ARTICLES
The Social Brain Hypothesis
Robin I.M. Dunbar
The concensus view has traditionally
been that brains evolved to process
information of ecological relevance.
This view, however, ignores an
important consideration: Brains are
exceedingly expensive both to evolve
and to maintain. The adult human
brain weighs about 2% of body weight
but consumes about 20% of total energy
intake.2 In the light of this, it is
difficult to justify the claim that primates,
and especially humans, need
larger brains than other species merely
to do the same ecological job. Claims
that primate ecological strategies involve
more complex problem-solving3,4
are plausible when applied to
the behaviors of particular species,
such as termite-extraction by chimpanzees
and nut-cracking by Cebus monkeys,
but fail to explain why all
primates, including those that are conventional
folivores, require larger brains
than those of all other mammals.
An alternative hypothesis offered
during the late 1980s was that primates’
large brains reflect the computational
demands of the complex social
systems that characterize the
order.5,6 Prima facie, this suggestion
seems plausible: There is ample evidence
that primate social systems are
more complex than those of other
species. These systems can be shown
to involve processes such as tactical
deception5 and coalition-formation,7,8
which are rare or occur only in simpler
forms in other taxonomic groups.
Because of this, the suggestion was
rapidly dubbed the Machiavellian intelligence
hypothesis, although there is a
growing preference to call it the social
brain hypothesis.9,10
Plausible as it seems, the social brain
hypothesis faced a problem that was
recognized at an early date. Specifically,
what quantitative empirical evidence
there was tended to favor one or
the other of the ecological hypotheses,1
whereas the evidence adduced
in favor of the social brain hypothesis
was, at best, anecdotal.6 In this article,
I shall first show how we can test
between the competing hypotheses
more conclusively and then consider
some of the implications of the social
brain hypothesis for humans. Finally,
I shall briefly consider some of the
underlying cognitive mechanisms that
might be involved.
TESTING BETWEEN ALTERNATIVE
HYPOTHESES
To test between the competing hypotheses,
we need to force the hypotheses
into conflict in such a way that
their predictions are mutually contradictory.
This allows the data to discriminate
unequivocally between
them. In the present case, we can do
this by asking which hypothesis best
predicts the differences in brain size
across the primate order. To do so, we
need to identify the specific quantitative
predictions made by each hypothesis
and to determine an appropriate
measure of brain size.
The four classes of hypotheses that
have been put forward to explain primate
brain evolution are epiphenomenal,
developmental, ecological, and
social in orientation (Table 1). The
epiphenomenal and developmental hypotheses
share the assumption that
evolution of the brain (or brain part) is
not a consequence of external selection
pressures but rather simply a
consequence of something to do with
the way biological growth processes
are organized. The epiphenomenal hypotheses
thus argue that brain evolution
is a mere byproduct of body size
evolution, and that brain part size is,
in turn, a byproduct of total brain
evolution.11
The developmental versions differ
only in that they provide a more specific
mechanism by presuming that
maternal metabolic input is the critical
factor influencing brain development.
This claim is given credibility by
the fact that the bulk of brain growth
in mammals occurs prenatally. Indeed,
it appears to be the completion
of brain development that precipitates
birth in mammals, with what little
postnatal brain growth occurs being
completed by the time an infant is
weaned. From this, the conclusion is
drawn that brain evolution must be
constrained by the spare energy, over
and above her basal metabolic requirements,
that the mother has to channel
into fetal development.12–14 Some evidence
in support of this claim comes
from the fact that frugivorous primates
have larger adult brains relative
Robin Dunbar is Professor of Evolutionary
Psychology and Behavioural Ecology at
the University of Liverpool, England. His
research primarily focuses on the behavioral
ecology of ungulates and human and
nonhuman primates, and on the cognitive
mechanisms and brain components that
underpin the decisions that animals make.
He runs a large research group, with graduate
students working on many different
species on four continents.
Key Words: brain size, neocortex, social brain
hypothesis, social skills, mind reading, primates
Conventional wisdom over the past 160 years in the cognitive and neurosciences
has assumed that brains evolved to process factual information about the world.
Most attention has therefore been focused on such features as pattern recognition,
color vision, and speech perception. By extension, it was assumed that brains
evolved to deal with essentially ecological problem-solving tasks.1
178 Evolutionary Anthropology
O
nly
for educational purposes
SR
T
II
to body size than do folivorous primates.1
This has been interpreted as
implying that frugivores have a richer
diet than folivores do and thus have
more spare energy to divert into fetal
growth. Large brains are thus seen as
a kind of emergent epigenetic effect of
spare capacity in the system.
Both kinds of explanations suffer
from the problem that they ignore a
fundamental principle of evolutionary
theory, which is that evolution is the
outcome of the balance between costs
and benefits. Because the cost of maintaining
a large brain is so great, it is
intrinsically unlikely that large brains
will evolve merely because they can.
Large brains will evolve only when the
selection factor in their favor is sufficient
to overcome the steep cost gradient.
Developmental constraints are undoubtedly
important, but rather than
being causal their role is that of a
constraint that must be overcome if
larger brains are to evolve. In addition,
Pagel and Harvey15 have shown that
the energetic arguments do not add
up: Precocial mammals do not have
higher metabolic rates than do altricial
mammals despite the fact that
they have neonatal brain sizes that
are, on average, twice as large. We
therefore do not need to consider either
epiphenomenal or developmental
hypotheses any further in the context
of this article.
This does not necessarily mean that
these explanations are wrong. Both
kinds of explanation may be true in
the sense that they correctly identify
developmental constraints on brain
growth, but they do not tell us why
brains actually evolved as they did.
They may tell us that if you want to
evolve a large brain, then you must
evolve a large body in order to carry
the energetic costs of doing so or a diet
that ensures sufficient energy to provide
for fetal brain development. No
such allometric argument can ever
imply that you have to evolve a large
brain or a large body. The large brain
or brain part is a cost that animals
must factor into their calculations
when considering whether or not a
large body or a large brain is a sensible
solution to a particular ecological
problem. Shifts to more energy-rich or
more easily processed diets may be
essential precursors of significant increases
in brain or brain part size.2
This would explain why frugivores
have larger brains than folivores do
and why hominids have larger brains
than great apes do.
This leaves us with just two classes
of hypotheses, the ecological and the
social. At least three versions of the
former can be identified, which I will
term the dietary, mental maps, and
extractive foraging hypotheses. In essence,
these argue, respectively, that
primate species will need larger brains
if (i) they are frugivorous because fruits
are more ephemeral and patchy in
their distribution than leaves are, and
hence require more memory to find;
(ii) they have larger ranges because of
the greater memory requirements of
large-scale mental maps; or (iii) their
diet requires them to extract resources
from a matrix in which they are embedded
(e.g., they must remove fruit
pulp from a case, stimulate gum flow
from a tree, extract termites from a
termitarium, or hunt species that are
cryptic or behave evasively).
For obvious reasons, I used the percentage
of fruit in the diet as an appropriate
index for the dietary hypothesis.
I used the size of the range area and
the length of the day journey as alternative
indices for the mental mapping
hypothesis, though I present the data
only for the first of these here. The first
index corresponds to the case in which
animals have to be able to manipulate
information about the locations of resources
relative to themselves in a
Euclidean space (an example would
be the nut-cracking activities of the
Taı¨ chimpanzees16); the second corresponds
to the possibility that the constraint
lies in the needs of some aspect
of inertial navigation. The extractive
foraging hypothesis is less easy to characterize
in quantitative terms because
there is no objective measure of the
degree to which diets vary in their
extractiveness. However, Gibson4 provided
a classification of primate species
into four categories of diet that
differ in their degree of extractiveness.
We can test this hypothesis by asking
whether there is a consistent variation
in brain size among these four categories,
with the species having the more
extractive diets having larger brains
than those with the less extractive
diets.
Finally, we need an index of social
complexity. In my original analyses, I
used social group size as a simple
measure of social complexity. Although
at best rather crude, this measure
nonetheless captures one aspect of the
TABLE 1. Hypotheses Used to Explain the
Evolution of Large Brains in Primates
Hypothesis Sources
A. Epiphenomenal
hypotheses
1. Large brains (or brain parts)
are an unavoidable consequence
of having a large
body (or brain) 11, 70
B. Ecological hypotheses
2. Frugivory imposes higher
cognitive demands than
folivory does 1, 65
3. Brain size constrains the size
of the mental map: 1
(a) constraint on size of
home range
(b) constraint on inertial
navigation (day journey
length)
4. Extractive foraging
hypothesis 3, 4
C. Social hypotheses
5. Brain size constrains size of
social network (group size):
6, 71,
72
(a) constraint on memory
for relationships
(b) constraint on social skills
to manage relationships
D. Developmental
hypotheses
6. Maternal energy constraints
determine energy
capacity for fetal brain
growth
12, 13
46, 55,
73
Because the cost of
maintaining a large brain
is so great, it is
intrinsically unlikely that
large brains will evolve
merely because they
can. Large brains will
evolve only when the
selection factor in their
favor is sufficient to
overcome the steep cost
gradient.
ARTICLES Evolutionary Anthropology 179
O
nly
for educational purposes
SR
T
II
complexity of social groups, the fact
that information-processing demands
can be expected to increase as the
number of relationships involved increases.
More importantly, perhaps,
this measure has the distinct merit of
being easily quantified and widely
available. Although it is possible to
conceive of a number of better measures
of social complexity, the appropriate
data are rarely available for
more than one or two species.
The second problem concerns the
most appropriate measure of brain
evolution. Hitherto, most studies have
considered the brain as a single functional
unit. This view has been reinforced
by Finlay and Darlington,11 who
argued that the evolution in brain part
size closely correlates with the evolution
of total brain size and can be
explained simply in terms of allometric
consequences of increases in total
brain size. However, Finlay and Darlington
failed to consider the possibility
that changes in brain size might
actually be driven by changes in its
parts rather than in the whole brain.
This is especially true of the neocortex,
for its volume accounts for 50% to
80% of total brain volume in primates.
Thus, changes in the volume of the
neocortex inevitably have a large direct
effect on apparent change in brain
volume that may be quite unrelated to
changes in other brain components.
This point is given weight by the fact
that Finlay and Darlington themselves
showed that neocortex size is an exponential
function of brain size, whereas
other brain components are not.
Finlay and Darlington11 notwithstanding,
there is evidence that brain
evolution has not been a history of
simple expansion in total volume.
Rather, brain evolution has been mosaic
in character, with both the rate
and the extent of evolution having
varied between components of the system.
MacLean17 pointed out many
years ago that primate brain evolution
can be viewed in terms of three major
systems (his concept of the triune
brain). These systems correspond to
the basic reptilian brain (hind- and
midbrain systems), the mammalian
brain (palaeocortex, subcortical systems),
and the primate brain (broadly,
the neocortex). A more important
point, perhaps, is that variations can
be found within these broad categories
in the rates at which different
components expanded, which, in at
least some cases, have been shown to
correlate with ecological factors.10 Partialling
out the effects of body size on
the size of brain components suggests
that the story may be more complex
than Finlay and Darlington11 supposed,
with some remodeling of brain
growth patterns occurring in the transitions
between insectivores, prosimians,
and anthropoids.18
The important point in the present
context is that, as Passingham19 noted,
relative to the more primitive parts of
the brain such as the medulla, the
neocortex shows dramatic and increasing
expansion across the range of primates
(Fig. 1). The neocortex is approximately
the same size as the
medulla in insectivores; however, it is
about 10 times larger than the medulla
in prosimians and 20–50 times
larger in the anthropoids, with the
human neocortex being as much as
105 times the size of the medulla.
This suggests that rather than looking
at total brain size, as previous
studies have done, we should in fact be
considering the brain system, namely
the neocortex, that has been mainly
responsible for the expansion of the
primate brain. From the point of view
of all the hypotheses of primate brain
evolution, this makes sense: The neocortex
is generally regarded as being
the seat of those cognitive processes
that we associate with reasoning and
consciousness, and therefore may be
expected to be under the most intense
selection from the need to increase or
improve the effectiveness of these pro-
cesses.
One additional problem needs to be
resolved. In his seminal study of brain
evolution, Jerison20 argued that brain
size can be expected to vary with body
size for no other reason than fundamental
allometric relationships associated
with the need to manage the
physiological machinery of the body.
What is of interest, he suggested, is
not absolute brain size, but the spare
brain capacity over and above that
needed to manage body mechanisms.
Figure 1. Neocortex volume as a ratio of medulla volume in different groups of primates (after
Passingham19). Source: Stephan et al29
. . . brain evolution has
not been a history of
simple expansion in total
volume. Rather, brain
evolution has been
mosaic in character,
with both the rate and
the extent of evolution
having varied between
components of the
system.
180 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II
For this reason, Jerison derived his
encephalization quotient. All subsequent
studies have used body size as
the appropriate baseline against which
to measure relative deviations in brain
size. However, a problem has since
emerged: Brain size is determined
early in development and, compared
to many other body systems, appears
to be highly conservative in evolutionary
terms. As a result, body size can
often change dramatically both ontogenetically
across populations in response
to local environmental conditions21
and phylogenetically22,23
without corresponding changes in
brain size. This is particularly conspicuous
in the case of phyletic dwarfs
(e.g., callitrichids and perhaps modern
humans and hylobatids22) and species
in which body size may have
increased in response to predation
pressure following the occupation of
more open terrestrial habitats (e.g.,
papionids24).
The lability of body size therefore
makes it a poor baseline, though one
that probably is adequate for analyses
on the mouse-elephant scale. Consequently,
it is necessary to find an internally
more consistent baseline for taxonomically
fine-grained analyses.
Willner22 suggested that either molar
tooth size or brain size may be suitable
because both are developmentally
conservative. Because we are concerned
with brain part size, some
aspect of brain size seems the most
appropriate.
At this point, three options are available.
One is to compare the neocortex,
the brain part of interest, with the
whole brain; the second is to use the
rest of the brain other than the part of
interest; the third is to use some less
variable primitive component of the
brain, such as the medulla, as a baseline.
Two options are in turn available
as mechanisms for controlling for
brain size in each of these cases. One
is to use residuals from a common
regression line against the baseline
(e.g., the residual of neocortex volume
on total brain volume or medula volume).
The other choice is to use ratios.
We have considered and tested all
these options10,24 (see Box 1). The results
are virtually identical irrespective
of which measure is used. One
explanation for this may be that all
these measures actually index the same
thing, absolute neocortex size, mainly
because the neocortex is such a large
component of the primate brain. Indeed,
the use of absolute neocortex
size produces results that are similar
to those obtained from relativized indices
of neocortex volume.24,25 This
makes some sense in computational
terms: As Byrne26 has pointed out, a
10% increase in the processing capacity
of a small computer is worth a
great deal less in information-processing
terms than is a 10% increase in a
large computer. Although residuals
from a common regression line would
conventionally be considered the safest
measure, and have been used in
many recent analyses,27,28 I shall continue
to use my original ratio index
because it provides the best predictor
(see Box 1).
Finally, it is now widely appreciated
that comparative analyses need to control
for the effects of phylogenetic
inertia. Closely related species can be
expected to have similar values for
many anatomical and behavioral dimensions
merely by virtue of having
inherited them from a recent common
ancestor. In such cases, plotting raw
data would result in pseudoreplication,
artificially inflating the sample
size by assuming that closely related
species are actually independent evolutionary
events. The ways of dealing
with this problem include plotting
means for higher taxonomic units, performing
nested analyses of variance
using phylogenetic levels as factors,
comparing matched pairs of species,
and making independent contrasts that
control directly for phylogeny. Each
method has its own advantages and
disadvantages, but the first and third
procedures are particularly associated
with loss of information and small
sample sizes. I shall use the first and
last method, the last because it allows
individual species to be compared, but
the first because it allows grade shifts
within data sets to be identified (a
problem that independent-contrasts
methods have difficulty dealing with).
I shall take the genus as a suitable
basis for analysis because genera typically
represent different reproductive
or ecological radiations and thus are
more likely to constitute independent
evolutionary events.
The resulting analyses are relatively
straightforward: Figure 2 presents the
data for neocortex ratio for the anthropoid
primate species in the data base
of Stephan, Frahm, and Baron.29 Neocortex
size, however measured, does
not correlate with any index of the
ecological hypotheses, but does correlate
with social group size. Similar
findings were reported by Sawaguchi
and Kudo,30 who found that neocortex
size correlated with mating system in
primates. Barton and Purvis31 have
confirmed that using both residuals of
neocortex volume on total brain volume
and the method of independent
contrasts yields the same result. Both
Barton10 and T. Joffe (unpublished)
have repeated the analyses using the
medulla as the baseline for comparison.
More importantly, Barton and
Purvis31 have shown that while relative
neocortex volume correlates with
group size but not the size of the
ranging area, the reverse is true of
relative hippocampus size. A correlation
between range area and hippocampus
size is to be expected because
of hippocampal involvement in spatial
memory.32,33 This correlation demonstrates
that it is not simply total brain
size that is important (a potential problem,
given the overwhelming size of
The neocortex is
generally regarded as
being the seat of those
cognitive processes that
we associate with
reasoning and
consciousness, and
therefore may be
expected to be under
the most intense
selection from the need
to increase or improve
the effectiveness of
these processes.
ARTICLES Evolutionary Anthropology 181
O
nly
for educational purposes
SR
T
II
the primate neocortex). Moreover, it
points to the specific involvement of
the neocortex.
The validity of this relationship
could be tested directly by using it to
predict group sizes in a sample of
species for which brain volumetric
data were not available in the original
sample of Stephan, Frahm, and
TABLE 2. Stepwise Regression Analysis
of Indices of Brain Component Volume
as Predictors of Group Size in
Anthropoid Primates, Based on
Independent Contrasts Analysis*
Independent Variable t P
Absolute brain volume Ϫ1.69 0.107
Residual of brain volume
on body mass Ϫ0.91 0.371
Absolute telencephalon
volume Ϫ1.72 0.101
Residual of telencephalon
volume on
body mass 0.61 0.546
Residual of telencephalon
volume on
brain volume 1.69 0.107
Absolute neocortex
volume Ϫ1.70 0.104
Residual of neocortex
volume on body mass 0.56 0.583
Residual of neocortex
volume on brain
volume 1.60 0.136
Neocortex ratio (against
rest of brain) 3.79 0.001
*Sample: 24 species of anthropoid primates
from Stephan, Frahm, and
Baron.29
Box 1. How to Measure Brains
R. Dunbar and Tracey H. Joffe
The different ways of measuring
relative brain size have raised doubts
as to the most appropriate technique
to use.65 Many researchers have preferred
to use residuals from the common
regression line of best fit for the
data set concerned. This provides a
measure of the extent to which brain
(or brain part) volume deviates from
what would be expected for an average
member of the relevant taxon of
the appropriate size. Although ratios
have been used to compare the relative
size of brain components,29,66 this
has been criticized on the grounds
that trade-offs within the brain may
mean that a given index simply measures
total brain size (or the size of a
brain part) and thus does not remove
the effects of absolute size. Ratios
may also be prone to autocorrelation
effects, especially when the baseline
is taken to be the whole brain and, as
in the case of the neocortex, the part in
question is a major volumetric component
of the brain.
Although there are likely to be
some trade-offs of this kind within the
brain, the fact that neocortex volume
increases progressively across the primate
order suggests that such constraints
are less likely to have a significant
effect on a ratio measure. Of
course the residuals procedure is itself
a ratio: Encephalization-type indices
are calculated as actual volume divided
by predicted volume (which,
when data are logged, becomes the
conventional actual minus predicted
values). Thus, ratio measures per se
may not be the problem. Rather, the
substantive objection is whether or not
a ratio partials out the allometric effects
of body size. In fact, it seems that
neocortex ratios are not correlated
with the basal brain (i.e., brain volume
excluding the neocortex) within major
taxonomic groups (unpublished analyses).
Consequently, this criticism has
less force than it might appear to have
on first sight. Moreover, any index that
uses the whole brain as its base is
likely to suffer from autocorrelation
effects. Because the neocortex is such
a large proportion of the brain in primates,
residuals of neocortex from
total brain size may simply be a measure
of neocortex plotted against it-
self.
To consider the problem in more
detail, we ran a stepwise regression
analysis on the 24 species of anthropoid
primates, including humans, on
the data base of Stephan, Frahm, and
Baron,29 with group size as the dependent
variable and nine indices of relative
brain or brain-part volume as independent
variables. In addition to
neocortex ratio, these included total
brain volume as well as telencephalon
and neocortex volume, each taken as
absolute volume and as a residual
from both body mass and brain volume.
All variables were log10-transformed
for analysis. In both cases,
neocortex ratio was selected as the
variable of first choice. We carried out
both regressions on generic plots and
independent contrast analyses. For
the contrasts analysis, the best fit
least-squares regression equation
through the origin was:
Contrast in log10 (group size) ϭ 3.834
* Contrast in log10 (neocortex ratio)
(r2 ϭ 0.395, F1,22 ϭ 15.39,
P ϭ 0.001).
With all other variables held constant,
none of the other eight indices made a
significant contribution to the variability
in group size in either analysis.
Table 2 gives the results for the independent
contrasts analysis.
Neocortex ratio is thus the single
most powerful predictor of group size
in these species. While the biological
significance of this variable remains
open to interpretation, the fact that it
provides the best predictor, independently
of all other confounding measures,
suggests that more detailed
consideration needs to be given to its
significance and meaning. It may be,
for example, that body size, rather
than being a determinant20 is simply a
constraint on neocortex size: A species
can evolve a large neocortex only
if its body is large enough to provide
the spare energy capacity through
Kleiber’s relationship for basal metabolic
rate to allow for a larger than
average brain. This interpretation is
implied by the Aiello and Wheeler2
‘‘expensive tissue hypothesis.’’It would
also be in line with Finlay and Darlington’s11
claim that in mammals the evolution
of brain-part size is driven, developmentally
at least, by the evolution of
the whole brain, thus generating very
tight correlations between brain-part
size and total brain size.
182 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II
Baron.29 I did this by exploiting the
fact that neocortex ratios can be predicted
from total brain volume,34 a
result that, in fact, follows directly
from the Finlay and Darlington11 findings.
The result was a significant fit
between predicted neocortex ratio and
observed mean group size for a sample
of 15 New and Old World monkey
species.35
Barton27 noted that the original
analyses of Dunbar24 seemed to imply
that variation in neocortex size was
much greater than variation in group
size in the prosimians. Using Dunbar’s24
data on group size, Barton suggested
that the relationship between
neocortex and group size did not apply
in the case of prosimians. However,
the data on prosimian group
sizes in this sample suffered from a
paucity of data, particularly for the
nocturnal species. Because many of
these are described as semi-solitary, it
was conservatively assumed in the
Dunbar24 database that their group
size was one. More recent field studies
have produced markedly improved estimates
of the sizes of social groups
and, in the case of the semi-solitary
species, daytime nest groups.36 Reanalysis
of the data for prosimians
using these improved estimates of social
group size suggests that these
species do in fact adhere to the same
relationship between neocortex and
group size as that which pertains for
other primates.25 More importantly,
the regression line for this taxon is
parallel to, but shifted to the left of,
that for other anthropoid primates
(Fig. 3).
This relationship has now been
shown to hold for at least four other
mammalian orders: bats,10 carnivores
and insectivores,37,38 and odontocete
cetaceans.39,40 In the case of the insectivores,
the data points are shifted far
to the left of those for the primates, as
might be expected of a taxonomic
group that is considered to be broadly
representative of the ancestral mammals.37
However, the relationship is
weak in this case, probably because
estimates of group size are particularly
uncertain for insectivores.
Surprisingly, the data for the carnivores
map directly onto those for the
simian primates, that is, the regression
lines for the two data sets do not
differ significantly. However, the carnivores
do not exhibit as wide a range of
neocortex ratios or group sizes as do
anthropoid primates. The fact that the
prosimians lie to the left of both these
taxonomic groups implies that the carnivores
represent an independent evolutionary
development along the same
principles as the anthropoid primates,
the difference being that they just have
not taken it as far as primates have.
One reason for this may be that the
carnivore social world is olfactiondominated
rather than vision-dominated,
as in the case of the primates.
Barton10,27,41 has pointed out that the
shift to a diurnal lifestyle based on
color vision, perhaps initially dietdriven,
but leading to a shift into vision-based
communication, may be
the key feature that has spurred on the
dramatic development of the primate
neocortex.
Figure 2. Relative neocortex size in anthropoid primates plotted against (a) percentage of fruit in the diet, (b) mean home-range size scaled as
the residual of range size regressed on body weight (after Dunbar24), (c) types of extractive foraging (after Gibson4), and (d) mean group size.
((a), (b), and (d) are redrawn from Dunbar24, Figures 6, 2 and 1, respectively; (c) is from Dunbar,35 Figure 2.)
ARTICLES Evolutionary Anthropology 183
O
nly
for educational purposes
SR
T
II
REFINING THE RELATIONSHIP
The social brain hypothesis implies
that constraints on group size arise
from the information-processing capacity
of the primate brain, and that
the neocortex plays a major role in
this. However, even this proposal is
open to several interpretations as to
how the relationship is mediated. At
least five possibilities can be usefully
considered. The constraint on group
size could be a result of the ability to
recognize and interpret visual signals
for identifying either individuals or
their behavior; limitations on memory
for faces; the ability to remember who
has a relationship with whom (e.g., all
dyadic relationships within the group
as a whole); the ability to manipulate
information about a set of relationships;
and the capacity to process emotional
information, particularly with
respect to recognizing and acting on
cues to other animals’ emotional states.
These are not all necessarily mutually
exclusive, but they do identify different
points in the cognitive mechanism
that might be the crucial informationprocessing
bottleneck.
Although visual mechanisms are
likely to be important for social interaction,
and may well have been the
initial kick for the evolution of large
brains in primates,10 it seems intrinsically
unlikely that the ultimate constraint
lies in the mechanisms of the
visual system itself.28 Although there
is a correlation between the relative
size of the visual cortex and group size
in anthropoid primates, the fit is much
poorer, and the slope significantly shallower
than that between the nonvisual
neocortex and group size (r2 ϭ 0.31 vs
r2 ϭ 0.61, respectively) (Fig. 4). Partial
correlation analysis indicates that only
the correlation for the nonvisual relationship
remains significant when the
other component is held constant28
(though this is not true for prosimians25).
A more important point is that
the volume of the lateral geniculate
nucleus, a major subcortical way station
in visual processing, does not
correlate with group size at all, indicating
that pattern recognition per se is
unlikely to be the issue.28 It may be of
some significance that the absolute
size of the visual cortex seems to reach
an asymptotic value in the great ape
clade, whereas the nonvisual neocortex
continues to increase in size. One
interpretation of this is that visual
processing does not necessarily continue
to improve indefinitely as the
size of the cortical processing machinery
increases, at least relative to the
opportunity cost of taking cortical neurons
away from other cognitive pro-
cesses.
It seems equally unlikely that the
problem lies with a pure memory constraint,
though memory capacity obviously
must impose some kind of upper
limit on the number of relationships
that an animal can have. There are
three reasons for this claim. First, in
humans at least, memory for faces is
an order of magnitude larger than the
predicted cognitive group size: Humans
are said to be able to attach
names to around 2,000 faces but have
a cognitive group size of only about
150. Second, there is no intrinsic reason
to suppose that memory per se is
the issue. The social brain hypothesis
is about the ability to manipulate information,
not simply to remember it.
Third, and perhaps most significantly,
memories appear to be stored mainly
in the temporal lobes,42 whereas recent
PET scan studies implicate the
prefrontal neocortex, notably Brodman
area 8, as the area for social skills
and, specifically, theory of mind.43
Frith44 has suggested that memories
and representations for objects or
events may involve interactions between
several levels of the neocortex
depending on the kinds of operations
Figure 3. Mean group size plotted against neocortex ratio for individual genera, shown
separately for prosimian, simian, and hominoid primates. Prosimian group size data, from
Dunbar and Joffe,25 include species for which neocortex ratio is estimated from total brain
volume. Anthropoid data are from Dunbar.24 Simians: 1, Miopithecus; 2, Papio; 3, Macaca; 4,
Procolobus; 5, Saimiri; 6, Erythrocebus; 7, Cercopithecus; 8, Lagothrix; 9, Cebus; 10, Ateles; 11,
Cercocebus; 12, Nasalis; 13, Callicebus; 14, Alouatta; 15, Callimico; 16, Cebuella; 17, Saguinus;
18, Aotus; 19, Pithecia; 20, Callicebus. Prosimians: a, Lemur; b, Varecia; c, Eulemur; d, Propithecus;
e, Indri; f, Microcebus; g, Galago; h, Hapalemur; i, Avahi; j, Perodictus.
The social brain
hypothesis implies that
constraints on group size
arise from the
information-processing
capacity of the primate
brain, and that the
neocortex plays a major
role in this. However,
even this proposal is
open to several
interpretations as to how
the relationship is
mediated.
184 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II
involved. These interactions could occur
between the sensory and association
cortices (perceiving an object),
between the association and frontal
cortices (remembering an object), and
among all three (being aware of perceiving
an object). It is worth noting in
this context that although social skills
are commonly disrupted by damage to
the prefrontal cortex, memory for
events and people is not.42
It seems unlikely that emotional responses
per se are the substantive
constraint. Although the correct emission
and interpretation of emotional
cues is of singular importance in the
management of social relationships,45
there is little evidence that the subcortical
areas principally associated with
emotional cuing (for example, the
amygdala in the limbic system) correlate
in any way with social group
size.28 Indeed, Keverne, Martel, and
Nevison46 point out that there has
been progressive reduction in the relative
sizes of the ‘‘emotional’’ centers in
the brain (the hypothalamus and septum)
in favor of the ‘‘executive’’ centers
(the neocortex and striate cortex)
during primate evolution. They interpret
this in terms of a shift away from
emotional control of behavior to more
conscious, deliberate control.
The only remaining alternative is
that the mechanisms involved lie in
the ability to manipulate information
about social relationships themselves.
This claim is supported by six additional
lines of evidence that point to
the fundamental importance of social
skills in the detailed management of
social relationships.
One is the fact that close analysis of
the data on group size and neocortex
volume suggests that there are, in fact,
distinct grades even within the anthropoid
primates (Fig. 3). Apes seem to lie
on a separate grade from the monkeys,
which in turn lie on a separate grade
from the prosimians. The slope coefficients
on these separate regression
lines do not differ significantly, but the
intercepts do. It is as if apes require
more computing power to manage the
same number of relationships that
monkeys do, and monkeys in turn
require more than prosimians do. This
gradation corresponds closely to the
perceived scaling of social complexity.
The second line of evidence is that
the rates with which tactical deception
are used correlate with neocortex
size.26 Species with large neocortex
ratios make significantly more use of
tactical deception, even when the differential
frequencies with which these
large-brained species have been studied
are taken into account.
Third, Pawlowski, Dunbar, and
Lowen47 have shown that among polygamous
primates the male rank correlation
with mating success is negatively
related to neocortex size (Fig. 5).
This is just what we would predict if
the lower ranking males of species
with larger neocortices were able to
use their greater computational capacities
to deploy more sophisticated social
skills, such as the use of coalitions
and capitalizing on female mate
choice, to undermine or circumvent
the power-based strategies of the dominant
animals.
The fourth line of evidence is Joffe’s48
demonstration that adult neocortex
size in primates correlates with the
length of the juvenile period, but not
with the length of gestation, lactation,
or the reproductive life span, even
though total brain size in mammals
correlates with the length of the gestation
period.49,50 This suggests that what
is most important in the development
of a large neocortex in primates is not
the embryological development of
brain tissue per se, which is associated
mainly with gestation length, but
rather the ‘‘software programming’’
that occurs during the period of social
learning between weaning and adult-
hood.
Fifth, Kudo, Lowen, and Dunbar51
have shown that grooming clique size,
a surrogate variable that indexes alliFigure
4. Independent contrasts in mean group size plotted against contrasts in the visual
cortex and the volume of the rest of the neocortex (nonvisual neocortex) for individual
anthropoid species. Note that the visual cortex is here defined as visual area V1; the nonvisual
cortex is the non-V1 volume of the neocortex and thus includes some higher order visual
processing components (e.g., visual area V2). Unfortunately, the data base of Stephan, Frahm,
and Baron29 does not allow us to define our measure of the nonvisual area any more finely than
this. (Reprinted from Joffe and Dunbar,28 Fig. 1.)
. . . there is no intrinsic
reason to suppose that
memory per se is the
issue. The social brain
hypothesis is about the
ability to manipulate
information, not simply to
remember it.
ARTICLES Evolutionary Anthropology 185
O
nly
for educational purposes
SR
T
II
ance size, correlates rather tightly with
relative neocortex and social group
size in primates, including humans
(Fig. 6). The human data derive from
two samples: hair-care networks
among female bushmen52 and support
cliques among adults in the United
Kingdom.53 What is remarkable is how
closely the human data fit with the
data from other primate species.
Grooming cliques of this kind invariably
function as coalitions in primate
groups. Coalitions are functionally crucial
to individuals within these groups
because they enable the animals to
minimize the levels of harassment and
competition that they inevitably suffer
when living in close proximity to others.54
Coalitions essentially allow primates
to manage a fine balancing act
between keeping other individuals off
their backs while at the same time
avoiding driving them away altogether
and thereby losing the benefits for
which the groups formed in the first
place. These results can thus probably
be interpreted as a direct cognitive
limitation on the number of individuals
with which an animal can simultaneously
maintain a relationship of sufficient
depth that they can be relied on
to provide unstinting mutual support
when one of them is under attack.
Because this is the core process that
gives primate social groups their internal
structure and coherence, this can
be seen as a crucial basis for primate
sociality.
Finally, Keverne, Martel, and Nevison46
have suggested that the neocortex
and striate cortex, those areas of
the primate brain that are responsible
for executive function, are under maternally
rather than paternally imprinted
genes (i.e., genes that ‘‘know’’
which parent they came from),
whereas the converse is true for the
limbic system, those parts of the brain
most closely associated with emotional
behavior. They interpret this in
relation to the cognitive demands of
the more intense social life of females
in matrilineal female-bonded societ-
ies.
IMPLICATIONS FOR HUMAN
GROUPS
The fact that the relationship between
neocortex size and what I will
term the cognitive group size holds up
so well in so many different taxonomic
groups raises the obvious question of
whether or not it also applies to humans.
We can easily predict a value for
group size in humans. Doing so, which
is simply a matter of using the human
neocortex volume to extrapolate a
value for group size from the primate
Figure 5. Independent contrasts in the Spearman rank correlation (rs) between male rank and
mating success plotted against contrasts in neocortex size for two different male cohort sizes (4
to 8, and 9 to 30 males) for individual species. The regression equations for the two cohort sizes
are significantly different from b ϭ 0. The species sampled are C. apella, P. entellus, C. aethiops,
M. fuscata, M. mulatta, M. radiata, M. arctoides, P. cynocephalus, P. anubis, P. ursinus, and P.
troglodytes. (Redrawn from Pawlowski, Dunbar, and Lowen,47 Fig. 1.)
Figure 6. Mean grooming clique size plotted against mean neocortex ratio for individual
primate genera. The square is Homo sapiens. Species sampled are L. catta, L. fulvus, Propithecus,
Indri, S. sciureus, C. apella, C. torquatus, A. geoffroyi, A. fusciceps, P. badius, P. entellus, P.
pileata, P. johnii, C. campbelli, C. diana, C. aethiops, C. mitis, E. patas, M. mulatta, M. fuscata,
M. arctoides, M. sylvana, M. radiata, P. anubis, P. ursinus, P. cynocephalus, P. hamadryas, T.
gelada, P. troglodytes, P. paniscus. (Redrawn from Kudo, Lowen, and Dunbar,51 Fig. 4a.)
186 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II
equation, produces a value in the order
of 150. The real issue is whether
humans really do go around in groups
of this size.
Identifying the relevant level of
grouping to measure in humans is
difficult because most humans live in
a series of hierarchically inclusive
groups. This, in itself, is not especially
unusual: Hierarchically structured
groups of this kind are characteristic
of primates54 and may be typical of
many mammals and birds.55 At least
in the case of the diurnal primates, it
seems that, with a few notable exceptions,
the various species’ grouping
patterns exhibit an overt level of stability
at roughly the same position in the
hierarchy across a wide range of taxa.
Moreover, because the various layers
of this hierarchy appear to be intimately
related to each other, probably
through being part of a series of causeand-consequence
chains,51 it would
not matter which particular grouping
level (for example, stable social group,
network, or grooming clique) was
taken to be the grouping criterion.
The problem with respect to humans
is that it is difficult to identify
which of the many potential grouping
levels is functionally or cognitively
equivalent to the particular level of
grouping that I happened to use for
primates. This difficulty is particularly
intrusive in this case because humans
live in a dispersed social system sometimes
referred to as a fission-fusion
system. In order to get around this
problem, I adopted the converse strategy
in my original analysis,56 asking
whether there was any group size consistently
characteristic of humans that
was of about the requisite size and, if
so, whether its intrinsic psychological
characteristics were similar to those
found in primate groups.
Because of the structural complexity
of postagricultural societies, I considered
only traditional hunter-gatherer
and small-scale horticultural
societies. Although census data on
such societies are limited, those that
are available suggest that there is indeed
a consistent group size in the
region of 150 individuals (Fig. 7). Except
among settled horticulturalists,
where the village seems to be the
relevant unit, this typically involves
the set of individuals from whom overnight
camps are easily and regularly
formed. Such groups are not often
conspicuous as physical entities (they
do not often appear together in one
place at one time), but they do invariably
have important ritual functions
for the individuals concerned. Among
Australian aboriginals, for example,
the relevant group is the clan, which
meets from time to time in jamborees
where the rituals of life (marriages
and rites of passage) are enacted and
tales of the old times are rehearsed to
remind everyone who they are and
why they hold a particular relationship
to each other. Indeed, this genuinely
seems to be the largest group of
people who know everyone in the
group as individuals at the level of
personal relationships. This is essentially
the definition that holds in the
case of primates.
A more extensive exploration of human
groups in other contexts suggests
that groupings of this size are widespread
and form an important component
of all human social systems, being
present in structures that range
from business organizations to the
arrangement of farming communities.56
Estimates of community size
for two traditional farming communities
in the United States, Hutterites
and an East Tennessee mountain community,
and of actual social network
sizes (from small-worlds experiments)
(shown as triangles on the right side of
Fig. 7) fit very closely within the relevant
range of group sizes.
It is easy, of course, to play the
numerologist in this context by finding
groups that fit whatever group size
one wishes to promote. The important
feature to note here, however, is that
the various human groups that can be
identified in any society seem to cluster
rather tightly around a series of
values (5, 12, 35, 150, 500, and 2,000)
with virtually no overlap in the variance
around these characteristic values.
They seem to represent points of
stability or clustering in the degrees of
familiarity within the broad range of
Figure 7. Mean sizes for different types of groups in traditional human societies. Individual
societies are ordered along the bottom, with data for three main types of social groups
(overnight camps, clans or villages, and tribes). Societies include hunter-gatherer and settled
horticulturalists from Australia, Africa, Asia, and North and South America. The triangles give
mean group sizes for three contemporary United States samples: mean network size from
small-worlds experiments (N ϭ 2),67 mean Hutterite community size,68 and the size of an East
Tennessee mountain community.69 The value of 150 predicted by the primate neocortex size
relationship (from Fig. 1d) is indicated by the horizontal line, with 95% confidence intervals
shown as dashed lines.
ARTICLES Evolutionary Anthropology 187
O
nly
for educational purposes
SR
T
II
human relationships, from the most
intimate to the most tenuous.
COGNITIVE MECHANISMS
The suggestion that the mechanisms
involved in these processes may
be concerned with social skills raises
the issue alluded to by the original
Machiavellian intelligence hypothesis,
namely to what extent cognitively sophisticated
mechanisms conferring the
ability to ‘‘mind-read’’ might be involved.
Tactical deception, in its strong
sense, implies the ability to hold false
beliefs and, thus, the presence of the
ability known as ‘‘theory of mind’’
(ToM). Of course, tactical deception as
practiced by primates on a daily basis
may not, as Byrne26 himself has
pointed out, be quite as sophisticated
as first impressions suggest. A more
conventional behaviorist account
based on simple associative learning
can invariably be given for almost all
examples reported in the literature.
Nonetheless, convincing evidence
suggests that humans at least do use
ToM in executing some of their more
manipulative social activities. And
while we may not wish to attribute full
ToM to all primates, at least circumstantial
evidence suggests that basic
ToM is present in great apes and that
monkeys may aspire to a level that
Byrne26 has described as level 1.5 intentionality
(full ToM being level 2
intentionality) (see Box 2). The difference
has been summed up rather
graphically by Cheney and Seyfarth’s57
observation that apes seem to be good
psychologists in that they are good at
reading minds, whereas monkeys are
good ethologists in that they are good
at reading behavior—or at least at
making inferences about intentions in
the everyday sense, even if not in the
philosophical sense of belief states.
Evidence that chimpanzees aspire
to at least a basic form of ToM is
provided by their performance on experimental
false-belief tasks.58–61 These
studies have attempted to develop analogues
of the classic false-belief tasks
used with children.62 Though it is clear
that chimpanzees do not perform to
the level at which fully competent
children perform, O’Connell’s61 experiments
at least suggest that they can
perform at the level of children who
stand on the threshold of acquiring
ToM. More importantly, chimpanzees
do better than autistic adults, one of
whose defining features is the lack of
ToM, on the same tests.
That mind-reading, the basis of ToM,
is difficult to do has been shown by
experiments on normal adults tested
on ‘‘advanced’’ ToM tasks, up to fifthorder
intentionality.62 These data suggest
that normal humans find tasks of
greater than fourth-order intentionality
exceedingly hard to do. The high
error rates at these levels do not reflect
a memory retention problem: All subjects
pass the tests that assess memory
for the story line. Moreover, the same
subjects show considerable competence
on reasoning tasks that involve
causal chains of up to the sixth order.
The difficulty seems genuinely to be
something to do with operating with
deeply embedded mental states.
One possibly significant observation
in this context is that the visual and
nonvisual components of the primate
neocortex do not increase isometrically.
Although initially there is a more
or less linear increase in the visual
area V1 with increasing size of the rest
of the neocortex, this drops off within
the great ape clade. From gorillas
through humans, increases in the size
of the visual area progress more slowly
than do increases in the size of the rest
of the neocortex.28 We interpret this as
implying that beyond a certain point
the acuity of the visual system does
Box 2. A Beginner’s Guide to Intensionality
Computers can be said to know
things because their memories contain
information; however, it seems
unlikely that they know that they know
these things, in that we have no evidence
that they can reflect on their
states of ‘‘mind.’’ In the jargon of the
philosophy of mind, computers are
zero-order intensional machines. Intensionality
(with an Ϫs) is the term
that philosophers of mind use to refer
to the state of having a state of mind
(knowing, believing, thinking, wanting,
understanding, intending, etc).
Most vertebrates are probably capable
of reflecting on their states of
mind, at least in some crude sense:
they know that they know. Organisms
of this kind are first-order intensional.
By extension, second-order intensional
organisms know that someone
else knows something, and third-order
intensional organisms know that someone
else knows that someone else
knows something. In principle, the sequence
can be extended reflexively
indefinitely, although, in practice, humans
rarely engage in more than
fourth-order intensionality in everyday
life and probably face an upper limit at
sixth-order (‘‘Peter knows that Jane
believes that Mark thinks that Paula
wants Jake to suppose that Amelia
intends to do something’’).
A minimum of fourth-order intensionality
is required for literature that
goes beyond the merely narrative (‘‘the
writer wants the reader to believe that
character A thinks that character B
intends to do something’’). Similar abilities
may be required for science, since
doing science requires us to ask
whether the world can be other than it
is (a second-order problem at the very
least) and then ask someone else to
do the same (an additional order of
intensionality).
. . . apes seem to be
good psychologists in
that they are good at
reading minds, whereas
monkeys are good
ethologists in that they
are good at reading
behavior . . .
188 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II
not increase linearly with size. Because
the total size of the neocortex is
limited by embryological and energetic
factors, this means that disproportionately
more capacity can be
dedicated to nonvisual areas of the
neocortex once the volume is above
the crucial threshold. This might explain
why apes appear to be capable of
the additional cognitive processing associated
with mind reading, whereas
monkeys are not. It might also explain
why humans are better at it than apes.
For humans, one important aspect
of ToM concerns its relevance to language,
a communication medium that
crucially depends on understanding
interlocutors’ mental states or intentions.
The kinds of metaphorical uses
of language that characterize not only
our rather telegraphic everyday exchanges
(in which ‘‘you know what I
mean?’’ is a common terminal clause)
but also lies at the very heart of the
metaphorical features of language. As
studies of pragmatics have amply demonstrated,63
a great deal of linguistic
communication is based on metaphor:
Understanding the intentions behind
a metaphor is crucial to successful
communication. Failure to understand
these intentions commonly results in
confusion or inappropriate responses.
Indeed, without these abilities it is
doubtful whether literature, notably
poetry, would be possible. Our conversations
would be confined to the banally
factual; those fine nuances of
meaning that create both the ambiguities
of politeness and the subtleties of
public relations would not be pos-
sible.64
REFERENCES
1 Clutton-Brock TH, Harvey PH (1980) Primates,
brains and ecology. J Zool Lond 190:309–323.
2 Aiello LC, Wheeler P (1995) The expensive
tissue hypothesis. Curr Anthropol 36:184–193.
3 Parker ST, Gibson KR (1977) Object manipulation,
tool use and sensorimotor intelligence as
feeding adaptations in great apes and cebus
monkeys. J Hum Evol 6:623–641.
4 Gibson KR (1986) Cognition, brain size and the
extraction of embedded food resources. In Else J,
Lee PC (eds), Primate Ontogeny, Cognition and
Social Behaviour, pp 93–104. Cambridge: Cambridge
University Press.
5 Whiten A, Byrne R (1988) Tactical deception in
primates. Behav Brain Sci 12:233–273.
6 Byrne R, Whiten A (eds) (1988) Machiavellian
Intelligence. Oxford: Oxford University Press.
7 Harcourt AH (1988) Alliances in contests and
social intelligence. In Byrne R, Whiten A (eds),
Machiavellian Intelligence, pp 142–152. Oxford:
Oxford University Press.
8 Harcourt AH (1989) Sociality and competition
in primates and non-primates. In Standen V,
Foley R (eds), Comparative Socioecology, pp 223–
242. Oxford: Blackwell Scientific.
9 Brothers L (1990) The social brain: A project
for integrating primate behaviour and neurophysiology
in a new domain. Concepts Neurosci 1:27–
251.
10 Barton RA, Dunbar RLM (1997) Evolution of
the social brain. In Whiten A, Byrne R (eds),
Machiavellian Intelligence, Vol. II. Cambridge:
Cambridge University Press.
11 Finlay BL, Darlington RB (1995) Linked regularities
in the development and evolution of mammalian
brains. Science 268:1678– 1684.
12 Martin RD (1981) Relative brain size and
metabolic rate in terrestrial vertebrates. Nature
Lond 293:57–60.
13 Martin RD (1983) Human brain evolution in
an ecological context. 52nd James Arthur Lecture,
American Museum of Natural History, New
York.
14 Hofman MA (1983) Evolution of the brain in
neonatal and adult placental mammals: A theoretical
approach. J Theoret Biol 105:317–322.
15 Pagel M, Harvey PH (1988) How mammals
produce large-brained offspring. Evolution 42:
948–957.
16 Boesch C, Boesch H (1984) Mental map in
chimpanzees. Primates 25:110–170.
17 MacLean PD (1982) On the origin and progressive
evolution of the triune brain. In Armstrong
E, Falk D (eds), Primate Brain Evolution, pp
291–310. New York: Plenum Press.
18 Joffe T, Dunbar RIM (n.d.). Primate brain
system evolution. Brain Behav Evol, submitted
for publication.
19 Passingham RE (1982) The Human Primate.
San Francisco: Freeman.
20 Jersion HJ (1973) Evolution of the Brain and
Intelligence. New York: Academic Press.
21 Dunbar RIM (1989) Environmental determinants
of intraspecific variation in body weight in
baboons. J Zool London 220:167–169.
22 Willner LA (1989) Sexual Dimorphism in Primates.
Ph.D. thesis, University of London.
23 Deacon TW (1990) Fallacies of progression
indices in theories of brain-size evolution. Int J
Primatol 12:193–236.
24 Dunbar RIM (1992) Neocortex size as a constraint
on group size in primates. J Hum Evol
20:469–493.
25 Dunbar RIM, Joffe TH (n.d.) Neocortex size
and social group size in prosimians. Primates,
submitted for publication.
26 Byrne RB (1995) The Thinking Primate. Oxford:
Oxford University Press.
27 Barton RA (1995) Neocortex size and behavioural
ecology in primates. Proc R Soc London B,
263:173–177.
28 Joffe TH, Dunbar RIM (1997) Visual and
socio-cognitive information processing in primate
brain evolution. Proc R Soc London B,
264:1303–1307.
29 Stephan H, Frahm H, Baron G (1981) New
and revised data on volumes of brain structures
in insectivores and primates. Folia Primatol
35:1–29.
30 Sawaguchi T, Kudo H (1990) Neocortical development
and social structure in primates. Primates
31:283–290.
31 Barton RA, Purvis A (1994) Primate brains
and ecology: Looking beneath the surface. In
Anderson JR, Thierry B, Herrenschmidt N (eds),
Current Primatology, pp 1–12. Strasbourg: University
of Strasbourg Press.
32. O’Keefe J, Nadel L (1978) The Hippocampus
as a Cognitive Map. Oxford: Oxford University
Press.
33. Krebs JH, Sherry DF, Healy SD, Perry VH,
Vaccarino AL (1989) Hippocampal specialisation
of food-storing birds. Proc Natl Acad Sci USA
86:1488–1492.
34. Aiello LC, Dunbar RIM (1993) Neocortex
size, group size and the evolution of language.
Curr Anthropol 34:184–193.
35 Dunbar RIM (1995) Neocortex size and group
size in primates: A test of the hypothesis. J Hum
Evol 28:287–296.
36. Bearder SK (1987) Lorises, bushbabies and
tarsiers: Diverse societies in solitary foragers. In
Smuts B, Cheney D, Seyfarth R, Wrangham R,
Struhsaker T (eds), Primate Societies, pp 12–24.
Chicago: Chicago University Press.
37 Dunbar RIM, Bever J (n.d.) Neocortex size
determines group size in carnivores and insectivores.
Ethology, in press.
38 Gittleman JH (1986) Carnivore brain size,
behavioural ecology and phylogeny. J Mammal
67:23–36.
39 Marino L (1996) What can dolphins tell us
about primate evolution? Evol Anthropol 5:81–86.
40 Tschudin A (1996) The Use of Neuroimaging
in the Assessment of Brain Size and Structure in
Odontocetes. MSc thesis, University of Natal
(Durban).
41 Barton RA, Purvis A, Harvey PH (1995) Evolutionary
radiation of visual and olfactory brain
systems in primates, bats and insectivores. Philos
Trans R Soc B 348:381–392.
42 Kolb B, Wishaw LO (1996) Fundamentals of
Human Neuropsychology. San Francisco: Free-
man.
43 Fletcher P, Happe´ F, Frith U, Baker SC, Dolan
RJ, Frakowiak RSJ, Frith CD (1996) Other minds
in the brain: A functional imaging study of ‘‘theory
of mind’’ in story comprehension. Cognition 00:
000–000.
44 Frith C (1996) Brain mechanisms for ‘‘having
a theory of mind.’’ J Psychopharmacol 10:9–16.
45 Armstrong E, Clarke MR, Hill EM (1987)
Relative size of the anterior thalamic nuclei differentiates
anthropoids by social system. Brain Behav
Evol 30:263–271.
46 Keverne EB, Martel FL, Nevison CM (1996)
Primate brain evolution: Genetic and functional
considerations. Proc R Soc Lond B 262:689–696.
47 Pawlowski B, Dunbar R, Lowen C (n.d.) Neocortex
size, social skills and mating success in
primates. Behaviour, in press.
48 Joffe TH (1997) Social pressures have selected
for an extended juvenile period in primates. J
Hum Evol, in press.
49 Bennett PM, Harvey PH (1985) Brain size,
development and metabolism in birds and mammals.
J Zool London 207:491–509.
50 Marino L (1997) The relationship between
gestation length, encephalisation and body weight
in odontocetes. Marine Mammal Sci 14:143–148.
51 Kudo H, Lowen S, Dunbar RIM (n.d.) Neocortex
size and social network size in primates.
Behaviour, submitted for publication.
52 Sugawara K (1984) Spatial proximity and
bodily contact among the central Kalahari San.
Afr Studies Monogr 3.
53 Dunbar RIM, Spoor M (1995) Social networks,
support cliques and kinship. Hum Nature
6:273–290.
54 Dunbar RIM (1988) Primate Social Systems.
London: Chapman & Hall.
55 Dunbar RIM (1989) Social systems as optimal
strategy sets. In Standen V, Foley R (eds), Comparative
Socioecology, pp 141–149. Oxford: Blackwell
Scientific.
ARTICLES Evolutionary Anthropology 189
O
nly
for educational purposes
SR
T
II
56 Dunbar RIM (1993) Coevolution of neocortical
size, group size and language in humans.
Behav Brain Sci 11:681–735.
57 Cheney DL, Seyfarth RM (1990) How Monkeys
See the World. Chicago: Chicago University Press.
58 Povinelli DJ, Nelson KE (1990) Inferences
about guessing and knowing in chimpanzees. J
Comp Psychol 104:203–210.
59 Povinelli DJ (1994) What chimpanzees (might)
know about the mind. In Wrangham R, McGrew
W, de Waal F, Heltne P (eds), Chimpanzee Cultures,
pp 285–300. Cambridge: Harvard University
Press.
60 O’Connell SM (1996) Theory of Mind in Chimpanzees.
Ph.D. thesis, University of Liverpool.
61 Perner J, Wimmer D (1985) ‘‘John thinks that
Mary thinks that. . . .’’ Attribution of secondorder
beliefs by 5 and 10 year-old children. J Exp
Child Psychol 39:437–471.
62 Kinderman P, Dunbar RIM, Bentall RP (n.d.)
Theory of mind deficits and causal attributions.
Br J Psychol, in press.
63 Desalles J-L (n.d.) Altruism, status and the
origin of relevance. In Hurford JR, StuddertKennedy
M, Knight C (eds), Evolution of Language.
Cambridge: Cambridge University Press.
64 Dunbar RIM (1997) Grooming, Gossip and the
Evolution of Language. Cambridge: Harvard University
Press.
65 Harvey PH, Krebs JR (1990) Comparing brains.
Science 249:150–156.
66 Stephan H (1972) Evolution of primate brains:
A comparative anatomical investigation. In Tuttle
R (ed), Functional and Evolutionary Biology of
Primates, pp 165–174. Chicago: Aldine-Atherton.
67 Killworth PD, Bernard HR, McCarty C (1984)
Measuring patterns of acquaintanceship. Curr
Anthropol 25:391–397.
68 Mange A, Mange E (1980) Genetics: Human
Aspects. W.B. Saunders.
69 Bryant FC (1981) We’re All Kin: A Cultural
Study of a Mountain Neighbourhood. Knoxville:
University of Tennessee Press.
70 Gould SJ (1975) Allometry in primates, with
emphasis on scaling and the evolution of the
brain. Contrib Primatol 5:244–292.
71 Jolly A (1969) Lemur social behaviour and
primate intelligence. Science 163:501–506.
72 Humphrey NK (1976) The social function of
intellect. In Bateson PPG, Hinde RA (eds) Growing
Points in Ethology, pp 303–317. Cambridge:
Cambridge University Press.
73 Armstrong E (1985) Relative brain size in
monkeys and prosimians. Am J Phys Anthropol
66:263–273.
௠ 1998 Wiley-Liss, Inc.
190 Evolutionary Anthropology ARTICLES
O
nly
for educational purposes
SR
T
II