ArticlePDF Available

The Ape That Kicked the Hornet's Nest: Response to Commentaries on "The Ape That Thought It Was a Peacock"



We respond to the commentaries on our target article, The Ape that Thought It Was a Peacock. We start with specific issues raised by the article. These relate to the magnitude of human sex differences; the evolution and relative importance of pair bonding, paternal care, and polygyny in our species; and the distinction between the males-compete/females-choose (MCFC) model of human sexual psychology and the mutual mate choice (MMC) model. We then evaluate two competing theories of human sex differences and similarities: Social Role Theory and Attachment Fertility Theory. We conclude with some thoughts about how to present and teach evolutionary psychological research and theories without conveying an exaggerated impression of the scale of human sex differences.
This article was downloaded by: [Steve Stewart-Williams]
On: 02 September 2013, At: 17:43
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Psychological Inquiry: An International Journal for the
Advancement of Psychological Theory
Publication details, including instructions for authors and subscription information:
The Ape That Kicked the Hornet's Nest: Response to
Commentaries on “The Ape That Thought It Was a
Steve Stewart-Williams
& Andrew G. Thomas
Department of Psychology , Swansea University , Swansea , United Kingdom
To cite this article: Steve Stewart-Williams & Andrew G. Thomas (2013) The Ape That Kicked the Hornet's Nest: Response
to Commentaries on “The Ape That Thought It Was a Peacock”, Psychological Inquiry: An International Journal for the
Advancement of Psychological Theory, 24:3, 248-271
To link to this article:
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
Psychological Inquiry, 24: 248–271, 2013
Taylor & Francis Group, LLC
ISSN: 1047-840X print / 1532-7965 online
DOI: 10.1080/1047840X.2013.823831
The Ape That Kicked the Hornet’s Nest: Response to Commentaries on
“The Ape That Thought It Was a Peacock”
Steve Stewart-Williams and Andrew G. Thomas
Department of Psychology, Swansea University, Swansea, United Kingdom
We respond to the commentaries on our target article, “The Ape That Thought It Was
a Peacock. We start with specific issues raised by the article. These relate to the
magnitude of human sex differences; the evolution and relative importance of pair
bonding, paternal care, and polygyny in our species; and the distinction between
the males-compete/females-choose (MCFC) model of human sexual psychology and
the mutual mate choice (MMC) model. We then evaluate two competing theories of
human sex differences and similarities: Social Role Theory and Attachment Fertility
Theory. We conclude with some thoughts about how to present and teach evolutionary
psychological research and theories without conveying an exaggerated impression of
the scale of human sex differences.
We were truly gratified by the quantity and quality
of the thoughtful responses to our target article. All the
commentaries raised important issues, and we enjoyed
mentally jousting with them all. In the following pages,
we report on some of the outcomes of this mental
jousting. First, though, we briefly recap the thesis of
the article in light of the commentaries. The thesis boils
down to five key claims:
1. Humans are a relatively monomorphic animal. Cer-
tainly, we exhibit some degree of dimorphism; it is,
after all, easy to tell men apart from women (Doug
Kenrick’s son, 2013, as cited in Kenrick, this issue).
And certainly we are more dimorphic for some traits
than for others (Kenrick, this issue; Wood & Eagly,
this issue). However, humans are less like polyg-
ynous peacocks or deer than we are like socially
monogamous foxes and robins (Gray, this issue),
or cooperatively breeding tamarins and marmosets
(Snowdon, this issue).
2. Our low general dimorphism is a consequence of
high levels of pair bonding and male parental care
in our ancestral past. This is not to say that hu-
mans are adapted solely for long-term pair bonding.
Like dunnocks,
human sexual psychology is com-
Dunnocks (also known as hedge sparrows) are LBBs—Little
Brown Birds—found throughout Europe and Asia. They are also
common in New Zealand.
patible with a range of reproductive arrangements,
including monogamy, polygyny, and promiscuity
(Roberts & Havl
cek, this issue). However, pair
bonding and biparental care were important enough
throughout human evolution to lead to a very gen-
eral decline in psychological dimorphism.
3. Contrary to the view that humans are relative-
ly monomorphic, evolutionary psychology (EP)
sometimes gives the impression that there are ex-
tremely large sex differences in our species. It
does this in two main ways. First, there are ver-
bal statements in the EP literature suggestive of
large differences, but for which the research actually
shows rather modest differences. Second, there is
research that appears to show very large differences,
but which on closer inspection probably overstates
these differences (e.g., Clark & Hatfield’s [1989]
“Would you go to bed with me?” study; the sexual
predilections of despotic leaders; see Betzig, 1986,
this issue).
4. The exaggeration of sex differences in EP is not
the product of a hidden agenda or the deliberate
cherry-picking of results (see Pham, Shackelford,
& Jeffery, this issue). Indeed, it is probably not
the product of any single cause (Buss, this issue;
G. F. Miller, this issue; Pound & Price, this issue).
However, one contributing factor might be EP’s
historical commitment to what we call the males-
compete/females choose (MCFC) model of sexual
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
selection—a model that emphasizes sexual dimor-
phism, female mate choice, and male competition
for mates.
5. The MCFC model is not the only model employed
in EP. It coexists with what we call the mutual mate
choice (MMC) model. This model emphasizes mu-
tual courtship and the centrality of pair bonding and
biparental care in our species. There is a great deal
of research in EP on these MMC phenomena. How-
ever, there is also a tendency to present humans as
a highly dimorphic MCFC species. Although some
commentators suggest that the MCFC and MMC
models are both correct and that humans faculta-
tively switch between the two (Pham et al., this is-
sue; Roberts & Havl
cek, this issue), we argue that
the models describe two different animals, only one
of which is us (see also Eastwick, this issue; Miller,
this issue).
We were heartened to see that several prominent
evolutionary psychologists were receptive to these
claims, including some of the claims that were some-
what critical of EP. Regarding the exaggeration of sex
differences, Buss (this issue) noted that “it’s possible
that some evolutionary psychologists, myself included,
may have contributed to these exaggerations (although
I appreciate that the authors do not attribute such ex-
aggerations to me)” (p. 174). Regarding the presence
of the MCFC model in EP, Campbell (this issue) re-
counted how her research on women’s competition for
mates initially met with a chilly reception, as a result
of the MCFC assumption that:
Women’s greater parental investment makes them a
limiting resource for male reproductive success and
so men must compete, often ferociously, for sexual
access. Women have no reason to compete since any
man will be only too happy to sexually oblige them.
(p. 178)
Also regarding the MCFC model, G. F. Miller (this
issue) noted that, “In my dissertation, I advocated an
MCFC ‘runaway brain’ model of human mental evo-
lution. ...So, for a while in the early 1990s, I was
indeed ‘an ape that thought it was a peacock”’ (p.
207). Only later did he make the arduous journey to
the MMC position. (We made the same journey, but
by that time we were lucky enough to have G. F.
Miller’s, 2000, Mating Mind as one of our road maps.)
Finally, several commentators provided additional ev-
idence for our claims regarding mutual mate choice
and reduced dimorphism (Eastwick; Gray; Snowdon),
and several provided additional examples of the ex-
aggeration of sex differences in EP (Eastwick; Harris;
But of course it wasn’t all agreement and harmony.
Various commentaries raised objections to the differ-
ent aspects of our thesis. Indeed, some of the commen-
taries were rather heated, leading us to wonder whether
we should have written on a less controversial topic,
such as religion or politics. Each critical commentary
pushed us to ask ourselves: Are we right? Is our posi-
tion defensible?” In every case, though, we ultimately
decided that the position is defensible, albeit with some
provisos and concessions which we will discuss. Need-
less to say, we cannot comment on every point raised.
However, in the following pages, we address some key
areas of disagreement and also attempt to clear up some
misunderstandings of our position. Our response is di-
vided into three main parts. First, we deal with some
specific issues raised by the target article. These relate
to the magnitude of human sex differences; the evo-
lution and relative importance of pair bonding, male
parental care, and polygyny; and the MCFC versus
MMC dichotomy. Second, we look at two alternative
theories of sex differences, namely, Social Role The-
ory and Attachment Fertility Theory. For those who
enjoy a good academic row, this section and the as-
sociated commentaries may be of interest (see L. C.
Miller, Christensen, Pedersen, Putcha-Bhagavatula, &
Appleby, this issue; Wood & Eagly, this issue.) Fi-
nally, we conclude with reflections on how to present
and teach EP research in a way that does not convey
an exaggerated impression of the magnitude of human
sex differences.
Specific Issues Raised by the Target Article
Cutting Sex Differences Down to Size
We need a new scale of standardized effect sizes. d =
0.1 ‘crap’ 0.2 ‘rubbish’ 0.3 ‘meh’ 0.4 ‘not bad’ ...and
so up to d = 2.0‘wellIllbe...’ —Tweeted by Neu-
roskeptic (8 June, 2013)
Love it or hate it, research on sex differences in-
evitably attracts attention. Of all the thousands of stud-
ies conducted by evolutionary psychologists over the
decades, it is the literature on sex differences that de-
fines EP for laypeople and for most psychologists. In
the target article, we provided two general arguments
for viewing sex differences as more modest than evo-
lutionary psychologists sometimes suggest. The argu-
ments went unnamed in the article, but here we will
call them the effect size argument and the tail-of-the-
distribution argument. Below we consider each in turn,
and ask whether the arguments can withstand the var-
ious criticisms raised against them by the commenta-
The Effect Size Argument
One of the central arguments in the target arti-
cle concerned effect sizes for human sex differences.
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
Evolutionary psychologists commonly argue that, al-
though most psychological sex differences are fairly
modest, those related to mating and reproduction are
often very large (Buss & Schmitt, 2011). One of the
largest relates to men’s and women’s willingness to en-
gage in casual sex, a trait sometimes known as socio-
sexuality (SO). The effect size for the SO sex difference
averages around d = .8 (Lippa, 2009; Schmitt, 2005).
This is conventionally described as a large effect (Co-
hen, 1988). The question we asked in the target article,
however, was, Is this really such a big difference? To
answer this question, we suggested comparing the SO
difference to the sex difference in height. The height
difference is generally viewed as rather modest, espe-
cially when humans are placed in a lineup with other
mammals. But here’s the key point: According to one
estimate, the effect size for the height difference is
around d = 1.63 (Lippa, 2009). This means that the
sex difference in SO is only around half the size of the
modest sex difference in height. This seems to suggest
that the SO difference—one of the largest sex differ-
ences psychologists have discovered—is not really so
large after all. In addition to height, we proposed com-
paring the SO difference to the size difference in a
species universally viewed as monomorphic: the lars
gibbon. The gibbon size difference is around d = .8
(calculated from data in Schultz, 1941, as reported by
Geissmann, 1993). This means that the SO difference is
only as large as the size difference in the monomorphic
gibbon. This suggests that the difference falls within
a range that can reasonably be considered monomor-
phic. Certainly, it dwarfs most other human sex differ-
ences. But this may just be because we’re a relatively
monomorphic animal.
In the original article, we extended this line of rea-
soning only to SO. However, it is much more broadly
applicable. The vast majority of sex differences fall
within the monomorphic range (Hyde, 2005). This in-
cludes the MCFC-type differences emphasized in EP,
such as sex differences in number of sexual partners
desired (d = .4 to .5; Schmitt & 118 Members of the
International Sexuality Description Project, 2003), sex
differences in the seeking of short-term mates (d =
.31–.67; Schmitt & 118 Members of the International
Sexuality Description Project, 2003), and sex differ-
ences in overall choosiness regarding short-term mates
(d = .5 for sexual relations; d = .76 for one-night
stands; calculated from data in Kenrick, Groth, Trost,
& Sadalla, 1993, Table 3). Evolutionary psychologists
often describe these differences as “profound, “stark,
or “fundamental” (e.g., Buss, 2003, pp. 256, 259; A. P.
C. Davies & Shackelford, 2008; Schmitt & 118 Mem-
bers of the International Sexuality Description Project,
2003, pp. 87, 95). In light of the effect size argument,
we might conclude that these are overstatements (or al-
ternatively, that the gibbon size difference is profound
and the human height difference doubly so). Note that
natural selection can produce strongly dimorphic psy-
chological differences; for example, the sex difference
in whether one is primarily attracted to men or women
is around d = 3.99 (Lippa, 2000). However, this is a
conspicuous exception to the rule. Overall, compared
to most mammals, humans are relatively monomorphic
in their sexual psychologies.
So, that’s the effect size argument. It is fair to say
that the argument met with mixed fortunes among the
commentators. Some liked it (e.g., Harris); others were
less enamored (e.g., Campbell, Kenrick, Pham et al.).
One recurring criticism was that the human height dif-
ference is an arbitrary yardstick by which to judge the
magnitude of other sex differences. As Campbell (this
issue) observed, as soon as we start making such judg-
ments, “we enter a relativistic debate as to whether the
glass is half full or half empty, reminiscent of Hyde’s
(2005) gender similarities hypothesis” (p. 181). Al-
though we agree that there is always some degree of
arbitrariness in this kind of exercise, we suggest that
it is still informative, and moreover that height is a
particularly useful yardstick to use. Unlike sex differ-
ences in psychology, or in physical traits like strength,
height is readily observable. As such, people have well-
informed intuitions about the magnitude of the sex dif-
ference and the variance within each sex (Pound &
Price, this issue). In addition, there is general agree-
ment that the sex difference in height in humans is small
or moderate compared to that found in most dimorphic
animals. Thus, rather than pushing us into a relativistic
debate, height provides a way to escape that debate.
Campbell does make the point that, if we had chosen
another yardstick—the sex difference in body mass,
fat mass, or upper body strength, for instance—the
psychological differences would have appeared even
smaller by comparison, thereby strengthening our case
(see Lassek & Gaulin, 2009). We suggest, however,
that using a more modest difference as our yardstick
has advantages of its own. Specifically, it allows us
to argue that not only are the psychological sex dif-
ferences much smaller than the very largest physical
differences, they’re smaller even than the modest sex
difference in height.
Of course, one could always argue that any attempt
to evaluate whether a given effect is small or large,
monomorphic or dimorphic, is still ultimately arbi-
trary, and thus should be excluded from the realm of
scientific discourse. Pham and colleagues bite the bul-
let and take this tack, suggesting that it is “meaningless
scientifically (p. 222) to assign any labels to any ef-
fect sizes. It is meaningless for the same reason that
it would be meaningless to assign the label “hot” to
the sun—after all, relative to much hotter stars, the
sun is rather cool. Although strictly speaking this is
true, we are not convinced that the point is decisive.
Note, to begin with, that it is a rather extreme way
to get around the effect size argument. It immediately
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
invalidates every claim in the EP literature—and in-
deed in all of psychology—that a given effect is large,
or profound, or fundamental, or anything else. More
important, though, it does not undermine our basic ar-
gument. Even if we abandon conventional labels such
as small, medium, and large, we still want some way
to get a handle on the relative magnitude of any differ-
ence. The comparison with height gives us that. Labels
or no labels, it tells us something important that the
sex difference in SO is only half the magnitude of
the sex difference in height. It also tells us something
important that the SO difference is around the same
magnitude as the sex difference in size in gibbons, and
less than one fourth the magnitude of the sex difference
in attraction to women versus men.
A very different criticism of the effect size argument
came from Kenrick (this issue). In his commentary, he
presented a nice reductio ad absurdum of our position:
Yes, an effect size less than 1.0 does allow for con-
siderable overlap in the distributions of two groups,
but if we started treating every effect less than 1.0 as
unworthy of our attention, I would have an easy time
doing the next revision on my social psychology text-
book: I could simply throw out most of the classic
experimental and correlational findings in the field,
including the bulk of findings on cultural influences,
hormonal influences, personality effects, persuasion
techniques, obedience to authority, conformity, and
on and on. Indeed, using the 1.0 effect size standard,
general psychology textbooks could also be, as a class,
reduced to very short pamphlets. (p. 203)
This is a great riposte, but it’s not, we think, fa-
tal to our position. After all, we’re not arguing that
small effects are unworthy of research attention or that
they should be barred from the textbooks; we’re argu-
ing that small effects shouldn’t be described as large
effects. Kenrick’s comment does raise an important
issue, though. In arguing that psychological sex differ-
ences in humans are generally small, we don’t mean to
imply that they are smaller or less important than other
findings in psychology. As Eagly (1995) observed, sex
differences fall within the same range of magnitudes
as most research findings: The vast majority are small
or medium and only a handful are large. However,
the conclusion to draw from this is not that sex dif-
ferences must therefore be important. The conclusion
to draw is that most effects in psychology are rela-
tively unimportant. That is, most variables, considered
in isolation, have relatively little impact on behavior.
This doesn’t mean that Kenrick should stop writing his
social psychology textbook or that we should all aban-
don psychology and become plumbers instead. The
small magnitude of most effects in psychology is it-
self a discovery of psychology. One might argue, in
fact, that it is one of the great metadiscoveries of the
field. Most variables have little impact, and thus most
of the phenomena studied by psychologists are prod-
ucts of a multiplicity of variables. Sex differences are
no exception to this rule.
The Tail-of-the-Distribution Argument
The second argument for downplaying sex differ-
ences applies to data taken from the tail of the distri-
bution. As a general rule, such data magnify any group
differences found at the mean. So, for instance, if we
look at the extremes of height in a population, we find
a very high ratio of men to women—at 6 feet tall,
for instance, the ratio is around 2000:1 (Pinker, 2005).
However, as we move closer and closer to the average
height, the ratio of men to women gets closer and closer
to 1:1. There is still a mean difference, of course, but
it’s much smaller than the difference at the tail. This
implies that if data from the tail are extrapolated to the
general population, we are likely to form an exagger-
ated impression of the typical difference. In the target
article, we applied this logic to several lines of evidence
in EP, namely, the utilization of prostitutes as an index
of the sex difference in interest in casual sex (Symons,
1979), the harem-holding habits of despotic leaders as
an index of the sex difference in polygamous inclina-
tions (Betzig, 1986, this issue), homicide rates as an
index of the sex difference in intrasexual competitive-
ness (Daly & Wilson, 2001b), and rates of premature
death as an index of the sex difference in risk taking
(Kruger & Nesse, 2006). In her commentary, Harris
(this issue) provided an additional example: homicide
statistics and other extreme behavioral syndromes as
an index of the sex difference in sexual jealousy. To
be clear, our argument isn’t simply that the “average
man manages to grow old without killing another per-
son, killing himself, or being married to more than
one woman (at the same time)” (Kenrick, this issue,
p. 202). The argument is that those few who do en-
gage in these behaviors come from the extreme of the
distribution of traits such as aggression or polygamous
inclination, and the sex ratio at the tail exaggerates the
difference in the wider population. The issue isn’t how
many men engage in these extreme behaviors but what
the sex differences at the tail say about the species as a
There were no direct attacks on the underlying
logic of the tail-of-the-distribution argument, but sev-
eral commentators had misgivings. Pound and Price
(this issue) noted that, although it is rare for people
to accumulate harems as large as villages, to murder
their fellow human beings, or to die in multicar pile-
ups, the sex differences seen in these various domains
still tell us something about sex differences in gen-
eral. Specifically, they tell us that, on average, human
males are more polygamous, more intrasexually com-
petitive, and more risk-prone than human females. If
we were to throw these data away, note Pound and
Price, we’d be losing some of the best data we’ve got.
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
Self-report measures are prone to social desirability
bias, and this compromises their accuracy. In contrast,
homicide statistics, road fatality statistics, and other
data from the tail are highly accurate and provide an
assay of sex differences that are otherwise very hard to
measure (see, e.g., Daly & Wilson, 1988).
We agree that sex differences at the tail of the dis-
tribution tell us something about the differences in the
rest of the population. However, for the reasons out-
lined, they probably exaggerate that something. The
tails sometimes tell a tale, as Pound and Price put it,
but to some extent it’s a tall tale: an embellished ac-
count of the facts. As far as we can tell, none of the
examples Pound and Price discuss escape the logic of
the tail-of-the-distribution argument. Take road fatali-
ties, for example. Between 1999 and 2007, men in the
United States were 2.3 times more likely than women
to die in a car accident. “The vast majority of jour-
neys end safely, Pound and Price observe, “but we
would argue that this 2.3 fold difference likely tells
us something about the activities and behavior of men
and women in the population at large” (p. 225). We
agree; it tells us that, on average, men are more risk
prone than women. But the average man is probably
not 2.3 times more risk prone than the average woman.
Although anyone can get unlucky and die in a car acci-
dent, people who are atypically high in risk-proneness
are at an increased risk. They will therefore be dispro-
portionately represented in the road fatality statistics.
Because these individuals come from the extreme of
the distribution for risk-proneness, the ratio of men to
women will be higher than it is in the general popula-
tion. Thus, although the road fatality statistics do tell
us something about the sex difference in risk prone-
ness, if we use them to estimate the difference in the
general population, our estimate will be an overesti-
mate. Consistent with this suggestion, questionnaire
and lab studies looking at less extreme forms of risk
taking in more typical samples tend to find a smaller
gender gap (Byrnes, Miller, & Schafer, 1999; Wood &
Eagly, 2012, this issue). Now, does this mean that we
should discard data from the tail of the distribution?
Definitely not! As Pound and Price note, these data
have been extremely useful in testing EP hypotheses.
Daly and Wilson’s (1988) homicide research in par-
ticular is some of the finest research in EP. Thus, our
point is not that we should never use such data; it is that
when we do, we need to take the tail-of-the-distribution
phenomenon into account before we draw conclusions
about the magnitude of sex differences closer to the
mean (i.e., for most people).
Buss (this issue) and Kenrick (this issue) raised a
rather different concern. They pointed out that even if
the mean sex differences in traits such as aggression
and short-term mate seeking are relatively modest, and
even if extreme manifestations of these traits—such
as homicide and sex trafficking—are relatively rare,
men are still massively overrepresented in these activ-
ities. Furthermore, despite their rarity, homicide and
sex trafficking create huge amounts of suffering. Thus,
the sex differences have real-world consequences and
these consequences really matter (Kenrick, this issue).
Our response is to say only that we completely agree.
We don’t deny for a minute that homicide and sex traf-
ficking matter or that the sex differences uncovered
by evolutionary psychologists explain men’s overrep-
resentation in these arenas. Our only claim is that the
sex differences found for these extreme behaviors are
not representative of the differences in traits like ag-
gression and short-term mate seeking in the species as
a whole.
What Have We Got Against Sex Differences,
At the end of his commentary, Kenrick poses two
questions for us:
Why do they want to sweep the differences under
the rug?
What makes it better to call humans monomorphic
than to acknowledge, and attempt to understand,
when and why we are dimorphic? (pp. 205–206)
Good questions! To begin with, notice that the first
question is premised on the assumption that our overall
argument is false. Given that we don’t share that as-
sumption, here’s how the question sounds to us. Imag-
ine that you overheard a colleague announce that men
are 4 times larger than women. You point out that al-
though there is certainly a sex difference in size in
our species, to claim that men are 4 times larger than
women is to exaggerate that difference. Your colleague
then turns to you and asks, “Why do you want to sweep
the difference under the rug?” Your answer, of course,
is that you don’t want to; you just don’t agree that
men are 4 times larger. And that’s our answer to Ken-
rick’s first question. We would only add that we’re
not just concerned about evolved sex differences be-
ing overstated; we’re also concerned about them be-
ing understated (see the later sections on Social Role
Theory and Attachment Fertility Theory). As for the
second question—why it might be better to call hu-
mans monomorphic than to acknowledge when and
why we are (somewhat) dimorphic—well, we don’t
agree that we have to choose between these options.
We can do both. We can acknowledge that human be-
ings exhibit varying degrees of dimorphism across dif-
ferent traits, while at the same time acknowledging
that most of this variation falls within the monomor-
phic range. In other words, we can locate humans on a
rough dimorphism–monomorphism continuum with-
out losing any information regarding precise levels
of dimorphism for particular traits. Of course, this
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
answer is premised on the assumption that our overall
argument is correct. We hope, though, that this section
has helped to make the case that it is.
The Three P’s: Pair Bonding, Paternal Care,
and Polygyny
The intimate male–female relationship ...which zo-
ologists have dubbed a “pair bond, is bred into our
bones. I believe this is what sets us apart from the apes
more than anything else.
—Frans de Waal (2005, p. 114)
The majority of husbands remind me of an orangutan
trying to play the violin.
e de Balzac.
If it is accepted that human sex differences are gen-
erally rather modest, the next question is why this might
be. The answer we presented in the target article is that
humans evolved to exhibit high levels of pair bonding
and biparental care, and that this substantially reduced
the sex difference in reproductive variability in our
species, leading to a general decline in dimorphism
(see also Gray, this issue). None of the commentators
denied that pair bonding is our most common mating
arrangement, and several agreed that it is (Pham et al.;
Pound & Price; Roberts & Havl
cek). Nevertheless,
there were several objections to the details of our pre-
sentation. These centered on two main issues: First, the
evolution of pair bonding in the hominin lineage, and
second, the relative importance of pair bonding ver-
sus polygyny in shaping the human psyche. We will
address each of these in turn.
Pair Bonding and Paternal Care: The Plot
There are many theories about the factors leading
to the evolution of pair bonding and paternal care in
the human line (e.g., Chapais, 2008; Lovejoy, 1981;
Winking & Gurven, 2011). In the target article, we
argued for a version of the provisioning hypothesis.
Specifically, we argued that pair bonding evolved pri-
marily to facilitate biparental care and that paternal
care was evolutionarily advantageous to men primar-
ily because it increased the survival and reproduc-
tive success of their offspring (see also L. C. Miller
et al., this issue). Most male mammals do not form
pair bonds or care for offspring, but human males
evolved to do both because of the extreme depen-
dency of our young. Several commentators challenged
this view (Campbell; Roberts & Havl
cek). Roberts
and Havl
cek, for instance, discussed research indi-
cating that the primary function of pair bonding for
men might instead have been mate guarding (i.e., keep-
ing other men away from the woman and vice versa).
They and Campbell also suggested that pair bonding
and paternal care may have evolved not because pa-
ternal care enhanced offspring survival but in response
to a female preference for investing males (see also
Campbell’s sexual conflict argument; Campbell, this
issue). We did acknowledge in the article that pair
bonding may have been favored in part as a male
mate guarding strategy (p. 146), and that paternal
care may have been favored in part by female choice
(p. 152). However, the commentators are correct in
suggesting that we gave paternal care and its impact on
offspring survival the starring roles in our evolutionary
drama, and thus there is indeed a case for us to an-
swer. The provisioning hypothesis has a long history
in biological anthropology (e.g., Lovejoy, 1981). As
the conventional view in the area, it might seem dated
or even a little old-fashioned. We argue, however, that
this old-fashioned view stacks up well against newer
Pair Bonding as a Mate Guarding Strategy
Consider, first, the idea that pair bonding is primar-
ily about mate guarding rather than paternal care—or
at any rate, that its original function was mate guarding
(e.g., Chapais, 2008). We see three problems with this
idea. First, it raises the question: What is the evolution-
ary advantage for the woman? If the man did not invest,
directly or indirectly, in the woman or her offspring, it
is not clear that the woman’s fitness would have been
enhanced by the pair bond. But if the woman’s fit-
ness was not enhanced, why would women evolve the
psychological disposition to fall in love, experience
jealousy, and engage in mate guarding of men? The
mate guarding hypothesis provides a stronger explana-
tion for men’s pair bonding psychology than it does for
Second, the mate guarding hypothesis does not ex-
plain why pair bonding evolved in our species but not
in any other Great Ape or in most mammals. Why are
we such an exception to the mammalian rule? The pro-
visioning hypothesis has a ready answer: We evolved
to form pair bonds because of the costliness of our off-
spring and the consequent need for allomaternal care.
The mate guarding hypothesis, in contrast, is forced
into the awkward position of saying that, although hu-
mans are the only Great Ape that exhibits pair bonding,
and although we are also the only Great Ape possess-
ing a conspicuous trait that could explain this fact (i.e.,
the costliness of our offspring), the former actually has
nothing to do with the latter. This is logically possible,
but it seems unlikely. More plausibly, the evolution of
human pair bonding is related primarily to the care of
our costly, big-brained young.
A third and final argument relates to men’s pa-
ternal psychology. Roberts and Havl
cek (this issue)
pointed out that “numerous mammalian species which
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
frequently form pair bonds do not show paternal care”
(see also Emlen & Oring, 1977). This is certainly
true. However, one reason to think that humans are
not one of these species is that humans do show pa-
ternal care (Marlowe, 2000). As detailed in the article,
there are strong reasons to think that the male psyche is
specifically adapted to bond with and care for children
(pp. 146–148). If pair bonding were wholly a matter
of mate guarding, we would not expect this. Thus, all
the evidence for men’s evolved paternal psychology
counts as evidence against the idea that pair bonding is
wholly a matter of mate guarding.
The Evolution of Paternal Care Through
Female Choice
None of the above arguments is applicable to the
second alternative explanation for pair bonding and
paternal care: the female choice hypothesis (Camp-
bell, this issue; Roberts & Havl
cek, this issue; see
also Buss & Schmitt, 1993). According to this hypoth-
esis, women evolved to mate preferentially with men
willing to invest in them and their offspring. This in
turn created a strong selection pressure on men, lead-
ing to the evolution of pair bonding and paternal care.
Thus, pair bonding and paternal care evolved in much
the same way as did the peacock’s tail: through the
mating decisions of females. Unlike the mate guard-
ing hypothesis, the female choice hypothesis provides
clear benefits of pair bonding for both women and men.
For women, the main advantage is a spare pair of hands
to help provide for her and her young (Campbell, this
issue). For men, in contrast, the main advantage comes
from gaining exclusive (or near-exclusive) sexual ac-
cess to the woman. In the absence of a female pref-
erence for investing males, men’s fitness might have
been better served by seeking multiple mates than by
pair bonding and caring for offspring (Winking & Gur-
ven, 2011). Given the female preference, however, the
latter strategy might have yielded a greater fitness in-
crement. Importantly, according to the female choice
hypothesis, this is because paternally inclined men
increased their chances of siring their mate’s future
children, not because paternal care selectively boosted
the survival and reproductive success of their own
It is certainly plausible that female choice could
shape paternal tendencies. Mathematical models indi-
cate that if females mate preferentially with parental
males, male parental care can evolve even when pater-
nity probability is low (Alonzo, 2012). But although
female choice might have played some part in the evo-
lution of paternal care, it seems unlikely that it was
the sole driver. The reason is simple: If the only repro-
ductive payoff for paternally-investing men was sexual
access to the woman, it would make no difference to
men’s fitness whether the children they cared for were
their own or someone else’s. As long as they had a
reasonable chance of siring their mate’s next child, pa-
ternal care would have achieved its goal. However, if
that were the case, there would be no reason to ex-
pect men to invest more in genetic offspring than in
step-offspring. In other words, there would be no rea-
son to expect the Cinderella effect (Daly & Wilson,
1988, 2001a; Marlowe, 1999). Thus, all the evidence
for the Cinderella effect counts as evidence against
the hypothesis that paternal care is solely a matter of
meeting female preferences and obtaining sexual ac-
cess. More than that, it tells us that men’s investment
in their genetic progeny must be an essential part of
the evolutionary story.
Problems With the Provisioning Hypothesis
We’ve pointed out some of the challenges facing
the mate guarding and female choice hypotheses. As
several commentators observed, though, our preferred
hypothesis faces some challenges of its own (Camp-
bell; Roberts & Havl
cek). One of the strongest comes
from an ambitious cross-cultural study by Sear and
Mace (2008). This involved a detailed review of 45
natural-fertility societies: societies that had not yet un-
dergone the demographic transition. Sear and Mace
were interested in children’s survival rates, and whether
the presence of a given category of family member
(e.g., mother, father, grandparent, sibling) was associ-
ated with better survival. The results were striking. In
every society for which there was evidence (28 out of
28), the presence of the mother increased the chances
that her offspring would survive. The presence of the
father, on the other hand, had a much less consistent
effect. Father presence was associated with increased
offspring survival in only 32% of societies (7 out of
Grandmothers and older siblings were associated
with increased survival in a greater number of soci-
eties than were fathers. Sear and Mace do not deny
that fathers sometimes invest in their young. They ar-
gue, however, that this is not as consistently important
or universal as many assume it is. Mothers always re-
ceive allomaternal support from someone, but it’s not
always the father (see also Hrdy, 2009).
We have some sympathy for this view; as we men-
tioned in the target article, paternal investment is facul-
tative rather than obligate, and this probably accounts
for some of the variability in paternal care across cul-
tures. We also agree that fathers are not the only allo-
maternal helpers (see Snowdon, this issue). Nonethe-
less, we suspect that Sear and Mace’s (2008) findings
underestimate the importance and impact of paternal
care. Indeed, it would be surprising if they did not;
after all, men in most societies engage in paternal care
Note that if we consider only the studies that Sear and Mace
judged to be statistically valid, the percentage of societies in which
father presence was associated with offspring survival rises to 47%
(7 of 15; see Sear & Mace, 2008, Table 3).
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
(Marlowe, 2000), and this would be rather puzzling
if paternal care usually had no effect. One issue with
the study, which we touched on in the article, is that
child survival is an extremely exacting measure of in-
vestment, and one that is blind to some of the sub-
tler ways that paternal care could enhance the father’s
fitness. This includes decreasing the woman’s inter-
birth interval, and thereby increasing the number of
offspring the man can sire with her (Campbell, this is-
sue). A second issue, also mentioned in the article, is
that most of the societies Sear and Mace surveyed were
agricultural societies, whereas for most of our evolu-
tionary history, humans lived as hunter-gatherers. The
fact that, in many agricultural societies, children do
just as well without an investing father does not imply
that this was necessarily the case among Pleistocene
hunter-gatherers (Draper & Harpending, 1987). Chil-
dren don’t need even a mother to survive in modern
welfare states, but we could not infer from this that
children did just as well without a mother in premodern
societies. In short, it is not clear that Sear and Mace’s
findings reflect the typical conditions in which humans
evolved, and thus it is not clear that the findings reflect
the typical rates or impact of paternal care throughout
most of our evolutionary history.
Even if we ignore these concerns, however, there
is still a genuine question about how to interpret the
findings. Sear and Mace (2008) seem to assume that
increased offspring survival tells us who’s providing
evolutionarily significant help: If the presence of the
father increases child survival, fathers are helping; if
it doesn’t, they’re not. We’re not so certain. Consider,
first, the fact that, in many of the societies surveyed
by Sear and Mace, there was little or no mother effect
after the child reached the age of 2—that is, young
children were just as likely to survive with or without a
mother. Does this mean that mothers in these societies
made no contribution to their children’s survival af-
ter their second birthdays? It seems unlikely. Two-year
olds cannot feed themselves or keep themselves safe.
Certainly, these responsibilities do not fall solely to
the mother; however, in most traditional societies, the
mother takes the leading role (Wood & Eagly, 2002).
Why, then, does the mother effect evaporate after chil-
dren hit 2? One possibility is that, if the mother dies,
other family members often take over and help keep
the child alive (Sear & Mace, 2008). This might some-
times mask the mother effect. However, the absence of
a mother effect would not then imply that living moth-
ers are not helping keep their children alive. Presum-
ably, they are. The same argument applies to fathers.
If a father dies (or absconds or otherwise vanishes),
other family members may take over his parental du-
ties, and this may often mask the father effect. But the
absence of a father effect would not then imply that
living fathers make no difference to their children’s
survival. They may well make a difference, but in their
absence, others may step in and make the same differ-
ence. If all the child’s allomaternal helpers vanished,
then the contribution of the father might show up more
Certainly, mothers are more reliably associated with
offspring survival than fathers. But in our view, this
implies only that mothers have a consistently bigger
impact on children’s welfare than fathers. It does not
imply that fathers rarely have any impact at all. Thus
Sear and Mace’s (2008) study does not undermine the
claim that paternal care is a central part of humanity’s
reproductive toolkit, or the claim that it was selected
because it boosted the survival and reproductive suc-
cess of the father’s genetic progeny. These hypotheses
are still live options.
Is Polygyny the Royal Road to Reproductive
Success for Men?
There is one more matter to consider: the relative
importance of pair bonding versus polygyny in shaping
men and women’s sexual psychologies. In the target
article, we argued that, although pair bonding is not
human beings’ only mating system, it is our primary
system. The pair bond is the most common context
for sex and reproduction in our species, and quite pos-
sibly has been for much of our evolutionary history
(Stewart-Williams & Thomas, this issue, pp. 145–148,
150–151). As such, this mode of mating has left a
deep imprint in our evolved nature. Some of the com-
mentators suggested that, in making this argument, we
might have overstated the importance of pair bonding
in our ancestral past, and correspondingly understated
the importance of polygynous or promiscuous mating.
Roberts and Havl
cek (this issue) point out that, al-
though social monogamy was indeed more common
than polygyny, polygyny was still a common-enough
feature of the human mating landscape that it too left an
imprint in our evolved nature. Furthermore, although
polygynous mating was infrequent, this does not im-
ply that men had no interest in it; instead, most men’s
polygynous ambitions may simply have been thwarted
by socio-ecological constraints (see also Kenrick, this
issue). In a similar vein, Pound and Price (this is-
sue) note that, although opportunities for polygynous
mating were usually rare, the reproductive benefits of
polygyny were so great for genes located in male bod-
ies that the male mind might still have evolved to take
advantage of those opportunities, if and when they did
arise. As a result, men may harbor strong polygamous
desires—much stronger than women’s—even if these
desires are frustrated for most men throughout most of
their lives (Symons, 1979). Pound and Price put it well:
because fat, salt, and sugar were nutritionally valuable
yet elusive in ancestral environments, people tend to
express unhealthily strong appetites for these nutrients
in environments in which they are abundant. ...Given
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
that polygyny was both elusive and reproductively
rewarding to ancestral men, they may have evolved
strong desires to achieve it. (p. 227)
We don’t have any major quarrels with any of these
arguments. We agree with the general point that an-
cestral humans engaged in nontrivial levels of polyg-
yny, and that this left an imprint in our evolved na-
ture, just as social monogamy did (Stewart-Williams
& Thomas, this issue, p. 150). We also agree that cer-
tain selection pressures would have pushed our species
toward polygyny. To balance out the picture, though,
we would like to add that there would also have been
selection pressures pushing in the opposite direction.
So, for instance, if we are right in thinking that an-
cestral men were often able to enhance their lifetime
fitness by forming pair bonds and contributing to the
care of their offspring, this would have selected against
a strongly polygynous or promiscuous male psychol-
ogy. It would have made it somewhat less profitable,
evolutionarily speaking, for most of our male ances-
tors to devote too much time or energy to pursuing
new wives or short-term flings. This is not to say that
natural selection has eliminated all polygynous incli-
nations from men’s minds; we know that it has not.
However, the commentators focused on the benefits of
polygynous mating for men but not on the costs, which
could foster an inflated expectation of the magnitude
of the sex difference in polygamous inclinations.
are happy to concede that polygyny may have been
more important than we implied in the target article.
Nonetheless, for the reasons given, we wouldn’t be too
surprised if we got it about right.
The Notorious MCFC versus MMC
Men ...are like verbal peacocks, using language
to impress women during courtship ...women need
good verbal skills in order to choose a good mate by
making an accurate judgement of a potential partner’s
—Workman and Reader (2008, p. 291)
The only model that makes sense for humans and
for other cooperatively breeding species is the Mutual
Mate Choice model.
—Snowdon (this issue, pp. 239–240)
Next on the agenda is our claim that there are two
contradictory conceptions of human sexual psychol-
ogy inhabiting the EP literature—the MCFC and MMC
models—and that the MCFC model has somewhat dis-
That said, we do take Campbell’s point that frequency-
dependent selection could produce a population consisting for the
most part of relatively monogamous men, but with a subpopulation
of rogues who cheat the system (see Bailey, Kirk, Zhu, Dunne, &
Martin, 2000).
torted the evolutionary psychologists’ picture of our
species. Of all the claims in the target article, this was
the one that most divided the commentators (see, e.g.,
Buss; Eastwick; G. F. Miller; Pham et al.; Pound &
Price; Roberts & Havl
cek). Among those who took
issue with it, the criticisms revolved around two main
points. The first was our assertion that there is a con-
tradiction in the EP literature between the MCFC and
MMC models. The second was the (supposed) impli-
cation that EP has entirely neglected core components
of the MMC model, including male mate choice, pair
bonding, and paternal care.
Is There a Contradiction in the EP
We begin with the idea that there is a contradiction
in EP between the MCFC and MMC models. The con-
tradiction, in our view, concerns claims about overall
levels of dimorphism in our species. On the one hand,
humans are presented as (primarily) a pair bonding,
biparental species with mutual mate choice—which
seems to imply low overall dimorphism. On the other
hand, humans are presented as a species with high
dimorphism—which is hard to reconcile with the idea
that we are primarily a pair bonding, biparental species
with mutual mate choice. Several commentators denied
that there is any real contradiction here (Pham et al.;
Roberts & Havl
cek). They agreed that there is mate-
rial in the EP literature that fits our description of the
MCFC model, and material that fits our description of
the MMC model. However, they argued that EP pro-
vides a way to reconcile these apparently competing
claims: Both are true; sometimes we are an MCFC
species and sometimes we are an MMC species. Ac-
cording to theories such as Sexual Strategies Theory
(Buss & Schmitt, 1993) and Strategic Pluralism Theory
(Gangestad & Simpson, 2000), humans facultatively
adjust their mating behavior in response to their cir-
cumstances (Betzig, this issue; Pham et al., this issue;
Pound & Price, this issue; Roberts & Havl
cek, this is-
sue). Sometimes, we form pair bonds; sometimes, we
engage in short-term mating. In the first case, we tend to
exhibit low dimorphism: Both sexes are choosy about
their mates and compete for the best mates available,
and both invest in any offspring (MMC model). In the
second case, we exhibit higher dimorphism: Women
are much choosier than men about their mates, and
men engage in vigorous competition to secure as many
sexual conquests as possible (MCFC model). Thus, EP
can account for all the claims made about our species,
and there is no contradiction in the literature.
Does this move work? First, let us say that there is
some merit in the suggestion. Humans are clearly more
dimorphic for some traits than for others, and as we
mentioned in the article, sex differences related to early
courtship and short-term mating tend to be larger than
those related to later courtship and long-term mating
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
(Buss & Schmitt, 1993; Kenrick et al., 1993; Kenrick,
Sadalla, Groth, & Trost, 1990; Li & Kenrick, 2006).
Thus, one might argue, as Pham et al. (this issue) did,
that some elements of the MCFC model sometimes
manifest in human behavior and psychology” (p. 222;
see also Roberts & Havl
cek, this issue).
Ultimately, though, we don’t think the proposed so-
lution dissolves the contradiction. In our view, pair
bonding and biparental care, although not our only mat-
ing options, were sufficiently important in our ancestral
past to lead to a very general decline in psychological
dimorphism. For most psychological traits—including
those related to sex and reproduction—the differences
are remarkably small. In other words, although we
sometimes exhibit trends in an MCFC direction, these
trends are weak. It is true that the sex difference in
choosiness is greater in a short-term than a long-term
context (Pham et al., this issue). However, as noted
previously, even in a short-term context, the difference
is still modest (see, e.g., Finkel & Eastwick, 2009;
Kenrick et al., 1993). Likewise, even for SO—one of
the largest and most theoretically significant sex dif-
ferences found in our species (Lippa, 2009; Schmitt,
2005)—the effect size is only around half that of the hu-
man height difference. Thus, although there is variation
across traits in the exact levels of dimorphism, almost
all this variation clusters together near the monomor-
phic end of the spectrum (see The Effect Size Argument,
pp. 249–251). This is consistent with the claim that pair
bonding and biparental care produced a general reduc-
tion in psychological dimorphism. If we accept this
claim, then there is immediately a contradiction in EP
between the pair bonding literature and the literature
suggesting high dimorphism.
Of course, this is precisely the claim that Pham et al.
(this issue) and Roberts and Havl
cek (this issue) tac-
itly reject. They argue that there is no contradiction
in the EP literature, which would seem to imply that
we cannot classify humans in terms of their overall
level of dimorphism (see also Kenrick, this issue). We
cannot say that humans tend toward monomorphism;
we’re monomorphic in some circumstances but dimor-
phic in others. However, this proposal deserves closer
scrutiny. For most species, we are happy to talk about
overall levels of dimorphism. Gibbons and penguins
are monomorphic; gorillas and peacocks are highly di-
morphic. Admittedly, gibbons and penguins are dimor-
phic when it comes to their reproductive anatomy, and
gorillas and peacocks are monomorphic when it comes
to perceptual capabilities. Nonetheless, it makes sense
to say that, overall, the former are monomorphic and
the latter dimorphic. On the face of it, it seems im-
plausible that humans would be an exception to this
rule—that we, unlike most animals, would be both di-
morphic and monomorphic, with no general tendency
in one direction or the other. It seems especially im-
plausible given that, when we look at the evidence,
most of it suggests that there is a general tendency
in one direction: Humans tend to have relatively low
dimorphism—again, even for traits like SO.
We do take Roberts and Havl
cek’s point that human
beings, like dunnocks, exhibit a fair amount of vari-
ability in their mating habits; indeed, we have made the
same comparison elsewhere (Stewart-Williams, 2005).
But this variability does not undermine our position.
Neither the MCFC nor the MMC model posits a single
human mating strategy. Both acknowledge the vari-
ability in men and women’s mating behavior; they sim-
ply construe it in different ways. So, for instance, pair
bonding is part of the MCFC model but is a “second-
best strategy” for men who fail to mate polygynously;
polygyny is part of the MMC model but is less central
than EP typically suggests (see Eastwick, this issue).
Thus, we disagree with Roberts and Havl
cek’s sug-
gestion that our article attempts “to define a species-
typical model based on one component of evolved na-
ture, while ignoring other components” (p. 233). On
a related note, we have some misgivings about how
Roberts and Havl
cek interpret the variability of mat-
ing patterns in both dunnocks and humans (see also
Pham et al., this issue; Pound & Price, this issue). Ac-
cording to N. B. Davies (1985, 1989), much of the
variability seen in dunnocks is the result of relatively
invariant male and female preferences interacting with
different ecological conditions (e.g., the local sex ratio,
the distribution of food), rather than facultative switch-
ing from one distinct sexual psychology to another. The
same may apply to humans. Thus, although humans are
somewhat more dimorphic for preferences and moti-
vations related to short-term than long-term mating, it
might be overstating the case to suggest that we faculta-
tively switch between a gibbon-like sexual psychology
(MMC) and a peacock-like sexual psychology (MCFC;
see Eastwick, 2009, this issue, for a perceptive discus-
sion of the short-term/long-term distinction). Instead,
different circumstances may simply give people more
or less scope to act on particular preferences and moti-
vations. In any case, the fact that humans exhibit some
flexibility in their mating behavior does not change
the fact that the psychological sex differences in our
species generally fall within the monomorphic range,
even in a short-term context.
Our conclusions, then, are as follows: (1) it is mean-
ingful to make statements about overall levels of di-
morphism in our species, (2) humans are relatively
monomorphic for most psychological traits, and (3)
there is thus a genuine conflict in EP between the lit-
erature suggesting low dimorphism and that suggest-
ing high dimorphism. It is reasonable to think that the
MMC and MCFC models represent distinct visions of
human nature, and that the apparent contradiction be-
tween them cannot be shooed away simply by arguing
that both are preprogrammed modes of operation of
the human animal.
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
Does EP Neglect Mutual Mate Choice?
Another point of contention relates to our sugges-
tion that the MCFC model has resulted in a somewhat
distorted EP picture of human sexual psychology and
has sometimes led EP to minimize the importance of
mutual mate choice. This suggestion drew fire from a
number of commentators. Pham et al. (this issue) were
concerned that the target article might persuade nonex-
perts that EP has neglected pair bonding and paternal
care, and has propagated an exclusively MCFC vision
of the human species. To demonstrate the falsity of this
view, they discussed various lines of EP research that
directly examine or tacitly assume the importance of
pair bonding and paternal care. (We thought it was a
nice touch that they included several studies on which
one of the present authors [SSW] was a collaborator!)
Likewise, Pound and Price (this issue) argued that the
MMC model “is already a major focus within evolu-
tionary psychology, noting, for instance, that “there
is in fact a rather substantial amount of research be-
ing conducted on male mating preferences” (p. 228).
Finally, Kenrick (this issue) pointed out that he and
other evolutionary psychologists have taken high male
parental investment into account in their research and
theorizing for several decades (see, e.g., Buss, this is-
sue; Kenrick et al., 1990).
We certainly hope that readers don’t get the im-
pression that EP has ignored male mate choice, pair
bonding, or paternal care, and perhaps we could have
been clearer about this. In our defense, though, we did
stress early in the piece, and at several points there-
after, that EP does conduct research on all aspects of
the MMC model. We even wrote that “EP has led the
field in documenting these aspects of human sexual
psychology” (p. 152). Our claim, remember, is not that
male choice, pair bonding, and paternal care are absent
from the literature but that there is an unnoticed contra-
diction in EP between the MMC and MCFC models.
We could not claim such a contradiction unless we rec-
ognized that EP studies elements of the MMC model
such as pair bonding and paternal care. And not only
did we recognize it; we cited a great deal of EP re-
search on these topics (see The Evolution of Mutual
Mate Choice, pp. 144–152). However, we also cited
a great deal of EP research reflecting the influence of
the MCFC model. The final section of the article was
devoted entirely to this task (see The MCFC Model in
Evolutionary Psychology, pp. 152–162). The fact that
we were able to draw on the EP literature in both the
MMC and MCFC sections neatly sums up our posi-
tion: There is plenty of research in EP premised on
the assumption of mutual mate choice, but there is
also plenty of research giving the distinct impression
that MCFC is the primary dynamic in human mating.
Arguably, the least ambiguous examples of the latter
are studies that treat courtship behaviors like humor,
language, and creative intelligence as if they were de-
signed solely for men to attract women and not also
the other way around (see the quote at the start of the
section). This represents a strong MCFC interpretation
that is erroneous even in a short-term mating context
(G. F. Miller, this issue). But the MCFC bias also leaks
out in subtler ways, as we hope we showed.
We share Pham et al.s concerns about the misrep-
resentation of EP. However, this is not the only thing to
be concerned about. If we are correct in thinking that
EP has overemphasized human sex differences, this too
is cause for concern and something worth pointing out.
Of course, it is also worth pointing out when scholars
underemphasize the sex differences. With that in mind,
we turn now to the next major section of the article.
Competing Theories
In this section, we consider two competitors to the
evolutionary psychological account of human sex dif-
ferences. These are Wood and Eagly’s Social Role
Theory and L. C. Miller et al.s Attachment Fertility
Theory. We start with Social Role Theory.
Social Role Theory
Stewart-Williams and Thomas’s mischaracterization
of our theory as invoking nurture alone apparently
arises from their inability to reason about nature and
nurture as interacting influences.
—Wood and Eagly (this issue, p. 243)
The gloves came off with Wood and Eagly’s com-
mentary! Wood and Eagly are the originators of a the-
ory of sex differences that stands in opposition to EP.
This is known variously as Social Role Theory, the
biosocial extension of Social Role Theory, and, more
recently, biosocial construction theory. We called it So-
cial Role Theory (SRT) in the target article, and will
continue to do so here. According to SRT, the contem-
porary division of labor in a society, and the associated
social roles, are the primary cause of the psychological
sex differences and similarities in our species. Ancient
selection pressures have little or nothing to do with it.
In our article, we provided several arguments against
this view. Wood and Eagly’s response was to declare
that we misunderstood their theory and to argue that,
contrary to our conclusion, SRT can explain the full
panoply of human sex differences. We disagree on both
of these points.
Is SRT a blank slate theory? To begin with,
Wood and Eagly (this issue) suggested that we misrep-
resented SRT by lumping it together with blank slate,
Nurture-Only theories of human psychology. They re-
ject such a designation, arguing that SRT is instead a
nature–nurture interactionist theory that “integrates the
broad classes of biological and social-cultural factors”
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
(p. 241). EP, in contrast—in their estimation—adheres
to a one-sided, Nature-Only view of sex differences.
(Ironically, evolutionary psychologists also describe
their approach as a nature–nurture interactionist the-
ory and claim that their competitors adhere to a one-
sided, Nurture-Only view; see, e.g., Tooby & Cos-
mides, 1992.)
Is SRT a blank slate theory? Our answer is: yes
and no. On the one hand, Wood and Eagly incorporate
some evolved sex differences into their model. As we
noted in the article, however, these are confined almost
entirely to physical differences—the differences in
height, strength, speed, and the ability to bear and
nurse offspring. According to SRT, these physical
differences help to determine the allocation of men and
women to different social roles, and these social roles
then shape people’s psychologies. In this way, evolved
physical sex differences indirectly help shape the psy-
chological sex differences we observe in the everyday
world. Based on these ideas, many have concluded that
SRT acknowledges evolved sex differences in men and
women’s bodies but retains a blank slate conception
of the human mind. Is this interpretation correct?
Again, the answer is yes and no. Wood and Eagly
accept that the mind has some built-in structure,
including the neural and hormonal systems underlying
“basic perceptual, sensory, and motivational systems”
(p. 241). As such, they don’t posit a blank slate mind
in the sense of a mind devoid of any native structure
(see also Eagly & Wood, 1999, p. 409). But for present
purposes, the question is, Do they accept that there are
any inherited psychological sex differences? If not,
they maintain a blank slate view with respect to sex
The overwhelming impression one gets from their
work is that that is exactly the view they maintain.
So, for instance, in a recent, detailed summation of
their theory, Wood and Eagly (2012) presented a fig-
ure outlining the key features of SRT (p. 58). This
included two broad sets of influences shaping the di-
vision of labor and thus the social roles of a society:
(1) cultural, economic, and ecological factors, and (2)
evolved physical sex differences. These were the only
factors they thought important enough to include. Cer-
tainly, if you look at the small print, Wood and Eagly
don’t entirely rule out inherited psychological differ-
ences. They note, for instance, that differences in early
temperament might help shape the division of labor,
perhaps because they make it easier to socialize boys
versus girls into certain roles (Wood & Eagly, 2012).
But as they note in their commentary, they reserve judg-
ment on both the origins and implications of tempera-
mental differences, and in their final assessment, “any
inherited psychological sex differences are not of over-
riding importance” (p. 243; see also Eagly & Wood,
1999, p. 414). We conclude, therefore, that when it
comes to the ultimate origins of sex differences, SRT
must be classed either as a blank slate theory or as a
virtually-blank-slate theory. If we’re wrong about this,
and Wood and Eagly accept that there are important
inherited psychological sex differences in our species,
they should directly and unequivocally say so.
Until that time, we reject the claim that we mis-
represented SRT. We also reject the claim that SRT
represents a genuinely integrative biosocial theory of
human sex differences. Granted, Wood and Eagly in-
clude evolved physical differences in their model, and
they describe how neurohormonal systems might be re-
cruited to enact social roles. But that’s all the bio they
include in their supposedly biosocial theory. There is
certainly no evolutionary bio. A truly integrative model
of sex differences would weave together sociobiologi-
cal theories and sociocultural theories. SRT simply re-
jects the sociobiological theories. Indeed, to a large ex-
tent, the theory is a reaction against those theories—an
attempt to frame an alternative explanation for the find-
ings of EP (see, e.g., Wood & Eagly, 2012, p. 102).
It seems reasonable to conclude, then, that Wood and
Eagly’s biosocial theory is essentially a traditional so-
cial theory disguised as a biosocial theory.
Our arguments against SRT. Wood and Eagly
devote a significant chunk of their commentary to re-
butting our arguments against SRT. Before determining
whether their efforts succeed, let us make clear our po-
sition on the evolutionary versus social causes of sex
differences in our species. Wood and Eagly assume
that we hold a Nature-Only view. This can be seen
in the title of their commentary: “Biology or Culture
Alone Cannot Account for Human Sex Differences
and Similarities.” Our response to this statement is: Of
course! Although we think that the evolutionary contri-
butions are sometimes more influential, and although
we focused on these in the article, we do not deny the
importance of social forces or cultural evolution (see,
e.g., Stewart-Williams, in press-a, in press-b). We do
not think that “inherited biological factors thoroughly
constrain sociocultural influences on men and women,
as Wood and Eagly charge (p. 242). On the contrary,
we imagine it would be possible in principle to set up
a society in which men and women were identical on
average, or in which the normal sex differences were
reversed, or in which the normal sex differences were
strongly exaggerated. (Indeed, the latter can arguably
be observed already in strongly patriarchal societies.)
We suggest, however, that it would take extreme so-
cial interventions to create and maintain societies like
To be fair, in some cases these explanations are plausible. Con-
sider, for instance, the fact that, on average, women rate earning
capacity as more important in a mate than do men (Buss, 1989).
Eagly and Wood (1999) showed that this gender difference is
larger in nations in which women have less independent access to
wealth, suggesting that the difference owes less to evolution than to
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
these, because they would go so strongly against the
grain of human nature. In the absence of continuing
social pressure, these societies would drift back to the
usual pattern of modest sex differences. That, in a nut-
shell, is how we conceptualize the interaction of nature
and nurture. Consistent with this view, our claim in the
article was not that social roles play no part in deter-
mining sex differences, but rather that they could not
be the complete explanation for the differences found
in human beings.
We provided several arguments in support of this
conclusion. First, we argued that certain sex differ-
ences found in our species—the differences in size,
physical aggression, sex drive, interest in multiple sex-
ual partners, parental inclinations, timing of puberty,
and life expectancy—are found in a very wide range of
species, and thus that the primary explanation for these
differences in humans (including the psychological dif-
ferences) is likely to be evolutionary forces common
to many species rather than social forces unique to our
own. Wood and Eagly basically ignored this argument.
They did argue that size dimorphism alone is not di-
agnostic of a species’ evolved mating pattern, but we
never suggested that it was. Our argument concerned
the cross-species trend across multiple traits and its im-
plications for the origins of a specified set of psycho-
logical sex differences seen in human beings. Given
that sociobiological theories offer an explanation for
the cross-species trend, whereas SRT does not, it would
be rash to accept Wood and Eagly’s conclusion that so-
ciobiological theories are “outmoded” or “outdated.
Wood and Eagly did address our next argument, but
they badly mangled it. The argument was this: Humans
evolved from an earlier species possessing inherited
sex differences in traits like aggression, parental in-
clination, and sex drive. SRT maintains that humans
no longer possess inherited sex differences in these or
other traits. This implies that, somewhere in our evo-
lutionary past, those inherited differences must have
been eliminated by natural selection, which in turn im-
plies that they must have become maladaptive—that is,
that they must have been selected against, relative to a
more flexible, less sexually differentiated psychology.
Wood and Eagly mounted a very poor response to this
argument. They first observed that they never said that
the sex differences became maladaptive or were re-
moved. And that’s true! We weren’t summarizing their
theory at that point; we were identifying an implication
of their theory that they have not adequately dealt with.
Wood and Eagly then denied that natural selection re-
moved any adaptations (e.g., neurohormonal systems)
shared with other animals. But we never claimed that
SRT implies that natural selection removed any adap-
tations. We claimed that it implies that natural selection
removed any inherited psychological sex differences.
This would almost certainly not have involved remov-
ing adaptations or neurohormonal systems shared with
other animals. Most likely, it would have involved tin-
kering with the neurohormonal systems directly, or it
would have involved the evolution of regulation by
higher order systems. Either way, SRT seems to imply
that natural selection must have nullified any preexist-
ing, inherited sex differences in traits like aggression,
parental inclination, and sex drive.
As we see it, there are two ways that Wood and
Eagly can respond to this argument. First, they can deny
that natural selection removed any inherited psycho-
logical differences—in which case, they agree with us
that those differences are still found in humans. Alter-
natively, they can deny that humans possess inherited
psychological sex differences—in which case, they are
committed to the view that natural selection removed
the differences (and, to reiterate our earlier conclusion,
they would then hold a blank slate view of human sex
differences). If they take the latter option (as we think
they must if they wish to preserve SRT), they then
need to explain why it happened. Why were these dif-
ferences suddenly selected against? It is not enough to
engage in vague hand waving about cumulative culture
and human adaptedness “not primarily to particular en-
vironmental features, but to variation itself” (Wood &
Eagly, this issue, p. 241). We agree that no other animal
possesses the kind of cumulative culture found in our
species. But many animals—from chimpanzees to ca-
puchin monkeys to sperm whales—engage in complex
social learning despite still possessing inherited sex
differences. These things are not mutually exclusive.
And certainly the dual-inheritance/gene-culture coevo-
lutionary theorists that Wood and Eagly align them-
selves with—Boyd and Richerson, Henrich, Laland,
Mesoudi—do not deny that humans possess evolved
psychological sex differences. For all these reasons,
we stand by our initial claim that Wood and Eagly
have yet to explain why natural selection might have
wiped out the male–female differences found in our
prehuman ancestors. In other words, they have yet to
square SRT with modern evolutionary theory.
Other problems with SRT. Wood and Eagly’s
commentary helped crystallize various other misgiv-
ings we have about their theory. First, although we
agree that a society’s roles and rules have some impact
on the sex differences found in that society, we suggest
that their impact is considerably stronger for overt be-
havior than for internal psychology. So, for instance, in
nations where women but not men can face execution
for adultery or premarital sex, there is likely to be a
large sex difference in these behaviors—not because
women have radically different sexual desires in those
nations, but because people generally inhibit behavior
that is likely to get them killed. Social roles no doubt
influence our desires as well; however, their impact on
desires is presumably weaker than their impact on be-
havior. Why else would women’s sexual behavior be
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
so strictly policed in these societies? More generally,
although people usually channel their behavior into the
framework of the local social roles, these roles may fit
their native psychologies to a greater or lesser degree.
Second, within the framework of SRT, the causal ar-
row only ever points in one direction: from social roles
to psychology. Wood and Eagly (this issue) suggest, for
instance, that the “division of labor yields the familiar
psychologies of women and men” (p. 242) but not that
the influence might also go the other way. This, we
suggest, is a rather extreme view. It seems much more
plausible to us that it’s a two-way street—that social
roles influence psychology but that psychology also
influences social roles. We can frame this in terms of
cultural evolution: Just as our tools evolved culturally
to fit our hands, so too our social roles evolved cultur-
ally to fit persisting aspects of the human mind. Roles
that jar too violently with human nature are unlikely
to persist for long, at least without the application of
significant social force. If this is correct, it raises the
possibility that some social roles might have evolved
culturally to fit traits that, although found in both sexes,
are more common in one than the other. This is em-
phatically not to say that there are some male roles and
some female roles. But it is to suggest that there might
be some social roles that suit more men than women,
and others that suit more women than men—not just
because of evolved physical differences but because of
evolved psychological differences as well. To be clear,
we are not proposing that socialization has no part to
play in fitting people to the social roles of their society.
Our proposal is much more modest: that socialization
is not the only contributing factor.
Third, Wood and Eagly attribute a huge amount of
power to social roles. Social roles are not merely one
influence among many; they are the primary determi-
nant of sex differences and similarities in our species
(Wood & Eagly, 2012). But various lines of evidence
point to the falsity of this conclusion. This includes re-
search on people who are born as members of one sex
but who are raised as members of the other. Such cases
effectively pit biology against social roles. Our reading
of the literature is that the outcome is often a compro-
mise, and that biology sometimes even trumps social
influences. Consider, for example, the famous—and
very sad—case of David Reimer (Colapinto, 2006; Di-
amond & Sigmundson, 1997). As an infant, Reimer’s
penis was mutilated in a botched circumcision. Un-
der the guidance of sexologist John Money, it was de-
cided that he should be surgically castrated and raised
as a girl. This was based on the assumption that sex
role socialization is all-important, and thus that, with
the appropriate socialization (plus a steady diet of fe-
male hormones), Reimer would turn out just like any
other girl. Early reports suggested that he was indeed
a normal, well-adjusted girl, just as SRT might pre-
dict (Money, 1975). For several decades, the case was
hailed as proof positive of the nurture theory. It was
eventually revealed, though, that Reimer had been a
deeply unhappy child, who constantly complained that
he felt like a boy. As a preschooler, he ripped off his
dresses and refused to play with dolls, opting instead
for guns and other stereotypically boy toys. At school,
he got into fist fights and insisted on standing while
urinating. Finally, at the age of 14, his parents told him
the full story and Reimer started living as a male. He
stopped taking the hormones, had an artificial penis
constructed, and eventually married a woman. Sadly,
the story ends with Reimer’s suicide at age 38.
Now this is a single case, and there is some de-
bate about what exactly we can conclude from it (see,
e.g., Jordan-Young, 2010). But other cases point in
the same direction. This includes several dozen cases
of otherwise-normal genetic males born with genital
anomalies (e.g., micropenises), who were surgically
reassigned as female while still infants. The psychia-
trist William Reiner (2004; Reiner & Gearhart, 2004)
reported that, despite being raised as girls, by adulthood
a majority of these individuals identified as males. Fur-
thermore, virtually all of them—including those who
identified as females—had typically masculine inter-
ests and preferences (see also Bradley, Oliver, Cher-
nick, & Zucker, 1998). These cases, like the Reimer
case, challenge the idea that social roles are the pri-
mary determinant of human sex differences. Evolution
matters too.
(Note that these cases also challenge Wood and
Eagly’s conception of the relationship between na-
ture and nurture. In their view, “nature and nurture are
not competing influences but instead are interacting
forces” (p. 242). The Reimer case and its kin suggest
that this is false: Nature and nurture may often work
together harmoniously, but they can also sometimes
come into conflict.)
Are human sex differences generally small?
Wood and Eagly (this issue) rejected our suggestion
that psychological sex differences in Homo sapiens
are generally rather modest and that EP sometimes ex-
aggerates those differences. They begin by stating that
we misconstrued the usual EP position on the issue,
which is that these differences are found only in do-
mains where men and women faced recurrently differ-
ent adaptive challenges (Buss’s well-known metathe-
ory of sex differences; Buss, 1995). We were surprised
by this argument. First, we were surprised that anyone
who read the article could take the impression that, in
our opinion, EP claims that sex differences are uni-
formly large across all domains. Aside from anything
else, we explicitly discussed the EP finding that the
differences tend to be larger in a short-term than a
long-term mating context (pp. 151, 157). More to the
point, though, the fact that EP does not claim large
sex differences across the board does not imply in any
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
way that evolutionary psychologists never exaggerate
the magnitude of particular sex differences. This ar-
gument can be safely ignored.
Next, Wood and Eagly argued that, contrary to our
view that sex differences are generally quite modest,
the differences are actually highly variable. We were a
little surprised by this argument too. As noted, Eagly
(1995) has written elsewhere that most sex differences
(and indeed most effects in psychology) are small or
medium in size and that only a few are large—which
is essentially our position (see also Hyde, 2005). Still,
perhaps Wood and Eagly object to our further sugges-
tion that even some of the larger differences (such as
the sex difference in SO) are not so large after all, and
that most differences fall within a range that can be
described as monomorphic. In their commentary, they
provide an example of the variability of sex differences,
so let’s see whether it contradicts our position. The ex-
ample concerns sex differences in different forms of
risk taking (Byrnes et al., 1999). For social risk taking,
they note, the sexes barely differ. In contrast, for risky
games involving physical skills, there is a sex differ-
ence with an effect size of d = .43: a medium effect
according to Cohen (1988). Finally, when it comes to
life-threatening acts of heroism, there is a very large
difference: Around 91% of the recipients of Carnegie
Hero medals are men (Becker & Eagly, 2004; see also
Johnson, 1996).
Does this variability undermine our claim that sex
differences tend to be modest? It does not. First, al-
though a d value of .43 constitutes a medium effect ac-
cording to Cohen’s arbitrary standard, it is only around
one fourth the magnitude of the modest human sex dif-
ference in height and only around half the magnitude of
the sex difference in size in the monomorphic gibbon.
It therefore fits easily with our generalization that most
human sex differences fall within the monomorphic
range (see The Effect Size Argument, pp. 249–251).
As for the Carnegie Hero medalists, it is probable that
these individuals come disproportionately from the ex-
treme right-hand tail of the distribution for risk taking.
As we noted several times in the target article, sex dif-
ferences are much larger at the tail than they are nearer
the mean, and thus data derived from the tail give an
inflated impression of the difference in the species as
a whole (see The Tail-of-the-Distribution Argument,
pp. 251–252). Thus, Wood and Eagly’s example is
perfectly consistent with the view that our hominin
ancestors underwent a general decline in psychologi-
cal dimorphism, and thus that most psychological sex
differences are fairly modest.
Can SRT explain variability in sex differences?
Wood and Eagly argue, quite rightly, that any complete
theory of sex differences must explain the variability
we see in the magnitude of these differences across
cultures and historical epochs. SRT, they suggest, does
exactly that. According to the theory, sex differences
and similarities are products of the social roles of the
society in which they’re found. The more differentiated
these roles are by sex, the larger the differences will be.
This leads to two sets of predictions. First, in nations
with high levels of gender equity, most sex differences
will be smaller than they are in nations with low levels
of gender equity. Second, as women’s status increased
in the West over the last century or so, sex differences
will have shrunk. Wood and Eagly (2012, this issue)
summarized various lines of evidence consistent with
these predictions.
Once again, we accept that social roles have an im-
pact on the minds of men and women, and thus on
the magnitude of any sex differences. We are not con-
vinced, however, that SRT provides a complete and
comprehensive explanation for the pattern of similar-
ities and differences across cultures and times. As we
noted in the target article, there is evidence that some
sex differences are actually larger in societies with
greater gender equity, rather than smaller (Schmitt,
2012; although see Harris, this issue; Wood & Eagly,
2012, this issue). Even leaving this aside, though, the
evidence that Wood and Eagly cite does not defini-
tively settle the question in favor of SRT. Consider, for
instance, the fact that, as women’s status has increased
in the Western world, women have become more agen-
tic and sexually adventurous, and thus that the sex
differences in these domains have become smaller (Pe-
terson & Hyde, 2010; Wood & Eagly, 2012). This is
consistent both with a pure SRT explanation and with
an EP/SRT hybrid explanation. The EP/SRT hybrid ex-
planation (which we favor) would be that women in the
West 50 or 100 years ago were less agentic and less
sexually adventurous because their natural tendencies
were being suppressed. As society has become freer,
these tendencies have emerged to a greater extent, and
thus women are now more similar to men in these traits.
The pure SRT explanation would be that women today
and women 50 or 100 years ago were simply socialized
into different roles and that neither role went against
the grain any more or any less than the other—after all,
within the framework of SRT, there is no grain; all the
normal-range variation that we see is attributable to the
social roles of the period. As we’ve made clear, the first
option seems much more plausible to us. However, the
point of the example is to show that evidence of smaller
sex differences in more gender-equitable societies or
historical epochs does not provide unambiguous sup-
port for SRT, as it fails to rule out significant evolved
psychological differences.
In sum, Wood and Eagly (this issue) argued that
we misrepresented their position; proffer an outdated,
one-sided view of sex differences; and fail to explain
important data. We argue that this has things precisely
back to front: They misrepresented our position; they
proffer an outdated, one-sided view of sex differences;
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
and they fail to explain important data. Our final ver-
dict on SRT is that, for some traits, social roles may
exert a significant influence on the magnitude of sex
differences, especially in overt behavior. However, so-
cial roles are merely one influence among many, and
SRT cannot replace EP as an explanation for the sex
differences we share with other animals.
Attachment Fertility Theory
Another challenge to our EP perspective came from
L. C. Miller et al. (this issue). The topic of their piece
was Attachment Fertility Theory (AFT; L. C. Miller
& Fishkin, 1997; L. C. Miller, Putcha-Bhagavatula,
& Pedersen, 2002; Pedersen, Putcha-Bhagavatula, &
Miller, 2011). As the name suggests, AFT is a theory
of human mating psychology based principally on at-
tachment theory. It has several things in common with
our approach. Specifically, it stresses the centrality of
pair bonding and biparental care, and it argues that
EP sometimes overstates the sex differences in human
sexuality. As will soon become clear, however, there
are also fundamental differences between the two ap-
proaches, and we argue that, on balance, AFT is not
a credible theory. In this section, we address several
issues raised by Miller et al.s commentary, including
the claim that we misconstrued their theory, that there
are negligible or no evolved sex differences in short-
term mating inclinations, and that recent research on
the sexual practices of men who have sex with men
provides unique support for AFT. Before going any
further, though, we need to address a peculiar allega-
tion in Miller et al.s commentary, namely, that we took
an argument from one of their publications but failed
to give them proper credit for it.
Credit where credit’s due? In the target article,
we summarized several lines of physiological evidence
suggesting that human beings are not an exclusively
promiscuous species, and that we probably had com-
paratively low levels of sperm competition through-
out our evolutionary history (for reviews, see Dixson,
1998; Gray & Anderson, 2010). In an earlier article,
L. C. Miller et al. (2002) surveyed some of the same re-
search and reached a similar (though not identical) con-
clusion. In that article, they provided some useful data
on relative testis size in humans and various other pri-
mates. We used those data in our article, and cited their
article accordingly (along with three other AFT arti-
cles). Apparently, though, Miller and colleagues think
they deserve more credit than this. In their commen-
tary, they say the following:
Stewart-Williams and Thomas’s position and review
covers very similar ground; in fact, they use the cal-
culations for relative testicle size (we computed) that
were part of our earlier review (the reader will note
the citation for that—but not the argument—in the
footnotes). (p. 214)
This has the clear implication that we simply
rewrote their summary of the research and cribbed their
argument. This allegation is false. First, we actually
made a rather different argument than Miller et al. Our
argument was that, although humans are adapted to en-
gage in some short-term mating, the physiological evi-
dence shows that this is not our primary evolved mating
pattern. They, in contrast, argued that humans are not
adapted for short-term mating at all—that according
to primatologists, we are “classified as long-term and
not short-term maters” (L. C. Miller et al., this issue,
p. 213). Second, the basic argument is not original
with L. C. Miller et al. (2002). Evolutionary biolo-
gists have debated for many decades whether human
physiological features place us among primates with
a long-term or a short-term mating system (Dixson,
1998; Gomendio, Harcourt, & Rold
an, 1998; Har-
court, Harvey, Larson, & Short, 1981; Short, 1979).
L. C. Miller et al. (2002) simply provided a brief sum-
mary (three paragraphs) of one position in this debate
(pp. 88–89). This does not make the argument theirs.
Third, we dispute the claim that we covered “very sim-
ilar ground” to them, at least in any sense that implies
wrongdoing. Admittedly, we and they both cite Dixson
(1998) and Nunn, Gittleman, and Antonovics (2000).
But that’s not because we regurgitated Miller et al.s
three-paragraph overview; it’s because we and they
were working from the same primary literature, and
these are seminal publications in the area. Note also
that we cited many studies that Miller et al. did not
cite (e.g., Gomendio et al., 1998; G. F. Miller, Tybur,
& Jordan, 2007; Short, 1979; Thornhill & Gangestad,
2008). There’s a simple explanation for the fact that
we didn’t credit Miller with the argument: We didn’t
get the argument from Miller. The allegation is false.
Has AFT said it all before? With that out of the
way, let’s move on to other matters. A common theme
in Miller et al.s commentary is that AFT has already
staked out the position we argued for in the target
article. The authors note, for instance, that “We find
ourselves in general agreement with the above basic
arguments. But, then again, as just suggested, many of
these and/or related arguments were made earlier in
AFT” (p. 212). We agree that there are several areas
of overlap between our work and theirs. For instance,
AFT proposes, as do we, that human beings evolved
to engage in high levels of pair bonding and biparental
care and that we are less sexually dimorphic than EP
sometimes implies. We suggest, however, that the over-
lap exists almost entirely at this very general level of
analysis. When we look at the nitty-gritty of the two
approaches, there are large and important differences
between them.
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
One difference concerns the theoretical basis of the
two positions. AFT’s approach to sex differences is
grounded in attachment theory (see L. C. Miller et al.,
2002, p. 89). Ours, in contrast, is grounded in Trivers’s
parental investment theory. Although these two theo-
ries are potentially reconcilable (Eastwick, this issue),
AFT theorists have not attempted to broker a recon-
ciliation. Indeed, they appear to eschew parental in-
vestment theory, suggesting (falsely, in our view) that
“Primatologists have argued that Trivers’s theory does
not apply well to primates” (L. C. Miller et al., 2002,
p. 89). Thus, the theoretical underpinnings of the two
approaches are fundamentally different.
A second difference, which we discuss in more de-
tail later, relates to AFT’s suggestion that short-term
mating is a by-product of the human pair bonding
system rather than an adaptation in and of itself (as
EP holds). As L. C. Miller, Pedersen, and Putcha-
Bhagavatula (2005) put it, AFT argues for universal,
sex-similar, evolved mechanisms leading up to and
affording pair-bonding. These could also quite natu-
rally ...produce short-term and other types of dating
as by-products (p. 290, emphasis added). This view of
short-term mating is arguably what distinguishes AFT
most clearly from other evolutionary theories of hu-
man sexuality. After all, most other theories argue, like
AFT, that human beings have evolved to form long-
term pair bonds (see Schmitt, Shackelford, & Buss,
2001, p. 231). It is AFT’s stance on short-term mat-
ing that differentiates it from other evolutionary ap-
proaches, our own included.
Finally, although both we and Miller et al. sug-
gest that sex differences in sexuality are smaller than
EP often implies, our evidence and arguments for this
conclusion are largely distinct from theirs (see, e.g.,
the effect size and tail-of-the-distribution arguments;
our dissection of particular studies such as Clark and
Hatfield, 1989). Moreover, AFT takes the whole line
of argument much further than we do, positing not
merely modest sex differences but negligible or even
no differences. As we discuss below, this is no longer
a credible position. In sum, L. C. Miller et al. (this
issue) suggest that we “make arguments very much
like those made previously by AFT” (p. 219). How-
ever, they overstate the similarities between our ap-
proach and theirs, and thus they overstate the extent to
which they have already put forward the position we
If we misrepresented AFT, so did its authors.
Next, Miller et al. argue that we misrepresented their
theory in a number of ways. They quote us as saying,
for instance, that according to AFT, “long-term pair
bonding is our solitary evolved mating pattern” and
that “short-term mating is merely a non-adaptive or
maladaptive by-product of these mechanisms operat-
ing in evolutionarily-novel conditions” (p. 151). They
agree with none of this. They argue, for instance, that
contrary to the view that long-term pair bonding is
our solitary evolved mating pattern, AFT can explain
“not only pair-bonding as a mating outcome, but ev-
ery other type of mating outcome” (p. 217), including
various forms of short-term mating. They also point
out that they never suggested that short-term mating
is nonadaptive or maladaptive. On the contrary, they
maintain that it can sometimes be adaptive for individ-
uals, perhaps functioning as a “secondary alternative
adaptive strategy” (p. 215).
We have considered these objections carefully, and
for the most part we stand by our initial description.
To some extent, we and Miller et al. may be talking at
cross-purposes. So, for instance, when we suggested
that, according to AFT, long-term mating is humans’
solitary evolved mating pattern, we were not suggest-
ing that AFT had no explanation for short-term mating.
Indeed, we outlined their explanation in the target ar-
ticle. Our point was that, according to AFT, short-term
mating is a by-product, rather than an adaptation in
and of itself, and thus that long-term mating is the one
mating outcome that the human attachment system has
been specifically designed to produce (see, e.g., L. C.
Miller & Fishkin, 1997, pp. 202, 228; L. C. Miller
et al., 2005, p. 290). As for our suggestion that, ac-
cording to AFT, short-term mating is nonadaptive or
maladaptive, that is a simple implication of the claim
that it is a by-product rather than an adaptation. AFT
theorists might not use the terms “non-adaptive” or
“maladaptive” themselves; however, that is what is im-
plied by their use of the evolutionary biologists’ term
But what about the claim that, rather than being
maladaptive or nonadaptive, short-term mating can ac-
tually be an adaptive alternative strategy for some indi-
viduals? This, we suggest, points to an area of confu-
sion within AFT. First, it is not clear that the adaptive-
strategy hypothesis is consistent with AFT’s claim that
short-term mating is “a ‘fallout’ of a failure to in-
terface with human’s adapted for social environment
(e.g., responsive paternal and maternal caregivers)”
(L. C. Miller & Fishkin, 1997, p. 228). More impor-
tant, though, it is not exactly clear what Miller and
colleagues mean by “adaptive. Do they mean that
short-term mating is adaptive in the strict technical
sense used by biologists—in other words, that engag-
ing in short-term mating led to higher average num-
ber of grandchildren in ancestral environments? Or
do they just mean that it is adaptive in the everyday
sense of the word—for example, emotionally or psy-
chologically beneficial? If they mean the latter, then
that does not contradict our assertion that, within the
AFT framework, short-term mating is a nonadaptive or
maladaptive by-product; a trait can be adaptive in the
everyday sense without being a biological adaptation.
However, if they mean the former—that short-term
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
mating is a distinct secondary strategy crafted by nat-
ural selection—then that would contradict their claim
that short-term mating is a by-product: an “emergent
outcome” of an attachment system designed, ideally,
to produce long-term relationships. It would imply
that humans, rather than being long-term maters (as
AFT has long claimed), instead have a variable mating
system incorporating both long-term and short-term
mating—the standard EP position (Buss & Schmitt,
1993). Thus, assuming that Miller et al. are using the
terms “adaptive” and “by-product” correctly, there is a
deep contradiction within AFT.
Are there sex differences in short-term mating?
As noted, one of the key distinctions between AFT
and EP relates to their stance on short-term mating.
AFT claims (or usually claims) that short-term mating
is a by-product of sex-similar long-term attachment
mechanisms. It is not an adaptation in and of itself
but instead “falls out” of the “system of mechanisms
leading up to, supporting, and enabling” pair-bonding
(L. C. Miller et al., this issue, p. 215). In the target ar-
ticle, we laid out an argument against this view. It goes
like this. There is good evidence that the sex difference
in short-term mate seeking is larger than the difference
in long-term seeking (Buss & Schmitt, 1993; Kenrick
et al., 1993). This pattern is even reported in at least one
AFT publication (L. C. Miller & Fishkin, 1997, p. 225,
Figure 8.4). But the pattern poses a unique difficulty
for AFT. If short-term mating is merely a by-product
of sex-similar attachment systems, why is the sex dif-
ference in short-term mating larger than the difference
in long-term mating? AFT has a hard time answering
this question. EP, in contrast, explains it effortlessly in
terms of average discrepancies in parental investment
in our ancestral past. This, we argue, provides a good
reason to favor EP over AFT.
L. C. Miller et al.s (this issue) main response to this
challenge was to deny that there is anything to explain:
The difference isn’t really there. “AFT,” they note, “ar-
gues that there is not credible evidence that humans
evolved sex-distinct short-term mating mechanisms”
(p. 214). If by “sex-distinct” mechanisms, they mean
two separate mechanisms, one found in men and one
found in women, then we agree—and so would all com-
petent evolutionary psychologists. But if they’re also
denying that there is an average difference in interest
in casual sex, then we strongly disagree—and so does
almost all the research on this topic. There is now a sub-
stantial body of work demonstrating this sex difference
in diverse samples across multiple nations (e.g., Buss
& Schmitt, 1993; Li & Kenrick, 2006; Schmitt, 2005;
Schmitt & 118 Members of the International Sexuality
Description Project, 2003). Given the imposing weight
of this research, and its consistency with the pattern
observed elsewhere in the animal kingdom, we would
need a very strong reason indeed to reject this finding.
Miller et al. cite various studies they suggest pro-
vide just such a reason. This includes their own re-
search on the topic (L. C. Miller & Fishkin, 1997;
Pedersen, Miller, Putcha-Bhagavatula, & Yang, 2002;
Pedersen et al., 2011). Most of this is aimed at debunk-
ing Buss and Schmitt’s (1993) number-of-partners-
desired measure (i.e., “How many different partners
would you ideally like to have sexual intercourse with
in the next 1 month, 1 year, 30 years, etc.?”) Miller
and colleagues argue that the data from this measure
are positively skewed, and thus that the appropriate
index of central tendency is the median rather than
the mean. When comparing medians, however, there
are typically no differences in the number of sex-
ual partners desired by women versus men—at least
when using the “appropriate” statistical analyses (L. C.
Miller & Fishkin, 1997; Pedersen et al., 2002; Peder-
sen et al., 2011). There are several things to say about
this. First, the number-of-partners-desired measure is
only one measure among many examining men and
women’s propensity to engage in casual sex and seek
sexual variety. Some of these measures, such as the
Sociosexual Orientation Inventory–Revised, produce
data that are not especially skewed, but nonetheless
detect the usual sex difference (Penke & Asendorpf,
2008). Second, Schmitt and 118 Members of the In-
ternational Sexuality Description Project (2003) con-
ducted a large-scale, cross-national study using the
number-of-partners-desired measure. Whereas AFT
studies typically involve several hundred American un-
dergraduates, this study included more than 16,000 par-
ticipants from more than 50 nations. To deal with the
potential problems associated with the use of means,
Schmitt and colleagues analyzed the data using various
nonparametric tests. As hypothesized, they found the
standard sex difference in every one of their 10 major
world regions.
Overall, attempts by Miller and others to make
the sex difference in short-term mating go away have
been unsuccessful (see also Conley, Moors, Matsick,
Ziegler, & Valentine, 2011, and response by Schmitt
et al., 2012). One might argue that their continued
efforts to do so make them examples of what Ken-
rick (this issue) called “the social science equivalent of
climate-change deniers” (p. 204).
Do MSM seeking UAI support AFT over EP?
Finally, L. C. Miller et al. (this issue) presented new
data which they suggest support AFT over its rivals.
The data came from a large sample of men who have
sex with men (MSM), all of whom were currently in
long-term relationships. There were two key findings:
First, men who were less emotionally close to their pri-
mary partners had more casual sex partners; and sec-
ond, these men engaged in higher rates of unprotected
anal intercourse (UAI). According to Miller et al., this
supports their view that humans are long-term pair
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
bonders and that short-term mating occurs mainly in
the context of faltering pair bonds.
The findings are interesting and potentially very
useful, and we agree that they are consistent with
AFT. However, we don’t believe that they are uniquely
consistent with AFT. They fit equally well with an
EP/MMC perspective. In his fascinating commentary,
Eastwick (this issue) discusses research on the deroga-
tion of alternatives (see, e.g., Maner, Rouby, & Gon-
zaga, 2008; Simpson, Gangestad, & Lerma, 1990).
This research shows that when people are involved
in a strong pair-bonded relationship, they often find
other potential romantic partners less alluring. This
applies not only to women but to men as well. Ac-
cording to Eastwick, the pattern suggests that men’s
short-term mating psychology is often deactivated in
the context of a strong pair bond, which he observes
fits well with the MMC model. The data that Miller and
colleagues present documents a related phenomenon:
MSM involved in strong pair bonds have fewer ex-
tracurricular partners. This is consistent with the AFT
view that short-term mating is a by-product. However,
it is equally consistent with the view that short-term
mating is an adaptation, but one that recedes into the
background when people are involved in a long-term
relationship. Thus, Miller et al.s data do not support
AFT over other theories of human mating psychology.
Taken together with the fact that AFT struggles to ex-
plain well-documented sex differences in sexual psy-
chology, we suggest that this theory should probably
be given little credence.
On the Teaching of Evolutionary Psychology
At this basic pedagogical level, [Stewart-Williams and
Thomas] are correct that EP over-emphasizes MCFC
logic and sex differences.
—G. F. Miller (this issue, p. 207).
Scientists carrying out primary empirical research—
and other academics reviewing and discussing this
work—should probably strive for greater discipline
when describing sex differences, particularly in com-
munication with lay audiences.
—Pound and Price (this issue, p. 228).
Before wrapping up, we would like to say a few
words about something close to our hearts: the teach-
ing of EP. In the target article, we argued that the data
gathered by evolutionary psychologists generally show
small sex differences but that the verbal descriptions of
these data often suggest that the differences are large.
By the time the research bubbles up to the level of pop-
ular discourse, the imagined differences have grown
larger still. This happens for a variety of reasons. As
several commentators noted, people have a natural
tendency to polarize the differences: to view a dif-
ference in two shades of gray as a black-and-white
difference (Kenrick; Pound & Price). On top of that,
the media often pounces on any sex differences that
EP happens to uncover, and qualifications, nuances,
and sex similarities are invariably lost in the process
(Buss; Kenrick; Pham et al.; Pound & Price). This all
raises the question of how we can present and teach
evolutionary psychological theories and findings in a
way that minimizes the misunderstandings. Pound and
Price made one proposal in their commentary; they
suggested that “One helpful antidote to these effects
may be, as suggested in the target article, comparative
analyses which demonstrate that many human sex dif-
ferences are in fact relatively small, compared to other
species and to our own perceptions and expectations”
(p. 227). Below, we suggest three additional tactics for
presenting EP research on sex differences in a way we
hope lessens the chances of inflating these differences
in the minds of students, nonspecialist audiences, and
perhaps even specialist audiences as well.
Avoiding Focusing Excessively on Differences
Imagine that a zoologist from Mars was sent to
Earth to study elephants, and that it had never seen
one before. Its initial observation upon seeing a herd
of elephants for the first time would presumably not
be: “Wow! On average, the males are somewhat larger
than the females!” It would be: “Wow! Those are large
animals!” A follow-up observation would be the av-
erage sex difference in size. However, this would be
a qualification to the initial observation—a peripheral
rather than a central claim about the morphology of ele-
phants. If, in its subsequent report, the Martian zoolo-
gist began by highlighting the sex difference and barely
mentioned that elephants are, first and foremost, large
animals, we should not be surprised if other Martians
got the wrong idea.
The distinction between central and peripheral
claims is applicable to many aspects of human sex-
uality. Consider, for instance, the emotion of jealousy.
Evolutionary psychologists place a strong emphasis on
sex differences in this domain. The standard claim is
that men are more worried by a partner’s sexual in-
fidelity than emotional infidelity, whereas women are
more worried by a partner’s emotional infidelity (e.g.,
Buunk, Angleitner, Oubaid, & Buss, 1996, p. 139).
As Harris (this issue) observed, however, such claims
overstate the actual differences. In studies employing
a forced-choice decision—for example, “Would you
be more upset by a partner’s sexual infidelity or emo-
tional infidelity?”—men are usually quite evenly split
in which option they choose (see, e.g., Buss, Larsen,
Westen, & Semmelroth, 1992; Buss, Shackelford, &
Kirkpatrick, 1999; Kuhle, 2011). Furthermore, de-
spite any small average sex differences in this area
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
(Pietrzak, Laird, Stevens, & Thompson, 2002; Sagarin
et al., 2012), the overwhelming trend is that most men
and most women are extremely upset by both sex-
ual and emotional infidelity (Lishner, Nguyen, Stocks,
& Zillmer, 2008). This suggests that the central EP
claim regarding jealousy should be “Human beings
evolved to experience jealousy in romantic relation-
ships” rather than “Men and women evolved different
patterns of jealousy. The latter statement is true but
should be considered a qualification to the former: a
peripheral rather than a central claim. To stress the
sex difference alone would be like observing that male
elephants are bigger than females while steadfastly ne-
glecting to mention that all adult elephants are large
compared to most terrestrial animals. It would almost
certainly foster an inaccurate view.
A similar analysis applies within the realm of mate
preferences. Several commentators pointed out that sex
differences in human mate preferences are generally
quite small (Eastwick, this issue; Eastwick, Luchies,
Finkel, & Hunt, in press; Snowdon, this issue). Con-
sider, for instance, the preference for physical attrac-
tiveness in a mate. In his international study of mate
preferences, Buss et al. (1990) had respondents rate
the importance of a mate’s looks on a 0-to-3 scale,
with the anchors irrelevant, desirable, important, and
indispensible. Collapsing across nations, the average
for both sexes fell between desirable and important.
The male average was close to important, whereas the
female average was right in the middle, leaving less
than half a point difference between the averages for
each sex. Likewise, a multinational survey of more than
200,000 men and women revealed only a moderate sex
difference in the preference for physical attractiveness
and a small-to-moderate difference in the preference
for facial attractiveness (Lippa, 2007). It seems, then,
that the sex difference in the preference for good looks
is rather modest (although see Kenrick, this issue; Li,
Bailey, & Kenrick, 2002). As such, the central claim
in EP should probably be “Human beings evolved to
put a fair amount of weight on good looks in a mate”
rather than “Men evolved to put more weight on good
looks than women.” Again, the latter statement is true
but potentially misleading. This sounds like a contra-
diction, but it is not; the statement is misleading if it is
given undue weight.
In addition to distinguishing central and peripheral
claims, it should always be emphasized that there are
important areas where sex differences in sexual psy-
chology are trivial or nonexistent. So, for instance,
of the 18 traits that Buss and colleagues (1990) exam-
ined in their cross-national investigation of mate prefer-
ences, the traits that people considered most important
in a mate were mutual love, dependability, emotional
stability/maturity, pleasing disposition, good health,
education/intelligence, sociability, and desire for home
and children. Most of these showed no consistent sex
differences. Similarly, the top-ranked traits in Lippa’s
(2007) study included intelligence, humor, honesty,
and kindness, and the sex differences for these traits
were small or negligible (see also Snowdon, this is-
sue). In other words, when it comes to the traits we
consider most important in a long-term mate, human
beings are largely monomorphic. This is one of the
most significant findings of these studies; however, it
is easily overlooked when the discussion becomes fix-
ated on traits that people consider less important but
where sex differences are found. By shining a spotlight
on these traits, we may create an inaccurate picture of
our species, even though the differences are real.Our
picture of human nature may be built on a foundation
of exceptions to the rule. The rule—the fact that males
and females in our species are surprisingly similar in
many ways—may be relegated to the background. By
taking genuine differences and then exaggerating their
importance, our picture of our evolved nature may be-
come a caricature: It may contain a recognizable grain
of truth but distort its object.
Presentation Strategies
Other practices could have the same effect. For
instance, a number of textbooks in EP divide long-
term mating preferences into women’s preferences
and men’s preferences (Buss, 2003, 2012; Gaulin &
McBurney, 2001). A natural interpretation would be
that women’s preferences are those that most women
possess but that most men do not, or that are important
to most women but not to most men. This is not the
case, however; women’s preferences are simply those
for which the average score for women is higher than
the average score for men, and vice versa. In some
cases, the trait in question may be reasonably impor-
tant to most men and most women (e.g., good looks);
in other cases, it may not be particularly important to
either (e.g., wealth for many people in the Western
world; Lippa, 2007). Taken together with the fact that
the differences are generally rather modest, and the
overlap between the sexes substantial, the practice of
dividing mate preferences into men’s and women’s is
likely to mislead. Indeed, given how prone people are
to stereotyping and thinking in dichotomous terms, it
would be surprising if it did not mislead people (see
Kenrick, this issue; Pound & Price, this issue).
A related point is that, in the teaching and report-
ing of findings in EP, there is often a tendency to start
by emphasizing the sex differences (perhaps because
these are more interesting and attention grabbing than
the similarities) and only later to mention that these
differences are generally quite small. This sequencing
of the material might foster an inflated impression of
the differences. To give a concrete example, the au-
thors of a widely used EP textbook noted that “Human
mate choice decisions reflect the essential differences
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
between males and females in terms of maximizing
fitness; females typically concentrate on the rearing
component of reproductive effort while males concen-
trate on the mating component” (Barrett, Dunbar, &
Lycett, 2002, p. 136). They did then qualify this state-
ment by noting that the differences are smaller in our
species than in most, as a result of male parental care.
However, they issued this brief qualification only after
first fixing the dichotomous vision of the sexes in the
readers’ minds. A better pedagogical strategy might
be to reverse the order of presentation: to start with
the fact that, as a species with biparental care, we are
relatively monomorphic in the allocation of mating ver-
sus parenting effort, and then to qualify this statement
by noting that, because there are average sex differ-
ences in parental investment and reproductive rates,
men exhibit a somewhat higher average level of mating
Keeping Sight of Within-Sex Variation
Last but not least, it is important not to lose sight of
individual differences. A danger in emphasizing mean
values for each sex is that these values may be projected
onto all or most normally developing men and women.
The mean may be treated as a description of the typical
group member, despite the fact that the majority of in-
dividuals fall above or below it. Psychologists do make
some effort to stress that means cannot be attributed to
all members of any group, as evidenced by the fact
that we often append the phrase “on average” to our
descriptions of mean differences. But is this enough?
Consider again the robust sex difference in willingness
to engage in casual sex: The mean SO score for men
is higher than that for women (Lippa, 2009; Schmitt,
2005). What does this tell us, though, about individual
men and women? It clearly does not tell us that all
men are interested in casual sex and that all women are
not. However, given the degree of overlap between the
male and female distributions (Simpson & Gangestad,
1991), it also does not tell us that a large majority of
men are more interested in casual sex than a large ma-
jority of women. That is, it is not accurate to say even
that “men are typically more interested in casual sex
than women, but there are of course exceptions.” Here
is what the data that the means are drawn from actually
tell us:
Men and women can be found at virtually every level
of interest in casual sex. At the right-hand tail of
the distribution, only a small number of people are
strongly interested in casual sex; however, of these
people, more are men than women. At the left-hand
tail, only a small number of people are strongly
disinterested in casual sex; however, of these people,
more are women than men. Most people—men and
women—fall somewhere in between. If you were to
choose one man and one woman at random, it would
be somewhat more likely that the man would have
higher SO. However, you wouldn’t want to bet your
life savings on it. Around a third of the time—i.e.,
closer to 50% than to 0%—the woman would have
higher SO. (Stewart-Williams & Thomas, this issue)
If this is not what springs immediately to mind as
soon as the words “on average” are appended to a
description of mean differences, then the words “on
average” have not rectified the damage done by the
use of means to describe populations of varied indi-
viduals. This is not to suggest, of course, that means
should never be used. The point is simply that we
should not lose sight of within-group variability and
the fact that male and female distributions almost al-
ways overlap for psychological traits. This is especially
important when addressing less statistically savvy au-
diences. Such audiences could perhaps be encouraged
to think of two normal distributions, one represent-
ing males and the other females. Instead of imagining
that natural selection creates two distinct psychologi-
cal types—a male type and a female type, described
by the mean values for each group—they could be en-
couraged to imagine that natural selection pushes the
male and female distributions closer together or fur-
ther apart. This simple expedient may help people to
visualize the effects of natural selection on average sex
differences without at the same time losing sight of
the variation within each sex. Admittedly, this formu-
lation glosses over many complications, including the
fact that the distribution for males is often wider and
flatter than that for females (Archer & Mehdikhani,
2003; Harris, this issue), the fact that the distributions
for many traits are highly skewed (Campbell, this issue;
Pedersen et al., 2002), and the fact that some portion of
the variation might itself be explicable in terms of nat-
ural selection (Buss & Greiling, 1999; Campbell, this
issue). Nevertheless, it is good place to start, and it may
help us avoid some of the more egregious misunder-
standings of what the data collected by evolutionary
psychologists are telling us about ourselves and our
One last thought. In their commentary, Pham and
colleagues suggested that the target article represents
an unfair and inappropriate attack on EP. We hope we
have shown that it represents instead a reasonable dif-
ference of opinion within EP. We grant, though, that
in one sense, the charge of unfairness may be true.
Evolutionary psychologists are not unique in making
unduly strong claims about their findings. No doubt,
researchers in all fields are prone to this. Why, then,
pick on EP, rather than any other field or on psychology
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
in general? The reason is that, in our view, EP repre-
sents a Great Leap Forward for psychological science.
Like researchers in every field, evolutionary psychol-
ogists sometimes push their theoretical framework too
far. However, evolutionary psychologists are genuinely
onto something. A lot of people have noticed this and
are paying attention. Thus, it’s important to get it right.
Now, maybe EP already has it right, and any criticism
of the field is inappropriate, misguided, and unfair. But
maybe not. Either way, though, perhaps we can agree
that everyone involved in this debate has the same mo-
tivation: to try to make our picture of our evolved nature
as accurate as possible.
Thanks go to Alan Beaton, Dave Benton, Andrew
Clark, Lauren Jackson, Rob Lowe, Geoffrey Miller,
Phil Reed, Jane Stewart-Williams, Jeremy Tree,
Christoph Weidemann, and Claire Williams.
Address correspondence to Steve Stewart-Williams,
Department of Psychology, Swansea University, Sin-
gleton Park, Swansea SA2 8PP Wales, United King-
dom. E-mail:
Alonzo, S. H. (2012). Sexual selection favours male parental care,
when females can choose. Proceedings of the Royal Society of
London Series B, 279, 1784–1790.
Archer, J., & Mehdikhani, M. (2003). Variability among males in
sexually selected attributes. Review of General Psychology, 7,
Bailey, J. M., Kirk, K. M., Zhu, G., Dunne, M. P., & Martin, N.
G. (2000). Do individual differences in sociosexuality repre-
sent genetic or environmentally contingent strategies? Evidence
from the Australian twin registry. Journal of Personality and
Social Psychology, 78, 537–545.
Barrett, L., Dunbar, R. I. M., & Lycett, J. (2002). Human evolutionary
psychology. Basingstoke, UK: Palgrave.
Becker, S. W., & Eagly, A. H. (2004). The heroism of women and
men. American Psychologist, 59, 163–178.
Betzig, L. L. (1986). Despotism and differential reproduction: A
Darwinian view of history. Hawthorne, NY: Aldine.
Bradley, S. J., Oliver, G. D., Chernick, A. B., & Zucker, K. J. (1998).
Experiment of nurture: Ablatio penis at 2 months, sex reas-
signment at 7 months, and a psychosexual follow-up in young
adulthood. Pediatrics, 102,15.
Buss, D. M. (1989). Sex differences in human mate preferences:
Evolutionary hypotheses tested in 37 cultures. Behavioral and
Brain Sciences, 12, 1–49.
Buss, D. M. (1995). Evolutionary psychology: A new paradigm for
psychological science. Psychological Inquiry, 6, 1–30.
Buss, D. M. (2003). The evolution of desire: Strategies of human
mating (Rev. ed.). New York, NY: Basic.
Buss, D. M. (2012). Evolutionary psychology: The new science of
the mind (4th ed.). Needham Heights, MA: Allyn & Bacon.
Buss, D. M., Abbott, M., Angleitner, A., Asherian, A., Biaggio,
A., Blanco-Villasenor, A., ...Yang, K.-S. (1990). International
preferences in selecting mates: A study of 37 cultures. Journal
of Cross-Cultural Psychology, 21, 5–47.
Buss, D. M., & Greiling, H. (1999). Adaptive individual differences.
Journal of Personality, 67, 209–243.
Buss, D. M., Larsen, R. J., Westen, D., & Semmelroth, J. (1992). Sex
differences in jealousy: Evolution, physiology, and psychology.
Psychological Science, 3, 251–255.
Buss, D. M., & Schmitt, D. P. (1993). Sexual strategies theory:
An evolutionary perspective on human mating. Psychological
Review, 100, 204–232.
Buss, D. M., & Schmitt, D. P. (2011). Evolutionary psychology and
feminism. Sex Roles, 64, 768–787.
Buss, D. M., Shackelford, T. K., & Kirkpatrick, L. A. (1999). Jeal-
ousy and the nature of beliefs about infidelity: Tests of com-
peting hypotheses about sex differences in the United States,
Korea, and Japan. Personal Relationships
, 16, 125–150.
Buunk, B. P., Angleitner, A., Oubaid, V., & Buss, D. M. (1996). Sex
differences in jealousy in evolutionary and cultural perspective:
Tests from the Netherlands, Germany, and the United States.
Psychological Science, 7, 359–363.
Byrnes, J. P., Miller, D. C., & Schafer, W. D. (1999). Gender differ-
ences in risk taking: A meta-analysis. Psychological Bulletin,
125, 367–383.
Chapais, B. (2008). Primeval kinship: How pair-bonding gave birth
to human society. Cambridge, MA: Harvard University Press.
Clark, R. D., & Hatfield, E. (1989). Gender differences in receptivity
to sexual offers. Journal of Psychology and Human Sexuality,
2, 39–55.
Cohen, J. (1988). Statistical power analysis for the behavioral sci-
ences (2nd ed.). Hillsdale, NJ: Erlbaum.
Colapinto, J. (2006). As nature made him: The boy who was raised
as a girl (Rev. ed.). New York, NY: Harper Perennial.
Conley, T. D., Moors, A. C., Matsick, J. L., Ziegler, A., & Valentine,
B. A. (2011). Women, men, and the bedroom: Methodologi-
cal and conceptual insights that narrow, reframe, and eliminate
gender differences in sexuality. Current Directions in Psycho-
logical Science, 20, 296–300.
Daly, M., & Wilson, M. (1988). Homicide. Hawthorne, NY: Aldine
de Gruyter.
Daly, M., & Wilson, M. (2001a). An assessment of some proposed
exceptions to the phenomenon of nepotistic discrimination
against stepchildren. Annales Zoologici Fennici, 36, 287–296.
Daly, M., & Wilson, M. (2001b). Risk-taking, intrasexual compe-
tition, and homicide. Nebraska Symposium on Motivation, 47,
Davies, A. P. C., & Shackelford, T. K. (2008). Two human natures:
How men and women evolved different psychologies. In C. B.
Crawford & D. L. Krebs (Eds.), Foundations of evolutionary
psychology (pp. 261–280). New York, NY: Erlbaum.
Davies, N. B. (1985). Cooperation and conflict among dunnocks,
Prunella modularis, in a variable mating system. Animal Be-
haviour, 33, 628–648.
Davies, N. B. (1989). Sexual conflict and the polygamy threshold.
Animal Behaviour, 38, 226–234.
de Waal, F. (2005). Our inner ape: A leading primatologist explains
whywearewhoweare. New York, NY: Riverhead.
Diamond, M., & Sigmundson, H. K. (1997). Sex reassignment at
birth: A long-term review and clinical implications. Archives of
Pediatric and Adolescent Medicine, 150, 298–304.
Dixson, A. F. (1998). Primate sexuality: Comparative studies of the
prosimians, monkeys, apes, and human beings. Oxford, UK:
Oxford University Press.
Draper, P., & Harpending, H. (1987). Parent investment and the
child’s environment. In J. B. Lancaster, J. Altmann, A. S. Rossi,
Downloaded by [Steve Stewart-Williams] at 17:43 02 September 2013
& L. R. Sherrod (Eds.), Parenting across the lifespan: Biosocial
dimensions (pp. 207–235). New York, NY: Aldine de Gruyter.
Eagly, A. (1995). The science and politics of comparing women and
men. American Psychologist, 50, 145–158.
Eagly, A. H., & Wood, W. (1999). The origins of sex differences
in human behavior: Evolved dispositions versus social roles.
American Psychologist, 54, 408–423.
Eastwick, P. W., Luchies, L. B., Finkel, E. J., & Hunt, L. L. (in press).
The predictive validity of ideal partner preferences: A review
and meta-analysis. Psychological Bulletin.
Eastwick, P. W. (2009). Beyond the Pleistocene: Using phylogeny
and constraint to inform the evolutionary psychology of human
mating. Psychological Bulletin, 135, 794–821.
Emlen, S. T., & Oring, L. W. (1977). Ecology, sexual selection, and
the evolution of mating systems. Science, 197, 215–223.
Finkel, E. J., & Eastwick, P. W. (2009). Arbitrary social norms
influence sex differences in romantic selectivity. Psychological
Science, 20, 1290–1295.
Gangestad, S. W., & Simpson, J. A. (2000). The evolution of hu-
man mating: Trade-offs and strategic pluralism. Behavioral and
Brain Sciences, 23, 573–587.
Gaulin, S. J. C., & McBurney, D. H. (2001). Psychology: An evolu-
tionary approach. Upper Saddle River, NJ: Prentice Hall.
Geissmann, T. (1993). Evolution of communication in gibbons (Hy-
lobatidae). (Unpublished doctoral dissertation). Z
urich Univer-
sity, Z
urich, Switzerland.
Gomendio, M., Harcourt, A. H., & Rold
an, E. R. S. (1998). Sperm
competition in mammals. In T. R. Birkhead & A. P. Moller
(Eds.), Sperm competition and sexual selection (pp. 667–756).
New York, NY: Academic Press.
Gray, P. B., & Anderson, K. G. (2010). Fatherhood: Evolution and
human paternal behavior. Cambridge, MA: Harvard University
Harcourt, A. H., Harvey, P. H., Larson, S. G., & Short, R. V. (1981).
Testis weight, body weight, and breeding system in primates.
Nature, 293, 55–57.
Hrdy, S. B. (2009). Mothers and others: The evolutionary origin
of mutual understanding. Cambridge, MA: Harvard University
Hyde, J. S. (2005). The gender similarities hypothesis. American
Psychologist, 60, 581–592.
Johnson, R. C. (1996). Attributes of Carnegie medalists performing
acts of heroism and of the recipients of these acts. Ethology and
Sociobiology, 17, 355–362.
Jordan-Young, R. M. (2010). Brain storm: The flaws in the science
of sex differences. Cambridge, MA: Harvard University Press.
Kenrick, D. T., Groth, G., Trost, M. R., & Sadalla, E. K. (1993).
Integrating evolutionary and social exchange perspectives on
relationships: Effects of gender, self-appraisal, and involve-
ment level on mate selection criteria. Journal of Personality
and Social Psychology, 64, 951–969.
Kenrick, D. T., Sadalla, E. K., Groth, G., & Trost, M. R. (1990).
Evolution, traits, and the stages of human courtship: Qualify-
ing the parental investment model. Journal of Personality, 58,
Kruger, D. J., & Nesse, R. M. (2006). An evolutionary life-history
framework for understanding sex differences in human mortal-
ity rates. Human Nature, 17, 74–97.
Kuhle, B. X. (2011). Did you have sex with him? Do you love
her? An in vivo test of sex differences in jealous interrogations.
Personality and Individual Differences, 51, 1044–1047.
Lassek, W. D., & Gaulin, S. J. C. (2009). Costs and benefits of
fat-free muscle mass in men: Relationship to mating success,
dietary requirements, and native immunity. Evolution and Hu-
man Behavior, 30, 322–328.
Li, N. P., Bailey, J. M., & Kenrick, D. T. (2002). The necessities and
luxuries of mate preferences: Testing the tradeoffs. Journal of
Personality and Social Psychology, 82, 947–955.
Li, N. P., & Kenrick, D. T. (2006). Sex similarities and differences
in preferences for short-term mates: What, whether and why.
Journal of Personality and Social Psychology, 90, 468–489.
Lippa, R. A. (2000). Gender-related traits in gay men, lesbian
women, and heterosexual