ARTICLE UDC 159.9.019.4 575.

Robin I.M. Dunbar

Why Humans Aren’t Just Great Apes1

Abstract: Although we share many aspects of our behaviour and bi-

ology with our primate cousins, humans are, nonetheless, different
in one crucial respect: our capacity to live in the world of the imagi-
nation. This is reflected in two core aspects of our behaviour that are
in many ways archetypal of what it is to be human: religion and
story-telling. I shall show how these remarkable traits seem to have
arisen as a natural development of the social brain hypothesis, and
the underlying nature of primate sociality and cognition, as human
societies have been forced to expand in size during the course of our
evolution over the past 5 million years.

Key words: human cognition, social brain, group size, intentional-

ity, religion, hominid evolution

Introduction

We share a long evolutionary history with the great apes, in particular: the
human lineage (including all the many fossil species that have existed over the
last 6 million years or so since our lineage parted company with that of the
chimpanzees) is firmly embedded within the African Great Ape family and
shares with them many aspects of their biology, genetics, psychology and
behaviour. More trivially, we also share a very high percentage of our DNA
with chimpanzees (though quite what this means is open to question). Yet, it
is surely obvious to everyone that we are not "just great apes". In several con-
spicuous ways, we are very different. That difference does not really seem to
concern the kinds of anatomical or cognitive differences – our bipedalism and
tool-making abilities – that anthropologists have tended to emphasise in the
past. Rather, I want to argue that the real difference lies in a much more in-

1
This paper is an edited version of the 6th Annual (2007) Psychology Lecture of the
British Academy and British Psychological Society, which will be published as: R.I.M.
Dunbar (2009). Mind the gap: or why humans aren’t great apes. Proceedings of the
British Academy 154: 000-000 (in press). I am grateful to the British Academy for per-
mission to republish the paper.

Issues in Ethnology and Anthropology n. s. Vol. 3. No. 3 (2008)

16 ROBIN I. M. DUNBAR

tangible set of competencies – the ability to live in the virtual world of the
mind. In a word, this is the world of culture. I will focus on two aspects of
human behaviour that are, in many respects, archetypal of human culture:
story-telling and religion. Both require us to be able to imagine worlds that do
not physically exist. I shall argue that, apes’ much-vaunted capacities for cul-
tural learning notwithstanding, no other living species is even on the same
page as humans in this respect – because all other animal species lack the
neuronal computational power required to make it possible. The key to under-
standing why this is so lies in the reasons why our brains have evolved.

The Evolution of the Social Brain

The social brain hypothesis was first proposed (as the Machiavellian Intel-
ligence Hypothesis) in the late 1980s by Byrne & Whiten (1988) as an expla-
nation for the fact that primates have much larger brains for body size than
any other taxonomic group (a point first noted by Jerison 1973). Most of this
increase in brain size is the result of an enlarged neocortex (Finlay & Darling-
ton 1990; Finlay et al. 2000), and especially an enlarged frontal lobe (though
some other subcortical areas such as the cerebellum are also differentially
enlarged in humans: McLeod et al. 2003). Primates rely on sociality as a tool
for solving the everyday ecological problems of survival and successful re-
production rather than solving these problems by individual trial and error. So
the proposal was that primates’ more complex social lives imposed signifi-
cantly greater cognitive demands on them than was the case for other non-
primate species.
Since this suggestion was originally proposed, considerable evidence has
been adduced in its support (for recent summaries, see Whiten and Byrne
1997; Emery et al. 2007; Dunbar and Schultz 2007a, b). One of the core find-
ings was a quantitative relationship between social group size and relative
neocortex size in primates (Fig. 1). On a double-log plot, mean species group
size is linearly related to relative neocortex size (indexed as the ratio of neo-
cortex volume to the volume of the rest of the brain). This has been inter-
preted as implying that some aspect of cognition imposes a constraint on the
number of relationships that an individual can maintain as a coherent social
network.
Although the essence of the social brain hypothesis is really about social
complexity and its cognitive demands (and considerable evidence is now
available to show that various indices of behavioural complexity correlate
with neocortex size: see Dunbar and Shultz 2007a), it is this quantitative rela-
tionship with group size that has mainly attracted attention. Indeed, it seems
that the social brain hypothesis is often seen as being synonymous with this


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 17

one core finding. However, it is important not to lose sight of the fact that this
group size effect is really only an emergent property of the underlying rela-
tionship which focuses on the cognitive demands of behavioural complexity.
This has recently been given added emphasis by new findings that have
emerged from attempts to test the social brain hypothesis on non-primate spe-
cies (Shultz and Dunbar 2007; Dunbar and Shultz 2007b) have shown that,
across a wide range of bird and mammal species (specifically carnivores,
ungulates and bats), the social brain hypothesis takes a qualitative rather than
a quantitative form. In all these taxonomic groups, it is pairbonded (i.e. repro-
ductively monogamous) species that have disproportionately large brains.
Anthropoid primates contrast strikingly with this pattern in that they exhibit a
quantitative relationship between group size and brain size.
These results suggest that it is the cognitive demands of pairbonding that
underpin the social brain (thereby leading to the initial enlargement of brain
size in several animal families), and that anthropoid primates (and maybe one
or two other numerically small groups such as elephants and the horse family)
have extended the use of these pairbond-like relationships to other non-
reproductive group members – thereby creating the quantitative relationship
that we observe in their case (Shultz and Dunbar 2007). In effect, they have
developed what we might well refer to as "friendships" (Smuts 1985; Silk
2002), relationships that have all the characteristics of pairbonds but without
the sexual connotation – although perhaps this may explain why friendships
can so easily acquire sexual overtones in humans. In anthropoid primates,
friendships function as alliances – reciprocal coalitions whose members relia-
bly come to each other’s aid (see for example, Dunbar 1980, 1989). These
serve the important function of buffering individuals against the costs of
group-living – competition during foraging, and the persistent forms of low
level harassment consequent on living in close proximity to others that results
in infertility in females in both captivity and the wild (Abbott et al. 1986;
Dunbar 1980).
Because pairbonding is often associated with bi-parental care, it is not in-
tuitively obvious whether the cognitive load is imposed by the demands of
pairbonding as such or by those of bi-parental care. Birds allow us to sort this
out because some species have one without the other: this dissociation dem-
onstrates unequivocally that it is species with pairbonds, irrespective of
whether or not they have bi-parental care, that have large brains. Birds also
emphasise a second point: it is species that have lifelong pairbonds that have
unusually large brains, not those that have annual pairbonds (i.e. those that
find new partners each year), indicating that it is something about the costs of
long term relationships that is cognitively demanding.
This raises two important issues. One is why pairbonds (and pairbond-
like relationships) are so costly in cognitive terms. The other is the more

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

18 ROBIN I. M. DUNBAR

difficult question of just what pairbonds actually are. In respect of the first,
it is perhaps obvious that pairbonds (and especially lifelong pairbonds) incur
significant risks for their members: a poor choice of partner will have a
massive effect on an individual’s fitness because of the opportunity costs of
having to start all over again with the business of mate-finding. Such cir-
cumstances can arise either because a mate is less fertile than other members
of the population or because it is less reliable (either in terms of providing
care for the young or in terms of infidelity, or both). The fact that poor
judgment could have catastrophic consequences for the individual’s fitness –
literally by reducing its lifetime reproductive output to zero in the worst case
– will inevitably impose intense selection for the ability to make fine
judgements about the suitability of prospective mates. However, the need to
maintain a high level of behavioural coordination and synchrony must also
impose significant demands on cognition. This is particularly obvious in
those cases where the pair has to share rearing duties, as in many bird spe-
cies where one member of the pair has to stay on the nest while the other
feeds. If the pair do not coordinate their behaviour effectively, the individual
left sitting on the eggs may eventually be forced to choose between starving
and abandoning the eggs. In effect, pairbonds are cooperative alliances for
joint reproduction and the pair members have to be willing to recognise and
accommodate the partner’s interests. In effect, they have to be able to sec-
ond-guess the partner’s needs and factor these into the scheduling of their
own requirements. It seems likely that it is this specific need that may have
provided the trigger for the evolution of those social cognitive skills associ-
ated with theory of mind in humans.
This does, however, raise a serious question: just what is a pairbond? We
tend to recognise pairbonds by virtue of the fact that joint reproduction is a
key component. But in reality, that is not the animals’ experience of it any
more than it is our own. Pairbonded species of birds have pairbonds in or-
der to be able to engage in bi-parental care, thereby generating significant
reproductive benefits. Bi-parental care is the ultimate goal, but to achieve
this they must find an effective solution to the proximate goal of creating
and maintaining a functional pairbond. Their perception of what is involved
(and hence the underpinning cognitive mechanisms) is the rather intangible
unknown. Indeed, we have difficulty trying to specify exactly what is in-
volved even in our own case: we know a relationship when we see or ex-
perience one, but we don’t have any adequate metric by which we can de-
fine it operationally. Relationships of this kind are something we feel rather
than cognise directly.
This much is perhaps obvious from the social psychology literature on
friendships: here, several decades of research have identified two key dimen-
sions to relationships – being close and feeling close (Sternberg 1997; Ber-


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 19

scheid et al. 1989). The first of these is easy to specify, because it simply has
to do with time spent together (in effect, spatial proximity) or the frequency of
interaction. The second is more difficult, because it has something to do with
our inner emotional experiences and we have great difficulty verbalising
these. We intuit them as a form of "hot" cognition, but they are not as acces-
sible to conscious verbal description as "cold" cognition. And for that reason,
we face a double dilemma in the case of animals, since we lack any means of
describing the inner experiences in other species. The time may now have
come to grapple directly with the thorny issue of animals’ mental experiences:
we may not be able to push it quietly under the behaviourist carpet any longer.

The Bonds that Bind

The social brain hypothesis, then, argues that maintaining and servicing
the kinds of intense relationships found in pairbonds (in most birds and
mammals) and friendships (among anthropoid primates) involves serious
cognitive work, which in turn is reflected in the size of a species’ brain. But it
is, at the same time, apparent, both from the human social psychology litera-
ture on friendships and from the ethological literature on how primates service
their relationships, that cognition is only part of the story. There is a deeply
emotional component to relationships that derives in both cases from doing
things together. This leads us, perhaps inevitably, into the issue of how pri-
mates bond their social groups.
For Anthropoid primates, and in particular Old World monkeys and apes,
social grooming is the principal mechanism used for social bonding. Groom-
ing is an intense activity in which one individual leafs through the fur of an-
other, removing bits of vegetation, dead skin and other debris. It is very much
a one-on-one activity, in which the groomer is often deeply concentrated on
its task. Grooming lowers the heart rate in the groomee and reduces the fre-
quency of signs of tension and stress (yawning, scratching, etc.) (Goosen
1981; Castles et al. 1999), to the point where the recipient of grooming can
become so relaxed it actually falls asleep. Many species devote considerable
proportions of their day to social grooming – in the limiting case, as much as
one fifth of the waking day. Grooming shares many of the features of mas-
sage: it is physically stimulating and mildly painful, and thus triggers the re-
lease of endorphins (the brain’s own painkillers) (Keverne et al. 1989). It is
endorphins that are probably responsible for these soporific effects.
However, the real significance of this is that these psychopharmacological
effects mediated by grooming seem to play a crucial role in the process of
building the trust and reciprocity that form the basis of primate social relation-
ships. We have no real idea how endorphins (or any of the other neuroendo-

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

20 ROBIN I. M. DUNBAR

crines that might be involved) give rise to bonded relationships, but one pos-
sibility is that they simply allow two individuals to feel relaxed enough in
each other’s company to spend time together. Time is a commodity that we
have to invest to create a relationship, and there seems to be a more or less
linear relationship between time spent together and the strength of a relation-
ship – at least in terms of functional consequences such as willingness to
come to a grooming partner’s aid (see Dunbar 1980, 1984, 1989). It may be
that endorphins simply provide the proximate reinforcer that makes it worth
spending time with someone else, or it may be that these neuroendocrines are
themselves intimately involved in the bonding process. Either way, it seems
that their role is critical.

Humans and the Social Brain

Modern humans have much larger brains (and especially neocortices) than
other primates, and we can legitimately ask what the relationship between
neocortex size and group size in primates can tell us about human group sizes.
As Figure 1 suggests, there are quite distinct grades in this relationship within
the primates: apes lie to the right of monkeys, and monkeys lie to the right of
prosimians, suggesting that servicing groups of a given size requires propor-
tionately more computational power as you pass from prosimians through
monkeys to the apes (Dunbar 1998). Hence, the appropriate regression line
from which to predict human group sizes is that for apes. Interpolating the
modern human neocortex ratio into the ape equation yields a predicted group
size of ~150 (Fig. 1).
A search of the ethnographic literature revealed that this is in fact the typi-
cal size of hunter-gatherer communities (Dunbar 1993; Hamilton et al. 2007).
More remarkably perhaps, this figure of ~150 appears frequently in many
aspects of historical and contemporary human organisation (Table 1). It was
the mean village size recorded for almost all English counties in the Dooms-
day Book as well as during the C18th, and is the typical size of the company in
most modern armies, the number of recipients of a typical Christmas card
distribution list in the Britain, and the size of the social network in reverse
"small world" experiments, amongst others. Thus, a wide range of contempo-
rary social phenomena seem to yield much the same kinds of grouping pat-
terns, despite marked differences in both scale and organisation. The only
substantive difference between social networks in traditional hunter-gatherer
and agricultural societies and modern post-industrial societies seems to be
that, in traditional societies, everyone in the community has more or less the
same network of 150 acquaintances, whereas in modern urban societies our
networks are highly fragmented – my 150 consists of a set of sub-networks


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 21

that barely overlap. You and I may share one small set of friends, say through
work, but there is no overlap at all in the remaining subsets – we do not share
any relatives, nor do we share hobby circles, church networks, spouses’
friends, schoolgate friends (the often temporary friendships built up through
one’s children’s school friends) or sports club friends. Networks in modern
societies are fragmented and dispersed (often over considerable geographical
distances), whereas in traditional societies they typically form a single cohe-
sive community – even though that community itself may be distributed over
a wide geographical area (as in many contemporary hunter-gatherers).
This figure of ~150 seems to mark a distinct limit for relationship quality:
there seems to be a marked difference in the quality of the relationships we
have with those who are inside the chosen circle versus those who are outside.
My informal definition for this limit to our social world is that it is everybody
whom we know as persons, everyone with whom we have a definable per-
sonal relationship. Those inside this circle are individuals towards whom we
feel some sense of obligation, whom we trust would help us out if we so re-
quested, who would reciprocate our sense of personal commitment. We know
where these individuals fit into our network of relationships, they know where
we fit into theirs, and our knowledge in both cases is based on personal ac-
quaintance. Sometimes, that knowledge can be indirect (friends of friends, or
a shared grandparent), but it defines those to whom we owe personal obliga-
tions; if we offend them, or spurn them in some way, that offence will come
to haunt us through the effect it has on the relationships that link us. In con-
trast, beyond this circle of 150, people cease to be individuals, at least in so
far as our relationships are concerned. Even though we recognise them as
individuals (i.e. we can put names to faces), our relationships with them are
less personal and more typological. We need rules of thumb to guide our in-
teractions with them rather than being able to rely on personalised knowledge.
In such cases, the rule is usually cued by some appropriate badge that signifies
the status of an individual and how we should address them, and this often
requires that we formally badge them in order to recognise them – social
roles, uniforms, badges of rank, styles of speech, and so on.
As with all primate social groups, human social networks are highly struc-
tured. We do not interact equally with all members of our immediate social
world. Rather, it seems that our social world consists of a series of hierarchi-
cally inclusive circles of acquaintanceship that are reflected in both the per-
ceived intimacy of the relationship and the frequency of interaction (Hill and
Dunbar 2003). These circles of acquaintanceship seem to have a very consis-
tent structure: each annulus includes about twice as many people as the one
immediately inside it, so that the cumulative numbers of individuals included
in successive circles exhibit a constant scaling ratio of approximately 3 (Zhou
et al. 2005; see also Hamilton et al. 2007). Roughly speaking, they progres-

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

22 ROBIN I. M. DUNBAR

sively include 5, 15, 50, 150, 500 and 1500 individuals, (see Dunbar 1992;
Zhou et al. 2005) and, for all we know, may extend beyond that in a further
series of circles that have the same ratios.

The Role for Cognition

The fact that brain size correlates with social group size implies that this
involves a cognitive limit. However, we know surprisingly little about the
kinds of cognition that might be involved in managing social relationships.
Although everyone probably agrees that this is some form of "social cogni-
tion", quite what that entails remains unclear. The only aspect of this that we
know much about is what has become known as theory of mind (Leslie 1987;
Perner 1991). Theory of mind is the ability to reflect on another individual’s
mind states. As such, it is one level in a potentially endless reflexive series of
mind states and beliefs about mind states known as the levels of intentionality
(Dennett 1989). We know a great deal about theory of mind (second order
intentionality) because developmental psychologists have explored it in con-
siderable depth. In simple terms, it is the cognitive rubicon that children pass
through at about the age of 4-5 years, although some individuals (autistic peo-
ple) never achieve this even as adults (Baron-Cohen et al. 1985). However,
the problem with theory of mind, is that we know a great deal about its natural
history, but as Roth and Leslie (1998) have pointed out, we have almost no
idea what it actually involves in cognitive terms.
Nonetheless, even though the exact processes involved may be somewhat
opaque, we can perhaps use the notion of intentionality to give us some pur-
chase on the problem of how humans differ from other primates since the
orders of intentionality form a natural scale, and thus seem to provide us with
an index of social cognitive competence. The claim that intentionality itself is
no more than some aspect of executive function (Ozonoff 1995; Mitchell
1997; Barrett et al. 2003; Stylianou 2007) would provide a justification for
this. My concern here is less with the debates about whether or not theory of
mind (or, more generally, intentionality) is modular or the outcome of execu-
tive function (although my inclination is towards the latter) than with the sim-
pler claim that the intentionality scale seems to provide us with a metric of
social cognitive competence (as indexed by the ability to hold several indi-
viduals’ mental states in mind at the same time).
This being so, our main interest at this point is what the natural limits of
intentional reasoning might be in humans. We have assayed normal adults in a
number of separate studies, and it seems that the limits of function for adults
is consistently fifth order ("I believe that you suppose that I imagine that you
want me to believe that….") (Kinderman et al. 1998; Stiller and Dunbar


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 23

2007). Around two-thirds of individuals have their limit at or below fifth or-
der intentionality, and three-quarters have their limit at or below sixth order.
There is considerable individual variation around this (Stiller and Dunbar
2007), and we have shown that the higher levels are lost during the active
phase of at least two well known clinical conditions (bipolar disorder: Kerr et
al. 2003; paranoid schizophrenia: Swarbrick 2000). These competencies de-
velop over a period of time between age 5 (when children first acquire theory
of mind, or second order intentionality) and the early teenage years (when
they finally acquire fifth order adult-level competencies) (Henzi et al. 2007).

Intentionality and the Virtual World

The issue of interest here is what can be achieved with different levels of
intentionality. If intentional competencies allow us to hold several different
individuals’ mind states in mind at the same time, then it seems likely that it
will impose constraints on cultural phenomena that require us to think inten-
tionally. This is perhaps most obvious in the case of imaginative play. Leslie
(1987) noted that theory of mind may be crucial for children to be able to
engage in fictive (i.e. pretend) play where they have to imagine that the world
is other than it really is (i.e. dolls can drink tea, the steering wheel on the back
of a chair is a real car). Leslie’s point can be extended to drama. Consider the
case of the audience watching Shakespeare’s Othello. They have to believe
that Iago intends that Othello imagines that Desdemona is in love with Cassio,
an activity involving four levels of intentionality. However, notice that, at this
point, the kind of story they are dealing with is not especially demanding (or,
for that matter, particularly enthralling). Why should Othello care if Desde-
mona has fantasies about Cassio, so long as they remain just fantasies? The
bottom line of everyday life is that very few of us would be anything but
mildly bemused by such a trivial phenomenon, and the story would end there
as a dull narrative. What gives Shakespeare’s play its bite is the fact that Iago
is able to persuade Othello that Cassio reciprocates Desdemona’s feelings,
thereby creating a romantic triangle and raising the stakes high enough for all
of us to be gripped by the drama (especially when, with the benefit of spec-
tator-sight, we are aware of Iago’s scheming plan). At this point, of course,
the audience is having to work at fifth order intentionality, and is thus at the
natural limits for the great majority of the population.
But, in putting this story together, Shakespeare himself has to go one level
higher than his audience, to sixth order: he has to intend that the audience
believes……etc. I suggest that this might explain why the capacity to enjoy
good literature is a widespread human universal, but the ability to compose

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

24 ROBIN I. M. DUNBAR

good literature is not – storytelling demands social cognitive competencies

that are beyond the normal range for the great majority of the population.
We can use the same framework for exploring the cognitive demands of
religion, because this too requires us to work with an imagined world – a
world that we cannot see or feel directly because it exists only in our heads.
The transcendental experiences that we have during religious events are un-
doubtedly very real, but they are not the stuff of real physical experiences
created by the world impinging on our senses. They belong, rather, to an inner
virtual world of the imagination.
A task analysis of the mentalising demands of religious beliefs suggests that
it is perfectly possible to have religion with any order of intentionality. How-
ever, the form of religion depends on the levels of intentionality at which you
can work. Crucially, there appears to be a critical difference between the forms
of religion possible at fourth and fifth order intentionality. Consider the follow-
ing kinds of religious statements that one could make with fourth and fifth order
intentionality (with the intentional components italicised and numbered):
A. You intend [1] that I believe [2] that we know [3] that god is willing [4]
{to intervene in the world on our behalf}. [4th order]
B. You intend [1] that I believe [2] that everyone else understands [3] that
we know [4] that he is willing [5] {to intervene ….}. [5th order]
At fourth order, you can have what I refer to as social religion: I can enjoin
you to believe what I believe to be true about some deity’s interests in us, but
you do not have to agree with that claim even though you may accept that I
am convinced by the truth of what I say. I can only make you agree with me
by using a police force, but in that case your level of personal and intellectual
commitment is likely to be very limited. However, fifth order seems to repre-
sent a crucial rubicon: it seems to entail an additional leap of faith that some-
how commits both of us to holding this belief about god. When I accept that
you believe this claim to be true, I ipso facto also commit myself to the verac-
ity of your claim. We now have what I term "communal religion" – a set of
beliefs that bind us into a single community whose members subscribe the
same world view. At this point, we have a very powerful mechanism enforc-
ing the communal will, for making us all sign up to the communal project
(whatever that may happen to be). We don’t need a police force to make us
behave in a religious way: we are all deeply and personally committed to it
and adhere to these beliefs of our own free will.

So What Makes Us Different?

I return to my opening question: why are humans not just great apes? I have
suggested that the answer lies in our capacity to live in a virtual mental world.


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 25

We can engage in activities that are well beyond the competencies of even the
great apes. Indeed, they are so far beyond the competencies of other species that
there is no chance that the proverbial chimpanzee sitting at a typewriter will
ever produce the works of Shakespeare. The one remaining question is why we
humans should have needed such extraordinary cognitive abilities.
From an evolutionary point of view, this is puzzling because brain tissue
is exceptionally expensive. Aiello and Wheeler (1995) pointed out some
time ago that brain tissue is among the most expensive matter in the body to
maintain, and that this provides a very steep gradient up which natural selec-
tion has to drive brain evolution if it is to increase brain size (their "expen-
sive tissue hypothesis"). There is some tentative comparative evidence
(Dunbar 2003a, b) that intentional competences are correlated with some
aspects of brain volume. If so, then it follows that the selection pressure for
the capacity to manage the higher orders of intentionality incurs a very sig-
nificant energetic cost to the individuals concerned. The advantages to be
gained from investing in such capacities must thus be considerable. What
might these be?
Primate societies are implicit social contracts. Like all social contracts,
their stability and functionality depends on the members trading off short and
long term benefits. As with pairbonds, a group will only remain stable as a
coalition providing members are willing to compromise on some of their short
term benefits in order to gain in the long run through group-level effects.
These group level effects come in the form of increased survival, and hence
higher reproductive rates, mainly as a result of reduced predation rates (Dun-
bar 1988; Shultz et al. 2004) though there may also be more direct benefits in
terms of the survival of individual offspring (Silk et al. 2003; Silk 2007).
However, all social contracts of this kind face the same problem: freerid-
ers. There is always a significant benefit to be gained by individuals who take
the benefits of sociality but do not pay all the costs (Enquist and Leimar 1993;
Nettle and Dunbar 1997; Dunbar 1999). Since the intrusiveness of freeriders
is proportional to the size of the population, the large communities that char-
acterise humans face a significantly greater challenge in this respect than
those of our primate cousins. As a result, more sophisticated mechanisms are
needed to ensure that freeriders do not overwhelm the community. While
there are a number of cognitive mechanisms for managing freeriders (Dunbar
1999), the absence of grooming on a large scale means that we inevitably lack
the endorphin-based mechanisms that lie at the root of primate sociality. This
is a serious issue, because it creates a "bonding gap" of quite significant pro-
portions (Dunbar, in press) that could seriously destablise the fragile basis on
which group-level collaboration depends. Something was needed to fill that
gap.

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

26 ROBIN I. M. DUNBAR

In fact, the pattern of brain size evolution suggests that this really only
became a serious issue in the later stages of human evolution following the
appearance of archaic humans (Homo heidelbergensis and allies) around
half a million years ago (Dunbar, in press). At this point, brain size took off
with a rapid rate of increase, implying a corresponding increase in the size
of communities that had to be bonded. Nonetheless, our capacity to create
the large communities that were presumably needed to ensure survival (and
the reasons why we needed such large communities remain unclear) must
have depended on solving the bonding issue satisfactorily. While social
conformity and "good behaviour" can always be imposed by punitive action
on backsliders (Orstrom et al. 1994; Clutton-Brock and Parker 1995; Fehr et
al. 2002), there are limits to which such action can really enforce social co-
operation. Since there are always benefits to be gained from freeriding and
there is a significant chance that any one freerider will escape detection
and/or punishment, the temptation to freeride will mean that some propor-
tion of individuals will always do so, no matter how effective punishment is
in reducing the absolute frequency. Social cooperation is far more effective
when individuals act willingly because they all voluntarily sign up to the
communal project (Orstrom et al. 1994). Mechanisms that create a sense of
communality are, thus, more likely to result in the required levels of altru-
ism than the use of social control.
As has long been appreciated in the social sciences, both religion and
story-telling play an important role in social bonding in all human cultures
(Durkheim 2001). Religion does so through the capacity of rituals to trigger
the release of endorphins, since many of these are of just the kind of mildly
stressful activities that are especially good at releasing endorphins. Religion,
of course, also has the advantage of having an intellectual dimension, and here
the cognitive demands become important in creating the kind of dual-process
bonding mechanism that we find in primate social grooming. To the extent
that the intellectual component of religion is a form, of story-telling, story-
telling itself enters the frame as an important mechanism for community
bonding. However, story-telling goes beyond mere religious explanations for
the way the world is and should be; it offers its own kind of opportunity to
develop a form of entertainment that can be just as powerful a bonding
mechanism in its own right.
One could argue that both of these are mere epi-phenomena – non-adaptive
by-products of the fact that we have big brains. Exactly this claim has, of
course, been made for music (Pinker’s [1998] "music as evolutionary cheese-
cake" argument). However, such a claim rests uncomfortably with the time,
money and effort that, irrespective of culture, humans invest in all three of
these phenomena. In fact, all human societies value story-telling for itself as a
form of entertainment, and many of these stories are intimately involved in


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 27

creating a sense of community: origin stories, tribal histories, and moral tales
are among the commonest of campfire "tellings", and all contribute directly to
providing a sense of community. Indeed, the very performance itself often
contributes directly, not least through laughter (another powerful releaser of
endorphins: Dunbar 2004). By way of confirmation, van Vugt et al. (submit-
ted) have recently demonstrated experimentally that laughing together causes
strangers (but not existing friends) to be significantly more generous towards
each other in public goods games. It is important to note that both religion and
story-telling both acquire their community-enhancing properties through lan-
guage: without language, it would not be possible to tell a "big enough story"
to create the bonding effects or to persuade people to take part in religious
rituals on a regular basis. It seems that regular participation in these commu-
nity-bonding exercises is essential, almost as though it was a form of inocula-
tion requiring boosters at set intervals to maintain a consistent level of per-
formance.
But aside from the trivially obvious answer of language, what is it that lim-
its these "activities of the mind" to humans alone? One answer has to be that
only humans can cope with the fifth order intentionality that is necessary to
allow these phenomena to produce the effects they do. The best that any non-
human species can do seems to be second order intentionality – and even that
is probably true only of great apes (O’Connell and Dunbar 2003), since the
consensus is that all other animal species can aspire only to 1st order. If men-
talising skills really are an emergent property of executive function compe-
tences (Barrett et al. 2003) and these in turn are a function of the com-
putational power (and hence size) of the brain (Dunbar 2003a), then the sim-
ple explanation for this striking difference between ourselves and all other
species lies in the size of our brain (and, perhaps more specifically, the frontal
lobes, since these are generally considered to be the locus operandi for those
capacities that we conventionally refer to as executive functions). In effect,
the differences between us and all other animals comes down to the fact that
the size of our brain allows us to do something that is simply not possible with
a smaller brain. While brain organisation and aspects of neural efficiency (see
for example Burki and Kaessmann 2004) must, of course, play a role, there
are good arguments for thinking that it is simply the size of the computer that
may be important (see for example, Duncan 2001). No matter what the dif-
ferences in structure and organisation may be, we do have a size issue to ex-
plain away and our over-large brains cannot be there merely by accident or as
a trivial by-product of something else.
In sum, the difference between us and our nearest cousins is not simply a
matter of having greater intelligence (however we want to measure that) but
lies in what that greater intelligence allows us to do (namely, live in a virtual
world), and in why we need that emergent capacity at all (i.e. to enable us to

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

28 ROBIN I. M. DUNBAR

bond much larger communities than would otherwise be possible for a mon-
key or an ape).

Table 1. Examples of human social groupings that conform to the predicted size of
~150 individuals
(Confidence intervals around the predicted mean are 100-200 (Dunbar 1993).

Grouping Typical size Source

Neolithic villages (Middle East, 6500-5500BC) 150-200 Oates (1977)

Maniple ("double century") (Roman army: 350-100BC) 120-130 Montross
(1975)
Doomsday Book (1085): Average county village size 150 Hill (1981),
Bintliff
(1999)
C18th English villages (mean of county means) 160 Laslett (1971)
Tribal societies (mean and range of communities; N=9) 148 (90-222) Dunbar
(1993)
Hunter-gatherer societies (mean clan size; N= 165 Hamilton et al
(2007)
Hutterite farming communities (Canada) (mean, N=51) 107 Mange &
Mange (1980)
"Nebraska" Amish parishes (mean, N=8) 113 Hurd (1985)
Church congregations (recommended ideal size) 200 Urban Church
Project (1974)
E. Tennessee rural mountain community 197 Bryant (1981)
Social network size (mean, N=2 "small world" experiments) 134 Killworth et
al (1984)
Goretex Inc: factory unit size 150 Gladwell
(2000)
Company (mean and range for 10 World War II armies) 180 (124-223) MacDonald
(1955)
Christmas card distribution lists (mean total recipients: N=43) 154 Hill & Dun-
bar (2003)
Research specialities (sciences and humanities) (mode, N=13) 100-200 Becher
(1989)


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 29

1000
Mean group size

100

10

1.0 2.0 3.0 4.0 5.0

Neocortex ratio

Figure 1. Mean social group size for different species of primates (prosimians, mon-
keys and apes) plotted against relative neocortex size (indexed as neocortex ratio, the
ratio of neocortex volume divided by the volume of the rest of the brain). Ape species
are distinguished as open symbols (lower left to top right: gibbons, gorillas, chimpan-
zees and modern humans) and a solid line of best fit. The point labelled for humans is
that predicted by the ape regression equation. Broken line: monkeys; dotted line:
prosimians. Note that the slopes of the regression lines are parallel, indicating the
presence of evolutionary grades. Redrawn from Dunbar (2007).

Bibliography

Abbott, D. H., E. B. Keverne, G. F. Moore, and U. Yodyinguad. 1986. Social sup-

pression of reproduction in subordinate talapoin monkeys. Miopithecus talapoin.
In: J. Else and P. C. Lee (eds.) Primate Ontogeny. Cambridge: Cambridge Univer-
sity Press, 329-341.
Aiello, L. C. and P. T. Wheeler. 1995. The expensive tissue hypothesis: the brain and
the digestive system in human evolution. Current Anthropology 36: 199-221.
Baron-Cohen, S., A. M. Leslie, and U. Frith. 1985. Does the autistic child have a
theory of mind? Cognition 21: 37-46.
Barrett, L., S. P. Henzi, and R. I. M. Dunbar. 2003. Primate cognition: from ‘what
now?’ to ‘what if?’ Trends in Cognitive Sciences 7: 494-497.

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

30 ROBIN I. M. DUNBAR

Becher, T. 1989. Academic Tribes and Territories. Milton Kaynes: Open University
Press.
Berscheid, E., M. Snyder, and A. M. Omoto. 1989. The relationship closeness inven-
tory: assessing the closeness of interpersonal relationships. Journal of Personality
and Social Psychology 57: 792-807.
Bintliff, J. 1999. Settlement and territory. In: G. Barker (ed.) Companion Encyclope-
dia of Archaeology. London: Routledge, 505-545.
Burki, F. and H. Kaessmann. 2004. Birth and adaptive evolution of a hominoid gene
that supports high neurotransmitter flux. Nature Genetics 36: 1061-1063.
Byrne, R. and A. Whiten (eds.). 1988. Machiavellian Intelligence. Oxford: Oxford
University Press.
Castles, D., A. Whiten, and F. Aureli. 1999. Social anxiety, relationships and self-
directed behaviour among wild female olive baboons. Animal Behaviour 58:
1207-1215.
Clutton-Brock, T. H. and G. A. Parler. 1995. Punishment in animal societies. Nature
373: 209-216.
Dennett, D. 1987. The Intentional Stance. Cambridge (MA): MIT Press.
Dunbar, R. I. M. (in press). Mind the bonding gap: constraints on the evolution of
hominin societies. In: S. Shennan (ed.) Pattern and Process in Cultural Evolution.
Berkely: University of California Press.
Dunbar, R. I. M. 1980. Determinants and evolutionary consequences of dominance
among female gelada baboons. Behavioural Ecology and Sociobiology 7: 253-265.
Dunbar, R. I. M. 1984. Reproductive Decisions: An Economic Analysis of Gelada
Baboon Social Strategies. Princeton: Princeton University Press.
Dunbar, R. I. M. 1988. Primate Social Systems. London: Chapman & Hall.
Dunbar, R. I. M. 1989. Reproductive strategies of female gelada baboons. In: A. Rasa,
C. Vogel and E. Voland (eds.) Sociobiology of Sexual and Reproductive
Strategies. London: Chapman & Hall, 74-92.
Dunbar, R. I. M. 1993. Coevolution of neocortex size, group size and language in
humans. Behavioral and Brain Sciences 16: 681-735.
Dunbar, R. I. M. 1998. The social brain hypothesis. Evolutionary Anthropology 6:
178-190.
Dunbar, R. I. M. 1999. Culture, honesty and the freerider problem. In: R.I.M. Dunbar,
C. Knight and C. Power (eds.) The Evolution of Culture. Edinburgh: Edinburgh
University Press, 194-213.
Dunbar, R. I. M. 2003b. The social brain: mind, language and society in evolutionary
perspective. Annual Review of Anthropology 32: 163-181.
Dunbar, R. I. M. 2004. Language, music and laughter in evolutionary perspective. In:
D. K. Oller and U. Griebel (eds.) Evolution of Communication Systems: A
Comparative Approach. Cambridge (MA): MIT Press, 257-274.
Dunbar, R. I. M. 2008. Cognitive constraints on the structure and dynamics of social
networks. Group Dynamics: Theory, Research, and Practice 12 (1): 7-16.
Dunbar, R. I. M. and S. Shultz. 2007a. Understanding primate brain evolution.
Philosophical Transactions of the Royal Society of London 362B: 649-658.
Dunbar, R. I. M. and S. Shultz. 2007b. Evolution in the social brain. Science 317:
1344-1347.


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 31

Duncan, J. 2001. An adaptive coding model of neural function in prefrontal cortex.

Nature Neuroscience 2: 820-829.
Durkheim, E. 2001 (1900). Elementary Forms of religious Life. Oxford: Oxford Uni-
versity Press.
Emery, N., N. Clayton, and C. Frith. 2007. Social Intelligence: From Brain to Cul-
ture. Oxford: Oxford University Press.
Enquist, M. and O. Leimar. 1993. The evolution of cooperation in mobile organisms.
Animal Behaviour 45: 747-757.
Fehr, E., U. Fishbacher, and S. Gächter. 2002. Strong reciprocity, human cooperation
and the enforcement of social norms. Human Nature 13: 1-25.
Finlay, B. L. and R. B. Darlington. 1995. Linked regularities in the development and
evolution of mammalian brains. Science 268: 1578-1584.
Finlay, B. L., R. B. Darlington, and N. Nicastro. 2001. Developmental structure in
brain evolution. Behavioral and Brain Sciences 24: 263-308.
Gladwell, M. 2000. The Tipping Point. London: Little, Brown and Co.
Goosen, C. 1981. On the function of allogrooming in Old-World monkeys. In: A. B.
Chiarelli and R. S. Corruccini (eds.) Primate Behaviour and Sociobiology. Berlin:
Springer, 110-120.
Hamilton, M. J., B. T. Milne, R. S. Walker, O. Burger, and J. H. Brown. 2007. The
complex structure of hunter-gatherer social networks. Proceedings of the Royal
Society, London 271B: 2195-2202.
Henzi, P., L. de Sousa Pereira, D. Hawker-Bond, J. Stiller, R. I. M. Dunbar, and L.
Barrett. 2007. Look who’s talking: developmental trends in the size of conversa-
tional cliques. Evolution and Human Behavior 28: 66-74.
Hill, D. 1981. An Atlas of Anglo-Saxon England. Oxford: Blackwell.
Hill, R. A. and R. I. M. Dunbar. 2003. Social network size in humans. Human Nature
14: 53-72.
Hurd, J. P. 1985. Sex differences in mate choice among the "Nebraska" Amish of
central Pennsylvania. Ethology and Sociobiology 6: 49-57.
Jerison, H. J. 1973. Evolution of the Brain and Intelligence. London: Academic Press.
Kerr, N., R. I. M. Dunbar, and R. P. Bentall. 2003. Theory of mind deficits in bipolar
affective disorder. Journal of Affective Disorders 73: 253-259.
Keverne, E. B., N. Martensz, and B. Tuite. 1989. Beta-endorphin concentrations in
cerebrospinal fluid of monkeys are influenced by grooming relationships. Psycho-
neuroendocrinology 14: 155-161.
Killworth, P. D., H. P. Bernard, and C. McCarty. 1984. Measuring patterns of ac-
qainatanceship. Current Anthropology 25: 385-397.
Kinderman, P., R. I. M. Dunbar, and R. P. Bentall. 1998. Theory-of-mind deficits and
causal attributions. British Journal of Psychology 89: 191-204.
Laslett, P. 1971. The World We Have Lost. London: Methuen.
Leslie, A. M. 1987. Pretense and representation – the origins of "theory of mind".
Psychological Review 94: 412-426.
MacDonald, C. B. 1955. Company. Encyclopedia Brittanica, 14th ed., 143-144.
Mange, A. and E. Mange. 1980. Genetics: Human Aspects. New York: Holt Rinehart
& Winston.

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)

32 ROBIN I. M. DUNBAR

McLeod, C. E., K. Zilles, A. Schleicher, J. K. Rilling, and K. E. Gibson. 2003. Ex-

pansion of the neocerebellum in Hominodea. Journal of Human Evolution 44:
401-429.
Mitchell, P. 1997. Introduction to Theory of Mind. London: Arnold.
Montross, L. 1975. Tactics. Encyclopedia Britannica. 15th ed.
Naroll, R. 1956. A preliminary index of social development. American Anthropologist
58: 687-715.
Nettle, D. and R. I. M. Dunbar. 1997. Social markers and the evolution of reciprocal
exchange. Current Anthropology 38: 93-99.
O’Connell, S. and R. I. M. Dunbar. 2003. A test for comprehension of false belief in
chimpanzees. Evolution and Cognition 9: 131-139.
Oates, J. 1977. Mesopotamian social organisation: Archaeological and philological
evidence. In: J. Friedman and M. J. Rowlands (ed.) The Evolution of Social
Systems. London: Duckworth, 457-485.
Orstrom, E., R. Gardner, and J. Walker. 1994. Rules, Games and Common-Pool Re-
sources. Ann Arbor: University of Michigan Press.
Ozonoff, S. 1995. Executive functions in autism. In: E. Schopler and G. B. Mesibov
(eds.) Learning and Cognition in Autism. New York: Plenum Press, 199-218.
Perner, J. 1991. Understanding the Representational Mind. Cambridge (MA): MIT Press.
Pinker, S. 1998. How the Mind Works. London: Allen Lane.
Roth D. and A. M. Leslie. 1998. Solving belief problems: Toward a task analysis.
Cognition 66: 1-31.
Shultz, S. and R. I. M. Dunbar. 2007. The evolution of the social brain: Anthropoid
primates contrast with other vertebrates. Proceedings of the Royal Society, London
274B: 2429-2436.
Shultz, S., R. Noe, S. McGraw, and R. I. M. Dunbar. 2004. A community-level evaluation
of the impact of prey behavioural and ecological characteristics on predator diet
composition. Proceedings of the Royal Society, London, 271B: 725-732.
Silk, J. B. 2002. The ‘F’-word in primatology. Behaviour 139: 421-446.
Silk, J. B. 2007. Social components of fitness in primate groups. Science 317: 1347-1350.
Silk, J. B., S. C. Alberts, and J. Altmann. 2003. Social bonds of female baboons
enhance infant survival. Science 302: 1232-1234.
Smuts, B. B. 1985. Sex and Friendship in Baboons. New York: Aldine.
Sternberg, R. J. 1997. Construct validation of a triangular love scale. European Jour-
nal of Social Psychology 27: 313-335.
Stiller, J. and R. I. M. Dunbar. 2007. Perspective-taking and social network size in
humans. Social Networks 29: 93-104.
Stylianou, M. 2007. Does Executive Function Training Improve Mentalising Ability?
PhD thesis, University of Liverpool.
Swarbrick, R. 2000. A Social Cognitive Model of Paranoid Delusions. University of
Manchester: PhD thesis.
Urban Church Project. 1974. Let My People Grow! Unpublished report to the General
Synod of the Church of England, London. Workpaper No.1.
Van Vugt, M., C. Hardy, J. Stow, and R. I. M. Dunbar. (submitted). Laughter as so-
cial lubricant: a biosocial hypothesis about the pro-social functions of laughter
and humor.


   . . . 3. . 3 (2008)
 WHY HUMANS AREN’T JUST GREAT APES 33

Whiten, A. and R. W. Byrne (eds.). 1997. Machiavellian Intelligence II: Extensions

and Evaluations. Cambridge: Cambridge University Press.
Zhou, W. -X., D. Sornette, R. A. Hill, and R. I. M. Dunbar. 2005. Discrete hierarchi-
cal organisation of social group sizes. Proceedings of the Royal Society, London
272B: 439-444.

Robin Danbar

Zašto lj udi ipak nisu s a mo veli ki maj muni

Mada delimo mnoge elemente ponašanja i fiziologije sa primatima, ljudi se razli-

kuju u jednoj krucijalnoj taki: sposobnosti da Xive u svetu mate. Dva kljuna
aspekta ponašanja, u mnogome tipina za ljude, to odraXavaju: prianje pria i re-
ligija. U ovom tekstu, pokazujem kako ove osobine nastaju kao posledica prirodnog
razvoja hipoteze o "društvenom mozgu", i prirode društvenosti i kognicije kod pri-
mata, kroz pove
anje broja lanova u ljudskim drutvima tokom poslednjih pet mil-
iona godina evolucije.

Kljune rei: kognicija, "društveni mozak", veliina grupe, intencionalnost, re-

ligija, evolucija hominida

Robin Danbar

Pourquoi après tout les hommes ne sont -ils pas que

de grands singes

Même si nous avons en commun avec les primates de nombreuses caractéristiques

de comportement et de physiologie, les hommes diffèrent d’eux sur un point essentiel:
la capacité de vivre dans le monde de l’imagination. Deux aspects fondamentaux du
comportement, sur bien des points typiques des hommes, le reflètent: la narration
d’histoires et la religion. Dans ce texte, je montre comment ces caractéristiques nais-
sent comme une conséquence du développement naturel de l’hypothèse du "cerveau
social", ainsi que de la sociabilité et de la cognition propres aux primates, et cela à
travers l’accroissement du nombre des membres des sociétés humaines au cours des
cinq derniers millions d’années d’évolution.

Mots-clés: cognition, "cerveau social", taille du groupe, intentionnalité, religion,

évolution des hominidés.

Issues in Ethnology and Anthropology n. s. Vol. 3. No .3 (2008)