The Social Brain Hypothesis
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
178 Evolutionary Anthropology
ARTICLES
The Social Brain Hypothesis
Robin I.M. Dunbar
Conventional wisdom over the past 160 years in the cognitive and neurosciences criminate unequivocally between
has assumed that brains evolved to process factual information about the world. them. In the present case, we can do
Most attention has therefore been focused on such features as pattern recognition, this by asking which hypothesis best
color vision, and speech perception. By extension, it was assumed that brains predicts the differences in brain size
evolved to deal with essentially ecological problem-solving tasks.1 across the primate order. To do so, we
need to identify the specific quantita-
tive predictions made by each hypoth-
The concensus view has tradition- tational demands of the complex so- esis and to determine an appropriate
ally been that brains evolved to pro- cial systems that characterize the measure of brain size.
cess information of ecological rel- order.5,6 Prima facie, this suggestion The four classes of hypotheses that
evance. This view, however, ignores an seems plausible: There is ample evi- have been put forward to explain pri-
important consideration: Brains are dence that primate social systems are mate brain evolution are epiphenom-
exceedingly expensive both to evolve more complex than those of other enal, developmental, ecological, and
and to maintain. The adult human species. These systems can be shown social in orientation (Table 1). The
brain weighs about 2% of body weight to involve processes such as tactical epiphenomenal and developmental hy-
but consumes about 20% of total en- deception5 and coalition-formation,7,8 potheses share the assumption that
ergy intake.2 In the light of this, it is which are rare or occur only in sim- evolution of the brain (or brain part) is
pler forms in other taxonomic groups. not a consequence of external selec-
difficult to justify the claim that pri-
Because of this, the suggestion was tion pressures but rather simply a
mates, and especially humans, need
rapidly dubbed the Machiavellian intel- consequence of something to do with
larger brains than other species merely
ligence hypothesis, although there is a the way biological growth processes
to do the same ecological job. Claims
growing preference to call it the social are organized. The epiphenomenal hy-
that primate ecological strategies in-
brain hypothesis.9,10 potheses thus argue that brain evolu-
volve more complex problem-solv-
Plausible as it seems, the social brain tion is a mere byproduct of body size
ing3,4 are plausible when applied to
hypothesis faced a problem that was evolution, and that brain part size is,
the behaviors of particular species, in turn, a byproduct of total brain
recognized at an early date. Specifi-
such as termite-extraction by chimpan- evolution.11
cally, what quantitative empirical evi-
zees and nut-cracking by Cebus mon- The developmental versions differ
dence there was tended to favor one or
keys, but fail to explain why all only in that they provide a more spe-
the other of the ecological hypoth-
primates, including those that are con- eses,1 whereas the evidence adduced cific mechanism by presuming that
ventional folivores, require larger brains in favor of the social brain hypothesis maternal metabolic input is the criti-
than those of all other mammals. was, at best, anecdotal.6 In this article, cal factor influencing brain develop-
An alternative hypothesis offered I shall first show how we can test ment. This claim is given credibility by
during the late 1980s was that pri- between the competing hypotheses the fact that the bulk of brain growth
mates’ large brains reflect the compu- more conclusively and then consider in mammals occurs prenatally. In-
some of the implications of the social deed, it appears to be the completion
brain hypothesis for humans. Finally, of brain development that precipitates
Robin Dunbar is Professor of Evolutionary I shall briefly consider some of the birth in mammals, with what little
Psychology and Behavioural Ecology at underlying cognitive mechanisms that postnatal brain growth occurs being
the University of Liverpool, England. His might be involved. completed by the time an infant is
research primarily focuses on the behav-
ioral ecology of ungulates and human and weaned. From this, the conclusion is
nonhuman primates, and on the cognitive drawn that brain evolution must be
mechanisms and brain components that TESTING BETWEEN ALTERNATIVE
constrained by the spare energy, over
underpin the decisions that animals make. HYPOTHESES and above her basal metabolic require-
He runs a large research group, with gradu-
ate students working on many different To test between the competing hy- ments, that the mother has to channel
species on four continents.
potheses, we need to force the hypoth- into fetal development.12–14 Some evi-
eses into conflict in such a way that dence in support of this claim comes
Key Words: brain size, neocortex, social brain their predictions are mutually contra- from the fact that frugivorous pri-
hypothesis, social skills, mind reading, primates dictory. This allows the data to dis- mates have larger adult brains relativeARTICLES Evolutionary Anthropology 179
TABLE 1. Hypotheses Used to Explain the the energetic arguments do not add This leaves us with just two classes
Evolution of Large Brains in Primates up: Precocial mammals do not have of hypotheses, the ecological and the
Hypothesis Sources higher metabolic rates than do altri- social. At least three versions of the
cial mammals despite the fact that former can be identified, which I will
A. Epiphenomenal
they have neonatal brain sizes that term the dietary, mental maps, and
hypotheses
1. Large brains (or brain parts) are, on average, twice as large. We extractive foraging hypotheses. In es-
are an unavoidable conse- therefore do not need to consider ei- sence, these argue, respectively, that
quence of having a large ther epiphenomenal or developmental primate species will need larger brains
body (or brain) 11, 70 hypotheses any further in the context if (i) they are frugivorous because fruits
B. Ecological hypotheses of this article. are more ephemeral and patchy in
2. Frugivory imposes higher This does not necessarily mean that their distribution than leaves are, and
cognitive demands than these explanations are wrong. Both hence require more memory to find;
folivory does 1, 65 kinds of explanation may be true in (ii) they have larger ranges because of
3. Brain size constrains the size
the sense that they correctly identify the greater memory requirements of
of the mental map: 1
developmental constraints on brain large-scale mental maps; or (iii) their
(a) constraint on size of
home range growth, but they do not tell us why diet requires them to extract resources
(b) constraint on inertial brains actually evolved as they did. from a matrix in which they are em-
navigation (day journey They may tell us that if you want to bedded (e.g., they must remove fruit
length) evolve a large brain, then you must pulp from a case, stimulate gum flow
4. Extractive foraging evolve a large body in order to carry from a tree, extract termites from a
hypothesis 3, 4 the energetic costs of doing so or a diet termitarium, or hunt species that are
C. Social hypotheses
cryptic or behave evasively).
5. Brain size constrains size of 6, 71,
For obvious reasons, I used the per-
social network (group size): 72
(a) constraint on memory Because the cost of centage of fruit in the diet as an appro-
for relationships priate index for the dietary hypothesis.
(b) constraint on social skills
maintaining a large brain I used the size of the range area and
to manage relationships is so great, it is the length of the day journey as alterna-
D. Developmental tive indices for the mental mapping
hypotheses intrinsically unlikely that hypothesis, though I present the data
6. Maternal energy con- large brains will evolve only for the first of these here. The first
straints determine energy 12, 13
index corresponds to the case in which
capacity for fetal brain 46, 55, merely because they animals have to be able to manipulate
growth 73
can. Large brains will information about the locations of re-
evolve only when the sources relative to themselves in a
Euclidean space (an example would
to body size than do folivorous pri-
selection factor in their be the nut-cracking activities of the
mates.1 This has been interpreted as favor is sufficient to Taı̈ chimpanzees16); the second corre-
implying that frugivores have a richer sponds to the possibility that the con-
diet than folivores do and thus have
overcome the steep cost straint lies in the needs of some aspect
more spare energy to divert into fetal gradient. of inertial navigation. The extractive
growth. Large brains are thus seen as foraging hypothesis is less easy to char-
a kind of emergent epigenetic effect of acterize in quantitative terms because
spare capacity in the system. there is no objective measure of the
Both kinds of explanations suffer that ensures sufficient energy to pro- degree to which diets vary in their
from the problem that they ignore a vide for fetal brain development. No extractiveness. However, Gibson4 pro-
fundamental principle of evolutionary such allometric argument can ever vided a classification of primate spe-
theory, which is that evolution is the imply that you have to evolve a large cies into four categories of diet that
outcome of the balance between costs brain or a large body. The large brain differ in their degree of extractiveness.
and benefits. Because the cost of main- or brain part is a cost that animals We can test this hypothesis by asking
taining a large brain is so great, it is must factor into their calculations whether there is a consistent variation
intrinsically unlikely that large brains when considering whether or not a in brain size among these four catego-
will evolve merely because they can. large body or a large brain is a sensible ries, with the species having the more
Large brains will evolve only when the solution to a particular ecological extractive diets having larger brains
selection factor in their favor is suffi- problem. Shifts to more energy-rich or than those with the less extractive
cient to overcome the steep cost gradi- more easily processed diets may be diets.
ent. Developmental constraints are un- essential precursors of significant in- Finally, we need an index of social
doubtedly important, but rather than creases in brain or brain part size.2 complexity. In my original analyses, I
being causal their role is that of a This would explain why frugivores used social group size as a simple
constraint that must be overcome if have larger brains than folivores do measure of social complexity. Although
larger brains are to evolve. In addition, and why hominids have larger brains at best rather crude, this measure
Pagel and Harvey15 have shown that than great apes do. nonetheless captures one aspect of the180 Evolutionary Anthropology ARTICLES
the brain such as the medulla, the
neocortex shows dramatic and increas-
ing expansion across the range of pri-
mates (Fig. 1). The neocortex is ap-
proximately the same size as the
medulla in insectivores; however, it is
about 10 times larger than the me-
dulla 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 look-
ing 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 neo-
Figure 1. Neocortex volume as a ratio of medulla volume in different groups of primates (after
Passingham19). Source: Stephan et al29
complexity of social groups, the fact showed that neocortex size is an expo-
. . . brain evolution has
that information-processing demands nential function of brain size, whereas not been a history of
can be expected to increase as the other brain components are not. simple expansion in total
number of relationships involved in- Finlay and Darlington11 notwith-
creases. More importantly, perhaps, standing, there is evidence that brain volume. Rather, brain
this measure has the distinct merit of evolution has not been a history of evolution has been
being easily quantified and widely simple expansion in total volume.
available. Although it is possible to Rather, brain evolution has been mo- mosaic in character,
conceive of a number of better mea- saic in character, with both the rate with both the rate and
sures of social complexity, the appro- and the extent of evolution having
priate data are rarely available for varied between components of the sys- the extent of evolution
more than one or two species. tem. MacLean17 pointed out many having varied between
The second problem concerns the years ago that primate brain evolution
most appropriate measure of brain can be viewed in terms of three major
components of the
evolution. Hitherto, most studies have systems (his concept of the triune system.
considered the brain as a single func- brain). These systems correspond to
tional unit. This view has been rein- the basic reptilian brain (hind- and
forced by Finlay and Darlington,11 who midbrain systems), the mammalian
argued that the evolution in brain part brain (palaeocortex, subcortical sys- cortex is generally regarded as being
size closely correlates with the evolu- tems), and the primate brain (broadly, the seat of those cognitive processes
tion of total brain size and can be the neocortex). A more important that we associate with reasoning and
explained simply in terms of allome- point, perhaps, is that variations can consciousness, and therefore may be
tric consequences of increases in total be found within these broad catego- expected to be under the most intense
brain size. However, Finlay and Dar- ries in the rates at which different selection from the need to increase or
lington failed to consider the possibil- components expanded, which, in at improve the effectiveness of these pro-
ity that changes in brain size might least some cases, have been shown to cesses.
actually be driven by changes in its correlate with ecological factors.10 Par- One additional problem needs to be
parts rather than in the whole brain. tialling out the effects of body size on resolved. In his seminal study of brain
This is especially true of the neocor- the size of brain components suggests evolution, Jerison20 argued that brain
tex, for its volume accounts for 50% to that the story may be more complex size can be expected to vary with body
80% of total brain volume in primates. than Finlay and Darlington11 sup- size for no other reason than funda-
Thus, changes in the volume of the posed, with some remodeling of brain mental allometric relationships associ-
neocortex inevitably have a large di- growth patterns occurring in the tran- ated with the need to manage the
rect effect on apparent change in brain sitions between insectivores, prosim- physiological machinery of the body.
volume that may be quite unrelated to ians, and anthropoids.18 What is of interest, he suggested, is
changes in other brain components. The important point in the present not absolute brain size, but the spare
This point is given weight by the fact context is that, as Passingham19 noted, brain capacity over and above that
that Finlay and Darlington themselves relative to the more primitive parts of needed to manage body mechanisms.ARTICLES Evolutionary Anthropology 181
For this reason, Jerison derived his thing, absolute neocortex size, mainly with this problem include plotting
encephalization quotient. All subse- because the neocortex is such a large means for higher taxonomic units, per-
quent studies have used body size as component of the primate brain. In- forming nested analyses of variance
the appropriate baseline against which deed, the use of absolute neocortex using phylogenetic levels as factors,
to measure relative deviations in brain size produces results that are similar
comparing matched pairs of species,
size. However, a problem has since to those obtained from relativized indi-
ces of neocortex volume.24,25 This and making independent contrasts that
emerged: Brain size is determined
early in development and, compared makes some sense in computational control directly for phylogeny. Each
to many other body systems, appears terms: As Byrne26 has pointed out, a method has its own advantages and
to be highly conservative in evolution- 10% increase in the processing capac- disadvantages, but the first and third
ary terms. As a result, body size can ity of a small computer is worth a procedures are particularly associated
often change dramatically both ontoge- great deal less in information-process- with loss of information and small
netically across populations in re- ing terms than is a 10% increase in a sample sizes. I shall use the first and
sponse to local environmental condi- large computer. Although residuals
last method, the last because it allows
tions21 and phylogenetically22,23 from a common regression line would
conventionally be considered the saf- individual species to be compared, but
without corresponding changes in
brain size. This is particularly con- est measure, and have been used in the first because it allows grade shifts
spicuous in the case of phyletic dwarfs many recent analyses,27,28 I shall con- within data sets to be identified (a
(e.g., callitrichids and perhaps mod- problem that independent-contrasts
ern humans and hylobatids22) and spe- methods have difficulty dealing with).
cies in which body size may have I shall take the genus as a suitable
increased in response to predation
The neocortex is
basis for analysis because genera typi-
pressure following the occupation of generally regarded as cally represent different reproductive
more open terrestrial habitats (e.g.,
papionids24).
being the seat of those or ecological radiations and thus are
cognitive processes that more likely to constitute independent
The lability of body size therefore
makes it a poor baseline, though one evolutionary events.
we associate with The resulting analyses are relatively
that probably is adequate for analyses
on the mouse-elephant scale. Conse- reasoning and straightforward: Figure 2 presents the
quently, it is necessary to find an inter- consciousness, and data for neocortex ratio for the anthro-
nally more consistent baseline for taxo- poid primate species in the data base
nomically fine-grained analyses. therefore may be of Stephan, Frahm, and Baron.29 Neo-
Willner22 suggested that either molar expected to be under cortex size, however measured, does
tooth size or brain size may be suit- not correlate with any index of the
able because both are developmen-
the most intense ecological hypotheses, but does corre-
tally conservative. Because we are con- selection from the need late with social group size. Similar
cerned with brain part size, some findings were reported by Sawaguchi
to increase or improve
aspect of brain size seems the most and Kudo,30 who found that neocortex
appropriate. the effectiveness of size correlated with mating system in
At this point, three options are avail- these processes. primates. Barton and Purvis31 have
able. One is to compare the neocortex, confirmed that using both residuals of
the brain part of interest, with the neocortex volume on total brain vol-
whole brain; the second is to use the ume and the method of independent
rest of the brain other than the part of tinue to use my original ratio index contrasts yields the same result. Both
interest; the third is to use some less because it provides the best predictor Barton10 and T. Joffe (unpublished)
variable primitive component of the (see Box 1). have repeated the analyses using the
brain, such as the medulla, as a base- Finally, it is now widely appreciated medulla as the baseline for compari-
line. Two options are in turn available that comparative analyses need to con- son. More importantly, Barton and
as mechanisms for controlling for trol for the effects of phylogenetic Purvis31 have shown that while rela-
brain size in each of these cases. One inertia. Closely related species can be
is to use residuals from a common tive neocortex volume correlates with
expected to have similar values for group size but not the size of the
regression line against the baseline
many anatomical and behavioral di- ranging area, the reverse is true of
(e.g., the residual of neocortex volume
mensions merely by virtue of having relative hippocampus size. A correla-
on total brain volume or medula vol-
ume). The other choice is to use ratios. inherited them from a recent common tion between range area and hippo-
We have considered and tested all ancestor. In such cases, plotting raw campus size is to be expected because
these options10,24 (see Box 1). The re- data would result in pseudoreplica- of hippocampal involvement in spatial
sults are virtually identical irrespec- tion, artificially inflating the sample memory.32,33 This correlation demon-
tive of which measure is used. One size by assuming that closely related strates that it is not simply total brain
explanation for this may be that all species are actually independent evolu- size that is important (a potential prob-
these measures actually index the same tionary events. The ways of dealing lem, given the overwhelming size of182 Evolutionary Anthropology ARTICLES
Box 1. How to Measure Brains
R. Dunbar and Tracey H. Joffe on first sight. Moreover, any index that dently of all other confounding mea-
The different ways of measuring uses the whole brain as its base is sures, suggests that more detailed
relative brain size have raised doubts likely to suffer from autocorrelation consideration needs to be given to its
as to the most appropriate technique effects. Because the neocortex is such significance and meaning. It may be,
to use.65 Many researchers have pre- a large proportion of the brain in pri- for example, that body size, rather
ferred to use residuals from the com- mates, residuals of neocortex from than being a determinant20 is simply a
mon regression line of best fit for the total brain size may simply be a mea- constraint on neocortex size: A spe-
data set concerned. This provides a sure of neocortex plotted against it- cies can evolve a large neocortex only
measure of the extent to which brain self. if its body is large enough to provide
(or brain part) volume deviates from To consider the problem in more the spare energy capacity through
what would be expected for an aver- detail, we ran a stepwise regression Kleiber’s relationship for basal meta-
age member of the relevant taxon of analysis on the 24 species of anthro- bolic rate to allow for a larger than
the appropriate size. Although ratios poid primates, including humans, on average brain. This interpretation is
have been used to compare the rela- the data base of Stephan, Frahm, and implied by the Aiello and Wheeler2
tive size of brain components,29,66 this Baron,29 with group size as the depen- ‘‘expensive tissue hypothesis.’’ It would
has been criticized on the grounds dent variable and nine indices of rela- also be in line with Finlay and Darling-
that trade-offs within the brain may tive brain or brain-part volume as inde- ton’s11 claim that in mammals the evo-
mean that a given index simply mea- pendent variables. In addition to lution of brain-part size is driven, devel-
sures total brain size (or the size of a neocortex ratio, these included total opmentally at least, by the evolution of
brain part) and thus does not remove brain volume as well as telencephalon the whole brain, thus generating very
the effects of absolute size. Ratios and neocortex volume, each taken as tight correlations between brain-part
may also be prone to autocorrelation absolute volume and as a residual size and total brain size.
effects, especially when the baseline from both body mass and brain vol-
is taken to be the whole brain and, as TABLE 2. Stepwise Regression Analysis
ume. All variables were log10-trans- of Indices of Brain Component Volume
in the case of the neocortex, the part in
formed for analysis. In both cases,
question is a major volumetric compo- as Predictors of Group Size in
neocortex ratio was selected as the
nent of the brain. Anthropoid Primates, Based on
variable of first choice. We carried out
Although there are likely to be Independent Contrasts Analysis*
both regressions on generic plots and
some trade-offs of this kind within the Independent Variable t P
independent contrast analyses. For
brain, the fact that neocortex volume
the contrasts analysis, the best fit Absolute brain volume 21.69 0.107
increases progressively across the pri-
least-squares regression equation Residual of brain volume
mate order suggests that such con- on body mass 20.91 0.371
through the origin was:
straints are less likely to have a signifi- Absolute telencephalon
cant effect on a ratio measure. Of Contrast in log10 (group size) 5 3.834 volume 21.72 0.101
course the residuals procedure is itself Residual of telen-
* Contrast in log10 (neocortex ratio)
a ratio: Encephalization-type indices cephalon volume on
are calculated as actual volume di- (r 2 5 0.395, F1,22 5 15.39, body mass 0.61 0.546
vided by predicted volume (which, Residual of telen-
P 5 0.001).
when data are logged, becomes the cephalon volume on
conventional actual minus predicted With all other variables held constant, brain volume 1.69 0.107
none of the other eight indices made a Absolute neocortex
values). Thus, ratio measures per se
volume 21.70 0.104
may not be the problem. Rather, the significant contribution to the variabil-
Residual of neocortex
substantive objection is whether or not ity in group size in either analysis. volume on body mass 0.56 0.583
a ratio partials out the allometric ef- Table 2 gives the results for the inde- Residual of neocortex
fects of body size. In fact, it seems that pendent contrasts analysis. volume on brain
neocortex ratios are not correlated Neocortex ratio is thus the single volume 1.60 0.136
with the basal brain (i.e., brain volume most powerful predictor of group size Neocortex ratio (against
excluding the neocortex) within major in these species. While the biological rest of brain) 3.79 0.001
taxonomic groups (unpublished analy- significance of this variable remains *Sample: 24 species of anthropoid pri-
ses). Consequently, this criticism has open to interpretation, the fact that it mates from Stephan, Frahm, and
less force than it might appear to have provides the best predictor, indepen- Baron.29
the primate neocortex). Moreover, it The validity of this relationship species for which brain volumetric
points to the specific involvement of could be tested directly by using it to data were not available in the original
the neocortex. predict group sizes in a sample of sample of Stephan, Frahm, andARTICLES Evolutionary Anthropology 183
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.)
Baron.29 I did this by exploiting the timates of the sizes of social groups Surprisingly, the data for the carni-
fact that neocortex ratios can be pre- and, in the case of the semi-solitary vores map directly onto those for the
dicted from total brain volume,34 a species, daytime nest groups.36 Re- simian primates, that is, the regres-
result that, in fact, follows directly analysis of the data for prosimians sion lines for the two data sets do not
from the Finlay and Darlington11 find- using these improved estimates of so- differ significantly. However, the carni-
ings. The result was a significant fit cial group size suggests that these vores do not exhibit as wide a range of
between predicted neocortex ratio and species do in fact adhere to the same neocortex ratios or group sizes as do
observed mean group size for a sam- relationship between neocortex and anthropoid primates. The fact that the
ple of 15 New and Old World monkey group size as that which pertains for prosimians lie to the left of both these
species.35 other primates.25 More importantly, taxonomic groups implies that the car-
Barton27 noted that the original the regression line for this taxon is
nivores represent an independent evo-
analyses of Dunbar24 seemed to imply parallel to, but shifted to the left of,
lutionary development along the same
that variation in neocortex size was that for other anthropoid primates
principles as the anthropoid primates,
much greater than variation in group (Fig. 3).
the difference being that they just have
size in the prosimians. Using Dun- This relationship has now been
not taken it as far as primates have.
bar’s24 data on group size, Barton sug- shown to hold for at least four other
gested that the relationship between mammalian orders: bats,10 carnivores One reason for this may be that the
neocortex and group size did not ap- and insectivores,37,38 and odontocete carnivore social world is olfaction-
ply in the case of prosimians. How- cetaceans.39,40 In the case of the insec- dominated rather than vision-domi-
ever, the data on prosimian group tivores, the data points are shifted far nated, as in the case of the primates.
sizes in this sample suffered from a to the left of those for the primates, as Barton10,27,41 has pointed out that the
paucity of data, particularly for the might be expected of a taxonomic shift to a diurnal lifestyle based on
nocturnal species. Because many of group that is considered to be broadly color vision, perhaps initially diet-
these are described as semi-solitary, it representative of the ancestral mam- driven, but leading to a shift into vi-
was conservatively assumed in the mals.37 However, the relationship is sion-based communication, may be
Dunbar24 database that their group weak in this case, probably because the key feature that has spurred on the
size was one. More recent field studies estimates of group size are particu- dramatic development of the primate
have produced markedly improved es- larly uncertain for insectivores. neocortex.184 Evolutionary Anthropology ARTICLES
size of the cortical processing machin-
ery increases, at least relative to the
opportunity cost of taking cortical neu-
rons away from other cognitive pro-
cesses.
It seems equally unlikely that the
problem lies with a pure memory con-
straint, though memory capacity obvi-
ously 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: Hu-
mans are said to be able to attach
names to around 2,000 faces but have
a cognitive group size of only about
Figure 3. Mean group size plotted against neocortex ratio for individual genera, shown The social brain
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 hypothesis implies that
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, constraints on group size
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, Propithe-
arise from the
cus; e, Indri; f, Microcebus; g, Galago; h, Hapalemur; i, Avahi; j, Perodictus. information-processing
capacity of the primate
REFINING THE RELATIONSHIP initial kick for the evolution of large
brains in primates,10 it seems intrinsi-
brain, and that the
The social brain hypothesis implies
that constraints on group size arise
cally unlikely that the ultimate con- neocortex plays a major
from the information-processing ca- straint lies in the mechanisms of the role in this. However,
pacity of the primate brain, and that visual system itself.28 Although there
the neocortex plays a major role in is a correlation between the relative even this proposal is
this. However, even this proposal is size of the visual cortex and group size open to several
open to several interpretations as to in anthropoid primates, the fit is much
poorer, and the slope significantly shal- interpretations as to how
how the relationship is mediated. At
least five possibilities can be usefully lower than that between the nonvisual the relationship is
considered. The constraint on group neocortex and group size (r2 5 0.31 vs
r2 5 0.61, respectively) (Fig. 4). Partial
mediated.
size could be a result of the ability to
recognize and interpret visual signals correlation analysis indicates that only
for identifying either individuals or the correlation for the nonvisual rela-
their behavior; limitations on memory tionship remains significant when the 150. Second, there is no intrinsic rea-
for faces; the ability to remember who other component is held constant28 son to suppose that memory per se is
has a relationship with whom (e.g., all (though this is not true for prosim-
the issue. The social brain hypothesis
dyadic relationships within the group ians25). A more important point is that
is about the ability to manipulate infor-
as a whole); the ability to manipulate the volume of the lateral geniculate
mation, not simply to remember it.
information about a set of relation- nucleus, a major subcortical way sta-
Third, and perhaps most significantly,
ships; and the capacity to process emo- tion in visual processing, does not
memories appear to be stored mainly
tional information, particularly with correlate with group size at all, indicat-
ing that pattern recognition per se is in the temporal lobes,42 whereas re-
respect to recognizing and acting on
unlikely to be the issue.28 It may be of cent PET scan studies implicate the
cues to other animals’ emotional states.
These are not all necessarily mutually some significance that the absolute prefrontal neocortex, notably Brod-
exclusive, but they do identify differ- size of the visual cortex seems to reach man area 8, as the area for social skills
ent points in the cognitive mechanism an asymptotic value in the great ape and, specifically, theory of mind.43
that might be the crucial information- clade, whereas the nonvisual neocor- Frith44 has suggested that memories
processing bottleneck. tex continues to increase in size. One and representations for objects or
Although visual mechanisms are interpretation of this is that visual events may involve interactions be-
likely to be important for social inter- processing does not necessarily con- tween several levels of the neocortex
action, and may well have been the tinue to improve indefinitely as the depending on the kinds of operationsARTICLES Evolutionary Anthropology 185
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 dif-
ferential frequencies with which these
large-brained species have been stud-
ied are taken into account.
Third, Pawlowski, Dunbar, and
Lowen47 have shown that among po-
lygamous primates the male rank cor-
relation with mating success is nega-
tively 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 capaci-
ties to deploy more sophisticated so-
cial skills, such as the use of coalitions
and capitalizing on female mate
choice, to undermine or circumvent
the power-based strategies of the domi-
nant animals.
Figure 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 . . . there is no intrinsic
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
reason to suppose that
this. (Reprinted from Joffe and Dunbar,28 Fig. 1.) memory per se is the
issue. The social brain
involved. These interactions could oc- pret this in terms of a shift away from
cur between the sensory and associa- emotional control of behavior to more hypothesis is about the
tion cortices (perceiving an object), conscious, deliberate control. ability to manipulate
between the association and frontal The only remaining alternative is
cortices (remembering an object), and that the mechanisms involved lie in
information, not simply to
among all three (being aware of per- the ability to manipulate information remember it.
ceiving an object). It is worth noting in about social relationships themselves.
this context that although social skills This claim is supported by six addi-
are commonly disrupted by damage to tional lines of evidence that point to
the prefrontal cortex, memory for the fundamental importance of social The fourth line of evidence is Jof-
events and people is not.42 skills in the detailed management of fe’s48 demonstration that adult neocor-
It seems unlikely that emotional re- social relationships. tex size in primates correlates with the
sponses per se are the substantive One is the fact that close analysis of length of the juvenile period, but not
constraint. Although the correct emis- the data on group size and neocortex with the length of gestation, lactation,
sion and interpretation of emotional volume suggests that there are, in fact, or the reproductive life span, even
cues is of singular importance in the distinct grades even within the anthro- though total brain size in mammals
management of social relationships,45 poid primates (Fig. 3). Apes seem to lie correlates with the length of the gesta-
there is little evidence that the subcor- on a separate grade from the monkeys, tion period.49,50 This suggests that what
tical areas principally associated with which in turn lie on a separate grade is most important in the development
emotional cuing (for example, the from the prosimians. The slope coeffi- of a large neocortex in primates is not
amygdala in the limbic system) corre- cients on these separate regression the embryological development of
late in any way with social group lines do not differ significantly, but the brain tissue per se, which is associated
size.28 Indeed, Keverne, Martel, and intercepts do. It is as if apes require mainly with gestation length, but
Nevison46 point out that there has more computing power to manage the rather the ‘‘software programming’’
been progressive reduction in the rela- same number of relationships that that occurs during the period of social
tive sizes of the ‘‘emotional’’ centers in monkeys do, and monkeys in turn learning between weaning and adult-
the brain (the hypothalamus and sep- require more than prosimians do. This hood.
tum) in favor of the ‘‘executive’’ cen- gradation corresponds closely to the Fifth, Kudo, Lowen, and Dunbar51
ters (the neocortex and striate cortex) perceived scaling of social complexity. have shown that grooming clique size,
during primate evolution. They inter- The second line of evidence is that a surrogate variable that indexes alli-186 Evolutionary Anthropology ARTICLES
gives primate social groups their inter-
nal structure and coherence, this can
be seen as a crucial basis for primate
sociality.
Finally, Keverne, Martel, and Nevi-
son46 have suggested that the neocor-
tex and striate cortex, those areas of
the primate brain that are responsible
for executive function, are under ma-
ternally rather than paternally im-
printed 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 emo-
tional 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
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 The fact that the relationship be-
to 8, and 9 to 30 males) for individual species. The regression equations for the two cohort sizes tween neocortex size and what I will
are significantly different from b 5 0. The species sampled are C. apella, P. entellus, C. aethiops, term the cognitive group size holds up
M. fuscata, M. mulatta, M. radiata, M. arctoides, P. cynocephalus, P. anubis, P. ursinus, and P. so well in so many different taxonomic
troglodytes. (Redrawn from Pawlowski, Dunbar, and Lowen,47 Fig. 1.)
groups raises the obvious question of
whether or not it also applies to hu-
mans. We can easily predict a value for
ance size, correlates rather tightly with ficient depth that they can be relied on group size in humans. Doing so, which
relative neocortex and social group to provide unstinting mutual support is simply a matter of using the human
size in primates, including humans when one of them is under attack. neocortex volume to extrapolate a
(Fig. 6). The human data derive from Because this is the core process that value for group size from the primate
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 invari-
ably function as coalitions in primate
groups. Coalitions are functionally cru-
cial 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 oth-
ers.54 Coalitions essentially allow pri-
mates 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 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, Propithe-
be interpreted as a direct cognitive
cus, Indri, S. sciureus, C. apella, C. torquatus, A. geoffroyi, A. fusciceps, P. badius, P. entellus, P.
limitation on the number of individu- pileata, P. johnii, C. campbelli, C. diana, C. aethiops, C. mitis, E. patas, M. mulatta, M. fuscata,
als with which an animal can simulta- M. arctoides, M. sylvana, M. radiata, P. anubis, P. ursinus, P. cynocephalus, P. hamadryas, T.
neously maintain a relationship of suf- gelada, P. troglodytes, P. paniscus. (Redrawn from Kudo, Lowen, and Dunbar,51 Fig. 4a.)ARTICLES Evolutionary Anthropology 187
such societies are limited, those that
are available suggest that there is in-
deed a consistent group size in the
region of 150 individuals (Fig. 7). Ex-
cept among settled horticulturalists,
where the village seems to be the
relevant unit, this typically involves
the set of individuals from whom over-
night 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 invari-
ably 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 relation-
ship to each other. Indeed, this genu-
Figure 7. Mean sizes for different types of groups in traditional human societies. Individual inely seems to be the largest group of
societies are ordered along the bottom, with data for three main types of social groups people who know everyone in the
(overnight camps, clans or villages, and tribes). Societies include hunter-gatherer and settled group as individuals at the level of
horticulturalists from Australia, Africa, Asia, and North and South America. The triangles give
personal relationships. This is essen-
mean group sizes for three contemporary United States samples: mean network size from
small-worlds experiments (N 5 2),67 mean Hutterite community size,68 and the size of an East tially the definition that holds in the
Tennessee mountain community.69 The value of 150 predicted by the primate neocortex size case of primates.
relationship (from Fig. 1d) is indicated by the horizontal line, with 95% confidence intervals A more extensive exploration of hu-
shown as dashed lines. man groups in other contexts suggests
that groupings of this size are wide-
spread and form an important compo-
nent of all human social systems, be-
equation, produces a value in the or- network, or grooming clique) was ing present in structures that range
der of 150. The real issue is whether taken to be the grouping criterion. from business organizations to the
humans really do go around in groups The problem with respect to hu- arrangement of farming communi-
of this size. mans is that it is difficult to identify ties.56 Estimates of community size
Identifying the relevant level of which of the many potential grouping for two traditional farming communi-
grouping to measure in humans is levels is functionally or cognitively ties in the United States, Hutterites
difficult because most humans live in equivalent to the particular level of and an East Tennessee mountain com-
a series of hierarchically inclusive grouping that I happened to use for munity, and of actual social network
groups. This, in itself, is not especially primates. This difficulty is particularly sizes (from small-worlds experiments)
unusual: Hierarchically structured intrusive in this case because humans (shown as triangles on the right side of
groups of this kind are characteristic live in a dispersed social system some- Fig. 7) fit very closely within the rel-
of primates54 and may be typical of times referred to as a fission-fusion evant range of group sizes.
many mammals and birds.55 At least system. In order to get around this It is easy, of course, to play the
in the case of the diurnal primates, it problem, I adopted the converse strat- numerologist in this context by find-
seems that, with a few notable excep- egy in my original analysis,56 asking ing groups that fit whatever group size
tions, the various species’ grouping whether there was any group size con- one wishes to promote. The important
patterns exhibit an overt level of stabil- sistently characteristic of humans that feature to note here, however, is that
ity at roughly the same position in the was of about the requisite size and, if the various human groups that can be
hierarchy across a wide range of taxa. so, whether its intrinsic psychological identified in any society seem to clus-
Moreover, because the various layers characteristics were similar to those ter rather tightly around a series of
of this hierarchy appear to be inti- found in primate groups. values (5, 12, 35, 150, 500, and 2,000)
mately related to each other, probably Because of the structural complex- with virtually no overlap in the vari-
through being part of a series of cause- ity of postagricultural societies, I con- ance around these characteristic val-
and-consequence chains,51 it would sidered only traditional hunter-gath- ues. They seem to represent points of
not matter which particular grouping erer and small-scale horticultural stability or clustering in the degrees of
level (for example, stable social group, societies. Although census data on familiarity within the broad range of188 Evolutionary Anthropology ARTICLES
Box 2. A Beginner’s Guide to Intensionality
Computers can be said to know mind, at least in some crude sense: wants Jake to suppose that Amelia
things because their memories con- they know that they know. Organisms intends to do something’’).
tain information; however, it seems of this kind are first-order intensional. A minimum of fourth-order inten-
unlikely that they know that they know By extension, second-order inten- sionality is required for literature that
these things, in that we have no evi- sional organisms know that someone goes beyond the merely narrative (‘‘the
dence that they can reflect on their else knows something, and third-order writer wants the reader to believe that
states of ‘‘mind.’’ In the jargon of the intensional organisms know that some- character A thinks that character B
philosophy of mind, computers are one else knows that someone else intends to do something’’). Similar abili-
zero-order intensional machines. In- knows something. In principle, the se- ties may be required for science, since
tensionality (with an 2s) is the term quence can be extended reflexively doing science requires us to ask
that philosophers of mind use to refer indefinitely, although, in practice, hu-
whether the world can be other than it
to the state of having a state of mind mans rarely engage in more than
is (a second-order problem at the very
(knowing, believing, thinking, wanting, fourth-order intensionality in everyday
least) and then ask someone else to
understanding, intending, etc). life and probably face an upper limit at
do the same (an additional order of
Most vertebrates are probably ca- sixth-order (‘‘Peter knows that Jane
pable of reflecting on their states of believes that Mark thinks that Paula intensionality).
human relationships, from the most graphically by Cheney and Seyfarth’s57 stand on the threshold of acquiring
intimate to the most tenuous. observation that apes seem to be good ToM. More importantly, chimpanzees
psychologists in that they are good at do better than autistic adults, one of
COGNITIVE MECHANISMS reading minds, whereas monkeys are whose defining features is the lack of
good ethologists in that they are good ToM, on the same tests.
The suggestion that the mecha- at reading behavior—or at least at That mind-reading, the basis of ToM,
nisms involved in these processes may making inferences about intentions in is difficult to do has been shown by
be concerned with social skills raises the everyday sense, even if not in the experiments on normal adults tested
the issue alluded to by the original
on ‘‘advanced’’ ToM tasks, up to fifth-
Machiavellian intelligence hypothesis,
order intentionality.62 These data sug-
namely to what extent cognitively so-
gest that normal humans find tasks of
phisticated mechanisms conferring the . . . apes seem to be greater than fourth-order intentional-
ability to ‘‘mind-read’’ might be in-
good psychologists in ity exceedingly hard to do. The high
volved. Tactical deception, in its strong
error rates at these levels do not reflect
sense, implies the ability to hold false that they are good at a memory retention problem: All sub-
beliefs and, thus, the presence of the
ability known as ‘‘theory of mind’’ reading minds, whereas jects pass the tests that assess memory
for the story line. Moreover, the same
(ToM). Of course, tactical deception as monkeys are good subjects show considerable compe-
practiced by primates on a daily basis
may not, as Byrne26 himself has ethologists in that they tence on reasoning tasks that involve
pointed out, be quite as sophisticated are good at reading causal chains of up to the sixth order.
as first impressions suggest. A more The difficulty seems genuinely to be
conventional behaviorist account
behavior . . . something to do with operating with
based on simple associative learning deeply embedded mental states.
can invariably be given for almost all One possibly significant observation
examples reported in the literature. in this context is that the visual and
philosophical sense of belief states. nonvisual components of the primate
Nonetheless, convincing evidence
Evidence that chimpanzees aspire neocortex do not increase isometri-
suggests that humans at least do use
to at least a basic form of ToM is cally. Although initially there is a more
ToM in executing some of their more
provided by their performance on ex- or less linear increase in the visual
manipulative social activities. And
while we may not wish to attribute full perimental false-belief tasks.58–61 These area V1 with increasing size of the rest
ToM to all primates, at least circum- studies have attempted to develop ana- of the neocortex, this drops off within
stantial evidence suggests that basic logues of the classic false-belief tasks the great ape clade. From gorillas
ToM is present in great apes and that used with children.62 Though it is clear through humans, increases in the size
monkeys may aspire to a level that that chimpanzees do not perform to of the visual area progress more slowly
Byrne26 has described as level 1.5 in- the level at which fully competent than do increases in the size of the rest
tentionality (full ToM being level 2 children perform, O’Connell’s61 experi- of the neocortex.28 We interpret this as
intentionality) (see Box 2). The differ- ments at least suggest that they can implying that beyond a certain point
ence has been summed up rather perform at the level of children who the acuity of the visual system doesARTICLES Evolutionary Anthropology 189
not increase linearly with size. Be- 8 Harcourt AH (1989) Sociality and competition 32. O’Keefe J, Nadel L (1978) The Hippocampus
in primates and non-primates. In Standen V, as a Cognitive Map. Oxford: Oxford University
cause the total size of the neocortex is Foley R (eds), Comparative Socioecology, pp 223– Press.
limited by embryological and ener- 242. Oxford: Blackwell Scientific. 33. Krebs JH, Sherry DF, Healy SD, Perry VH,
getic factors, this means that dispro- 9 Brothers L (1990) The social brain: A project Vaccarino AL (1989) Hippocampal specialisation
portionately more capacity can be for integrating primate behaviour and neurophysi- of food-storing birds. Proc Natl Acad Sci USA
ology in a new domain. Concepts Neurosci 1:27– 86:1488–1492.
dedicated to nonvisual areas of the 251. 34. Aiello LC, Dunbar RIM (1993) Neocortex
neocortex once the volume is above 10 Barton RA, Dunbar RLM (1997) Evolution of size, group size and the evolution of language.
the crucial threshold. This might ex- the social brain. In Whiten A, Byrne R (eds), Curr Anthropol 34:184–193.
plain why apes appear to be capable of Machiavellian Intelligence, Vol. II. Cambridge: 35 Dunbar RIM (1995) Neocortex size and group
Cambridge University Press. size in primates: A test of the hypothesis. J Hum
the additional cognitive processing as- Evol 28:287–296.
11 Finlay BL, Darlington RB (1995) Linked regu-
sociated with mind reading, whereas larities in the development and evolution of mam- 36. Bearder SK (1987) Lorises, bushbabies and
monkeys are not. It might also explain malian brains. Science 268:1678– 1684. tarsiers: Diverse societies in solitary foragers. In
Smuts B, Cheney D, Seyfarth R, Wrangham R,
why humans are better at it than apes. 12 Martin RD (1981) Relative brain size and
Struhsaker T (eds), Primate Societies, pp 12–24.
metabolic rate in terrestrial vertebrates. Nature
For humans, one important aspect Chicago: Chicago University Press.
Lond 293:57–60.
of ToM concerns its relevance to lan- 13 Martin RD (1983) Human brain evolution in
37 Dunbar RIM, Bever J (n.d.) Neocortex size
guage, a communication medium that determines group size in carnivores and insecti-
an ecological context. 52nd James Arthur Lec-
vores. Ethology, in press.
crucially depends on understanding ture, American Museum of Natural History, New
38 Gittleman JH (1986) Carnivore brain size,
York.
interlocutors’ mental states or inten- behavioural ecology and phylogeny. J Mammal
14 Hofman MA (1983) Evolution of the brain in
tions. The kinds of metaphorical uses neonatal and adult placental mammals: A theo-
67:23–36.
of language that characterize not only retical approach. J Theoret Biol 105:317–322. 39 Marino L (1996) What can dolphins tell us
about primate evolution? Evol Anthropol 5:81–86.
our rather telegraphic everyday ex- 15 Pagel M, Harvey PH (1988) How mammals
produce large-brained offspring. Evolution 42: 40 Tschudin A (1996) The Use of Neuroimaging
changes (in which ‘‘you know what I in the Assessment of Brain Size and Structure in
948–957.
mean?’’ is a common terminal clause) Odontocetes. MSc thesis, University of Natal
16 Boesch C, Boesch H (1984) Mental map in (Durban).
but also lies at the very heart of the chimpanzees. Primates 25:110–170.
41 Barton RA, Purvis A, Harvey PH (1995) Evolu-
metaphorical features of language. As 17 MacLean PD (1982) On the origin and progres- tionary radiation of visual and olfactory brain
studies of pragmatics have amply dem- sive evolution of the triune brain. In Armstrong systems in primates, bats and insectivores. Philos
E, Falk D (eds), Primate Brain Evolution, pp Trans R Soc B 348:381–392.
onstrated,63 a great deal of linguistic 291–310. New York: Plenum Press.
communication is based on metaphor: 42 Kolb B, Wishaw LO (1996) Fundamentals of
18 Joffe T, Dunbar RIM (n.d.). Primate brain Human Neuropsychology. San Francisco: Free-
Understanding the intentions behind system evolution. Brain Behav Evol, submitted man.
a metaphor is crucial to successful for publication. 43 Fletcher P, Happé F, Frith U, Baker SC, Dolan
communication. Failure to understand 19 Passingham RE (1982) The Human Primate. RJ, Frakowiak RSJ, Frith CD (1996) Other minds
San Francisco: Freeman. in the brain: A functional imaging study of ‘‘theory
these intentions commonly results in of mind’’ in story comprehension. Cognition 00:
20 Jersion HJ (1973) Evolution of the Brain and
confusion or inappropriate responses. Intelligence. New York: Academic Press. 000–000.
Indeed, without these abilities it is 21 Dunbar RIM (1989) Environmental determi- 44 Frith C (1996) Brain mechanisms for ‘‘having
doubtful whether literature, notably nants of intraspecific variation in body weight in a theory of mind.’’ J Psychopharmacol 10:9–16.
baboons. J Zool London 220:167–169. 45 Armstrong E, Clarke MR, Hill EM (1987)
poetry, would be possible. Our conver- Relative size of the anterior thalamic nuclei differ-
22 Willner LA (1989) Sexual Dimorphism in Pri-
sations would be confined to the ba- mates. Ph.D. thesis, University of London. entiates anthropoids by social system. Brain Be-
nally factual; those fine nuances of hav Evol 30:263–271.
23 Deacon TW (1990) Fallacies of progression
meaning that create both the ambigu- indices in theories of brain-size evolution. Int J 46 Keverne EB, Martel FL, Nevison CM (1996)
Primate brain evolution: Genetic and functional
ities of politeness and the subtleties of Primatol 12:193–236.
considerations. Proc R Soc Lond B 262:689–696.
public relations would not be pos- 24 Dunbar RIM (1992) Neocortex size as a con-
47 Pawlowski B, Dunbar R, Lowen C (n.d.) Neo-
straint on group size in primates. J Hum Evol
sible.64 20:469–493.
cortex size, social skills and mating success in
primates. Behaviour, in press.
25 Dunbar RIM, Joffe TH (n.d.) Neocortex size 48 Joffe TH (1997) Social pressures have selected
and social group size in prosimians. Primates, for an extended juvenile period in primates. J
REFERENCES submitted for publication. Hum Evol, in press.
1 Clutton-Brock TH, Harvey PH (1980) Primates, 26 Byrne RB (1995) The Thinking Primate. Ox- 49 Bennett PM, Harvey PH (1985) Brain size,
brains and ecology. J Zool Lond 190:309–323. ford: Oxford University Press. development and metabolism in birds and mam-
2 Aiello LC, Wheeler P (1995) The expensive 27 Barton RA (1995) Neocortex size and behav- mals. J Zool London 207:491–509.
tissue hypothesis. Curr Anthropol 36:184–193. ioural ecology in primates. Proc R Soc London B, 50 Marino L (1997) The relationship between
3 Parker ST, Gibson KR (1977) Object manipula- 263:173–177. gestation length, encephalisation and body weight
tion, tool use and sensorimotor intelligence as 28 Joffe TH, Dunbar RIM (1997) Visual and in odontocetes. Marine Mammal Sci 14:143–148.
feeding adaptations in great apes and cebus socio-cognitive information processing in pri- 51 Kudo H, Lowen S, Dunbar RIM (n.d.) Neocor-
monkeys. J Hum Evol 6:623–641. mate brain evolution. Proc R Soc London B, tex size and social network size in primates.
4 Gibson KR (1986) Cognition, brain size and the 264:1303–1307. Behaviour, submitted for publication.
extraction of embedded food resources. In Else J, 29 Stephan H, Frahm H, Baron G (1981) New 52 Sugawara K (1984) Spatial proximity and
Lee PC (eds), Primate Ontogeny, Cognition and and revised data on volumes of brain structures bodily contact among the central Kalahari San.
Social Behaviour, pp 93–104. Cambridge: Cam- in insectivores and primates. Folia Primatol Afr Studies Monogr 3.
bridge University Press. 35:1–29. 53 Dunbar RIM, Spoor M (1995) Social net-
5 Whiten A, Byrne R (1988) Tactical deception in 30 Sawaguchi T, Kudo H (1990) Neocortical devel- works, support cliques and kinship. Hum Nature
primates. Behav Brain Sci 12:233–273. opment and social structure in primates. Pri- 6:273–290.
6 Byrne R, Whiten A (eds) (1988) Machiavellian mates 31:283–290. 54 Dunbar RIM (1988) Primate Social Systems.
Intelligence. Oxford: Oxford University Press. 31 Barton RA, Purvis A (1994) Primate brains London: Chapman & Hall.
7 Harcourt AH (1988) Alliances in contests and and ecology: Looking beneath the surface. In 55 Dunbar RIM (1989) Social systems as optimal
social intelligence. In Byrne R, Whiten A (eds), Anderson JR, Thierry B, Herrenschmidt N (eds), strategy sets. In Standen V, Foley R (eds), Com-
Machiavellian Intelligence, pp 142–152. Oxford: Current Primatology, pp 1–12. Strasbourg: Univer- parative Socioecology, pp 141–149. Oxford: Black-
Oxford University Press. sity of Strasbourg Press. well Scientific.You can also read
