013:75). How-
ever, the way subjects reacted to inequality
POWER IN CULTURAL EVOLUTIONDecember 2018 315
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
does not make them attractive partners for
cooperation. Paying a personal cost in order
to reduce the payoff of luckier individuals
simply reduces the expected payoff from en-
gaging in cooperation. Unlike a tax,which
redistributes payoffs and thereby benets the
recipient, the negative tokens administered
by 68% of subjects reduced both their own
and the recipientspayoffs without beneting
anyone. Thus, the disposition that motivates
negative-token giving does not make us at-
tractive partners in cooperation. But the be-
havior is easily explained if we assume that
we are adapted to living in hunter-gatherer
bands where a certain degree of egalitarian-
ism was enforced, and people were prohib-
ited from accumulating signicantly more
resources than their fellows.
Discussion
Mainstream cultural evolutionary theory in
the Boyd and Richerson tradition assumes that
cultural variants spread as a consequence of
individualscontent and context learning bi-
ases, Darwinian selection, and random forces
in a way that is amenable to mathematical
modeling. Durham (1991), Fracchia and Le-
wontin (2005), and Lewontin (2005) raise
the challenge that often cultural variants are
spread through the exercise of power, imply-
ing that the target of explanation for cultural
evolution should be the behavior of power-
ful individuals and groups. This paper argued
that Durham and Fracchia and Lewontins
critique may be valid as far as the evolution of
morality and prosociality goes (at least since
250,000 bp), but less so for the spread of non-
moral cultural variants (perhaps until the
adventofsedentarism,food storage, and agri-
culture).
Cultural evolutionists emphasize that their
models are not intended to capture every-
thing that happens in real life (Richerson
and Boyd 1987; McElreath and Boyd 2007:
46). Like all models, they are meant to be
simplications that isolate some of the main
forces at play. But if something like Boehms
(1999, 2012) account of the evolution of mo-
rality is correct, cultural evolutionary models
of the evolution of morality and the coevolu-
tion of norms and norm psychology are distor-
tions,notsimplications. A legitimate model
may isolate one or a few forces among the
multitude that exist in the real world. It can-
not postulate forces that are not operative at
all, or ignore those that have a decisive inu-
ence on the phenomena under investigation.
In light of the anthropological evidence
reviewed in this paper, we can see how Boyd
et al.s (2010) model of the evolution of co-
ordinated punishmentdistorts history. Ac-
cording to Boyd et al., there was a genetic
mutation(s) associated with the behavior to
punish cheaters if τother people in my group
(honestly) signal that they are prepared to punish
cheaters (again, this mutation can survive bet-
ter than the mutation associated with always
punish cheaters, since always punishers are
disadvantaged when they are a small minor-
ity). In a population containing some condi-
tional punishers, by chance a few groups will
have the threshold number (i.e., τ+ 1), and
they will prosper and replace other groups,
thereby increasing the number of conditional
punishers.
The story that Boehm (1999, 2012) tells is
very different. He says that coalitions within
human groups formed and explicitly agreed
to enforce certain kinds of prosocial behav-
iorthroughpunishment.Thepeopleinvolved
did not have any preexisting disposition to
punish. The practice of punishing people to
enforce prosocial norms arose all at once.
In contrast, on Boyd et al.s (2010) account
the genetically based disposition to punish
arose rst, then people developed a social
norm to punish. On Boehms (1999, 2012)
account, because groups started practicing
collective punishment, this created selection
pressures for people with a genetic disposi-
tion to punish norm transgressors.
To be clear, the point of this paper is not to
say that there are real-life complexities that
cultural evolutionary models fail to capture.
The point is that some of the fundamental
forces that drive cultural evolution, and drove
the coevolution of norms and norm psychol-
ogy, cannot be accommodated by the sorts
of models used in cultural evolutionary the-
ory. Most of the mathematical models em-
ployed in cultural evolutionary theory are
best adapted to cases whereforces act in regu-
lar, iterative ways. Perhaps evolution in some
domains of culture does work like thatcul-
tural evolutionary modeling in these domains
316 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
is safe from Durham (1991) and Fracchia and
Lewontins (2005) challenge. In other do-
mainsincluding the morala more tradi-
tional historical/anthropological/sociological
approach may be more fruitful (and legiti-
mate).
acknowledgments
I am grateful to Andreas Mogensen for valuable con-
versations on this topic and feedback on multiple drafts
of this paper. I received helpful comments on previous
drafts from Christopher Boehm, Neven Sesardić, and
two anonymous reviewers.
REFERENCES
Bachrach P., Baratz M. S. 1970. Power and Poverty: The-
ory and Practice. New York: Oxford University Press.
Baumard N., André J.-B., Sperber D. 2013. A mutualis-
tic approach to morality: the evolution of fairness by
partner choice. Behavioral and Brain Sciences 36:5978.
Boehm C. 1993. Egalitarian behavior and reverse dom-
inance hierarchy. Current Anthropology 34:227240.
Boehm C. 1997. Impact of the human egalitarian syn-
drome on Darwinian selection mechanics. American
Naturalist 150:S100S121.
Boehm C. 1999. Hierarchy in the Forest: The Evolution
of Egalitarian Behavior. Cambridge (Massachusetts):
Harvard University Press.
Boehm C. 2012. Moral Origins: The Evolution of Virtue,
Altruism, and Shame. New York: Basic Books.
Bowles S., Gintis H. 2011. A Cooperative Species: Human
Reciprocity and Its Evolution. Princeton (New Jersey):
Princeton University Press.
Boyd D. J. 2001. Life without pigs: recent subsistence
changes among the Irakia Awa, Papua New Guinea.
Human Ecology 29:259282.
Boyd R., Richerson P. J. 1985. Culture and the Evolutionary
Process. Chicago(Illinois):Universityof Chicago Press.
Boyd R., Richerson P. J. 1992. Punishment allows the
evolutionofcooperation(oranythingelse)insizable
groups. Ethology and Sociobiology 13:171195.
Boyd R., Richerson P. J. 2002. Group benecial norms
can spread rapidly in a structured population. Jour-
nal of Theoretical Biology 215:287296.
Boyd R., Gintis H., Bowles S. 2010. Coordinated punish-
ment of defectors sustains cooperation and can pro-
liferate when rare. Science 328:617620.
Cashdan E. A. 1980. Egalitarianism among hunters and
gatherers. American Anthropologist 82:116120.
Chudek M., Henrich J. 2011. Culture-gene coevolution,
norm-psychology and the emergence of human pro-
sociality. Trends in Cognitive Sciences 15:218226.
Chudek M., Heller S., Birch S., Henrich J. 2012. Prestige-
biased cultural learning: bystanders differential at-
tention to potential models inuences childrens
learning. Evolution and Human Behavior 33:4656.
Chudek M., Zhao W., Henrich J. 2013. Culture-gene
coevolution, large-scale cooperation, and the shap-
ing of human social psychology. Pages 425457 in
Cooperation and Its Evolution, edited by K. Sterelny,
R. Joyce, B. Calcott, and B. Fraser. Cambridge (Mas-
sachusetts): MIT Press.
Dahl R. A. 1957. The concept of power. Behavioural Sci-
ence 2:201215.
Darwin C. 1871. The Descent of Man, and Selection in Re-
lation to Sex, Volume 1. London (United Kingdom):
John Murray.
Dawes C. T., Fowler J. H., Johnson T., McElreath R.,
Smirnov O. 2007. Egalitarian motives in humans.
Nature 446:794796.
Durham W. H. 1991. Coevolution: Genes, Culture, and Hu-
man Diversity. Stanford (California): Stanford Uni-
versity Press.
Erdal D., Whiten A. 1994. On human egalitarianism: an
evolutionary product of Machiavellian status escala-
tion? Current Anthropology 35:175178.
Fracchia J., Lewontin R. C. 2005. The price of metaphor.
History and Theory 44:1429.
Goodall J. 1986. The Chimpanzees of Gombe: Patterns of
Behavior. Cambridge (Massachusetts): Harvard Uni-
versity Press.
Gürerk Ö., Irlenbusch B., Rockenbach B. 2006. The
competitive advantage of sanctioning institutions.
Science 312:108111.
Gurven M. 2004. To give and to give not: the behavioral
ecology of human food transfers.Behavioral and Brain
Sciences 27:543559.
Henrich J. 2006. Cooperation, punishment, and the evo-
lution of human institutions. Science 312:6061.
Henrich J. 2016. The Secret of Our Success: How Culture Is
Driving Human Evolution, Domesticating Our Species, and
Making Us Smarter. Princeton (New Jersey): Princeton
University Press.
Henrich J., Boyd R. 2001. Why people punish defectors:
weak conformist transmission can stabilize costly en-
forcement of norms in cooperative dilemmas. Jour-
nal of Theoretical Biology 208:7989.
Henrich J., McElreath R. 2003. The evolution of cultural
evolution. Evolutionary Anthropology 12:123135.
Horner V., Proctor D., Bonnie K. E., Whiten A., de
Waal F. B. M. 2010. Prestige affects cultural learning
in chimpanzees. PLOS ONE 5:e10625.
Knauft B. M. 1991. Violence and sociality in human evo-
lution. Current Anthropology 32:391409.
Lewens T. 2007. Darwin. London (United Kingdom):
Routledge.
Lewens T. 2015. Cultural Evolution: Conceptual Chal-
lenges. Oxford (United Kingdom): Oxford Univer-
sity Press.
POWER IN CULTURAL EVOLUTIONDecember 2018 317
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
Lewontin R. C. 20 October 2005. The wars over evolu-
tion. New York Review of Books. http://www.nybooks
.com/articles/2005/10/20/the-wars-over-evolution/.
Lukes S. 2005. Power: A Radical View. Second Edition.
New York: Palgrave Macmillan.
Mameli M. 2013. Meat made us moral: a hypothesis on
the nature and evolution of moral judgment. Biology
& Philosophy 28:903931.
McElreath R., Boyd R. 2007. Mathematical Models of So-
cial Evolution: A Guide for the Perplexed. Chicago (Illi-
nois): University of Chicago Press.
McGrew W. C. 2004. The Cultured Chimpanzee: Reections
on Cultural Primatology. Cambridge (United Kingdom):
Cambridge University Press.
Ostrom E., Walker J., Gardner R. 1992. Covenants with
and without a sword: self-governance is possible.
American Political Science Review 86:404417.
Oxoby R. J., Spraggon J. 2008. Mine and yours: prop-
erty rights in dictator games. Journal of Economic Be-
havior & Organization 65:703713.
Richerson P. J., Boyd R. 1987. Simple models of com-
plex phenomena: the case of cultural evolution.
Pages 2752 in The Latest on the Best: Essays on Evolu-
tion and Optimality, edited by J. Dupré. Cambridge
(Massachusetts): MIT Press.
Richerson P. J., Boyd R. 2005. Not by Genes Alone: How
Culture Transformed Human Evolution. Chicago (Illi-
nois): University of Chicago Press.
Richerson P. J., Henrich J. 2012. Tribal social instincts
and the cultural evolution of institutions to solve col-
lective action problems. Cliodynamics: The Journal of
Theoretical and Mathematical History 3:3880.
Rufe B. J. 1998. More is better, but fair is fair:tipping in
dictator and ultimatum games. Games and Economic
Behavior 23:247265.
Service E. R. 1975. Origins of the State and Civilization:
The Process of Cultural Evolution. New York: Norton.
Singh M., Wrangham R., Glowacki L. 2017. Self-interest
and the design of rules. Human Nature 28:457480.
Turnbull C. M. 1961. The Forest People: A Study of the Pyg-
mies of the Congo. New York: Simon and Schuster.
Associate Editor: Michael Ruse
Handling Editor: Daniel E. Dykhuizen
318 Volume 93THE QUARTERLY REVIEW OF BIOLOGY
This content downloaded from 080.042.128.018 on November 12, 2018 06:45:58 AM
All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
... To varying degrees we are able to critically evaluate alternative beliefs, values, and practices and adopt those that we believe advance our explicitly represented short-and long-term goals. Institutions that are the product of conscious design may create selection pressures that cause people to evolve teleologically to conform to their demands (Boehm 2012;Cofnas 2018). The capacity for culture-while being itself mostly the product of natural selection-allows us to transcend the forces of blind selection and direct even our biological evolution. ...
Article
Full-text available
According to the standard formulation, natural selection requires variation, differential fitness, and heritability. I argue that this formulation is inadequate because it fails to distinguish natural selection from artificial selection, intelligent design, forward-looking orthogenetic selection, and adaptation via the selection of nonrandom variation. I suggest adding a no teleology condition. The no teleology condition says that the evolutionary process is not guided toward an endpoint represented in the mind of an agent, variation is produced randomly with respect to adaptation, and selection pressures are not forward looking.
... The evolution of social norms is an important topic of study in research on cultural evolution. Cultural evolution theory has been fruitful in showing the importance of biased transmission in the spread and maintenance of cultural traits (Boyd and Richerson, 1985;Henrich and Boyd, 1998;Boyd and Richerson, 2002;McElreath et al., 2003;Efferson et al., 2008;Mesoudi, 2011b), but limited consideration has been given to power dynamics and their role in the spread of ideas (Singh et al., 2017;Cofnas, 2018) and this field of inquiry has somewhat ignored gender (Lawson et al., 2023). Instead, social influence is mostly considered in the form of deference and voluntary copying of the behaviour of high-status individuals rather than coercion by those in power. ...
Article
Full-text available
Sexual conflict theory has been successfully applied to predict how in non-human animal populations, sex ratios can lead to conflicting reproductive interests of females and males and affect their bargaining positions in resolving such conflicts of interests. Recently this theory has been extended to understand the resolution of sexual conflict in humans, but with mixed success. We argue that an underappreciation of the complex relationship between gender norms and sex ratios has hampered a successful understanding of sexual conflict in humans. In this paper, we review and expand upon existing theory to increase its applicability to humans, where gender norms regulate sex ratio-effects on sexual conflict. Gender norms constrain who is on the marriage market, how they are valued, and may affect reproductive decision-making power. Gender norms can also directly affect sex ratios, and we hypothesize that they structure how individuals respond to market value gained or lost through biased sex ratios. Importantly, gender norms are in part a product of women's and men's sometimes conflicting reproductive interests, but these norms are also subject to other evolutionary processes. An integration of sexual conflict theory and cultural evolutionary theory is required to allow for a full understanding of sexual conflict in humans.
... Evolutionary scientists studying social and cultural evolution have proposed a multitude of mechanisms by which cultural change can be effected: random change, biased transmission, and natural selection on cultural traits (Boyd & Richerson, 1992); selective imitation and selective migration (Richerson et al., 2016); cultural group selection (Richerson et al., 2016;Turchin, 2006); genetic group selection (Bowles et al., 2003); self-interested design (Gavrilets & Duwal Shrestha, 2021;Singh et al., 2017); evoked culture (Krasnow & Delton, 2016;Tooby & Cosmides, 1989); cognitive niche (Pinker, 2010;Tooby & deVore, 1987;Whiten & Erdal, 2012); and demographic swamping (Henrich, 2004). Which ones are more important is a controversial topic (Cofnas, 2018;Gavrilets & Duwal Shrestha, 2021;Krasnow & Delton, 2016Richerson et al., 2016Singh et al., 2017;Smith, 2020). ...
Article
Full-text available
Evolutionary scientists studying social and cultural evolution have proposed a multitude of mechanisms by which cultural change can be effected. In this article we discuss two influential ideas from the theory of biological evolution that can inform this debate: the contrast between the micro- and macro-evolution, and the distinction between the tempo and mode of evolution. We add the empirical depth to these ideas by summarizing recent results from the analyses of data on past societies in Seshat: Global History Databank. Our review of these results suggests that the tempo (rates of change, including their acceleration and deceleration) of cultural macroevolution is characterized by periods of apparent stasis interspersed by rapid change. Furthermore, when we focus on large-scale changes in cultural traits of whole groups, the most important macroevolutionary mode involves inter-polity interactions, including competition and warfare, but also cultural exchange and selective imitation; mechanisms that are key components of cultural multilevel selection (CMLS) theory.
... Fracchia and Lewontin (2005) remarked that power would be needed to raise pro-sociality in a society. This discussion is basically about whether power asymmetries follow on pro-sociality in societies (Cofnas, 2018) or vice versa. In this section, I will discuss theoretically the relation between pro-sociality and power asymmetries. ...
Article
Full-text available
Social hierarchies exist in democracies as well as in authoritarian societies. However, their nature is different. Democratic hierarchies are built bottomup through election while autocratic hierarchies are built top-down through domination. Both, however, have power asymmetries between the weaker citizens and the stronger politicians, which are amplified the stronger the hierarchies are. This manuscript introduces a model that combines pro-/anti-social behavior with different degrees of hierarchies. It is argued that this model has the power to categorize countries according to these criteria and indicate when and how societies move between democracy and authoritarianism. Importantly, I illustrate that the balance between power asymmetries and principal-agent chains is key for understanding when and why democracies sometimes transcend into authoritarianism.
Article
Recently there has been a rise in the application of concepts and methods from biological evolutionary theory to human cultures and societies where the aim is to explain these by describing them as population-level phenomena reducible to individual-level processes. I argue against this type of view by using Mesoudi's Cultural Evolution as a case study. I claim that Mesoudi’s ontological assumptions about cultures and societies are dubious and his methodological assumptions inadequate when it comes to addressing cultural and social phenomena. A consequence is that this approach to studying culture is, at the very least, incomplete and of limited application.
Article
Full-text available
I argue that our general disposition to make moral judgments and our core moral intuitions are likely the product of social selection—a kind of gene–culture coevolution driven by the enforcement of collectively agreed-upon rules. Social selection could potentially track mind-independent moral truth by a process that I term realist social selection: our ancestors could have acquired moral knowledge via reason and enforced rules based on that knowledge, thereby creating selection pressures that drove the evolution of our moral psychology. Given anthropological evidence that early humans designed rules with the conscious aim of preserving individual autonomy and advancing their collective interests, the theory of realist social selection appears to be attractive for moral realists. The goal of evolutionary debunking arguments should be to show not that our moral beliefs are the product of natural selection, but that realist social selection did not occur.
Article
Full-text available
Groupishness is a set of tendencies to respond to group members with prosociality and cooperation in ways that transcend apparent self-interest. Its evolution is puzzling because it gives the impression of breaking the ordinary rules of natural selection. Boehm's solution is that moral elements of groupishness originated and evolved as a result of group members becoming efficient executioners of antisocial individuals, and he noted that self-domestication would have proceeded from the same dynamic. Self-domestication is indicated first at ~300,000 years ago and has probably gathered pace ever since, suggesting selection for self-domestication and groupishness for at least 12,000 generations. Here I propose that a specifically human style of violence, targeted conspiratorial killing, contributed importantly to both self-domestication and to promoting groupishness. Targeted conspiratorial killing is unknown in chimpanzees or any other vertebrate, and is significant because it permits coalitions to kill antisocial individuals cheaply. The hypothesis that major elements of groupishness are due to targeted conspiratorial killing helps explain why they are much more elaborated in humans than in other species.