BEHAVIORAL SCIENCES (1992) 15, 368-421

Printed in the United States of America

e
e
1a and Nancy Wiimsen Thornhillb
department of Biology and bDepartments of Biology and Anthropology,
University of New Mexico, Albuquerque, NM 87131-1091
Electronic mail: nthorn@unmvm.bitnet

Abstract? Psychological adaptation underlies all human behavior. Thus, sexual coercion by men could either arise from a rape-specific
psychological adaptation or it could be a side-effect of a more general psychological adaptation not directly related to rape.
Determining the specific environmental cues that men's brains have been designed by selection to process may help us decide which
of these rival explanations is correct. We examine six testable predictions against existing data: (1) Both coercive and noncoercive sex
will be associated with high levels of sexual arousal and performance in men. (2) Achieving physical control of a sexually unwilling
woman will be sexually arousing to men. (3) Young men will be more sexually coercive than older men. (4) Men of low socioeconomic
status will likewise be more sexually coercive. (5) A man's motivation to use sexual coercion will be influenced by its effects on his
social image. (6) Even in long-term relationships men will be motivated to use coercion when their mates show a lack of interest in or
resistance to sex because these are interpreted as signs of sexual infidelity. Current data support all six predictions and are hence
consistent with the rape-specific hypothesis, but this does not eliminate the side-effect hypothesis, which is likewise compatible with
the findings, as well as with the further evidence that forced matings increased the fitness of ancestral males during human evolution.
We suggest some research that may help decide between the two hypotheses.

Keywords^ aggression; comparative psychology; evolution; psychological adaptation; rape; sex differences; sexual coercion; sexual
dimorphism; sexual selection; socialization; sociobiology

information on the natural history of human sexual be-

Is rape just a side-effect of psychological adaptations to
circumstances other than rape, such as a desire for sex
coupled with a general coercive tendency, or does it arise
directly from an evolutionary adaptation to sexual coer-
cion itself?1 The answer to this question has practical as
well as theoretical consequences: It can help us identify
the kinds of social contexts that discourage coercive sex.
We hypothesize that sexual coercion by men reflects a
sex-specific, species-typical psychological adaptation to
rape: Men have certain psychological traits that evolved
by natural selection specifically in the context of coercive
sex and made rape adaptive during human evolution.
The hypothesis that there is a psychological adaptation
specific to sexual coercion by men was implicit in some
earlier evolutionary writings (Alexander 1979a; Alex-
ander & Noonan 1979; Shields & Shields 1983; Thomhill
& Thornhill 1983). These authors focused directly on rape
and dealt only superficially with the mechanisms that
regulate men's sexual arousal and their inclination to use
force. Sexual coerciveness alone is not evidence of a rape-
specific psychological adaptation, nor do men have any
specialized morphology that might have evolved for rape.
One form of positive evidence would be if environmental
cues that were specific to the costs and benefits of sexual
coercion during human evolution still regulated contem-
porary coerciveness.
The rape-specific hypothesis is examined in the light of
havior and published laboratory studies of men's sexual
arousal in response to sexually-explicit stimuli. The hy-
pothesis is not contradicted by current information on
human coercive sexuality, but neither is the rival side-
effect hypothesis. We outline some research that may
help decide between the two.
2a Adaptationism
In this target article and in theoretical biology in general
adaptation refers to complexly organized, purposefully
designed, phenotypic features of individual organisms
that exist because they solved a specific environmental
problem during evolutionary history (e.g., Burian 1983;
Darwin 1859; Dawkins 1986; Symons 1979; Williams
1966; see R. Thornhill 1990 for review). Four natural
processes are known to cause evolution or changes in
gene frequencies of populations, but selection is the only
one that can create an adaptation. The other three -
mutation, drift, and gene flow - lack the necessary
creativity because their action is random relative to indi-
viduals' environmental problems. Selection, when it acts
in a directional, cumulative way over long periods of
geological time, creates complex phenotypic designs out
of the simple, random genetic variation generated by the
other three natural processes. Selection itself is not a
random process. It consists of differential survival and

1
1992 Cambridge University Press 0140-525X192 $5.00+.00 363
Thornhill & Thornhill: Evolution of sexual coercion
reproduction by individuals because of differences in
their phenotypic design.
3. Psfchoiogicai adaptationism and rape
Human psychological adaptations are information-
processing mechanisms providing solutions to problems
that influenced the survival and reproductive perfor-
mance of individuals during our species' evolutionary
history. Human psychological adaptations are specially
engineered to process specific environmental information
and to guide feelings, emotions, learning, and behavior
toward specific reproductive ends. 2
Perception, memory storage and retrieval, cognitive
analysis, and so forth are evolved information-processing
mechanisms and must be characterized in functional
evolutionary terms, that is, by the kinds of information
they are designed by selection to process rather than by
eeurophysiology or neuroanatomy (also see Cosmides &
Tooby 1987; 1989; Tooby & Cosmides 1989). Psychologi-
cal adaptation must somehow causally underlie all human
feelings, emotion, learning, and behavior (see Cosmides
& Tooby 1987; Symons 1987a).
To determine whether or not there is a rape-specific
adaptation and to begin to characterize its design we must
examine how men's sexual information-processing mech-
anisms regulate their sexual arousal and behavior in the
context of sexual coercion. An empirical focus on rape
behavior rather than on rape psychology is not appropri-
ate, because neither the fact that men do rape nor the
coercive and sexual components of an act of rape can be
taken as evidence for (or against) the hypothesis that men
are adapted to rape per se. There is nothing about rape
behavior itself (in humans or nonhuman species) that
could indicate whether or not there is rape-specific de-
sign. The physical force associated with human rape did
not evolve in the context of forced copulations; aggressive
behavior is not limited to rape and is used by children and
adults of both sexes. Likewise, it is unlikely that men's
tendencies to use threats of physical force or other forms
of coercion to gain sexual access to unwilling partners
were specifically selected because they provided a re-
productive advantage in the context of rape.
Nor, as noted earlier, do men have a morphological
specialization that is a candidate for rape adaptation.
Phenotypic features specific to sexual coercion have been
demonstrated in males of only a few species ol nonhuman
animals, and in all cases these involve morphology (e.g.,
an abdominal clamp for holding the wings of sexually
unwilling females, Thornhill 1984a; see Thornhill &
Sauer, 1991, for review).
Whether or not men are specifically adapted to do it,
rape cannot be fully understood without explicitly exam-
ining men's sexual psychology because the act of rape
involves psychological changes (emotions, arousal, etc.)
and behavior, and all psychological changes and behavior
are the result of information processing by the nervous
system. The discovery of the environmental cues that
men's brains are designed to process will help us under-
stand the environmental conditions influencing all as-
pects of men's sexuality.
We are implicitly assuming that men and women differ
in sexual psychological adaptation. It is often assumed in
364 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
the social and behavioral sciences that human psychology
is the same for both sexes and consists of only a few
general-purpose learning adaptations (Symons 1987a).
However, given current knowledge of the functional
specificity of the vast numbers of adaptations whose
design is understood (e.g., the human heart is specially
designed to pump blood in a human body), as well as
theoretical advances in understanding how selection
works in molding adaptations (they are specific solutions
to specific environmental problems), it is most likely that
psychological adaptation to sexuality is sexually di-
morphic and includes many, highly specialized adapta-
tions. Some involve learning and are designed to process
information specific to the very different sexual problems
faced by males and females in human evolutionary history
(see Symons 1987a; 1987b, for a detailed analysis and
critique of the idea that human sexual psychology is
monomorphic and consists only of a few general-purpose
adaptations).
3.1. Rap® and! social learning. "Learning" is often invoked
in the social science literature on rape as if it were a
complete explanation for men's and women's sexual be-
havior (for a useful review, see Ellis 1989; also see Russell
1982; 1984). Such a view is based on the misunderstand-
ing of two fundamental facts of developmental biology: (1)
"Learned" behaviors result from causal gene-environment
interactions during ontogeny. (2) A vast number of experi-
ences necessarily occur during the ontogeny of any be-
havior; only a very small number of these would be
"learned." The term "learned" is far too simplistic to
warrant its use as the major or only causal explanation of
any human behavior (see also Alexander 1990; Cosmides
& Tooby 1987; Daly & Wilson 1987; 1988; Symons, in
press). Furthermore, learning and socialization are never
alternatives to evolutionary hypotheses about psychologi-
cal adaptation (e.g., see Cosmides & Tooby 1987).
The "social learning theory of rape" and the "feminist
theory of rape" are very similar (for a review of literature,
see Ellis 1989). According to both, rape is primarily or
solely caused by arbitrary differences in the way men and
women are socialized about heterosexual conduct. If their
socialization were the same, they would be equally likely
to rape. Both theories posit that rape results from general
sex differences - such as men being more motivated than
women to use aggression and to be noncommittal in
sexual relations — and that these general sex differences
are solely or primarily caused by contingent differential
socialization. But general sex differences similar to those
outlined by these theories of rape - for example, that
males are more aggressive, sexually assertive and eager to
copulate, and less discriminating of mates - occur not
only in humans, but in all animal species with an evolu-
tionary history of polygyny (see Daly & Wilson 1983;
Thornhill 1986; Thornhill & Alcock 1983; Trivers 1972;
1985).3 In the vast majority of these species there is no
sexual training of juveniles by other members of the
group (e.g., all polygynous invertebrates), and none have
the extensive sexual socialization seen in humans (see
Low 1989). Hence the human sex differences common to
polygynous species cannot be parsimoniously fully ex-
plained by a difference in sexual socialization in humans.
The hypothesis of rape-specific adaptation does not
assume that there is no learning or sex-specific socializa-

tion in the development of human sexuality. Indeed,
learning might be important during the ontogeny of the
rape-specific adaptation itself. If so, the learning would
not be arbitrary but would be guided by special-purpose
psychological adaptations that influence perception, cog-
nition, memory, and information evaluation in a way that
is specific to rape.
3.2. Rape's current effects onreproduction*The extent to
which sexual coercion currently promotes the reproduc-
tive success of men is not of central importance to the
approach emphasized in the target article. Indeed, data
about offspring production by victims of sexual coercion
are irrelevant to the rape-adaptation hypothesis. Thus our
approach is very different from that of Ellis (1989) and
Thiessen (1989), who consider the current reproductive
consequences of rape to be of fundamental importance.
An adaptation is identified and characterized by its evolu-
tionary function, what it has been designed by selection to
do. The relationship between an adaptation and current
reproduction depends on the similarity between the
environment in which the adaptation is expressed cur-
rently and the environmental features that generated the
selection that designed the adaptation. Often this correla-
tion no longer exists for contemporary organisms (perhaps
especially humans). As a result, adaptations are not distin-
guished from nonadaptations by their effects on current
reproduction: A phenotypic trait can influence current
reproduction negatively but still be an adaptation - a
product of long-term directional selection for a solution to
a fitness-related ecological problem - and a trait can
influence current reproduction positively and not be an
adaptation. In addition, the effect of an adaptation on
current reproduction does not identify the evolutionary
function of the adaptation because the current environ-
ment may differ from the evolutionary environment that
generated the selection that designed the adaptation (see
also Williams 1966; R. Thornhill 1990).
3.3. Rape in nonhuman animals. One cannot derive evi-
dence for or against the existence of human rape adapta-
tion from male sexual adaptations in nonhuman species
because the selective environments that produced hu-
man psychological adaptations were unique to our spe-
cies. This is why we do not discuss here the few nonhu-
man species in which there is evidence for the existence of
rape adaptation (e.g., certain morphological structures in
male scorpionflies, Thornhill 1984a; Thornhill & Sauer
1991). The nonhuman evidence shows that selection in
the context of sexual coercion by males can, under some
conditions, produce rape adaptation, but it cannot dem-
onstrate that such selection was effective in designing
rape adaptations in other species (e.g., Homo sapiens) in
which rape adaptation has not been demonstrated. The
only way to provide evidence about the selection pres-
sures that designed humans is to identify and characterize
human adaptations (see also Symons 1982; Tooby &
Devore 1987).
3.4. Heritabiiity* In the literature on human rape there is
considerable interest in rape heritability (see Ellis 1989,
especially pp. 86-88). Heritability refers to the extent to
which the variation in a phenotypic trait among individ-
Thornhill & Thornhill: Evolution of sexual coercion
uals is caused by genetic as opposed to environmental
variation.4 The concept of heritability is of value primarily
in programs to improve domestic plants and animals by
artificial selection (Falconer 1981; also see R. Thornhill
1990). [See also Wahlsten "Insensitivity of the Analysis of
Variance to Hereditary-Environment Interaction" BBS
13(1) 1990; Plomin & Bergman "The Nature of Nurture"
BBS 14(3) 1991.]
We emphasize that the rape-adaptation hypothesis
does not imply that rape is heritable. That is, it does not
predict that variation among men in inclination to rape
reflects genetic differences. Instead, it implies that men
have psychological traits whose underlying genes are
virtually fixed or invariant in the human gene pool; this is
what is meant by "sex-specific" and "species-typical."
Thus, we predict that the heritability of the psychological
adaptation for rape will be near or at zero. This is not to
say that male personality features that may contribute to
rape (e.g., aggressiveness) are not heritable; some may be
(see Ellis 1989, pp. 86-97), but the heritability of person-
ality features is not relevant to the question of whether
there is a specific adaptation to rape.
According to the rape-adaptation hypothesis, men's
use of sexual coercion will not vary with differences In
genetic conditions but with environmental conditions
(see also Shields & Shields 1983; Thornhill & Thornhill
1983). The hypothetical rape-specific adaptation should
regulate men's use offeree in a facultative way by process-
ing environmental information that was associated with
engaging in rape during human evolution (Crawford &
Anderson 1989; Tooby & Cosmides, 1990, provide de-
tailed discussions of why psychological adaptation will
have negligible heritability).
That rape adaptation is species-typical does not exclude
the possibility of minor geographical variation in the
design of its underlying psychological traits. For example,
selective differences between human populations could
have led to differences in how easily the rape adaptation is
activated by environmental cues. Even for minor differ-
ences between populations to evolve by selection, how-
ever, it is necessary that there have been a long period of
low or no migration between populations and cumulative
directional selections that differed between populations.
4. Coercion and men's mating strategy
We dissect the overall human male reproductive strategy
to point out where we feel rape fits into it (see Thornhill &
Thornhill, 1983, for detailed discussion). The reproduc-
tive strategies of all organisms, including men and
women, have the following two components: mating
effort (the reproductive effort associated with conceiving
offspring) and nepotistic effort (the reproductive effort
associated with aiding offspring and other relatives).
Men's mating effort consists of obtaining matings, keep-
ing mates, and enhancing paternity confidence with
mates (see also Buss 1988; Daly & Wilson 1983; Flinn
1988). [See also Hartung: "Matrilineal Inheritance" BBS
8(4) 1985. ] The effort associated with obtaining a mate can
be called men's mating strategy. This can be conceptually
partitioned into three tactics (Shields & Shields 1983): (1)
honest advertisement and courtship, (2) deceptive adver-
tisement and courtship, and (3) coercion. Coerced mat-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 365
Thornhill & Thornhill: Evolution of sexual coercion
ings or rapes are achieved by physical force or by explicit
or implicit threat of physical harm or negative social
consequences.
In humans, there is a large sexual asymmetry in the
minimal reproductive effort required for the production
of offspring. The minimum for a man is a few minutes of
time and an energetically cheap ejaculate; the minimum
for a woman is nine months of pregnancy and a long
period of lactation. Because of this sexual asymmetry,
during human evolution males who could gain sexual
access to multiple females out-reproduced males who
could not and thus were favored by sexual selection.
Because women's sexuality was not molded by the same
selection as male sexuality, three complementary human
sex differences are observed: Compared to women, men
are less discriminating about sexual partners, more moti-
vated to seek copulation with many partners, and more
eager to include copulation as part of an interaction with
the opposite sex. Sexual selection on females in human
evolutionary history favored individuals who could gain
access to males whose resources and genetic endowment
could promote the survival of offspring. Selection pres-
sure on females to compete for multiple mates was weak
relative to that on males, because sexual access to multi-
ple males has little effect on the number of offspring a
female can produce, whereas access to multiple females
can have a great effect on the number of offspring a male
can produce. Women are more selective of mates than
men because in human evolutionary history females
made a larger minimal investment in offspring (thus
losing more reproductive potential than males from a
poor mate choice), which resulted in stronger natural
selection on females than on males for mate choice.
Furthermore, females are the objects of more sexual
competition than males and thus have a greater oppor-
tunity to choose effectively (see Alexander 1979a; Buss
1987; Daly & Wilson 1983; Smith 1984; Symons 1979;
1987b; Symons & Ellis 1989; Townsend 1987; 1989).
Because of the different ways that selection acted on
the sexes during human evolutionary history, evolution-
ary psychologists believe human sexual psychology is
dimorphic, that is, the respective adaptations differ in
men and women (Symons 1979; 1987b).
The evolved sex difference in mating strategy leads to
differences in how men and women feel about whether,
when, and how often it is in their interest to mate.
Because women are more selective about mates and more
interested in evaluating them and delaying copulation,
men, to get sexual access, must often break through
feminine barriers of hesitation, equivocation, and re-
sistance (see Kirkendall, 1961, for a review of human
heterosexual sexual interactions). Men get women to
copulate by using all three tactics mentioned above. As
we show later, these tactics can be used either singly or in
combination and they often blend into one another.
According to the rape-adaptation hypothesis, during hu-
man evolutionary history there was enough directional
selection on males in favor of traits that solved the prob-
lem of forcing sex on a reluctant partner to produce
psychological inclinations specifically toward rape. Here
366 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
are some reasons why it seems reasonable to assume that
ancestral human males sometimes increased their re-
productive success by rape. 5
1. There is strong evidence that the following mas-
culine traits are evolved ones: Compared to women, men
are more aroused by visual and fantasized sexual stimuli;
men are more indiscriminate about mating partners; and
men are more likely to infer the presence of sexual
interest in a potential mate when there is no such interest
(Abbey 1982; Abbey & Melby 1986; Kinsey et al. 1948;
1953; Saal et al. 1989; Symons 1979; Townsend 1987).
2. There is strong evidence that women's selectivity
reflects adaptation to the circumstances of mate choice
per se (see Buss 1987; 1989a; Symons 1979; Townsend
1987; 1989). Women's traits and the male traits men-
tioned in (1) evolved, indeed coevolved, in the same
evolutionary environment. Thus, sexually attractive
females who were sexually uninterested and resisted the
sexual advances of males are likely to have been a consis-
tent feature of the human evolutionary environment; this
produced directional selection favoring males whose sex-
ual arousal did not depend on the sexual interest or
arousal of a potential mate.
3. There are also costs to males associated with trying
to mate by force that, all else equal, would have made
rape maladaptive in the human evolutionary environ-
ment. As discussed by Shields and Shields (1983) and
Thornhill and Thornhill (1987), potential injury to the
male and to his family members, loss of status or re-
sources, and other factors inevitably contribute to the
costs of rape. Given the cost of forcing matings, one might
have expected selection to counteract (1) above and in-
stead favor males who were sexually interested only in
sexually interested females. This is not the inclination
that has evolved in men, however; as we show below men
use sexual coercion and show high levels of sexual arousal
in response to both coercive and mutually consensual
sexual materials in laboratory settings. Hence there
seems to be no specific psychological trait that is spe-
cialized to prevent men from using force to mate. Thus,
despite the opportunity for selection to act against males
who were sexually coercive and despite the costs that
should have made any such selection strong and effective,
men are commonly sexually coercive, and their arousal to
depictions of sexual acts in the laboratory is unaffected by
whether or not force is used.
In evaluating the potential advantage of forced sexual
intercourse during human evolution, it is important to
consider the survival prospects of the offspring from such
matings. Psychological adaptation for parental care in
both men and women confirms that offspring required a
large amount of parental care during human evolution
(see Daly & Wilson 1988).
Alexander and Noonan (1979) have suggested one way
rape may have contributed to successful offspring produc-
tion in human evolutionary history: A female's sexual
unwillingness could be evidence that she was pair-
bonded, and thus that the rapist's offspring would be
unwittingly reared by the cuckold. Raped females would
be reluctant to reveal to their mates that they had been
raped because this would make their mates cooperate less
in the care of existing offspring (for an analysis of paternity
and cuckoldry in humans, see Alexander 1979b; Buss
1988; Daly & Wilson 1988; Daly et al. 1982; Dickemann

1979a; 1979b; Flinn 1988; Smith 1984). Alexander and
Noonan's scenario can be extended to include the selec-
tive advantage to men of sexual coercion even with their
long-term mates. Sexual uninterest or resistance may
reflect interests outside the pair-bond; if so, men's moti-
vation to use forced copulation within mateships may be
an evolved strategy for ejaculate competition with the
sexual rival.
It is important to consider the alternative possibility
that men's ready capacity for sexual arousal with both
willing and unwilling (see sect. 7 and 8) sexual partners
evolved as a result of selective factors other than success-
ful reproduction through rape. Sexual coercion might be a
side effect, not adaptive in its own right (perhaps even
maladaptive), of two more general adaptations: (a) the
psychological mechanisms underlying men's general de-
sire for sexual intercourse, and (b) the mechanisms under-
lying our species' general tendency to use force to attain
any reward that was correlated with successful survival
and reproduction in our evolutionary history. According
to this side-effect hypothesis, there are no psychological
mechanisms specifically designed for processing informa-
tion about rape.
S- Predictions
In this section, we outline several predictions that follow
from the rape-specific hypothesis. The list is not exhaus-
tive; it focuses on what can be tested against the existing
literature on human sexuality.
Prediction 1. Men will exhibit high levels of sexual
motivation and performance in both coercive and non-
coercive mating situations. If men have a rape-specific
adaptation, then their sexual arousal and ability to copu-
late and ejaculate should not be affected by whether the
woman has given her consent. In contrast, the rape-
adaptation hypothesis would be falsified by evidence that
men are sexually aroused and competent only or pri-
marily when they perceive that a woman is interested or
not resisting coitus.
Data supporting the first prediction would suggest an
adaptation for rape, but demonstrating it would require
direct evidence of rape-specific design so as to eliminate
the rival side-effect hypothesis.
In evolutionary theory, individual "interests" are based
on evolved reproductive interests (Alexander 1987; Daly
& Wilson 1988). This suggests that men will pursue
copulation byforcewhen it serves their evolved interests;
a woman's interests should be considered by a man only to
the extent that they coincide with his own. Interests may
coincide to varying degrees. The smaller the overlap of
interests with the partner, the greater the probability of
sexual coercion by men. Predictions 2 to 6 are based on
the conflicting reproduction interests of the sexes.
Prediction 2* Gaining physical control over an unwill-
ing sexual partner byforceshould be sexually arousing to
men because it facilitates forced copulation. As long as the
victim of sexual assault is capable of resistance, there are a
risk of injury to the assailant and a potential for thwarting
of the assault. Rape-specific selection would have gener-
ated mechanisms for assessing such risks and regulating
male motivation and behavior accordingly. In the context
of coercion, male sexual arousal has a cost; it might
Thornhill & Thornhill: Evolution of sexual coercion
interfere with his ability to detect the presence of danger-
ous conspecifics or predators. In contrast, the side-effect
hypothesis does not predict that gaining control of an
unwilling partner will be sexually arousing.
Prediction 3* A man's age should affect his willingness
to use sexual coercion. Because in our evolutionary past
young men (mid-teens to early 20s) were trying to enter
the breeding population for the first time, they engaged
in the most intense mate competition, and therefore took
the highest risks and experienced the highest mortality
(Alexander 1979b; Thornhill & Thornhill 1983; Trivers
1985, Chapter 12; Wilson & Daly 1985). Young males'
difficulty in entering the breeding population probably
arose because older males controlled the resources, sta-
tus, and politics, and thus monopolized the females. The
opportunity cost of engaging in risky behavior is lower for
younger men and therefore mortality is greaterforyoung
men than for older men (see also Thornhill & Thornhill
1983).
The side-effect hypothesis predicts the same age-
dependence of sexual coercion as the rape-adaptation
hypothesis. Many types of male sexual activity show a
major peak in the mid-teens and early 20s (Gagnon 1977;
Kinsey et al. 1948); young men are more quickly and
easily aroused by explicit sexual stimuli in laboratory
settings (e.g., Langevin 1983). Young men are also most
prone to exhibit violence in general (Alexander 1979b;
Daly & Wilson 1988; Wilson & Daly 1985). Hence greater
sexual coerciveness in young men could be interpreted as
a side-effect of their greater libido and their greater
tendency to use violence to get what they want.
Prediction 4* Men's willingness to use sexual coercion
should be related to their social status. Women-prefer as
mates men of high social and economic status (Buss 1987;
1989a; Symons 1979; Townsend 1989). Sexual access to
preferred mates (young and attractive) is thus positively
correlated with the status, resource holdings, and pres-
tige of a man. This correlation has been demonstrated
repeatedly in industrial societies (Buss 1987; 1989a),
traditional societies (Betzig 1985; Betzig et al. 1988), and
in the historical record (Betzig 1985; Voland & Engel
1990). We have predicted that men of low socioeconomic
status will be more inclined to rape (Thornhill & Thornhill
1983) because they have less access to preferred mates,
but because the poor are also more prone to crime and
other forms of violence, a positive outcome would like-
wise be consistent with the side-effect hypothesis.
Prediction 5, Sexual coerciveness will be very sensitive
to the probability of detection and negative social conse-
quences or punishment. Men's willingness to use force
will be constrained by the countervailing desire of dis-
playing sexual "morality." Lack of concern for social image
will promote sexual coercion regardless of age and social
status (also Shields & Shields 1983; Thornhill & Thornhill
1983). Prediction 5, like Predictions 1, 3, and 4, follow
from both the rape-specific and the side-effect hypoth-
eses. Evidence supporting Prediction 5 could be inter-
preted as a side-effect of men's adaptation for copulation
and species-typical adaptation for regulating the personal
use offeree in general.
Predictions 2 to 5 are based on the conflicting re-
productive interests of the sexes. For example, we expect
men who are young and of low socioeconomic status to be
more coercive sexually than older men of high socioeco-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 367
Thornhill & Thornhill: Evolution of sexual coercion
nomic status because the interests of the former conflict
more with those of women, who prefer mates with re-
sources and status. Also, a man's reputation is determined
by how others perceive him, which is in turn influenced
by what they know of his social exploitativeness. Hence a
man's interest in enhancing his social reputation may
sometimes be served by not coercing sexually unin-
terested women.
Now consider congruence and noncongruence in the
reproductive interests of pair-bonded mates. Two impor-
tant factors that promote similarity of interests between
men and women in mateships are shared genetic offspring
and the woman's sexual fidelity; these two factors are
interrelated. Men must typically provide resources and
other benefits, because women's sexual interest is tied to
men's status and resources (e.g., Buss 1987; 1989a; Sy-
mons 1979; Townsend 1989). When a man provides
abundant resources to a mate and her offspring her
reproductive interests are served. Under such conditions
conflicts of interest (thus coercion of one party by the
other) should be minimal. Infidelity by a pair-bonded
woman leads to uncertainty about paternity in her mate;
their reproductive interests consequently diverge. When
women perceive that their reproductive interests are not
being served by a long-term mate, they may show him
less emotional commitment and sexual interest and seek
other mates. Men may find women's sexual arousal excit-
ing because it is a sign of exclusive sexual access and hence
a high probability of siring that mate's offspring. Pair-
bonded men may have accordingly evolved to view a lack
of sexual interest from a mate as a sign that she may have
another mate. This leads to the last prediction in this
section.
Prediction 6* A pair-bonded man will be most likely to
use sexual coercion with an unwilling pair-bonded mate
when he suspects or discover infidelity. Again, if this
prediction is met, rape adaptation is not yet demonstrated
because the outcome could be a side-effect of adaptation
to enhance paternity reliability and to use force to attain
desired goals.
In sections 7 to 11 we examine the available data on
the six predictions just outlined. These predictions put
the rape adaptation hypothesis in jeopardy; although the
failure of even one of them would falsify the hypothesis,
only Prediction 2 can falsify its rival, the side-effect
hypothesis, providing direct evidence of adaptation to
rape. The final section of the target article outlines some
additional predictions that, if met, could eliminate the
side-effect hypothesis and provide decisive evidence of
rape-specific adaptation.
sxual behawior by men
In this section we analyze the natural history of coercive
sexual behavior. We show that coercion is a significant
component of men's sexual activity in general. Men
frequently pursue sexual access by using a combination of
coercion and noncoercion as indicated by the following:
1. Homosexual rape is a common occurrence in men's
prisons and is not usually perpetrated by men with a
strong homosexual orientation; it is usually the act of
heterosexually oriented men who lack access to preferred
368 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
sexual partners (for reviews, see Lockwood 1980; Scacco
1975).
2. Rapists who become visible to police and end up in
crime statistics are frequently found to have pair-bonded
with one woman and raped another woman, a nonpair-
bond mate. Among convicted rapists 19% to 43% are
married at the time of the rape or have been previously
married, and a much larger percentage have been in-
volved in (presumably less coercive) pair-bond mateships
(see Thornhill & Thornhill 1983). (The majority of rapes
that become visible to police involve strangers; McDer-
mott 1979; Russell 1984).
3. Many men use both noncoercive and coercive
means to copulate with pair-bond mates (on rape by
husbands and boyfriends, see Finkelhor & Yllo 1983;
1985; Russell 1982; Shields & Hanneke 1983).
4. The self-reported sexual histories of young men and
the literature on date and acquaintance rape in general
indicate an abundance of sexual coercion (Kirkendall
1961; also reviews by Malamuth 1981a; 1984; Pirog-Good
& Stets 1989; Russell 1984).
There is no question that many men's sexual reper-
toires include a mixture of noncoercive and coercive
tactics including physical force. In addition, men often
pursue an individual copulation with a mixture of tactics.
Thus, it is wrong to dichotomize copulations as involving
honest versus deceptive courtship or as unforced versus
forced. Courtship and the interactions associated with
maintaining pair-bonds include explicit and implicit
promises about commitment. In part these promises may
be broken because there was no intention to keep them
(see Kirkendall 1961). The forced versus unforced dichot-
omy may apply only to a small subset of human copula-
tions: An example is the action of a man, without any
sexual negotiation or honest or even deceptive courtship,
using physical force or the threat of physical harm to
capture and copulate with a woman against her will. The
legal system in modern societies is concerned with
whether or not an allegedly forced copulation was legally
forced, because the answer determines whether the
crime of rape was committed. The difficulty that the legal
system has in distinguishing forced from unforced sexual
intercourse illustrates the conceptual problem arising
from the fact that men tend to use mixtures of the three
mating techniques (described previously in sect. 4) in a
single attempted mating.
The common use of mixed coercive and noncoercive
approaches is also revealed by the recent literature on
acquaintance rape, including sexual coercion in dating
(Kirkendall 1961; see also reviews by Belknap 1989; Ellis
1989; Malamuth & Briere 1986; Pirog-Good & Stets 1989;
Russell 1984). Young women have a significant risk of
sexual victimization by men with whom they are ac-
quainted, including dates. Men tend to date with the
hope of mutually consensual sex; they use both honest
and deceptive courtship to get dates. Women tend to date
because of their interest in courtship (for a review of the
human courtship literature, see Grammer 1989). Sexual
coercion is apparently a common occurrence on dates,
though it typically is used only when noncoercion fails
(e.g., Kirkendall 1961; see, Ellis 1989, for a review). The
sexual coercion with dates and acquaintances can include
clearly physical force. Alternatively, the woman may

protest that she does not want to be sexually intimate, but
the man persists nonviolently until he achieves sexual
intimacy. Compliance in the latter case is gained by
coercion in the form of social pressure or intimidation that
leads the woman to feel that if she does not comply, there
will be unpleasant consequences.
As in the literature on acquaintance rape, the data on
rape by husbands and boyfriends and on rape of children
(Finkelhor 1984; Thornhill, in press) demonstrate that
men often combine noncoercive and coercive tactics,
including physical force.
7.1. Sexual coerclon/noncoerclon: A continuum* Not
only are the three mating tactics (honest advertisement,
deceptive advertisement, and coercion - see sect. 4) used
in combination to obtain single copulations, making the
dichotomies of honest versus deceptive courtship and
forced versus unforced copulation simplistic, the three
tactics often grade into each other, with only arbitrary
boundaries between them. Others have also pointed out
that the forced versus unforced dichotomy is inappropri-
ate because they are on a continuum (e.g., Clark & Lewis
1977; Medea & Thompson 1974; Russell 1982). We sug-
gest that it is the continuity between forced and unforced
copulations, even more than the combination of sexually
coercive and noncoercive tactics in the pursuit of single
copulations, that creates the great difficulty in all en-
deavors to deal with the concept of rape, whether in the
legal context or in everyday life.
Many sexual interactions between men and women
probably grade into coercion. As discussed earlier, the
difference in the action of selection on males and females
in human evolutionary history has resulted in women's
being attracted sexually to partners with high status and
resource-providing potential (Buss 1987; 1989a; Symons
1979; Townsend 1989). The association between women's
erotic responsiveness and the resources provided by
mates means that in general men must purchase consen-
sual sexual access to women (Symons 1979) and that much
of intersexual conflict and bargaining will end with male
coercion because men's ability or willingness to pay will
often be inconsistent with women's needs or expectations.
In humans, courtship and the interactions between
pair-bonded mates accompanying copulation may include
male violence toward the mate or her offspring or explicit
or implicit threats of violence that grade into displays and
vows of emotional commitment. Most commonly, how-
ever, explicit or implicit threats of unpleasant nonviolent
consequences (e.g., a man's withdrawal of financial sup-
port or emotional involvement) rather than actual or
threatened male violence grade into noncoercive
behavior.
Buss's (1989b) findings on sexual conflict in mateships
bear on the coercion/noncoercion continuum. Buss stud-
ied undergraduates who had been involved in a recent
heterosexual relationship and newlywed couples who had
been married one year or less. He found highly significant
sex differences in the incidence of anger about sexual
rejection for both the undergraduate and newlywed sam-
ples: Men were more upset and angry about women
withholding sex than vice versa. We suggest that the
greater anger and distress shown by men are often per-
ceived as threats of violence or of the withdrawal of
Thornhill & Thornhill: Evolution of sexual coercion
resources or emotional commitment. Conflict about sex-
ual access during dating is to be expected, but its occur-
rence between newlyweds when sexual conflict is rela-
tively limited (see, Buss 1989b, for empirical support and
discussion) is surprising and intriguing. If men's emo-
tional reactions toward mates do reflect threats, even
mateships involving relatively high congruence in the
interests of mates may often include sexual coercion. Buss
found that male anger and distress occurred at about the
same rates among undergraduates and newlyweds.
The literature could have falsified the rape-specific
hypothesis if it had shown that sexual coercion is used
infrequently or only by a small percentage of men. In-
stead, consistent with the specificity hypothesis, men
tend to persist in seeking sex and can become sexually
aroused regardless of whether a woman is interested in or
resists their advances.
8. Viewing coerciwe and noncoerciwe stimuli in
the laboratory
8.1. Rapists. To analyze further the environmental cues to
which men are sexually responsive we examined the
literature on men's arousal in response to audio and video
stimuli in the laboratory showing coercive and noncoer-
cive sex.
Because penile erection is the only physiological re-
sponse in men that occurs almost exclusively during
sexual arousal, researchers generally agree that measures
of penile tumescence best reflect men's sexual arousal and
preference for various sexual stimuli (Zuckerman 1971;
also Abel et al. 1977; Earls & Marshall 1983; Langevie
1983; Quinsey & Chaplin 1982). In addition to penile
tumescence, self-reported sexual arousal is assessed in
some research. Although phallometric data increase ob-
jectivity, their drawback is that volunteers for such re-
search sometimes differ significantly from nonvolunteers
in their sexual behavior and personality. In contrast, no
differences have been found between the general popula-
tion and volunteers for studies that use only self-report of
sexual arousal (see Farkas et al. 1978; Malamuth & Check
1983, for review and discussion).
Malamuth (1981a) provides a useful review of the
earlier laboratory studies on sexual arousal in men. All
five studies by G. Abel and associates that were reviewed
by Malamuth involved incarcerated, heterosexual, and
mostly mentally ill rapists; these showed equally high
levels of penile tumescence when they listened to au-
dio tape depictions of rape and mutually consenting inter-
course. This finding was replicated by Barbaree et al.
(1979) and Quinsey et al. (1981).6 To assess the generality
of incarcerated rapists' similar responses to rape and
noncoercive sex it is necessary to test the general male
population as well.
8.2. Nonrapists. Malamuth's (1981a) review lists only two
studies comparing the sexual arousal of incarcerated rap-
ists and nonrapists to portrayals of coercive and noncoer-
cive sex. Both studies can be criticized on scientific
grounds. Abel et al. (1977) and Barbaree et al. (1979)
found that, unlike incarcerated rapists, incarcerated non-
rapists showed little sexual arousal in response to depic-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 369
Thornhill & Thornhill: Evolution of sexual coercion
tions of rape compared to consensual sex. Abel et al. 's
study used very small samples, however, comparing only
18 rapists and seven nonrapists. In addition, all their
subjects were patients in the same mental institution, but
otherwise the nonrapists were not an appropriate com-
parison group because the majority of the rapists were
heterosexual whereas four of the nonrapists were known
to be aroused by homosexual stimuli, and about the same
number were pedophiliacs. Barbaree et al.'s study was
also not controlled adequately; although they compared
the sexual responses of rapists with those of heterosexual
male college students, the responses were measured
under what were probably very inhibitory conditions for
male college students: the treatment and incarceration
areas of a psychiatric center. Malamuth (1981a), Langevin
(1983), and others have also criticized Abel et al. and
Barbaree et al. on methodological grounds.
Four other studies reviewed by Malamuth (1981a)
show that for undergraduate university men simulations
of rape can be just as stimulating as simulations of consen-
sual sex (Farkas 1979; Malamuth 1981b; Malamuth &
Check 1980a; Schmidt 1975). The sexual responses of
men to rape in these studies varied with the victim's
sexual response in the depictions. If the victim was
portrayed as eventually becoming "involuntarily sexually
aroused" by the sexual assault the subject became just as
aroused (both in terms of self-reports measures and penile
tumescence) as by consensual sex (see also Malamuth &
Check 1980a; 1980b; Malamuth et al. 1980a). On the
other hand, when the victim behaved continuously as if
the sexual assault were abhorrent to her, men showed
significantly less arousal (Malamuth et al. 1980b; Mal-
amuth & Check 1980a; 1980b). Malamuth (1981a) also
pointed out that the two studies (Abel et al. 1977; Barba-
ree et al. 1979) that had found nonrapists to be more
aroused by consensual sex than rape used tapes that
emphasized the victim's abhorrence during the sexual
assault.
A more recent study of college undergraduates by
Malamuth et al. (1986) supports the finding that when the
rape victim shows signs of sexual arousal, men find it as
stimulating as noncoercive sex. In this study each subject
read two sexually explicit stories. One portrayed a hetero-
sexual rape. In describing the rape scene the investigators
said, "We sought to make it representative of rape depic-
tions typically found in pornography. Consequently,
there was some suggestion that despite her nonconsent
and pain, the victim showed some sexual arousal." The
other story was a sexually explicit account of a man and
woman having mutually consenting intercourse. The sub-
jects' reports of their sexual arousal to the two stories were
significantly different: Mutually consenting sex was more
exciting than rape; for penile tumescence, however, the
difference between the two stories was small and not
statistically significant.
Much of the research by Malamuth and his colleagues
(see reviews in Malamuth 1981a; 1984; also Tieger 1981)
has been focused on identifying variation among under-
graduate college men in "propensity to rape." He found
considerable variation in each of a number of large sam-
ples derived from different colleges in North America.
Malamuth's subjects were asked to rate themselves on a
five-point scale of likelihood of committing heterosexual
rape if they could be assured of not being caught or
370 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
punished. The scale ranged from "not at all likely to rape"
(scored as 1) to "very likely to rape" (scored as 5). Thirty-
five to 44% of the sample, depending on the study, scored
2 or higher and were accordingly interpreted by Mal-
amuth as likely to rape. It has been shown repeatedly that
high-likelihood-of-rape subjects show greater sexual
arousal (in both self-report and penile tumescence data) to
rape depictions than low-likelihood-of-rape subjects, but
that the two groups typically exhibit similar sexual arousal
to depictions of mutually consenting intercourse. It has
also been repeatedly shown that the self-reported likeli-
hood of rape across the range of ratings 1-5 is positively
correlated with sexual arousal to rape depictions, but not
with arousal to consenting depictions (Malamuth 1984,
for review of studies).
Malamuth and Check (1983) conducted an experiment
using sexually explicit audiotapes that varied according to
whether or not the woman consented and whether the
outcome of the sexual interaction for her was arousal or
disgust. Both self-report and penile erection revealed
that when the woman was portrayed as experiencing
disgust, both low- and high-likelihood-of-rape subjects (as
measured by the 5-point scale, above) were less aroused
sexually by the nonconsensual than the consensual inter-
actions. If the woman was depicted as eventually showing
some sexual arousal, however, low-likelihood-of-rape
subjects (56-65% of the various samples) found both
kinds of interactions highly arousing (as measured by self-
report and penile tumescence). These findings suggest
that sexual arousal in the victim in simulated rape disin-
hibits the sexual arousal of nonrapist men. We now turn to
studies that indicate that not even sexual arousal in the
victim is required to produce high arousal in nonrapist
men hearing or viewing rape simulations.
Part of Barbaree et al.'s (1979) study suggested that the
woman's sexual arousal or consent was not critical in
nonrapist men's reactions to erotica. They studied the
penile reactions of male university students to sexually
explicit audiotapes that varied in how they portrayed
female consent and arousal. In the first tape the woman
initiated sexual foreplay and was enthusiastic throughout;
in the second, the woman was passive, neither rejecting
nor accepting the man's sexual advances; in the third, the
woman was resistant at first, but was finally seduced. The
authors concluded that neither the woman's sexual con-
sent nor pleasure was a critical variable for the subjects in
the study, who responded similarly to all three stimuli.
Briddell et al. (1978) studied the effect of actual alcohol
consumption and beliefs about it on the sexual responses
of male undergraduate social drinkers. Half were given an
alcoholic beverage, the other half a nonalcoholic one; half
of each group was told they had been given alcohol, half
that their drink was nonalcoholic. All were then pre-
sented with taped narratives portraying (1) mutually en-
joyable coitus, (2) rape, and (3) sadistic aggression. The
rape tape contained no indication of an enjoyable sexual
response of the victim. In the sadistic aggression, the
aggressor was a man portrayed as using deadly force, and
the victim was a woman, but there was no reference to
sexual involvement.
Briddell et al. found that the average penile tumes-
cence to consensual sex and the rape were the same for
the subjects who believed they had consumed alcohol,
regardless of whether they actually had. The subjects who

believed they had consumed a nonalcoholic beverage, on
the other hand, showed significantly less arousal to the
rape than the consensual intercourse even when they had
actually consumed alcohol. All four groups were signifi-
cantly less aroused by the sadistic aggression than by the
other two tapes, suggesting that men are not sexually
aroused by violence per se (even when it occurs between
a man and a woman; see also sect. 8.4).
A study by Quinsey et al. (1981) examined the effect of
the belief that sexual responsiveness to unusual themes
was expected in the testing situation. The study included
four groups of men. Two groups were composed of insti-
tutionalized mentally ill subjects: (1) rapists and (2)
nonsex-offenders. The other two groups were composed
of men from the local community, most of them unem-
ployed. One group of community men was given "regu-
lar" instructions: They were asked to listen closely and
imagine that they were the man in the tape. The other
group of community men was given these instructions
plus three sentences of "permissive" instructions that
stated that sexual arousal to unusual sexual stimuli is
normal. All men listened to the following kinds of au-
dio tapes: (1) neutral tapes describing nonsexual and non-
aggressive interactions between a man and woman, (2)
tapes explicitly describing foreplay and heterosexual
coitus with a willing partner, (3) tapes explicitly describ-
ing situations in which significantforcewas used by a man
in achieving sexual intercourse with a female stranger
who remained unwilling throughout the interaction, and
(4) tapes involving no sexual activity but describing a
woman being beaten and injured by a man.
There were no significant differences among the four
groups of men in penile responses to tapes depicting the
neutral heterosexual interactions, nonsexual physical
abuse, or mutually consenting intercourse. All four
groups showed significantly more arousal to rape and
mutually consenting intercourse narratives than to physi-
cal abuse and neutral narratives. The rapists responded
significantly more to rape narratives than the nonsexual-
offender patients and the community men with regular
instructions, but the community men who had received
permissive instructions showed a significantly greater
response to rape narratives than the community men with
regular instructions, and they did not differ significantly
from the rapists in their response to the rape narratives.
Quiesey et al.'s study suggests that the social inhibitions
of normal men affecting their sexual arousal to depictions
of rape (without any sexual arousal in the victim) can be
easily modified by brief instructions to the effect that
sexual arousal to unusual sexual stimuli is normal
Malamuth's (1981b) study of the sexual responses of
undergraduates examined further the effect of the con-
tent of portrayals of heterosexual rape on men's sexual
response. In the first part of the study, the subjects
viewed 16 slides accompanied with a narrative by a male
voice. The slides and text were similar to a recent issue of
an erotic magazine. There were two versions of the series.
The first 11 slides and narrative were identical, but the
last five went in the direction of either consensual sex or
rape. The study also varied the subjects' instructions: Half
received permissive instructions and half neutral ones, as
above. All men showed high levels of sexual response.
Most important, there was no significant difference in
penile tumescence or self-reported arousal between men
Thornhill & Thornhill: Evolution of sexual coercion
viewing the rape and the consensual sex. Furthermore, in
this study, in contrast to Quinsey et al.'s (1981) study
discussed above, permissive versus neutral instructions
did not have a significant effect on men's arousal in
response to rape, which was equally high in both cases.
Blader and Marshall (1984) have also shown in a study of
male university students that groups given "normative"
versus "nonnormative" (i.e., permissive and neutral) in-
struction do not differ significantly in their response to
depictions of rape. In the discussion of the results from
the first part of his study, Malamuth emphasizes that "the
data suggest that within the context of explicit sexual
stimuli, the manipulation of the woman's consent alone
does not significantly affect nondeviants' arousal" (p. 41).
In the second part of Malamuth's (1981b) study, the
men were presented with an audio narrative (no pictures)
describing an armed rapist violently raping a woman who
clearly abhors the assault during the entire experience.
Malamuth emphasizes that the rape narrative was "vir-
tually identical" to the one used by Abel et al. (1977) and
Barbaree et al. (1979) except that in his study it was read
by a woman instead of a man. As we discussed earlier, the
studies of Abel et al. (1977) and Barbaree et al. (1979)
suggested that nonrapists exhibit less sexual arousal to
rape depictions than rapists do. In the second part of
Malamuth's study, the men responded with the same
high arousal to the audio-only rape narrative read by a
woman as to the audiovisual consensual and rape series in
the first part of the study. Hence, even a rape depiction
that emphasizes the victim's abhorrence can stimulate
high sexual arousal in normal men when read by a woman
rather than by a man (see also Farkas 1979).
Blader and Marshall (1984) found that reporting arousal
while being exposed to erotic stimuli can reduce arousal.
Penile tumescence was measured in male undergradu-
ates in response to audiotape stimuli. Half the subjects, in
addition to having penile tumescence measured con-
tinuously, reported their subjective arousal throughout
the period of exposure to the stimuli by pushing a lever.
There were four audiotape stimuli: consensual sex, rape
with restraint, rape with gratuitous physical violence, and
assault against a woman with no ostensible sexual content.
All stimuli were read by a male voice. The rapes were not
portrayed as producing sexual arousal in the victim. After
the experimental stimuli, a consensual sexually explicit
videotape was presented to all subjects to elicit maximum
penile tumescence, for comparison with their responses
to each audiotape stimulus. In addition, for each subject
an index of relative arousal to rape was computed by
dividing the response to rape by the response to consen-
sual sex.
The subjects showed a higher relative response to the
restrained rape (index = 66%) than the violent rape
(45%). This pattern was then, broken down in terms of
whether or not they were asked to report their sexual
arousal during the measurement. There was a significant
difference in the rape index between the report (34%) and
no-report (58%) groups for violent rape but not for re-
strained rape (65% and 68%, respectively). Self-report of
arousal during the period of stimulation lowered penile
response with violent rape only, not with restrained rape,
consensual sex, or nonsexual assault. Blader and Marshall
(1984, p. 629) say:
The critical element here seems to be the addition of
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 371
Thornhill & Thornhill: Evolution of sexual coercion
violence. It appears that cognitive factors involved in
the self-report task can influence physiological re-
sponses, but only at a point beyond a threshold of
inappropriateness, i.e., the introduction of clearly gra-
tuitous violence. Male sexual arousal to coercive stim-
uli may be a more robust effect than has been pre-
viously believed.
iS= It appears from the
laboratory studies discussed above that not only incarcer-
ated rapists but many other men (the studies collectively
implicate young men in general) are sexually aroused to
similar degrees by stimuli explicitly portraying consen-
sual sex and rape. The inhibition of sexual response to
depictions of rape experienced by many young men can
be eliminated by any of the following factors: (1) signs of
"involuntary" sexual arousal in the rape victim, (2) the
belief that one has consumed alcohol, (3) instructions that
sexual response to unusual stimuli is normal, (4) the
narration of a rape by a woman rather than a man, and (5)
not requiring men to report their sexual arousal while it is
being measured by phallometry. Note that sexual arousal
in the women is only one of several variables that can
disiehibit men's sexual arousal in response to rape depic-
tions, and not a necessary one.
Both the ease with which men's arousal to forced sexual
interactions can be disinhibited and the similarity of their
responses to consensual sex and rape under these labora-
tory conditions are consistent with the hypothesis of a
specific adaptation to rape: Ancestral men were not sex-
ually aroused only by sexually willing women; consent
was not a prerequisite for arousal. What is the evidence
that men's sexual responses in the laboratory, however,
are related to their actual sexual behavior, including
sexual coercion? First, research by Malamuth and his
colleagues shows that the laboratory measures of arousal
in response to depictions of rape are positively correlated
with: (a) self-reports of personal use of force against
women in sexual relations and (b) self-reports of the
likelihood of using sexual coercion in the future (for
review see Malamuth 1984; also Malamuth 1986). Sec-
ond, measuring penile tumescence in response to audio
and video stimuli is felt to be the best method of identify-
ing the sexual preferences of men who seek psychiatric
assistance for homosexuality, bisexuality, or pedophilia
(Greer & Stuart 1983; Langevin 1983). Third, the sexual
arousal of men with homosexual, bisexual, or pedophilic
preferences to laboratory depictions likewise seems to be
correlated with their history of actual sexual behavior
(Langevin 1983). Finally, the typical mate preference of
heterosexual men is young women (Buss 1987; 1989a;
Felson & Krohn 1990; Symons 1979; 1987b; Thornhill &
Thornhill 1983; Townsend 1987; 1989), and laboratory
studies of sexual arousal confirm that heterosexual men
show the greatest penile response to women between the
ages of 18 and 25 and significantly less to pubescent or
prepubescent girls (reviewed by Langevin 1983).
It was predicted earlier in the target article that men
would be sexually responsive and capable in both coer-
cive and noncoercive sexual interactions. The laboratory
studies discussed above examine only penile erection, not
the actual sexual motivation involved in initiating and
performing sexual coercion. There are no reliable data on
the comparative sexual competence of men in actual
372 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
coerced and noncoerced sexual settings. Thornhill and
Thornhill (1983; see also Palmer 1989) have critically
evaluated the literature on "sexual dysfunction" of rapists;
the most relevant point is that many men do achieve
sexual access by coercion, including physical force, ac-
cording to the literature on acquaintance rape, rape by
boyfriends and husbands, rape of children, and homosex-
ual rape in prisons (see references above). Indeed, we
argued above that male coercion in one form or another is
involved in some (perhaps large) proportion of human
copulations. This widespread use of force suggests that
men are quite capable of sexual performance in the
context of coercion.
s role of wioience* Rape contains both sexual and
violent stimuli. Are men aroused only by the sexual
stimuli and not by the violence in rape, and does this
account for the data on their similar reactions to coerced
and consensual sex in the laboratory? The study of under-
graduate men by Blader and Marshall (1984) discussed
above specifically addressed this question. First, the
subjects showed low arousal to depictions of nonsexual
violence by a man to a woman compared to their arousal to
sexual stimuli. Second, the men were less aroused by
depictions of gratuitous violence accompanying rape than
by depictions of "restrained rape," in which the level of
violence was no higher than what was "necessary to
accomplish rape."
Men's sexual arousal to violence per se has been exam-
ined in other studies of nondeviants as well as incarcer-
ated rapists. Quinsey et al. (1981) reported that most
rapists do not show much sexual arousal to depictions of
nonsexual violence toward women, not differing from
normal men in that respect (also, Briddell et al. 1978; see
sect. 8.2). Overall, men's penile responses to nonsexual
violence are about the same as their responses to neutral
stimuli (see review in Langevin 1983).
Quinsey et al. (1984), in contrast, reported that incar-
cerated rapists showed substantially more sexual arousal
(penile tumescence) than nonrapists to depictions of non-
sexual violence involving a woman victim and man per-
petrator. (The authors suggested that the discrepancy
between their 1981 and 1984 findings might have
stemmed from the chance inclusion of more sadistic
individuals in the rapist group, but they could not test this
possibility.) In the 1984 study, however, the rapists were
not responding to violence per se, because their arousal to
nonsexual violence involving men only was no different
from their responses to neutral stimuli (men and women
in nonsexual, nonaggressive heterosocial settings). In
Quinsey et al. (1984) as well as in Abel et al. 1977, the
degree of arousal of individual rapists to nonsexual male-
on-female violence was positively correlated with their
arousal to rape depictions but not with their arousal to
nonsexual violence involving men only. As Quinsey et al.
emphasized, their results indicate that the rapists were
sexually responding to violence against women, not to
violence itself.
9. Aggressiwe control of the sexual partner
We suggest that violence per se will not be sexually
stimulating for most men, but that, as predicted earlier,
men will tend to be sexually aroused by achieving physi-

cal control of an unwilling sexual partner through force
because it represents an evolved cue that mating can now
be successfully obtained or completed.
That men find aggressive control of a sexual partner
during rape sexually arousing is suggested by aspects of
Quinsey et al.'s (1984) study. They recorded the penile
responses of rapists and nonrapists to audiotaped stories
about heterosexual bondage and spanking but without
sexual content. Two categories of bondage and spanking
stories are relevant here: (1) consenting bondage and
spanking with a female partner (woman tied up and
spanked), and (2) nonconsenting bondage and spanking
with a female victim. The subjects in the study inter-
preted these depictions as implicitly sexual: The rapists
and nonrapists were significantly more aroused sexually
by both bondage and spanking stories than by neutral
stimuli. Rapists and nonrapists responded identically to
both consensual and noncoercive stories.
The prediction that men will find aggressive, even
violent, control of unwilling women stimulating is also
supported by the common inclusion in hardcore erotica
(Dietz & Evans 1982) of sexual violence with women as
victims and men as perpetrators. According to the re-
search report of the President's Commission on
Obscenity and Pornography (1970), the pornographic
magazine and movie business caters to the "average" man
and not just to men with "anomalous" sexual preferences.
This implies that many men are sexually motivated by
vicariously assuming physical control over their sexual
partners while fantasizing with pornographic material.
Serial murders may be relevant to the idea that physical
control of sexually unwilling women is sexually stimulat-
ing to men. Serial murderers kill many victims often over
several years; they are almost always men and the victims
are almost always women, especially young women. 'Se-
rial murderers seem to be motivated to murder, in part,
to achieve sexual arousal from the aggression accompany-
ing the murder (Langevin 1983; Leyton 1986). Serial
murders may be a pathological expression of the hypothe-
sized psychological adaptation that makes physical con-
trol of unwilling women facilitate men's sexual arousal.
Many incarcerated rapists report that physically dominat-
ing their victims is sexually stimulating and motivates
rape (Groth's, 1979, "sadistic" and "power" rapists; Lan-
gevin 1983, Chapter 12).
A relevant note pertains to pedophiles, men with
sexual preferences for children. It seems that the sexual
arousal of pedophiles is frequently and perhaps primarily
facilitated by the vulnerability, lack of dominance, and
helplessness of their child sexual partners (studies re-
viewed in Howells 1981; Langevin 1983, Chapter 8;
Quinsey 1977). Pedophiles have a low general sense of
efficacy and self-esteem in their social relationships with
both women and men. In the literature on pedophilia, it is
argued that relating to children sexually gives pedophiles
a feeling of power and control that is otherwise absent in
their lives. Finkelhor (1984) is critical of this argument as
an explanation for the sexual interest in children shown by
pedophiles. He points out that those who propose the
argument apparently feel that sexual arousal automat-
ically follows from the pedophile's dominance and control
of a child. Finkelhor agrees that the argument could
account for the nonsexual motivation surrounding ped-
ophilic behavior, but not a sexual preference for children.
Thornhill & Thornhill: Evolution of sexual coercion
But if selection has designed the men to find physical
control of unwilling mates sexually facilitating, the ped-
ophile's ability to dominate children would be a salient
feature in their sexual arousal.
In summary, there is some evidence that aggressively
dominating unwilling mates facilitates the sexual arousal
of men. Additional studies might provide direct evidence
of psychological adaptation specifically for rape itself. It
would be difficult to-explain men's sexual arousal in
response to the physical domination of an unwilling mate
as a side-effect of a specific noncoercive sexual adaptation,
along with a general adaptation that is designed to secure
species-specific goals by force yet is not regulated by goal-
specific cues.
The results from the study of bondage and spanking,
discussed above, are especially intriguing. Apparently,
men find the theme of nonconsenting bondage and spank-
ing (women tied up and spanked) sexually arousing de-
spite the absence of sexual content. Why should such
themes be interpreted as sexual at all, unless there is an
adaptation to rape? Any future studies of sexual arousal to
depictions of bondage should include women's reactions
to tied-up men. The rape-adaptation hypothesis would
predict that women will not find the achievement of
physical control of a man sexually arousing.
10- Coerciwe sexuality of men: Additional
influences
It was predicted that sexual coercion would be related to
age and social status in men: Young men and socioeco-
nomically deprived men are indeed the most likely to use
force in connection with sex (reviewed in McDermott
1979; Russell 1984; Thornhill & Thornhill 1983), but
behavior could just reflect differential opportunity. Bet-
ter evidence would come from studies that examined the
influence of men's age or social status on sexual arousal to
viewing rape and consensual sex. Such studies have not
yet been done, but they could provide more direct
evidence of adaptation to rape. If researchers found that
young and middle-aged men were similarly aroused by
viewing consensual sex but young men were more highly
and quickly aroused by rape, this result would be difficult
for the side-effect hypothesis to explain. A similar effect
with low- and high-socioeconomic-status men would also
favor the rape-specific hypothesis. As L. Cosmides has
pointed out to us, these effects would be even more
decisive if the groups of men compared were engaging in
similar amounts of sexual intercourse.
It was predicted that regardless of social status and age,
men will pursue sex coercively when it will not harm their
social reputation. That men are sensitive to the effect of
rape on social reputation is indicated by the data from the
studies mentioned above showing that rape is considered
personally acceptable by many men when there is no
chance of being caught or punished. Moreover, the fre-
quency of rape of strangers in war (Brownmiller 1975;
Shields & Shields 1983), and even in peacetime in West-
ern societies, supports this prediction. War and Western
societies provide a degree of anonymity that allows men
to try to get sex by force with no negative effects on their
reputation for moral sexuality. Shields and Shields's
(1983) analysis of rape by U.S. and Vietnamese soldiers
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 373
Thornhill & Thornhill: Evolution of sexual coercion
during the Vietnam War indicates that anonymity and
other social factors that can affect the probability of
punishment for rape (e.g., if rape is condoned or pro-
moted by a warring government) are important determi-
nants of men's motivation to use force. Even acquaintance
rape may often be motivated by a sense of anonymity and
impunity in large societies. Rape of nonstrangers is more
common than rape of strangers (e.g., Russell 1984), yet
the literature on sexual harassment in the workplace
(e.g., Report of the U.S. Merit Systems Protection Board
1981; see Studd & Gattiker, in press, for excellent review)
reveals that men rarely rape women whom they interact
with regularly in situations in which detection of rape or
rumor of rape could damage men socially.
As discussed earlier men's arousal to rape depictions is
disinhibited by (a) instructions to the effect that arousal to
unusual stimuli is normal in the context of the experiment
and by (b) measuring sexual arousal only by phallometry
rather than by requiring concurrent self-report. Thus
men's motivation to appear moral appears to serve as a
restraint on their tendency toward sexual coercion. It is
still conceivable that this arises as a side-effect of psycho-
logical adaptation for regulating coercion together with an
adaptation for consensual sex.
11D Sexual conflict within mateships
Here we briefly summarize the evidence that, as pre-
dicted by the rape-specific hypothesis, conflict of re-
productive interests resulting from women's sexual in-
fidelity is an important cause of rape within pair-bonded
mateships. There is considerable evidence that men
equate sexual unwillingness and resistance in their long-
term mates with infidelity, and that the rape of mates is
motivated by sexual jealousy. First, the rape of a long-
term mate is particularly likely to occur during or after the
breakup of a mateship (Finkelhor & Yllo 1985; Russell
1982, in which concern about infidelity is directly impli-
cated). Second, the strong relationship between wife
battering and sexual jealousy of husbands has been dis-
cussed by several authors (see especially Daly & Wilson
1988). There also appears to be a strong relationship
between battering and raping the wife. "Sexual matters"
were the major source of marital conflict in Finkelhor and
Yllo's (1985) study of marital rape victims. Russell's (1982)
study reveals that sexual matters, including sexual jeal-
ousy on the part of the husband, played some role in 53%
of the wife beatings reported by victims of wife rape.
Frieze (1980) studied 137 chronically battered wives, 34%
of whom reported rape as well as battering. She found
that the wives of battering husbands who also raped
perceived their husbands as more sexually jealous than
did wives of husbands who battered but did not rape. This
study also showed that the degree and frequency of
violence toward wives by husbands who had both bat-
tered and raped their wives was significantly greater as
was their likelihood of beating their wives when they
were pregnant and the lengths to which they would go to
control the behavior of their wives, even to the point of
extreme claustration.
If future research were to reveal that coercive sex in
mateships is induced by cues that indicate the woman's
374 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
sexual infidelity, there are still two possible interpreta-
tions: (a) the rape-specific hypothesis and (b) the side-
effect hypothesis - an adaptation to enhance paternity
reliability and a general adaptation to pursue goals by
force. Either way, further research on this issue would
undoubtedly yield important information about the de-
sign of men's sexual psychology.
There is evidence that men in pair-bonded mateships
adjust ejaculate size and possibly ejaculate composition as
well, to conditions of ejaculate competition (Baker &
Bellis 1989b). Thus if female sexual unwillingness was
indicative of infidelity in pair-bonded females in human
evolutionary history (see sect. 5), it might be found that
ejaculates produced during rape have features designed
for effectiveness in ejaculate competition. This finding
would not necessarily demonstrate adaptation to rape,
however, because the ejaculate properties would reflect
selection in the context of ejaculate competition, and
ejaculate competition is not specific to rape.
and suggestions for future
The hypothesis that men have psychological traits that are
designed for the specific purpose of rape has survived the
following tests: It is consistent with (a) what is known of
the natural history of men's sexual coerciveness; (b) the
results of laboratory studies of arousal to depictions of
sexual coercion; (c) the results of laboratory studies and
other evidence indicating that men are sexually aroused
by physical control of an unwilling mate through force; (d)
data suggesting that men's desire to give the appearance
of having a moral sexuality is an important condition
regulating men's use of sexual coercion, and (e) informa-
tion on rape in mateships. In addition, there is reason to
infer that our male evolutionary ancestors sometimes
enhanced their reproductive success through rape. There
is still insufficient evidence, however, to demonstrate
phenotypic design specifically for rape. Thus, it must be
concluded that all current data on the sexuality of men
(with the possible exception of [c]) can also be explained
by the hypothesis that rape is a side-effect of more general
adaptation.
The data discussed in the target article also clarify the
rival hypothesis because certain side-effect explanations
for rape are falsified by the findings. In the literature, one
explanation of coercive sex is that it stems from a patho-
logical psychological condition (for references and discus-
sion, see Russell 1984; Shields & Shields 1983; Thornhill
& Thornhill 1987), a hypertrophied side-effect of a more
general adaptation either to consensual sex or to attaining
goals by force or both. Coercive sex seems-too widespread
to be explained as a pathological side-effect, however.
Another side-effect theory is that heterosexual men in
general are sexually aroused by almost any context involv-
ing a woman, and this unregulated libido, coupled with
men's violent nature, leads to rape (see Russell 1984).
According to evidence in this paper, however, men's
libido is not unregulated in the presence of women as
stimuli. Most men do not find depictions of neutral or
gratuitously violent man-on-woman interactions without
explicit sexual context sexually arousing. Also, it seems

that both men's arousal to laboratory depictions of
coerced sex. and their actual use of sexual coercion are
influenced by concern about social reputation. Men's
sexual arousal and behavior in the context of coercive sex
are clearly regulated, but is the regulation specific to
sexual coercion? The answer to this question is crucial to
distinguishing between the two rival hypotheses.
The evidence in this paper does not contradict the
hypothesis that rape is a side-effect of the interaction of
two types of adaptations, neither of which evolved in the
context of sexual coercion: (1) men's adaptation for copula-
tion and (2) a species-typical adaptation to pursue by force
goals (such as comfort, sexual satisfaction, a full stomach,
status, etc.) that consistently promoted high reproductive
performance during human evolutionary history. Accord-
ing to this view, men's pursuit of these goals by coercive
means is not regulated by goal-specific psychological
mechanisms and environmental cues, only by very gen-
eral goal-independent ones: That is, the same environ-
mental cues and the same psychological mechanisms
influence men's use of force to obtain any goal, whether
sexual or otherwise.
Laboratory experiments like those discussed in this
paper can control confounding variables and help deter-
mine the actual cues that men are adapted to process.
Although slides and videotapes of sexually explicit activity
typically evoke greater arousal in men than audiotapes
(Laws & Osborn 1983), audio narratives read by research
assistants have the advantage that presumptive causal
variables can be easily and effectively manipulated.
Other approaches, such as self-reports of arousal to
sexually explicit stimuli and naturalistic studies of sexual
conflict between pair-bonded men and women (see Buss
1989b) could also yield important information. Variables
that may influence the risks and reward of rape could also
be manipulated in laboratory experiments; rape should
be most likely under low risk/high reward conditions. We
discussed earlier how physical domination of a woman
influences the costs of rape for the rapist and suggested
further research on whether it might be a cue directly
affecting men's sexual arousal. Other risk factors to vary
may include (1) the social power of the rape victim's mate
and family, (2) the probability of detection or punish-
ment, as in peacetime versus wartime; (see also Shields &
Shields 1983), and (3) the age of the victim (because of its
correlation with fertility).
If men showed arousal to depictions of rape primarily or
only when rape-specific risks were minimal and rewards
maximal but were less sensitive to the same conditions
when they accompanied depictions of consensual sex, this
would favor the rape-specific hypothesis. For example,
men might show equal arousal to consensual sexual sce-
narios regardless of whether the woman was portrayed as
having a powerful family; or they might even show max-
imum arousal when the consenting partner had a power-
ful family (which is expected reversal of the kin-group
variable: positive for "honest" matings, negative for "dis-
honest" ones). Coercive sex might also be more stimulat-
ing when portrayed as occurring in wartime than in
peacetime, with no corresponding differences for consen-
sual sex.
The rape-specific hypothesis predicts that the more the
evolved interests of a man and woman diverge the greater
Thornhill & Thornhill: Evolution of sexual coercion
the likelihood of sexual coercion. This suggests laboratory
experiments in which the age of the woman is held
constant and the familiarity and expectations of future
interactions between the partners are varied.
Regardless of whether rape turns out to be a manifesta-
tion of a rape-specific adaptation or a side-effect of adapta-
tion to other circumstances research on the evolved
design of men's sexual psychology could lead to a much
better understanding of rape. It could identify the en-
vironmental cues that men's brains are designed to pro-
cess and therefore the circumstances that determine their
use of sexual coercion. This is not just a matter of theoreti-
cal importance; it is practically important, too, because it
can help identify the specific kinds of contexts that will
discourage rape.
ACKNOWLEDGMENTS
We acknowledge the support of the H. F. Guggenheim Founda-
tion. R. T. also acknowledges the support of P. Risser, vice
president for research at the University of New Mexico. We
thank G. Barlow, R. Bixler, D. Buss, L. Cosmides, C.
Crawford, M. Daly, J. Dupre, M. Ghiselin, B. Gladue, S.
Hamad, V. Quinsey, W. Shields, N. Simon, D. Symons, D.
Thiessen, J. Tooby, R. Trivers, M. Wilson, and several anony-
mous readers for their useful criticisms on the manuscript. For
useful discussion or correspondence or both about the ideas
presented, we thank L. Betzig, D. Buss, L. Ellis, A. Kodric-
Brown, L. Cosmides, M. Daly, N. Malamuth, P. Stacey, D.
Symons, J. Tooby, J. Townsend, and P. Thornhill. As always,
A., M., and P. Thornhill were inspirational. S. Andrew's help
with library work, and A. Rice's help with word processing and
useful suggestions on the manuscript are greatly appreciated.
NOTES
1. The terms sexual coercion and rape are used as synonyms
in this paper.
2o Adaptations promoted fitness in our evolutionary past; we
are not assuming that they continue to be adaptive in our
current environment; see section 3.2.
3. Polygyny here means greater sexual selection on males
than on females; there is no implication that male-female pair-
bonds are absent.
4o Note that the term heritahility does not apply to the traits
of an individual The dichotomy "genetic versus environmental"
cannot be legitimately applied to the features of an individual.
5o We thank D. Symons for help with these three points.
6o There is some evidence, however, that the relatively small
number of incarcerated criminally insane rapists who have used
extreme gratuitous violence during rape incidents (e.g., bru-
tally beating and killing victims and mutilating them after death)
are consistently more aroused by depictions of rape than by
consensual sex (Abel et al. 1977; 1978; Quinsey & Chaplin 1982;
1984).
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 375
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
Commentary submitted by the qualified professional readership of this
journal will be considered for publication in a later issue as Continuing
Commentary on this article. Integrative overviews and syntheses are
especially encouraged.
Just science?
Kathleen A. Akinsa and Mary E. Windhamb
^Xerox Palo Alto Research Center, Palo Alto, CA 94304 and ^Department
of Philosophy, Miami University, Oxford, OH 45056
Electronic mail: akins@parc.xerox.com
It was the second to last day of our kayaking trip, and as we
peered out glumly from under the flapping tarpaulin at the
pouring rain, the ocean swells, we knew we weren't going
anywhere that day. "Well, as you can see," said Ilya (our
fearless leader), pausing to clear his throat as he did before
any official pronouncement, "as you can see, this will be a day
of R, B, and F." We all looked puzzled. Silently, we each
enumerated the possibilities. "You mean *R and R', Ilya?" I
asked. (He was, I had long since realized, a gentle soul, the
kind of man who turns crimson at the slightest sexual sugges-
tion.) "Ah, well, uh, no" he stammered. "You know, like, a
tent day . . . a, um, day spent in the tent . . . Reading,
Breeding, and Feeding!" he finally blurted. We all laughed,
as much at Ilya as at the expression itself. Suddenly, Louise
looked put upon. "Now why is everyone staring at us?" she
demanded, for sure enough, four pairs of eyes had solidly
fixated on the only couple in the group. "Oh, Louise," I said,
trying to ease the situation, "it's nothing. Think of it as pure
envy - you know, that we should all be so lucky." That
seemed to do the trick - we all thought that was pretty funny.
Then, as the general mirth trailed off, Ilya again cleared his
throat, his eyes twinkling. This was it. I could tell. The
revenge of the shy person. "Well, you know, Kathleen," he
said, staring me dead in the eye, pronouncing each word
slowly, "as a woman, there's something you should remem-
ber: You do have a choice."
Personal reminiscence
What is the nature of human sexual experience and what kind of
explanation(s) do we need in order to understand it? This, we
think, is an exceptionally difficult question, one for which there
now exists but a hint of an answer. Take the story above, which is
given to serve as a reminder of the complexity arid subtlety of
human sexual interaction. In that story, there were no "explicit"
sexual acts - of intercourse, reproduction, or even courtship. In
some sense, "nothing" happened. But of course, in another,
broader sense, in the sense relevant to understanding human
sexuality, much occurred - sexual events that hinged upon the
complex sexual relations between members of that group, their
moral and political views, their views about each other, and so
on. What, one wonders, would fully explain that?
We do not purport to know the answer to this question, but
one thing seems clear at the outset: Human sexual behavior,
unlike that of frogs and snails, is complicated by two nontrivial
properties. First, our sexual behavior is tightly intertwined with
intentional mental events, with our thoughts about our part-
ners, their beliefs and feelings, with our views about sexual
practices, with our norms of moral behavior and so on. Our
sexual feelings are not modular or "cognitively impenetrable" (as
they say) in the standard sense. (Try, for example, to think of a
"pure" sexual stimulus - one that, no matter what the circum-
stances, would evoke sexual arousal in a man or a woman, a
376 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
response unaffected by that person's various intentional states.
What candidates might there be? Viewing a pair of breasts may
be arousing to a man, but certainly not invariably so; among
other things, who the breasts belong to often has a significant
effect. The same can be said for any sort of sexual caress; what is
arousing "all things considered," can be genuinely repulsive
under the right conditions.) Second, in the course of sexual
interactions, we interpret the physical acts of other people as
pieces of behavior: We assign intentional descriptions to those
acts. A penis enters a vagina - a physical act - but this, we know,
can be many things. It might be an act of mutual passion, of
attempted conception, of defilement, of adultery, of contrition,
of sexual control, of why-not?-apathetic-boredom, of prostitu-
tion, . . . or any combination of the above. In interacting with
others, we interpret human behavior, sexual and otherwise, not
just as physical events but as acts endowed with meaning.
These two points are not, we take it, particularly novel, but
they are crucial to deciding what kind of theory is appropriate to
human sexual behavior. For one, we must set aside any be-
havioristic or quasibehavioristic forms of explanation and with
them, the inherited terms of that tradition. For example, one
cannot speak, as Thornhill & Thornhill (T & T) do, of "the
circumstances of rape itself" or of the "environmental cues" that
are detected by a man's "sexual information processing mecha-
nisms," mechanisms that serve to "activate" rape behavior.
Certainly this way of speaking ("environmental cues," "activa-
tion," and so on) is intelligible and appropriate in other contexts,
such as when we explain how and why a moth follows a
pheromone trail. There we can say, in physical terms, exactly
what the "environmental cues" are and how the behavior comes
about. Here, however, the same terminology amounts to no
more than jargon. The "circumstances of rape" are intentional
contexts, when a woman has an intentional psychological state of
a particular kind1 - when she does not want to have sexual
intercourse - and this is a state that she can express in any
number of ways, or can fail to express at all. Hence, if a man is to
perceive the "circumstances of rape," he must necessarily inter-
pret the woman's behavior (or lack thereof) as indicating this
mental state - he must come to believe that she does not
want . . . Needless to say, this is the kind of complex psycholog-
ical process about which present neuroscience (or any other
cognitive science) has little information. Certainly there is no
physiological evidence of any "mechanisms" that "show special
design" to interpret "rape circumstances" or to "activate" rape
behavior; nor do we have any idea how such an interpretive
process might work. In other words, such quasibehavioristic
terminology is essentially empty in this context.
Second, if we wish to understand a specific piece of human
behavior, we must be careful in deciding what class of actions we
want to explain - what the particular act is an act of, given the
motivations and intentions of the subject. A man puts his bank
card into an automatic teller and withdraws $400. This, we
know, can be many things: an act of embezzlement, of fiscal
irresponsibility, of generosity (he intends to send it to a charita-
ble cause), of hoarding (he intends to put in under his mattress),
of revenge (it is his wife's account), of theft . . . or again, of any
combination of the same. How we explain his act will depend on
how exactly the bank transaction is intensionally construed. This
same principle holds for any explanation (evolutionary or other-
wise) of sexual behavior: Rape, we reckon, is at least as inten-
tionally complex as savings account withdrawal. For example,
we know that (1) a central motivation for very violent rapists
resides in the combination of the terror/pain of the victim plus a
(perceived) control over her sexual response (a man who says to
his victim: "Say you like it, cunt, or 111 cut your tits off!" has more
in mind than reproduction); (2) rape often occurs during war as
an act of revenge and defilement directed primarily against
one's male enemies; (3) it is common for men who exhibit
pathological jealousy to commit rape; (4) gang rape is usually
 Commentartf /Thornhill & Thornhill: Evolution of sexual coercion
more about "being one of the boys" than about uninhibited
sexual desire; (5) in recent cases of murder/rape (as opposed to
rape/murder) the central reason for the murder was not to inflict
pain but to have intercourse with a corpse. These acts, at least
prima facie, are of many different types. What T & T have failed
to establish is that all rapes have a common (adapted) psycholog-
ical cause (one that arises out of reproductive interests) or even
one cause plus a host of pathological expressions. They have
failed to establish that rape is one kind of act.
Third, any explanation of human sexuality must contain a
reasonable account of the complex interaction between inten-
tional mental states and the human feelings of sexual arousal,
sexual urges, desires, and dispositions. Now, classical philo-
sophical theories of human nature from Plato through Kant cast
this problem as a conflict between the various parts of the soul,
in particular, between the faculty of reason or intellect, on the
one hand, and "the passions" or feelings on the other - "cold"
reason being charged with the task of reigning in, controlling or
"educating" the animal passions. On this view of things, men's
sexuality is seen as a constant and seething desire for sexual
intercourse, one that the higher considerations of morality, self-
preservation, and social sanctions just barely manage to re-
strain. Moreover, should those passions somehow come to be
disinhibited, the views goes, we would see male sexuality laid
bare - man's true and innate sexual desires. This classical
account of the relation of "reason and passion" seems to be the
one implicitly adopted by T & T. For example, they take the
following as primary evidence for their theory: Experiments
that are premised on attempts to disinhibit rape behavior;
experiments that show that men will have erections in response
to a depiction of rape if these subjects are told that they have
consumed alcohol or that arousal from unusual sexual stimuli is
normal, or if the sexual narrative depicts the rape victim as
experiencing sexual pleasure; experiments in which men have
admitted that they would commit rape if there were no possibil-
ity of social consequences, and so on.
We do not ourselves know how to explain the interweaving of
sexual feelings/desires with belief - indeed, we are not sure
whether this way of putting things starts us off on the right foot.
It seems unlikely, however, that the inhibition of sexual re-
sponse is the only possible effect of intentional states upon
sexual processes: If there can be inhibition, why not enhance-
ment - or redirection or even the creation of new sexual
dispositions? One need only think here of the many genres of
pornography available - kiddie, infant, Asian, amputee, "con-
centration camp," sado-masochistic, genital self-mutilation,
necrophiliac, pregnant woman porn (not to mention the inno-
cent shoe catalogue) - to make one wonder what, if anything, is
not potentially an object of sexual arousal. And if our thoughts
and beliefs can affect the very nature of our sexual/psychological
dispositions, exactly what kind of psychological states does T &
T's theory take to be the product of evolutionary adaptation?
This is a crucial question, one on which the very coherence of
their theory depends.
Even on the assumption that the classical account of reason
and passion is correct, however, there is no reason to think that a
disinhibition of a sexual response reveals the "true" product of
selective adaptation. Consider the human motor system, for
example, which works through a set of interconnected push and
pull relations, through mutual inhibitory connections between
the tensor and flexor muscles. Thus, in some cases of cortical
damage, the tensor muscles become disinhibited and spas.ticity
results. But this does not mean that there has been a selective
adaptation for tensor rigidity, that spasticity is the "real" nature
of human motor behavior. What was selected for in this case was
the functional whole, the entire set of inhibitory and excitatory
relations which, when taken together, form motor control.
Similarly, we cannot infer that if the desire for nonconsensual
sexual intercourse can be disinhibited from the fear of social
reprisal that the disposition to rape is the true result of selective
pressure. This behavior may occur - but we cannot conclude
that it was selected for unless we know something more about
the general organizational principles of the system.
Although we have presented arguments against T & T's
evolutionary explanation of rape, this does not adequately
convey our views about the target article. We think, first, that
this research constitutes bad science. For the reasons given
above and for many others as well, the evidence does not justify
the hypothesis. 2 Then again, we do not think that this is just bad
science either. The hypothesis under consideration is that rape
is a "natural" or evolved behavior, one that is "activated" by
"environmental cues." Rape is treated, in other words, like any
other naturally occurring phenomenon (say lightning); hence
the goal of scientific research is construed as the identification of
those conditions under which rape "strikes." "If you don't want
to be hit by lightning, don't play golf in a thunderstorm" - this, it
seems, could be the only practical social implication of the
Thornhill view. Rape is explained as a phenomenon that lies
outside the realm of moral judgment.
Most people, we trust, will find this an offensive conclusion -
one that has to make a person wonder. We are not claiming, of
course, that there are no patently offensive facts in nature (there
are), or that science would be wrong to uncover them (it would
not). Rather, if the legitimate scientific interests of a researcher
lie within a morally sensitive area, such an investigation cannot
pretend to duck the moral issues behind a pretense of "just
doing science." If one wants to prove that genocide is an
adaptive behavior or that the practice of slavery has enhanced
societal fitness or that the rape of women by men is "natural,"
the very subject matter carries with it obligations - to examine
one's own conscience carefully about the ideological motivations
for selecting a particular topic, approach or hypothesis and to
address the question with unusual scientific care and sensitivity.
These things matter - and we cannot avoid treating them with
respect.
NOTES
1. Or would have, were she capable of having intentional states at
that moment.
2, For example: (a) the predictions do notfollowfrom the hypothesis
alone, but only together with auxiliary assumptions about past re-
productive strategies that are equally as dubious; (b) there are no telling
predictions, ones that we would not have expected to be true butforthe
hypothesis (would it notfollow,we wondered, that a husband should
find his wife more arousing when she fails to show sexual interest?); and
(c) despite T & T's claims, the "natural history" of male sexuality is almost
completely ignored.
Ewidence for an ewoSwi adaptation to rape?
Not f et
Elizabeth Rice Allgeier and Michael W. Wiederman
Department of Psychology, Bowling Green State University, Bowling
Green, OH 43403
Electronic mail: eallgei@trapper.bitnet
We share the Thornhills' desire for greater understanding of
men's (and women's) sexual psychology, and we agree that
evolutionary theory Is a productive framework in which to
conduct this work. We believe, however, that the Thornhills'
attempt to amass evidence to support a specific adaptation to
rape Is flawed by selective inclusion of data, misinterpretation of
findings, and neglect of research that conflicts with their
predictions.
The Thornhills assert that there Is strong evidence that men
are more likely than women to infer sexual interest by a
potential partner when no such interest exists (sect. 5, para. 2).
Citing the work of Abbey (1982) and Saal et al. (1989) to buttress
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 377
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
this conclusion is not persuasive. In Abbey's work, men and
women provided ratings of the flirtatiousness, promiscuity, and
seductiveness of couples they observed through one-way glass
who were engaging in a conversation. They provided these
ratings on 7-point scales with high numbers indicating the
presence of the quality and low numbers indicating its absence.
Because both men and women gave very low ratings (e.g., mean
well below the midpoint), the most appropriate conclusion was
that neither gender inferred sexual interest on the part of the
women. The same issue applies to the findings of Saal et al.
(1989) in that both men and women gave ratings that were below
the midpoint of the scales. In another study by Abbey (1987),
not cited by the Thornhills, surveys indicated that the majority
of both men and women (60% and 72%, respectively) reported
having their friendliness misinterpreted as sexual interest by
others, and equal percentages (40%) of men and women re-
ported that they had misinterpreted someone else's friendliness
as indicative of sexual interest.
In their first two predictions the Thornhills hypothesized that
both coercive and noncoercive sex are correlated with high
levels of sexual arousal and performance in men. According to
their second hypothesis, achieving sexual control of a woman by
force is sexually arousing to men. As support for these predic-
tions the Thornhills portrayed erotica as commonly including
sexual violence with women as victims and men as perpetrators
(sect. 9, para. 3), yet a review of 650 X-rated films found that
only 10% contained sadistic, violent, or victimized sex (Rimmer
1984). Indeed, the depiction of women in X-rated movies most
often corresponds to the male fantasy of "a woman who finds
[him] irresistible and who is eager to perform every type of
sexual act [he] may desire" (Reiss 1990, p. 138).
Sexual fantasies also lend clues to what is most arousing for
men and women. In Hunt's (1974) sample of American adults,
13% of men and 3% of women reported fantasizing forcing
someone to have sex, but nearly as many men (10%) fantasized
about being forced to have sex. Considerably larger percentages
of men fantasized about consensual activities: intercourse with a
loved one (75%); intercourse with strangers (47%); and group
sex (33%). With a college student sample, Sue (1979) also
obtained self-reports of fantasies involving forcing someone to
have sex (24% of men and 16% of women). However, 21% of
men reported fantasies of being forced to have sex, and the
majority of men (55%) fantasized about others finding them
sexually irresistible. The key element in men's arousal seems to
be cues related to the certainty of sexual activity, hence the male
fantasy, and corresponding erotic depictions, of insatiable
females who find men sexually irresistable and are willing to
have sex in any and all settings. This also explains the finding, as
noted by the Thornhills, that in general men experience greater
arousal in response to depictions involving aroused consensual
sexual partners than to depictions of forced sex (see Palmer
1991, p. 378).
Some research on arousal to coercive versus consensual sex
has focused on the degree of victim abhorrence and the realism
of rape depictions (Malamuth & Donnerstein 1984; Stock 1982).
As Stock (1983) pointed out, many so-called rape depictions
focus on erotic aspects of the situation or on the perpetrator's
sensations rather than on the victim's negative response. For
example, we considered the rape description used by Abel et al.
(1977), and subsequently used by Barbaree et al. (1979) and
Malamuth (1981b). We counted a total of 382 words in the
depiction, 37 (10%) of which were used to indicate some form of
victim resistance or negative experience. Most of the descrip-
tion focused on the visual appearance of the woman, the man's
sexual arousal, and what the man was doing to her. Similarly, of
the 183 words comprising the rape depiction used in Quinsey et
al. (1984), 37 (20%) were used in phrases describing the victim's
resistance and/or negative reactions. Stock's (1983) position is
that men would become much less aroused to realistic depic-
tions of rape, and that the portrayal of rape in erotica, and in the
378 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
existing studies of men's arousal to rape depictions, involves an
eroticized version of rape with the emphasis removed from the
aggressive nature of the act. Stock (1983) also presented data
demonstrating that women were more aroused, both phys-
iologically and subjectively, to a "distorted misrepresentation of
rape in which the victim does not surfer and no harm is done"
than to more realistic descriptions of rape. Further research
should examine whether such is the case for men.
Returning to the issue of the ThornhihV prediction of high
levels of sexual functioning for men in coercive situations, the
Thornhills concluded that there are no reliable data on the
comparative sexual competence of men in coercive versus
noncoercive sexual settings (sect. 8.3, para. 3). We are not sure
what the basis of that conclusion was. Groth and Burgess (1977)
interviewed 170 men incarcerated for sexual assault. Of the 135
men for whom data were available, sexual dysfunction was
reported more often than successful completion. Specifically,
32% reported no dysfunction during the rape, whereas 43%
reported sexual dysfunction (20% erectile dysfunction, 19%
retarded ejaculation, and 4% premature ejaculation). In con-
trast, practically none of these rapists reported similar dysfunc-
tions in their sexual relations with consenting partners; the
dysfunctions appeared to be specific to the coercive situation.
With regard to the ubiquity of male sexual coercion, the
Thornhills conclude that the rape-specific hypothesis would
have been falsified if the literature "had shown that sexual
coercion is used infrequently or only by a small percentage of
men" (sect. 7.1, para. 5). The Thornhills propose that most
sexual interactions contain an element of male coercion. Evi-
dence presented as supportive of this stance included Buss
(1989b) who found "highly significant sex differences in the
incidence of anger about sexual rejection" (sect. 7.1, para. 4). In
actuality, this gender difference was that 14% of the men
surveyed, versus 6% of the women, reported being irritated or
annoyed by their partner withholding sex at least once during
the past year (Buss 1989b). Respondents also rated on a 7-point
scale how upset they were over their partner's sexual withhold-
ing (higher numbers indicating being more upset). The means
were 5.03 (SD = 0.99) for men and 4.29 (SD = 1.16) for women
(Buss 1989b, p. 743), and although this gender difference was
statistically significant, it differed by less than one full-scale
point. Also, in the same study only 3% of the women and 2% of
the men reported being upset over partner sexual aggression
during the past year (Buss 1989b, p. 74). It appears that sexual
withholding and sexual aggression by one's mate occur in a small
minority of couples. The distress generated by sexual withhold-
ing is moderate in nature, and is similar in reported intensity in
men and women.
In their review of the literature, the Thornhills did not
include research such as that conducted by Byers and Lewis
(1988), who had college students keep daily diaries of their
dating experiences over a four-week period. They reported that
disagreements in which the man desired more intimate sexual
activity than did the woman occurred in only 7% of dates, and of
those respondents who reported such disagreements, 90% re-
ported only one or two over the four-week period. When these
disagreements did occur, less than 11% of the respondents
reported that the man continued the unwanted sexual advances
(Byers & Lewis 1988, p. 23). This is hardly evidence that sexual
coercion is used frequently (incidence) or by a large percentage
of men (prevalence). In addition, in the numerous studies
conducted by Malamuth and his colleagues on self-reported
likelihood by males to rape if assured of not being caught (see
Malamuth & Donnerstein 1984 for review), it should be remem-
bered that roughly two-thirds of men report no likelihood even
under such conditions.
The contention that men's greater willingness to use sexual
coerciveness should be related to their lower social status
(prediction 4, sect. 6, para. 8) has a number of problems with it.
There are certainly ample data indicating that women are more
 C'ommentary/Thornhill & Thornhill: Evolution of sexual coercion
likely than men to place importance on high social and economic
status (SES) in mate selection (e.g., Allgeier & Wiederman
1991; Buss 1989a; Townsend & Levy 1990), so men of low SES
may have a more difficult time attracting a mate. Low SES men
are indeed overrepresented among incarcerated rapists, but
given that most rapes are not reported (Koss et al. 1987), studies
of unidentified rapists may be important to consider in reaching
conclusions about the relationship between SES and likelihood
to rape. Studies of unidentified assailants versus nonassailants
from adolescent and adult populations have failed to find social
class or educational differences (Polk et al. 1981; Smithyman
1979). Ageton (1983) conducted a particularly remarkable five-
year prospective study of a nationally representative sample of
adolescents and also found no relationship between social class
and likelihood of committing sexual aggression. That is, contrary
to the Thornhills' contention, rapists were not notably of low
SES.
Finally, there are the recently published findings indicating
that men experience sexual coercion by women (Muehlenhard
& Cook 1988; S truckman- John son 1988). In S truckman-
Johnson's survey, a higher percentage of women (24%) than of
men (16%) reported having been coerced by the other gender to
have sex against their will. Nonetheless, does the fact that men
are experiencing sexual aggression by women suggest that
women who coerced during our evolutionary past may have
experienced reproductive success as a result of that behavior, or
that there are underlying psychological mechanisms leading to a
tendency to engage in that behavior?
ire complex and inwoiwe
John Archer
Department of Psychology, Lancashire Polytechnic, Preston PR1 2TQ,
England
The argument in Thornhill & ThornhuTs (T & T's) target article
takes the form: If X (rape is an evolved conditional mating tactic)
then Y (six specific hypotheses concerning rape being a general
propensity of men today). Most of the article is taken up with
arguing that Y is the case. Even if this were achieved satisfac-
torily, it would not demonstrate the link with X, because
alternative hypotheses could account for Y (e.g., feminist hy-
potheses linking rape and power). In the limited space available,
I shall not elaborate on this major criticism, but instead concen-
trate on demonstrating that the other part of the hypothesis, the
functional argument (X above), does not stand up to scrutiny.
If rape is to be viewed as an evolved male tactic, it needs to be
considered along with alternative tactics and the circumstances
affecting their fitness (ideally within an ESS-game theory frame-
work: Maynard Smith 1982 [see also Maynard Smith: "Game
Theory and the Evolution of Behavior" BBS 7(1) 1984.]). There
are a wide range of such mating tactics in nonhumans. In some
species, tactics are fixed in individuals during a breeding season
or lifetime, whereas in others males can show different tactics at
different times. The first case includes specialisations within a
polymorphic population, for example, sneakers and fighters
(Gross 1985) or satellite and territorial males (van Rhijn 1974). It
also includes tactics conditional on developmental conditions,
which predict aspects of the adult environment (for an applica-
tion to humans, see Draper & Harpending 1982).
In other cases, the same males show different tactics. In the
target article, rape was viewed as a conditional tactic depending
on assessment of current costs and benefits. This is not the only
possible strategy. Others include rape being a less beneficial
tactic adopted when the main one is unavailable, or one that is
adopted by individuals poorly fitted to undertake the main tactic
(Dominey 1984). Eliminating such alternatives in favour of the
preferred explanation would have provided a sharper focus than
the contrast between a single adaptive explanation and a non-
adaptive one made in the target article.
Alternative mating tactics have different costs and benefits
depending on the social system in which they operate. Where
there is polygyny, the tactic of sneak copulations will entail the
benefit of avoiding intermale competition. In apparently mo-
nogamous species, "extrapair copulations" (EPCs) also occur
(Mock & Fujioka 1990). The benefits for males are to sire further
offspring while avoiding parental care. For females, they are
probably to promote sperm competition and foster genetic
variety in the offspring (Bellis & Baker 1990; Smith 1984).
Whether or not copulation is forced is independent of
whether or not it occurs with a stable partner. Men rape both
partners and nonpartners, but any suggested functional expla-
nation will need to be different in the two cases. To understand
why this is the case, it is necessary to consider sperm competi-
tion (Parker 1970; 1984), a common form of intermale
competition.
EPCs are widespread in most monogamous species (Mock &
Fujioka 1990) and are common in humans (Bellis & Baker 1990).
It is not necessary to presume that EPCs have evolved as a result
of forced copulations because they carry fitness benefits for both
sexes (see above). Some EPCs may be forced, however, but
these would entail additional costs, for example, the likelihood
of detection by the main partner. In view of the prevalence of
mate guarding following insemination (Parker 1974), forced
copulations with already mated females may be possible in only
a minority of situations. In species where fertile females are not
guarded by males and cannot resist forced copulation, it would
be a low-cost tactic for males, but its effectiveness would be
diminished by sperm competition, as multiple matings could
occur (Smith 1984).
T & T suggest a different selective advantage for the rape of an
existing partner, that is, to facilitate mating at a fertile time
when paternity is uncertain owing to the possibility of EPCs.
Again, the effectiveness of rape as a tactic needs to be con-
sidered in relation to sperm competition. It is predicted (Baker
& Bellis 1988; 1989a; 1989b) that in species where there are
stable pairs, and there is the possibility of EPCs, males will
evolve the ability to adjust sperm numbers (and the proportion
of competitive sperm morphs) in the ejaculate according to the
likelihood of EPCs. This likelihood can be assessed on the basis
of the time the pair have spent apart. Baker and Bellis (1989b)
found that amongst a small sample of human couples, males did
show the predicted negative correlation between the time spent
together since previous intercourse and the numbers of sperm
in the ejaculate during subsequent intercourse. (There was,
however, no effect of time since last ejaculation on sperm
number.) A survey of women's sexual activities (Baker et al.
1989) indicated that the time spent away from the main sexual
partner indicated the probability of an EPC, as the theory
predicted. The conditional male response provides a mecha-
nism to maximise sperm competition when a female is likely to
have mated elsewhere. It is a much more subtle means of
ensuring paternity than rape by the main partner, especially
when ovulation is concealed.
Forced copulation by a male removes a female's choice, and
would be to the detriment of her fitness. The target article was
male-oriented in not considering countertactics by the female,
which would be expected if, as argued, rape were widespread
among evolving human males. Countertactics could either seek
to prevent rape, or to minimise the possibility of conception. As
in other mammalian species, there is ejection of semen from the
human vagina after intercourse. Analysis of flowback samples by
Baker and Bellis (1989c) showed that both the volume and
composition differ according to whether the woman had experi-
enced orgasm: Essentially sperm competition is promoted by
orgasm, hence increasing the chance of conception for the
current mating. This finding suggests a mechanism for distin-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 379
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
guishing between preferred and nonpreferred partners. If so,
intercourse that the woman finds sexually arousing would be
more likely to result in fertilisation than forced intercourse, and
indeed any intercourse that does not produce orgasm.
Apart from such differential retention of sperm, it will gener-
ally be in a female's interest to promote sperm competition. If
she is fertilized by the most competitive sperm, her sons will
have enhanced fitness. Bellis and Baker (1990) found evidence
from a large sample of young women who had a main sexual
partner that EPCs were more likely to occur during the fertile
phase of the cycle than were IPCs (intrapair copulations), and
that EPCs that occurred within 5 days of the last IPC (i.e., likely
to involve sperm competition) were more likely to occur at the
fertile phase of the cycle than EPCs not involving such "double-
matings." They argued that this pattern was the result of female
choice, and reflects the selective advantages to the female of
promoting sperm competition.
In conclusion, the circumstances under which rape was an
adaptive tactic in human evolution are likely to have been more
limited than indicated in the target article. In order to specify
precisely what they were, a more complex analysis involving
both male and female tactics and a range of possible pay-offs
need to be entered into a game theoretic analysis of the sort used
for analysing fighting strategies (Maynard Smith 1982). Only
then might it be worth making functional predictions about
social psychological studies.
The dereaiizati@ii ©f rape
Betty M. Bayer and Robert S. Sleele
Department of Psychology, Wesleyan University, Middletown, CT 06459
Electronic mail: bhayer@eagle.wesleyan.edu
He forced me into the ladies' room and raped me on a dirty cot
next to the wall while he kept the knife at my throat.
Afterward he said he was "sorry" and he "had to do it that
way". . . . The policewoman who talked to me made it clear
she didn't believe me. The hospital attendants also said they
thought I hadn't resisted enough . . . is a woman supposed to
be killed before they will believe she was afraid for her life?
Woman who was raped.
(Gager & Schurr 1976, p. 20)
There he was, a man who had the physical power to lock me
up and rape me, without any real threat of societal punish-
ment, telling me that I was oppressive because I was a woman!
Then he started telling me he could understand how men
sometimes go out and rape women. . . . He looked at me and
said, "Don't make me hurt you," as though I was, by not
giving in to him, forcing him to rape me. That's how he
justified the whole thing. He kept saying women were forcing
him to rape them by not being there when he needed them.
Woman who was raped.
(Stoltenberg 1989, p. 19)
I emerged from the warmth and conviviality of a group of
friends at a restaurant into the darkness and emptiness of a
city street late at night. . . . [a] man approached, walked past
me, and continued down the street. . . . Did he know how
frightened I was? If I had been raped, would I have been
blamed for being out alone in a dark, dangerous place? Would
I have blamed myself?
jane, talking about women's fears.
(Gordon & Riger 1989, p. 1)
When women talk about what it means to be raped and what it
means to live in fear of being raped by men, they are articulating
an understanding of what it means to live in a rape culture
(Griffin 1979). Thornhill & Thornhill's (T &T's) analysis removes
380 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
and distances men's raping women from women's experiences of
being raped. They accomplish this derealization of rape by
constructing an academic explanation of men's raping women
from an androcentric theoretical perspective ("derealization" is
the symbolic transformation of the real into the speculative in
order to protect the ego from anxieties that it fears would be
overwhelming, see Freud 1936/1964; also see Fine 1989; Smith
1989). From their sociobiological vantage point, rape is not
about a terrifying experience in women's lives, but rather a
men's mating strategy. By omitting women who are the object of
rape from their hypotheses, T & T make rape and men their dual
subjects. This phallocentric focus leads them to the tautological
conclusion in their hypothesis that "sexual coercion by men
reflects . . . psychological adaptation to rape." In this hypoth-
esis the authors construe the context of rape as only having to do
with men as acting subjects and as subjects of the act.
T & T enlist the idea of psychological adaptation to make their
evolutionary story of male sexual coercion work. Psychological
adaptation must function as cause and effect; it must lend
credence and intent to this evolutionary tale; it must move the
story beyond the tautology of "sexual coercion is rape" (Note 1);
it must rationalize social meanings in terms of natural historical
givens and thereby absolve men who rape women of their
personal responsibility for their actions. Psychological adapta-
tion functions as the transition term between environmental
events and male sexuality: "The discovery of the environmental
cues that men's brains are designed to process will help us
understand the environmental conditions influencing all aspects
of men's sexuality." Because it substitutes the scientific jargon of
"environmental cues" for the actual victims of rape who are
women, T & T's delineation of environmental events or "cues" is
but a further derealization of rape (see Stoltenberg 1989).
As the mediator between environmental events (or "cues")
and men's sexual coerciveness, the concept of psychological
adaptation is also used by T & T to locate the cause of rape
outside or beyond specific men's choices to harm women.
Indeed, the authors' discussion is of the harm rape causes, not to
women, but to men; dangers include things like status loss and
injury to men during the assault. While these costs are balanced
by some token benefits, they turn out to be costs only in theory
because T & T admit that men who commit rapes do not seem in
the least concerned with the dangers. T & T's cost-benefit
analysis, then, does not have much explanatory value but does
serve the purposes of derealization. It moves our attention away
from the real harm rape does to women and invites us to enter
into economic speculation about rape's costs and benefits to
men.
Although the consequences of harm and injury to women
have been left out of T & T's economics, women are brought into
their consideration of environmental cues shaping men's psy-
chological adaptation to rape them. Women are objectified as
both relevant and irrelevant environmental cues. As causative
environmental cues, "attractive women," be they real, fan-
tasized, or pictorial images, represent visual stimuli for male
sexual arousal. Women's nonconsent or consent is also ab-
stracted as an environmental cue - or experimental condition of
male sexual arousal, just as experimentally contrived responses
of women to rape are used as environmental cues for male sexual
arousal (sect. 8). Similarly, men's perception of women's in-
fidelity is used as just another environmental cue directing this
male psychodrama of rape (Hall 1983; Smith 1990).
Contrary to T & T's premise, however, men's rape of women
seems not to depend on men's processing of these very same
environmental events or cues. Sexually attractive women who
are sexually uninterested in or "resistant" to men's sexual
advances are presented by T & T as both involved in the
evolution of male sexual coercion and as being unimportant or
Irrelevant environmental cues because men are said to be
sexually indiscriminate and incapable of perceiving sexual Inter-
est or disinterest (sect. 5). In fact, attention to environmental
 Commentary/Thornhill & Thornhill: Evolution of sexual coercion
cues seems to coincide with a loss of male sexual arousal (sect. 8),
and male restraint with "men's desire to give the appearance of
having a 'moral' sexuality" (sect. 12). Women are made into
psychological abstractions: They are "barriers of hesitation,
equivocation, . . . resistance, . . . [and] reluctance" and "the
object of a theory of nonconsent" (Bumiller 1991, p. 102), which
together rationalize in evolutionary terms men's rape of women.
T & T thus reinscribe in their text the very strategies men have
relied upon historically to deprive rape of its brutal reality and to
de-emphasize the horror of rape. Only in this way can the
difference between women's consent and nonconsent be made
unclear. Only so can T & T erase the meaning and experience of
rape for women while they rely on a sexual objectification of
women as environmental cues to arouse and disinhibit men's
sexual behavior.
Psychological adaptation to environmental cues allows male
sexual objectification of women, and sexual objectification dis-
engages male rape from its ethical meaning: "The dirty little
secret about sexual objectification is that it is an act that cannot
be performed with any attention to its ethical meaning" (Stolten-
berg 1989, p. 55). T & T's sociobiological perspective on male
rape sustains the construction of male subjectivity as "con-
tingent upon the unreality of someone else" (Stoltenberg, p. 55).
Because T & T admit that the evidence does not allow them to
choose between their two competing hypotheses, what could
-their marshalling of all this evidence actually accomplish? Per-
haps this work serves the purpose it identifies: men's historical
domination of women. As such, the work perpetuates male rape
culture by strategically reproducing and reifying it.
Men: h genetically invariant predisposition
to rape?
Ray H. Bixler
Department of Psychology, University of Louisville, Louisville, KY 40292
Thornhill & ThornhilTs (T & T's) target article is a peculiar
mixture of the excellent and the not-so-good.1 First the excel-
lent: (1) the assumption, explicit or implicit, that the sexual
adaptations of males and females differ; (2) the criticism of
classical learning theory and the conclusion that learning results
"from causal gene-environment interactions"; and (3) the
characterization of "adaptation" as identified by its evolutionary
function and the recognition that the functional relationship
may no longer exist, especially in our species - hence, the
possibility that a behavior we are predisposed to express be-
cause it was beneficial during much of our species' existence
may no longer be adaptive.
The not-so-good: (a) "Adaptation refers to . . . purposefully
designed, phenotypic features." It is unwise to use the meta-
phor "purposefully designed" without quotation marks or a
general disclaimer that, although ideological terminology is
used as a matter of convenience, neither design nor purpose is
implied. The Thornhills know natural selection is devoid of
purpose, that it simply produces effects from environmental
causes interacting with existing genetic patterns; some readers,
however, will not be so sophisticated.
(b) "The only way to provide evidence about the selection
pressures that designed humans is to identify and characterize
human adaptations." As a generalization (b) is incorrect. The
comparative method (Mayr 1982, especially pp. 31-32; Thorn-
hill & Alcock 1983, pp. 21-26) has been of central importance in
our understanding of play (Fagen 1981; Smith 1984, Symons
1978), sex differences in behavior (Bixler 1980; 1986; Daly &
Wilson 1983; 1988; Freedman 1979; Hrdy 1981), infanticide
(Hausfater & Hrdy 1984), and incest avoidance (Bixler 1992a).
However, (b) might be an appropriate conclusion regarding the
specific behavior of human rape (which I doubt). To establish
that it would be desirable to demonstrate how and why the
comparative method fails for the behavior being analyzed (e.g.,
see Bixler 1992b, on sexual abuse of children).
(c) "Men have psychological traits whose underlying genes are
virtually fixed or invariant" and (d) "men's use of sexual coercion
will not vary with differences in genetic conditions." These
hypotheses appear to involve typological thinking. Further-
more, to suggest that the genetic foundations of a complex
psychological predisposition are "virtually fixed or invariant"
across any complex species represents a striking departure from
modern evolutionary and genetic theory. The Thornhills' basi-
cally correct position regarding learning stems from a recogni-
tion of the wide range of intraspecies and cross-species differ-
ences in genetic adaptive predispositions. In contrast, their
view of psychological adaptation assumes that all, or nearly all,
men are equally fit genetically insofar as the inclination to rape is
concerned. Their hypothesis implies a genetic pattern that has
been immune not only to selection but to mutation as well. This
apparent separation of the psychological (c) from the behavioral
(d) is so complete that it is reminiscent of dualism.
In analyzing adaptations it is important to remain aware that
we do not observe learning, motivation, and other psychological
processes. We infer them. A universally invariant psychological
adaptation to rape is here inferred from varying arousal patterns
generated in males in the laboratory. Some of the rape stimuli
that arouse men also arouse women (Stock 1982). In this in-
stance, I leave the making of inferences to the reader!
It is certainly reasonable to assume that rape is one form of
sexual behavior produced by natural selection (but see Palmer
1991b). Rape, like infanticide and xenophobic hate crimes, has
certainly promoted the relative reproductive success of some of
its perpetrators. Genetic predispositions to such behavior have
been demonstrated to vary widely in infrahuman animals,
however. Because men appear to differ dramatically in their
rape behavior, would it not, in the absence of clear evidence, be
parsimonious to assume that some of that variance is a product of
hereditary, psychological, and behavioral predispositions?
NOTE
1. Palmer (1991b) provides a critique of an earlier draft of the
Thornhills' article that can profitably be read in conjunction with this
published version because he challenges its basic tenets. [See his
accompanying commentary, this issue.]
A feminist response to
in men
Susan Brownmillera and Barbara Mehrhofb
a
61 Jane Street, New York, NY 10014 and bDepartment of Anthropology,
New York University, New York, NY 10012
Thornhill & Thornhill (T & T) propose that sometime during the
course of human evolution there was a specific adaptation for
males to rape. In other words, selection acted upon males to
produce a psychological proclivity for coercion in order to
increase their reproductive success. Skirting the trap of heri-
tability, the authors leap to an unsupported hypothesis of
asymmetric brain traits, and present a fanciful narrative of
reluctant, sexually attractive females pursued by aroused males
"cued" to overcome their resistance.
Sexual coercion and rape are not synonymous terms. We take
issue with T & T's interchangeable usage, which equates violent
assault with wily, dogged persistence, and their failure to
consider those acts of sexually inspired intimidation that are
different from actual copulation. For example, a woman who is
fondled on the street, verbally harassed at the office, or flashed
at by an exhibitionist may reasonably view these manifestations
of hostile behavior as part of a continuum of threatened sexual
aggression. Because they bear only a symbolic connection to
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 381
C ommentary/Thornhill & Thornhill: Evolution of sexual coercion
forced copulation, and hence to a strategy of male reproductive
success, they fall outside the Thornhill definition.
We agree there is no evidence that during the course of
evolution rapists left more offspring than men who did not rape.
Looking at extant nonhuman primate species, we find that only
among the orangutan has forced copulation been observed. No
reported case resulted in pregnancy (Mitani 1985). Ovulation
occurs in the human female only a few days of every month. In
terms of successful reproductive strategy, the hit or miss
ejaculations of a single-strike rapist are a form of Russian
roulette compared to ongoing consensual mating. Only in war-
fare, when large numbers of women are raped in a short period
of time, would conception rates increase.
Malamuth (1981b), Malamuth & Check (1980a), and Mal-
arauth et al. (1986) in their studies of male response to pornogra-
phy report a significant incidence of penile arousal to images of
rape. This in itself is not surprising. Visual depictions of females
as sexual playthings are created and marketed to reflect and
promote a particular phallocentric view of sexuality within the
culture. Neither Malamuth nor his coresearchers propose that
the male response to pornography has an inherent basis in the
brain. The leap to this conclusion is a Thornhill invention.
The central insight of the feminist theory of rape identifies the
act as a crime of violence committed against women as a
demonstration of male domination and power. The sexual moti-
vation, orgasmic release, is a secondary component. T & T
misread feminist theory. Brownmiller (1975) did not suggest
that rape is "primarily or solely" caused by arbitrary differences
in the way men and women are socialized about heterosexual
conduct. She stated, fairly irrefutably, that intrinsic differences
in size, strength, and reproductive anatomy permit men to rape
but deny an equal opportunity to women.
"Attractiveness" is not a dynamic in rape except in myth, most
particularly in pornographic myth, it might be noted. Victims
range in age from 3-month-old infants to 87-year-old grand-
mothers. Ten percent of rape victims are themselves male,
forced to play a surrogate female role. Ejaculation, when and if it
occurs, does not necessarily take place within the vagina of
female victims; any orifice that can be intruded upon will do.
Sticks, bottles, lead pipes, the barrel of a rifle, and other objects
of penetration are additional devices of dehumanization. In gang
rapes the splashing phenomenon, urine and semen, is common.
Far from being aroused by resistance, real-life offenders go to
considerable lengths to target victims who are least able to
resist. They use such varied means as an isolated setting, the
element of surprise, a knife, a gun, a verbal death threat,
punching, kicking, choking, and the cooperation of more than
one assailant (pairs and gangs) to ensure a terrorized, submis-
sive, compliant victim. More than 25% of reported rape offenses
are logged by police as "attempted rape" or "assault with intent
to commit rape." No less traumatic to victims, these assaults
either failed to achieve conventional penetration or were termi-
nated when the victim escaped or effectively resisted. Women
evade, thwart, or fight off their attackers two times out of three,
according to a Department of justice survey that went beyond
reported complaints (McDermott 1979). In rape as in other
violent crimes, many assailants respond to resistance (struggle
and screams that might be heard) with a flight reaction.
The Thornhills posit that "men will tend to be sexually
aroused by achieving physical control of an unwilling sexual
partner through force because it represents an evolved cue that
mating can now be successfully obtained or completed." The
antiseptic use of "partner" and "mating" to describe an act of
violence is not accidental in a theory that presents male re-
productive strategy as the key. But leaving the emotional effect
of language aside, such an "evolved cue" would be in conflict
with other cues in the male brain that presumably are coded for
copulation with willing partners, and coded for flight when faced
with resistance.
Ninety percent of violent crimes — rape, robbery, assault, and
382 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
murder - are committed by men, and largely by young men at
the height of their physical strength and athletic prowess, who
prey on those they perceive as weaker and more vulnerable.
Poverty, ignorance, alcohol, and above all, distorted concepts of
masculinity play a contributory role. Violent offense is a loser's
game. The forcible rapist (who is also the burglar, the robber,
the mugger) retires from the field when his legs give out. Social
policy must address itself across the board to the problem of
confused young males who see their manhood in physical
aggression. It is reductive and reactionary to isolate rape from
other forms of violent antisocial behavior and dignify it with
adaptive significance.
Blinded by "scienee": IHtaw not to think
about social problems
John Dupre
Department of Philosophy, Stanford University, Stanford, CA 94305
Electronic mail: dupre@csli.stanford.edu
The target article provides a striking illustration of how the
supposition that only one particular approach to a complex
problem is properly "scientific" can lead to a narrowness of focus
that verges at times on self-parody. The particular blinkers that
the Thornhills have elected to don are announced proudly in the
first sentence of their abstract: "Psychological adaptation under-
lies all human behavior." There is, no doubt, some innocuous
interpretation that makes this true: Human behavior is pro-
duced by beings that have evolved, and their ability to do
anything at all depends on a range of psychological capacities
that have also evolved. So one might equally well remark that
neurological organization underlies all human behavior or,
more to the present point, socialization underlies all the be-
havior of normal humans past infancy. The acceptable readings
of any of these slogans trade on the vague and unexplained
(except for the even vaguer phrase "somehow causally under-
lies" [sect. 3]) metaphor of "underlying," with its unfortunate
insinuation that things are not, at any rate, what they seem. But
taken as a filter on what views should be taken seriously on a
problem that is in reality social, psychological, moral, and
political, this initial slogan is disastrous.
Consequences of this tunnel vision abound in theoretical
claims, dismissals of competing approaches, and interpretation
of "data." The Thornhills dismiss (sect. 3.1) "the social learning
theory of rape" and the "feminist theory of rape," which they
take to be essentially the same, on the basis of one citation of a
review article (contrast the literally dozens of citations of a small
clique of sympathetic sociobiologists). I no more know than do
the Thornhills what the social learning theory of rape, let alone
the feminist theory of rape, may be. The opening paragraph of
section 3.1 appears to gesture toward an argument for the
conclusion that "the term learned is far too simplistic to warrant
its use as the major or only causal explanation of any human
behavior" (though I defy anyone to interpret the Thornhills'
premise (2)). This claim is so patently false - think of playing
chess,,making up sociobiological stories, engaging in a romantic
correspondence, or indeed almost any interesting human be-
havior - that some deeper explanation is needed of why anyone
should be moved to assert it. 1 The only possible explanation is
that it provides an excuse for ignoring the important questions
about rape, questions that are at least addressed by the litera-
ture the Thornhills ignore, about the social conditions that tend
to promote sexual violence.
Especially characteristic of the far reaches of scientism is the
Thornhills' embarrassment in the face of an issue with an
obvious moral dimension. They do mention the moral question
twice (sects. 6 and 12) in terms of the "desire to give the
appearance of having a 'moral' sexuality" (sect. 12). Both times
 Commentary/Thornhill & Thornhill: Evolution of sexual coercion
the term "moral" (or "morality") appears in sneer quotes. But
why nobody might even have the desire to be "moral," or even
be moral, is quite unexplained. If this is not the kind of 1930s
positivism that is still occasionally found in areas of science most
concerned with defending a dubious scientific status, one might
surmise that it reflects merely the radical and reductive egoism
that is part of sociobiology's patrimony from its economic
forefathers.
The largely meaningless claim about universal adaptationism
at the beginning of the Thornhills' target article is in fact used to
motivate a series of a priori evolutionary speculations about the
benefits of rape. 2 The Thornhills argue that it is likely that men
have evolved an adaptation specifically designed for sexual
coercion and, moreover, that almost all men's sexuality is to
some degree coercive. Although some of the evolutionary story-
telling is presented as if it were a kind of evidence (especially in
sect. 5), the Thornhills do seem aware that their speculations
about men's sexuality require some buttressing with empirical
facts. Unfortunately, this leads them into an even shadier and
more bizarre area of scientific research.
The "phallometric" studies of male responses to various kinds
of sexual, violent, and sexually violent material are so riddled
with obvious difficulties of interpretation as to be every bit as
worthless as the evolutionary storytelling that motivates their
present discussion. Obviously the data derived from studies of
prison and insane asylum inmates are irrelevant for drawing
general psychological conclusions. The investigations of the
relevance of female enjoyment suggest a conception of rape
derived from Hustler magazine, and make that part of the
discussion irrelevant to real questions about rape. More gener-
ally, the assumption that the measurement of penile tumes-
cence in response to depictions of rape is evidence of a disposi-
tion to perform violent sexual acts has all the compelling force of
the inference that overweight middle-aged men who reveal
objective signs of excitement on watching televised sports
events are disposed to play professional football. And, most
important, even if the data are admitted to bear on the question
of how widespread is men's tendency to rape, none of this has
the slightest relevance to the question of whether rape reveals
the proper functioning of an innate psychological organ or is
instead socially conditioned.
Finally, it is essential to note that this is not just bad science
but harmful science. The deeply political nature of this whole
area is somewhat humorously illustrated by the mandatory
reference (in a field in which most participants remain male and
ambitious) to the claim that "women prefer as mates men of high
social and economic status" (sect. 6). Needless to say it is young,
poor men who, lacking these advantages, will be driven to rape
(sect. 10). (The reader can be left to fill in other features of the
stereotypical rapist.) Despite the Thornhills' suggestion that the
knowledge they seek will show how to reduce sexual violence, it
seems quite clear that the biologicization of rape and the
dismissal of social or "moral" factors will both tend to legitimate
rape and to deflect attention from social factors (such as the
depiction of women in the media and advertising, and the
compliance of the legal system) that are plausibly hypothesized
to promote sexual violence. Whatever the intent of the authors,
their claims will undoubtedly be taken to show that since rape is
a "natural" phenomenon, its reduction or elimination is an
unrealistic goal. If such claims really were established, then we
would just have to accept these possibly harmful consequences;
but because, as far as I can see, the Thornhills say nothing that
even affects the probability of their truth, the propagation of
such claims should be strongly resisted.
NOTES
1. I do not mean to deny that other kinds of explanation, for some-
what different questions, are quite compatible one with another. Why
does he play chess? To pass the time/ impress his friends, etc. Why is he
able to play chess? Because he has learned/ been taught, etc. Which is
the major causal explanation is an obviously silly question.
2. The Thornhills entirely ignore the devastating critiques that have
been directed against the value of such speculations by biologists and
philosophers such as Gould, Lewontin, and Kitcher (e.g., Gould &
Lewontin 1979; Kitcher 1985; and the articles by Lewontin, Kitcher,
and others in Dupre 1987), not to mention a number of feminist
critiques directed specifically at its use in the present context (e.g.,
Fausto-Sterling 1985, Ch.6). [See also BBS multiple book review of
Kitcher's Vaulting Ambition, BBS 10(1) 1987.]
Coereiwe sexuality and dominance
Irenaus Eibl-Eibesfeldt
Forschungsstelle fur Humanethologie, Max-Planck Institut fur
Verhaltensphysiologie, Andechs, Germany
The term "coercive sexuality" as used by Thornhill & Thornhill
(T & T) encompasses a number of sexual behaviors, some of
which may be adaptive in certain circumstances. Among others,
it includes relationships of (1) dominance-submission sexuality,
(2) acts such as the rape of the defeated by warriors upon victory
that are not so much a part of men's sexual repertory as they are
acts of aggression, and (3) situations in which a female leads a
male on to the point of no return, particularly when alcohol is
involved. It is the first two aspects that I would like to discuss
from the point of view of our phylogenetic heritage.
Reptile sexuality is characterized by dominance and submis-
sion and the reptile brain, as is well known, is still with us (Bailey
1987). In reptiles, males display and then females accept by
assuming submissive postures. There is no coercion because
females accept or refuse as a function of whether they are
prepared to copulate and probably also the male's quality of
display. Dominance and submission are the patterns involved.
Male nuptial behaviors in reptiles are derived from threat
displays and females must behave submissively for copulation to
occur. There are no nurturant or affiliative behaviors observed
in courtship. Affiliative behaviors of this sort evolved in birds
and mammals, probably independently of each other. With the
evolution of behavior patterns of parenting such as feeding,
grooming, and brood defense, the motivation for caretaking
came into the world. In the infant, furthermore, the corre-
sponding motivation to seek nurturance as well as the infantile
signals to trigger such responses in the parents evolved. These
parent-child adaptations proved to be preadapted for bonding
among adults (Eibl-Eibesfeldt 1970; 1971; 1989).
Many courtship patterns, such as (1) courtship feeding in
birds and kiss-feeding in humans, (2) patterns of social grooming
and the like, which clearly originated from parental behaviors
and occur in a derived form during courtship, and (3) rituals of
greeting, serve the function of strengthening already existing
bonds. Furthermore, infantile appeals are used to appease and
attract during the various stages of courtship. In black-headed
gulls, for example, both sexes are characterized by a black face
mask, a signal that triggers aggression. Males therefore show
ambivalent behaviors toward females. They call them to their
territorial site but often attack them once they have landed. The
females escape attack by assuming a begging posture and per-
forming begging behaviors. This triggers regurgitation and
feeding of the approaching female by the male (Tinbergen
1959).
With the evolution of individualized brood care, further-
more, the ability for individual bonding came into the world. It
was of selective advantage for mothers or parents to recognize
their youngsters individually and for these in turn to recognize
their parents. The establishment of an individualized bond is
basically what characterizes what we call love in humans. The
individualized bond and affiliative motivations and behaviors
derived from the parent-child relationship are prime charac-
teristics of human courtship and sexual behavior. The old
reptilian heritage is still with us, however (Bailey 1987; Eibl-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 383
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
Eibesfeldt 1990). The human male also displays in an aggressive
way during courtship. This agonistic behavior can be seen as
functional when rivals strive to outdo each other in competition,
displaying health and strength through their performance as
indicators of a good genetic constitution. Since human males
contribute to the defence of the offspring through aggressive
display, they also demonstrate their ability for family defence.
This still has nothing to do with coercive sex. In a functioning
society, rape is rightfully classified as deviant behavior and as
such does not seem to be particularly adaptive because rapists
are persecuted and offspring produced this way are often
aborted or killed after birth.
Normal sexual relations are under the control of the phy-
logenetically newer affiliative behaviors and characterized by
individual bonding. The urge to dominate also plays a certain
role, as does submissiveness. But if either of these predominates
in sexual behavior, this should be considered deviant, as in the
cases of sadism and masochism (Eibl-Eibesfeldt 1990). This may
be seen as partial regression to an earlier vertebrate pattern, but
only to a limited degree because reptiles certainly are not
rapists. Males try to achieve dominance by display but they do
not copulate by coercion. To my knowledge, any such attempts
would not work on reptiles because the female must cooperate.
In humans, there is a close link between the achievement of
dominance and male sexual physiology. If males win a tennis
match, their blood testosterone level increases significantly
within 24 hours (Mazur & Lamb 1980). If they lose a match, the
level drops. The same holds for mental achievements. If male
medical students pass an exam with success, their blood testos-
terone level goes up; if they fail, their hormone level drops.
In addition, we know that male sexual presenting in primates
is a dominance threat, a ritualized gesture derived from domi-
nance mounting. In quite a number of monkeys, males stand
guard to protect their group against intrusions of conspecifics of
other groups. They expose their genitalia, and if foreign group
members approach, many of the guardian males exhibit erec-
tions (Wickler 1967).
In this context it is also interesting to note that phallic displays
with similar functions occur in humans. One universal expres-
sion is guardian figures exhibiting phallic displays. We find
them as grotesques on old churches, as guardian figures in
Indonesia, Africa, and other parts of the world (Eibl-Eibesfeldt
& Siitterlin 1992; Eibl-Eibesfeldt 1989; Wickler 1967). Humans
also verbalize phallic threats in agonistic encounters.
There are many ways in which the close relationship between
human sexuality and aggression manifests itself. When the last
Consul of Algier was captured by the rebels, he was homosexu-
ally raped. Well-known also are the rape orgies of victorious
troups as expressions of male dominance. In these situations
rape is more of an aggressive act than another form of human
sexual behavior. The adaptive aspect seems to be the expression
of dominance and the spiritual subjugation of the defeated.
Whether there is also a reproductive aspect has to be investi-
gated. Enough data to do this should be available from some
areas of the world.
sex plus wiolence?
Aurelio J. Figueredo
Department of Psychology, University of Arizona, Tucson, AZ 85721
Electronic mall: ajf@rvax.ccit.arizona.edu
Thornhill & Thornhill (T & T) propose two alternative hypoth-
eses regarding human rape. One is that human rape is not a
specifically selected psychological adaptation, but an epi-
phenomenal or incidental "side-effect" of two other, more gen-
eral, adaptations: (1) the male-typical sexual strategy, and (2) the
male-typical (or perhaps species-typical) coercive strategy. For
384 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
several reasons that will be made clear below, I prefer to call this
the "additive model" rather than the "side-effect" hypothesis,
because it implies that human rape represents a simple summa-
tion of two separate impulses. The alternative is that human
rape is, indeed, a specifically selected adaptation. I prefer to call
this the "nonadditive model," because it implies that the com-
bination of sexual motivation and violent coercion represents a
synergistic composite specially favored by natural selection.
To test these two competing models, T & T adopt a simple
Popperian "falsifications t" strategy. They review the available
evidence for results that are inconsistent with the predictions of
either theory, thus exposing them to the risk of falsification.
They review two major sources of evidence: (1) demographic,
and (2) experimental. The demographic information concerning
the highly suggestive natural histories of both rape perpetrators
and rape victims is consistent with either of these explanations
(and several others). Because these circumstantial congruities
have been relatively well known for some time, their status as
confirmations of a priori theoretical "predictions" is dubious and
open to characterization as post hoc. I will therefore direct my
comments to the interpretation of the experimental evidence.
To summarize T & T's review of this literature, many studies
show both rapist and nonrapist human males to be aroused
equally by coercive and noncoercive sexual stimuli, but not by
nonsexual coercive stimuli. T & T conclude that these results are
also consistent with both models and that neither model can be
thereby falsified.
T & T are not very clear, however, about what differential
predictions can be expected from the two models, other than
morphological specializations for rape (which are absent in
humans). If the models make the same general predictions, it is
not surprising that much evidence can be found to be consistent
with both of them. A simple strategy of theory falsification is
insufficient to resolve conflicts between theories that share
many empirical predictions. Neo-Popperian philosophers of
science such as Lakatos (1970) emphasize instead competition
among theories rather than "naive falsificationsm" based on
maximizing predictive capacity or "adding empirical content,"
while minimizing requisite assumptions or "employing fewer
primitives" and thus preserving model parsimony. Thus, the
specific adaptation hypothesis, or nonadditive model, should be
deemed the less parsimonious of the two, because it invokes an
extra rape-specific adaptation, over and above two already quite
familiar human impulses. This does not necessarily imply that it
is wrong, but that it should be required to explain more data
than the simpler additive model. This is a common situation in
the analysis of variance, where an additive model is given
priority over a competing model that contains an additional
interaction term. The significance of the synergistic joint effect
is evaluated on the basis of what additional variance it can
account for, over and above that of the additive main effects.
There has been some recent controversy over this general point
(see Wahlsten: "Insensitivity of the Analysis of Variance to
Heredity-environment Interaction" BBS 13(1) 1990), but that is
a different argument.
Failure to reject the null hypothesis is not therefore equally
consistent with both theoretical hypotheses. Because nonsexual
coercive stimuli were generally found to be nonarousing (al-
though using sexual arousal as the sole criterion makes this point
somewhat moot), the side-effect hypothesis, or "additive
model," would predict no difference between coercive and
noncoercive sexual stimuli; the specific adaptation hypothesis or
"nonadditive model," on the other hand, should predict an extra
interactive effect, in excess of what is predicted by the arithme-
tic sum of the additive main effects. Because no such extra
increment has been reported, the evidence therefore favors the
additive model. If rape truly represented a specific psychologi-
cal adaptation, coerced sexual stimuli should be more arousing
than noncoerced sexual stimuli. That they apparently are not is
disconfirmatory.
 Commentary/ThornhiW & Thornhill: Evolution of sexual coercion
That having been said, this partial solution remains un-
satisfactory. Several other major issues need to be addressed.
For example, much of the foregoing depends on the supposition
that the difference between rapists and nonrapists is in differ-
ential excitation by rape stimuli, whereas the difference could
be due to differential inhibition of rape responses. The experi-
mental results only show that there is no differential excitation
by rape stimuli; differential inhibition was not adequately
tested, because penile erection is not specifically a rape re-
sponse. A few studies were cited addressing such issues as the
nature of the environmental conditions, such as expected social
costs, or personality characteristics, such as differential risk-
aversion, that might inhibit or disinhibit rape behavior. More
research along these lines is needed. Another problem is pre-
cisely what is meant by the postulated "general coercive strat-
egy" in the additive model. If this generalized coercion is
proposed as a species-typical social strategy, then rape behavior
is produced in human males merely by the addition of the male-
typical sexual strategy. If generalized coercion is proposed as a
male-typical strategy, however, we encounter certain prob-
lems. In humans, as in many other species exhibiting varying
degrees of polygyny, males have more morphological and psy-
chological adaptations for the use of physical force than females.
Because this excess coercive adaptation on the part of males is
sexually selected (and thus specifically sexual in function to the
degree that it exceeds the species-typical level), the meaning of
a "general coercive strategy" becomes unclear.
To the extent that coercion is a male-typical sexual strategy,
rather than a species-typical social strategy, the only difference
in male strategies may reside in the immediate or direct target
of coercion, the ultimate or indirect target always being
the female. In intermale competition, which is thought to be the
principal function of male-typical coercive adaptations, the
direct targets are other males. The indirect targets, however,
are still the females, because female mating behavior is indi-
rectly constrained by male coercive behavior in restricting
female sexual access to competing males. Even in intermale
resource competition, as opposed to pure dominance competi-
tion, the indirect targets are females, because female mating
behavior is indirectly constrained by male coercive behavior in
restricting female economic access to resources dominated by
different males. Except in mating systems of pure unrestricted
female choice, based exclusively on male courtship displays that
advertise their genetic quality, the question becomes not
whether male sexual coercion is present, but what form it will
take. Whether the females are the direct or the indirect targets
of male sexual coercion - although this difference is no mere
subtlety - should not blind us to the near universality of some
form of male coercion that is specifically sexual in function.
Instead of the single continuum of sexual coercion proposed by
T & T, we may find a complex array of multiple and graded
coercive and noncoercive mating strategies. The additive model
of human rape, although apparently supported by the data, may
not represent an incidental "side-effect" at all, but one possible
manifestation, among many others, of the various combinations
of alternative sexual strategies available, many of them either
directly or indirectly coercive in nature.
w ewoiutionary psychology of priesthood
Jennifer J. Freyda and J. Q. Johnsonb
department of Psychology, bOffice of University Computing, University of
Oregon, Eugene, OR 97403
Electronic mall: ajjf@dynamic.uoregon.edu; bjqj@oregon.uoregon.edu
Instead of getting tangled in the powerful relationship between
socialization and adult aggression (for example, the enormous
correlation between exposure to childhood sexual abuse and
adult rape or the existence of essentially rape-free societies
[Sanday 1.979]), Thornhill & Thornhill (T & T) go right to the
heart of the matter.
As a sample of the sorts of hypotheses this kind of approach
inspires we explore the question: How can a society create and
maintain ecclesiastical celibacy? Following T & T, we hypothe-
size that this phenomenon relates to the fact that psychological
adaptation underlies all human behavior. By cultural definition,
priests remain celibate. This means that males do not need to
guard their females when left alone with a male priest. But of
course studies have shown that priests really do not remain
celibate at all. Furthermore, they are inclined toward adul-
terous affairs with the married female members of their flock
(see, e.g., Brundage 1987). To the extent that these priests
successfully impregnate their mistresses, they have the ideal
opportunity for perpetuating their own genes. Not only do they
have access to a large population of women, but better yet, they
are guaranteed that the offspring will be provided for by other
males (the husbands). The celibate clergy are therefore seen as a
straightforward institutionalization of the cuckold or "sexy son"
mating strategy (e.g., Buss 1988).
Many questions await further research, but we are confident
that this novel approach to clerical celibacy, inspired by T & T's
target article, will prove revealing concerning the human
condition.
Men are not
Andrew Futtermana and Sabrina Zirkelb
Department of Psychology, College of the Holy Cross, Worcester, MA
01610
Electronic mail: afutterman@hlycross.hitnet; b
zirkel@hlycross.hitnet
The Thornhills' account of male coercive sexuality is yet another
attempt to attribute the cause of a complex human social
behavior, in this case rape, to uncontrollable biological forces.
The authors hypothesize that coercive sexuality is the product of
evolution and is therefore an adaptation. Evolutionary theory
tells us that adaptations reflect natural selection and are deter-
mined by genes. To validate their argument the Thornhills must
demonstrate, independently of the theory of evolution, that
rape is an adaptation and is determined by the genes. They do
not do so.
What the Thornhills do instead is present a selective sampling
of studies suggesting that differences in sexual behavior be-
tween men and women, and among men, follow a regular
pattern. They then attribute this pattern to a "genetic leash" that
tugs men, particularly young, poor men, in the direction of
coercive sexuality. There does not seem to be systematic testing
of a theory of the causes of rape. No experimental evidence is
presented that specifically supports their evolutionary account.
To show that male coercive sexuality is a result of natural
selection requires demonstrating the presence of three condi-
tions: First, a pattern of clear phenotypic differences must be
present among individuals within a population. Second, these
phenotypic differences must be shown to be inherited according
to the laws of genetics. Third, the differences must be shown to
lead to differential reproductive success. In humans, these
three conditions have been demonstrated in the case of sickle-
cell anemia: The sickle-cell trait is clearly defined, the trait is
inherited according to known genetic law, and individuals with
sickle cell have fewer viable offspring than nonanemic
individuals.
With respect to coercive sexuality, however, the Thornhills
do not clearly define the trait, they present no evidence of a
gene or genes associated with rape, and do not in any way
demonstrate differential reproductive success of coercive versus
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 385
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
noncoercive individuals. In fact, we would argue that data
relating to conditions 2 and 3 do not exist, and that differences
exist between human social behavioral traits (such as coercive
sexuality) and sickle-cell anemia that make it unlikely that
evidence of conditions 2 and 3forsuch behavioral traits will ever
exist.
Unlike sickle-cell anemia, coercive sexuality or rape can only
be defined in terms of a specific social context with a specific
history. The Thornhills provide only the most general and
superficial definition of coercive sexuality and rape and ignore
the social and historical context in which such terms are mean-
ingful. By citing nonhuman animal studies, the authors imply
that a precise definition is not necessary, that human rape is a
special case of a more general natural phenomenon found in
animals. Yet, characterizing nonhuman behavior as rape blurs
qualitative differences between human and nonhuman species
(e.g.,flexibleuse of symbolic language and the ability to change
the environment in goal-directed ways). The authors totally
overlook what humans have done over thousands of years to
create societies.
Furthermore, unlike sickle-cell anemia, coercive sexuality
and rape do not admit of single definitions and therefore do not
reflect homogenous phenotypes. Definitions of what is coercive
vary across cultures and social contexts (cf. Sanday 1981).
Multiple definitions of rape have existed and presently exist in
our culture (Loh 1981). More concretely, difficulties in defining
coercive sexuality and rape are evident in the rate of rape
convictions (at least 20% of rape cases that go to trial result in
jury verdicts of not guilty; U.S. Bureau of Census 1988) and in
the general disagreement over what constitutes coercion (par-
ticularly when the woman has been drinking alcohol) in "ac-
quaintance rape" on college campuses (Muehlenhard et al.
1985; Muehlenhard & Lin ton 1987). If rape is not a single
homogenous phenotype but rather means different things to
different people in varying social and historical contexts, then
what exactly is the rape-specific adaptation about which the
Thornhills speak?
Methodological problems preclude demonstrating the sec-
ond condition, that rape (or any human social behavior) is
inherited according to the laws of genetics. To obtain even the
most rudimentary data about how genes influence human be-
havior, experiments involving the exact reproduction of particu-
lar human genotypes and the systematic manipulation of se-
quences of environments would have to be conducted with
individuals or groups. Only in this manner can the relative
influence of particular genotypes and particular environmental
sequences on the development of behavior be assessed. Func-
tions that describe the relationship between phenotypes, geno-
types, and environmental sequences, and thus assess the rela-
tive importance of genes and environment, are called "norms of
reaction." Because necessary controls over gene reproduction
and human developmental environments are lacking, norms of
reaction have not been demonstrated even for human anatomi-
cal traits (Lewontin 1984). Needless to say, data pertaining to
the extent of genetic influence on an ill-defined, complex social
behavior like coercive sexuality are likewise unavailable.
Demonstrating the reproductive advantage of coercive sex-
uality (condition 3) presupposes that conditions 1 and 2 obtain.
Specifically, individuals having genes associated with the clearly
defined phenotype "coercive sexuality" must be shown to pro-
duce more viable offspring over many generations than individ-
uals without such genes. In light of the difficulties in demon-
strating conditions 1 and 2 noted above, it is not possible to
demonstrate condition 3. Since no available data exist, the
Thornhills instead provide an "adaptive story" (Lewonton 1984)
drawn from a paper by Alexander and Noonan (1979) in support
of the assertion that coercive sexuality confers a reproductive
advantage. This story proposes that "raped females would be
reluctant to reveal to their mates that they had been raped
because this would make their mates cooperate less in the care
386 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
of existing offspring." In this manner, sexually coercive males
produce high numbers of viable offspring. This speculation
seems farfetched at best. -Certainly just as plausible would be
the story that men who raped would receive particularly harsh
punishment (perhaps even death) because they pose a risk to
socially powerful pair-bonded men, and consequently sexually
coercive men would have produced very few offspring.
In sum, the Thornhills present no data that specifically
support their adaptationist model of human coercive sexuality.
When confronted by difficult facts they simply use evolutionary
theory to explain everything away. For example, rape is not
heritable among men. The theory of natural selection demands
individual differences that are heritable. The Thornhills there-
fore conclude that natural selection must have produced a sex-
linked adaptation in humans. Just like that, no loose ends: The
theory explains everything and nothing.
In light of the lack of supporting evidence, why is the
Thornhills' "Evolutionary psychology of male coercive sex-
uality" found interesting at all? One reason these kinds of
arguments generate interest is that they reduce intractable
social problems, in this case rape, to the action of molecules in
the individual. As a consequence, we are led to assume that
there are substantial limits to the treatment of rapists, and we
are directed away from developing more comprehensive social
programs that redress systematic causes of rape (e.g., sexism).
More comprehensive educational, training, and treatment pro-
grams are expensive and require financial sacrifice (see Cantor,
in press). Furthermore, by focusing attention away from the
wide cultural variability in the sexual coerciveness of men in
different societies (see Sanday 1981) and the malleability of
recidivism rates among rapists in this country (e.g., in the state
of Washington, recidivism among rapists dropped at least 50%
following psychological treatment, CNBC Real Story, Septem-
ber 24, 1991), a model of complex social problems such as rape
suggests that the status quo is somehow inevitable and that
attempts at social change are ultimately futile. Such social
problems are not inevitable, however, and we should not allow
ourselves to be lulled into complacency about them by those
who would suggest otherwise.
Rape: The perfect adaptationist storf
Nicola J. Gaveya and Russell D. Grayb
department of Psychology, University of Auckland, Auckland, New
Zealand and bDepartment of Psychology, University of Otago, Dunedin,
New Zealand
Electronic mall: apsyjigavey@ccnov2.ac.nz and brdgray@otago.ac.nz
Biologists and social scientists have not always enjoyed an
amicable exchange of views. The nature/nurture and sociobiol-
ogy debates have left a legacy of bitterness and mutual misun-
derstanding [see BBS multiple book review of Kitcher's "Vault-
ing Ambition" BBS 10(1) 1987; and Lumsden & Wilson's
"Genes, Mind and Culture" BBS 5(1) 1982]. In recent years,
however, there has been an encouraging trend away from the
crude biological determinism and adaptationist stories that have
given biologists such a bad name. There has been a shift instead
toward nondichotomous views of behavioral development
(Bateson 1983; Oyama 1985; Johnston 1987 [see also Johnston:
"Contrasting Approaches to a Theory of Learning" BBS 4(1)
1981, and "Developmental Explanation and the Ontogeny of
Birdsong" BBS 11(4) 1988]) and an emphasis on a plurality of
evolutionary processes and explanations (Gray 1988; Lewontin
1983; Stearns, in press). Unfortunately, Thornhill &ThornhilTs
(T & Ts) adaptationist account of rape is not an example of this
newer trend. Their target article suffers from naive adaptation-
ism, a dichotomous view of development, and a superficial
analysis of the social science and feminist literature on rape.
 Commentary/Thornhill & Thornhill: Evolution of sexual coercion
Naiwe adBptationism* The Thornhills assert, as an article of
faith, that "psychological adaptation underlies all human be-
havior" (abstract). In a trivial sense, this is partly true. We are
indeed the products of an evolutionary history, part of which
would have included natural selection. For this claim to be
nontrivial, however, specific psychological traits must them-
selves be adaptations, rather than merely linked in an unspec-
ified way to some vague underlying adaptations. In an effort to
avoid the charge of vacuous adaptive story-telling, contempo-
rary evolutionary biologists typically use a combination of op-
timality models (see Gray 1987; Stephens & Krebs 1986) and
comparative methods (see Harvey & Pagel 1991) to test hypoth-
eses in a rigorous way. [See also: Houston & McNamara: "A
Framework for the Functional Analysis of Behavior" BBS 11(8)
1988; Clark: "Modeling Behavioral Adaptations" BBS 14(1)
1991; and Schoemaker: "The Quest for Optimally" BBS 14(2)
1991. ] The Thornhills' claim that men may have specific psycho-
logical traits that have evolved by natural selection for coercive
sex is not based on these kinds of analyses. Instead, in section 5
of the target article they present a loose cost-benefit argument,
"why it seems reasonable to assume that ancestral human males
sometimes increased their reproductive success by rape," and in
section 6 they make some predictions based on this hypothesis.
On close inspection, their cost-benefit argument boils down to
the claim that the advantages of sexual coercion must have
outweighed the many costs because sexual coercion exists to-
day. Establishing the existence of an adaptation by stating that
the feature exists and evolved is circular and vacuous.
According to the Thornhills, "the failure of even one of [their
predictions] would falsify the [rape adaptation] hypothesis.
Predictions 1 and 4, at least, are probably false. Although they
themselves find no data to falsify these predictions, a close
reading of one of the data sources they use to support their
argument actually suggests that prediction 4 is false. Prediction
4 suggests that men of lower socioeconomic status will be more
likely to rape than men of higher socioeconomic status. The
Thornhills cite Russell's (1984) data as evidence that most rapists
are of lower socioeconomic status (see sect. 10). There is no
doubt that Russell's research is probably the most thorough,
carefully conducted, and comprehensive available on the epi-
demiology of rape. Although Russell cautions that her data on
the social class of rapists should only be regarded as tentative,
she notes In the conclusion to her book that "the more com-
prehensive and realistic view of rape provided by Russell's
random sample survey suggests that by several measures, rape
is actually overrepresented In the upper middle class" (p. 285).
Prediction 1 (and the related prediction 2), which holds that
"men will exhibit high levels of sexual motivation and perfor-
mance in both coercive and noncoercive mating situations" is
also not clearly supported by the data. There are a lot of data to
support the claim that many men find rape stimuli sexually
arousing, but there is no similar body of data to verify that men
actually "perform" well sexually in such situations. In fact, there
Is frequent reference to "sexual dysfunction" among rapists,
Including inadequate penile erection, inability to ejaculate, and
premature ejaculation before penetration (see Harding 1985).
T & T present a somewhat confusing scenario for what would
constitute confirmatory data for prediction 5 that "sexual coer-
clveness will be very sensitive to the probability of detection and
negative social consequences or punishment." The Thornhills
seem to consider that the anonymity of Western societies results
In less risk of being caught, hence making sexual coercion more
likely (sect. 10). If this Is the case, however, stranger rape
(where anonymity is higher) should be far more prevalent than
nonstranger rape - in fact, the opposite is true (e.g., Gavey
1991a; Russell 1984).
In addition to these false predictions, there Is considerable
evidence that male sexual psychology is badly designed in terms
of reproductive interests. On the face of it, the Thornhills'
argument that male sexual coercion is an adaptation is at odds
with evidence of the following kinds: males perpetrate high
levels of child sexual abuse (e.g,. Russell 1984); sexual violence
is also perpetrated against pre- and post-reproductive age
women; similarly, men have consensual sexual interactions with
pre- and post- reproductive age women; men sometimes rape in
"gangs"; some men rape women and then kill their victims; men
also rape boys and men; forced oral and anal intercourse, and
many other forms of noncoital sexual aggression, are prevalent
(e.g., Gavey 1991a; Koss et al. 1987). We have no doubt that the
Thornhills could explain away this problematic evidence, for
example, by appealing to unspecified changes from ancestral to
current environments. For any such ad hoc maneuver to be
convincing, however, they would need to specify the exact
selective features that have changed. This might be difficult,
bearing in mind that they have already asserted that "sexually
attractive females who were sexually uninterested and resisted
the sexual advances of males are likely to have been a consistent
feature of the human evolutionary environment." Given the
virtual impossibility of estimating the costs and benefits of
sexual coercion for men in past and present environments, such
an exercise would degenerate into the spinning of "just-so"
stories. As the late Don Rosen (1982) remarked, there are only
two factors that constrain these kinds of stories - the inventive-
ness of the author and the gullibility of the audience.
Dewelopmental dichotomies. Despite assertions elsewhere
(Thornhill & Thornhill 1989, p. 96) "that all complex traits of all
organisms, . . . are the products of genes and environment"
and that "It is erroneous to speak of a trait of an individual
organism as either genetic or environmental," there is an im-
plicit nature/nurture dichotomy In the target article. This im-
plicit dichotomy Is achieved by a subtle conflation of evolution-
ary and developmental causation. The Thornhills' emphasis on
the evolutionary design of aspects of men's brains leads them to
mistakenly see these features as having no contingent develop-
mental history. Thus terms like "psychological adaptation" and
"human information processing mechanism" are euphe-
mistically used as codes for innate structures - structures that
are ontologically prior to the social environment. From this
perspective the environment has only a secondary role in
development, "activating" or "disinhibiting" these rape adapta-
tions with appropriate cues. Similarly, if sex-specific differences
in socialization exist, they will be guided by these more basic
"special-purpose psychological adaptations." Assigning the en-
vironment only a secondary role is in effect a diluted form of
genetic determinism.
To criticize this dichotomous view of development is not to
argue for an environmentalist perspective, but rather to claim,
as Oyama (1989) and Lickliter and Berry (1990) have, that
developmental causation cannot be realistically cleaved into a
primary evolutionary/biological component and a secondary
environmental/social component. Neural structures, like all
other phenotypic features, have a developmental history that is
contingent upon certain environmental factors, including the
social environment (see Blakemore & Cooper 1970; Horn11991;
Meaney et al. 1988). The Thornhills' argument that it is not
parsimonious to explain sexual coercion in humans as a partial
consequence of sex-specific differences in socialization (because
sexual coercion is found in species that do not have social
learning) is at odds with their claim that they are discussing an
adaptation that is unique to our species (see section 3.3). If
sexual coercion is due to different psychological adaptations in
different species (i.e., it is not homologous), then there is no
reason to suggest that it develops in the same way.
Glib analysis of the social science and feminist literature. T &
T's misrepresentation and ready dismissal of feminist and social
science theories of rape is disappointing, particularly as much of
the data they cite in support of their predictions originate from
social science and feminist research. Evolutionary biological
and social scientific/feminist accounts of rape need not be
mutually exclusive (see Gowaty, in press). Much of the data that
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 387
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
T & T imply are important for understanding rape have pre-
viously been cited in support of feminist theories of rape (see,
Check & Malamuth 1985; Gavey 1991b) - for example, the
pervasiveness of sexual victimization, the continuum of coercive
heterosexuality, and the proclivity toward rape in "normal"
nonrapist men. Furthermore, feminist and social scientific theo-
ries offer much more than merely pointing to "arbitrary differ-
ences" (sect. 3.1) in gender socialization. For example, they
emphasize that sex is a social as well as a biological activity for
humans. Sex is not neutral; it is imbued with many powerful
meanings and "functions" other than reproduction (see Foucault
1981). Some feminists would contend that it is at this level of the
cultural construction of meaning that sex and power/violence
have become fused. This sort of feminist analysis can explain
some of the data (such as, the prevalence of child sexual abuse,
the prevalence of noncoital forms of sexual violence, and the
existence of homosexual rape), that are problematic for the
Thornhills' adaptationist story. The Thornhills' conflation of sex
with reproduction provides them with an inadequate starting
point from which to explore the complexity of human sexual
behavior, including the possibility that various sexual practices
may have multiple function.
Like T & T, we too are concerned about the problem of rape,
and we applaud their attempt to understand it better. Any
serious analysis of the evolution of human behavior must be
based on modern developmental and evolutionary theory. Such
an approach to rape would avoid naive adaptationism and the
opposition of biological to cultural explanations. It would take
social science and feminist literature seriously, generate novel
and nontrivial predictions, and consider the sociopolitical im-
plications of the analysis. Thornhill & Thornhill's target article
does none of these.
behawior
Michael T. Ghiselin
Department of invertebrates and Geology, California Academy of Sciences,
Golden Gate Park, San Francisco, CA 94118
Electronic mail: caslih@cmsa.herkeley.edu
Good canons of evidence exist that allow us to establish that
some phenotypic traits have a genetic basis. We can find
Mendelian ratios and construct linkage-maps. We can study
pedigrees. We can understand biochemical pathways. Color
blindness is a sex-linked recessive.
Thornhill & Thornhill (T & T) rule out some of the best
evidence we might have by proposing that human coercive
sexuality is not only adaptive with a genetic basis, but that it is
"virtually fixed" in the human population. The assumption that
it is fixed would seem to be gratuitous. If it is not fixed, however,
then it must be variable and subject to natural selection, and
could vary geographically and otherwise. T & T tell us that if
genetic diversity were to evolve, then populations would have
to be separate for a long period of time with little or no migration
and under divergent selection pressure. But from what we know
about human evolution in many other traits, considerable diver-
sification has in fact occurred. Where are the data?
Likewise gratuitous is the assumption that human popula-
tions are monomorphic with respect to the adaptations in
question. But given the obvious point that human beings are
gonochorists, not hermaphrodites, how can we rule out the
possibility that human populations are genetically polymorphic
for inherited coercive sexual behavior? How do we know that
rapists are not recombinants that are selected against while their
genes are maintained in the population because of some advan-
tage to the heterozygotes? Are there no mutants or modifiers?
388 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
T & T use a lot of metaphors about "design" and such, which
can be excused because these have become fashionable, but
perhaps for no other reason. In anatomy we do find all sorts of
specialized structures, and their very existence does support an
adaptive significance for them. On the other hand, labor is
sometimes divided, sometimes combined, depending upon
balances or tradeoffs among various selection pressures. It is not
true that the human heart is specifically designed to pump blood
in a human body. The heart has been evolving for over half a
billion years, and we make do with what our ancestors have
passed down to us.
The target article intentionally provides an example of a
failure to distinguish between two competing hypotheses. As a
purely logical exercise there is nothing wrong with that, but
what does one do, given such negative results? The work of
Darwin (1872) on emotional expression provides a very helpful
example with respect to a similar problem. He found that
although many parts of the body express emotional states, they
exist for other reasons. And although some expressions and
gestures are used communicatively, they did not evolve because
they were useful in communication. Pre-existing parts and
actions were pressed into service in a new function. Only with
respect to language did Darwin conclude that there had been
some adaptation, and not much at that.
We might elaborate a general principle for such circum-
stances. If there is a selective advantage to special adaptations
that further a given function, then those adaptations may indeed
evolve. But then again they may not evolve. The requisite
mutations may not appear. And even if there is some advantage
to a particular trait, the overall costs may outweigh the benefits.
In the particular case under examination, the existence of
general adaptations that further survival and reproduction make
special ones seem superfluous. So we have good reason not to
invoke their existence unless we have sound, positive results.
biosocial behawior and coerci¥e
sxualir
Brian A. Gladue
Department of Psychology, North Dakota State University, Fargo, ND
58105
Electronic mail: nuO94246@ndsuvm1 .bitnet
There is nothing fundamentally new in what Thornhill & Thorn-
hill (T & T) propose in this target article. Building upon their
important earlier theoretical contributions to the understanding
of evolved human sexuality (Thornhill & Thornhill 1983), T & T
continue to generate "predictions" and hypotheses about this
aspect of human sexual interaction. In that earlier paper, they
offered a set of hypotheses and predictions about categorical
factors associated with assailants, victims, and situations; they
evaluated the literature and reams of demographic and criminal
justice data to make a compelling case for at least considering
that "rape" is not merely an act of conscious hate, but has
possible reproductive strategic consequences for both assailants
and victims. That was a landmark, if controversial work, appear-
ing alongside another influential article in that same issue of
Ethology and Sociohiology (Shields & Shields 1983). As a result,
the scholarly debate on the evolutionary psychology of sexual
assault quickened and deepened.
In the subsequent decade, what has emerged from this
debate, however, is yet another surplus of ideas and "predic-
tions," but so few data. The latest construct, succinctly outlined
by T & T in their target article has it that sexual coercion by men
either arose from a rape-specific adaptation (the "rape-specific
hypothesis"), or that sexual assault could be a side-effect of a
more general psychological adaptation not directly related to
 Commentary/ThornhiW & Thornhill: Evolution of sexual coercion
rape or even to sexuality (the "side-effect hypothesis"). The
proximal psychophysiological (i.e., penile arousal) data re-
viewed to test and differentiate these hypotheses stem primarily
from atypical groups of men (pedophiles, serial murderers,
sadists) or from North American college undergraduates tested
in social psychology laboratories. T & T conclude that these data
are "consistent with the rape-specific hypothesis, but this does
not eliminate the side-effect hypothesis" (abstract). Ultimately,
the authors, who appear to support the "rape-specific" hypoth-
esis end up arguing that, objectively and scientifically, the issue
is still open for discussion, and the data can be argued either
way. Hence, regarding the long-standing question of what the
basis and origin of sexual coercion is, there is no answer in this
article. A recent article by Palmer (1991b) summarizes much of
this specific vs. byproduct debate and finds that the evidence
can go either way [see also Palmer's accompanying commen-
tary, this issue]. Given the devotion that scientists ought to have
for parsimonious explanations over convoluted ones, the "lean"
is distinctly toward a byproduct explanatory model of sexual
coercion.
There are however some interesting gems of ideas and re-
search directions that T & T have brought to the surface. One of
these builds upon the finding that men can be readily aroused in
response to a woman's sexual, arousal whether or not the sexual
activity is coercive. In that context, do men become sexually
aroused if the woman is portrayed as being not aroused? In the
laboratory (or real world) are men aroused by coercive scenarios
where women show no affect at all, where the woman is
unconscious, or nonresistant yet also nonaroused? In our labora-
tory we have found that no matter what the situation or scenario
- coercive, noncoercive, "kinky," or mundane - men more so
than women are more likely to rate a description of sexual
activity as erotic (Gladue, unpublished observations), but that
some degree of arousal must be taking place in the woman for an
"erotic rating" to emerge. If rape is truly a specific psychological
adaptation in men, then men should engage in coercive inter-
course with sleeping, unconscious, unaroused women.
Another issue T & T introduce is the "ejaculate size and
composition" factor: If sexual assaultive intercourse results in
less ejaculate or probability of conception than consenting
intercourse, the notion of a "rape-specific hypothesis" begins to
wane. Their third interesting point concerns the curious data
from studies of bondage and spanking in which men apparently
find such activity arousing even in the "absence" of sexual
content, yet women do not. Before arguing that such scenarios
lack sexual content, however, one should demonstrate arousal
in men to scenes of nonconsenting bondage and spanking
wherein women are fully and nonerotically clothed and gener-
ate no overt sexual arousal stimuli. In any case, T & T offer sex-
dimorphic potential explanations for arousal in response to such
scenarios: If men are aroused by them, it is evidence for an
adaptation to rape. If women are aroused by them, it implies
they are interested in the achievement of physical control of a
man. Why two different explanations: Could not male arousal to
bondage and spanking scenarios be related to a desire for
physical control of a woman (or man, as in the case of homosexual
activity and imagery)?
Despite these minor conceptual snags, T & T ought to be
congratulated for once again framing the controversial argument
regarding the evolutionary psychology of sexual assault. Since
the database can lead toward either hypothesis, however, and
since even the most compelling verbal engagement still cannot
settle the issue without supportive facts, we are stuck in a
theoretical quagmire likely to generate more intellectual heat
than light. With luck and hard work, we might know more when
T & T's next theoretical overview comes out.
What if within-sex wariation is greater than
between-sex wariation?
Patricia Adair Gowaty
Department of Biological Sciences, Clemson University, Clemson, SC
29634-1903
Electronic mail: pagwty@clemson.bitnet
The Thornhill & Thornhill (T & T) target article is sexist in that it
relies on and perpetuates an unjustified narrow objectification of
both men and women in terms of sex acts and attitudes.
I am making this charge in the hopes of stimulating the
authors to consider seriously socially important deficiencies in
their scientific approaches to human sexuality. The study of sex
is important and needs to be done well in full cognizance of the
possibility of abuse. The Thornhills seem to have resisted the
idea that the very study of sex, despite being a valid scientific
goal, is also a way to socially construct meaning in our gendered
world; I think they fail for not recognizing that the way they
approach the study of sex may be part and parcel of the
mechanisms for the social control of women within Western
patriarchal institutions. This failure of imagination limits their
critique of the methods and the nature of alternative explana-
tions available to them in their quest to understand human
sexuality.
The Westernized North American cultural view espoused in
T & T s target article is suspect from other perspectives as well,
including that it is limited (the view from other parts of the world
might be significantly different) and seems to reflect the status
quo of gendered societies in which women are constructed as
second-class citizens. The possibility that this sociobiological
view of the nature of sex is a socially constructed artifact of the
power dynamics of facultatively sexist society, and not merely
about hypotheses for the evolution of sexuality, is not con-
sidered here. This seems unsatisfactory to me in that the
research traditions that have suggested this powerful alternative
are not fringe elements in the academy but widely respected
and well known.
Despite the apparent logic of the posited sexual asymmetries
between women and men, the sociobiological research para-
digm that yields this view has been flawed and remains flawed in
that it has failed to ask questions about variation in sexual
behavior among women and among men. Most of us who are
interested in adaptation and function begin by describing the
variation in the traits of interest and then we posit potentially
testable functional or adaptive hypotheses to explain the trait
and its variation. Hence the unanswered first question about
human sexuality is: Under what proximate circumstances do
women and men copulate? How do rape and other coercive
practices fit into a comprehensive picture of human sexuality?
We would do well to ask this question in terms of the range of
sexual alternatives available to humans - both men and women.
These questions expose the main flaw in this research paradigm,
namely, the lack of attention to the fact that not all men and not
all women fit this sociobiological proscriptive for "sex-specific,
species-typical" psychological adaptation. Sexual attitudes are
notoriously variable, and it is difficult to get an unbiased view of
sexual behavior, so the claims that "compared to women, men
are less discriminating about sexual partners, more motivated to
seek copulation with many partners, and more eager to include
copulation as part of an interaction with the opposite sex" (sect.
4, para. 2) immediately arouses suspicion, especially for the
relatively large number of us who, despite our heterosexuality,
for example, do not fit these sociobiological stereotypes.
Does this mean that I would reject entirely an effort to study
the biology of human sexuality? No, but I would ask that it be
done within robust frameworks, which include, first, a descrip-
tion of variation in the expression of the trait and its variants. In
this case, the trait would not be rape or rape adaptation, but
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 389
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
copulation and the proximate conditions under which vaginas
and penises come into contact, an operational definition of
copulation that would include rape. The proximate preludes of
interest would include where on the continuum of mutual
consent/force the actors fit, and we would ask questions about
this continuum for women as well as men. Such a study would no
doubt be fraught with difficulties: Perhaps it cannot be accom-
plished without bias, but it is where we ought to start. Against
the backdrop of objectively described variation in the circum-
stances of copulations, we would probably have data that al-
lowed us to posit sociobiologically fascinating and reasonable
hypotheses about the adaptive value of different modes of
copulation and their psychological underpinnings, from mutu-
ally consenting to forced. Before we have a reasonable descrip-
tion of the proximate circumstances of copulation, however, it
seems to me that most attempts to understand the adaptive
significance of any behavior (its function over evolutionary time)
or its current function are bound to be inadequate.
This failure to pay attention to the range of variation in
behavior and attitude among women and among men is bound
to lead to controversy. When sexuality is objectified as it is by
many sociobiologists, many readers are going to experience
cognitive dissonance, because "we are not that way." Are those
men who would not rape even given the "perfect opportunity"
mutants? Are women who are not attracted to high-status men
deviant, or merely free of economic constraint, able to support
themselves, and therefore able to express their own true, best
choices in terms of sexual partners or circumstances? What will
it mean if the within-sex variation in psychological traits under-
pinning sexual congress is greater than the between-sex varia-
tion? Until we have an adequate picture of the fundamental
variation in the traits we are interested in, I fear that we will
continue to speak at cross purposes.
Setting real about rape
John Hartung
Health Science Center, State University of New York, Brooklyn, NY 11203
After theoretical arguments and general evidence have been
presented, old-style social scientists like to see an example - a
case study that illustrates how the hypotheses at hand function
on the ground - preferably from a not-too-obscure culture that
is either extant or ancestral to an extant population (the latter
consideration is especially important for evolutionary hypoth-
eses). When applied to a hypothesis I like, this requirement has
always struck me as irritating. When applied to a hypothesis of
my own, I consider it insulting. Accordingly, I understand that
Thornhill & Thornhill (T & T) may have considered such an
anthropological diversion too distracting, but I would like to
offer a case study for the sake of completeness and because the
particulars involved fit neatly with the rape-specific perspective
but could only be written onto the tabula rasa after going
through unspeakable contortions of logic.
Consider the following account of gaining sexual access, by
rape, to 32,000 nonpregnant women:
They warred against Midian, as the LORD commanded Moses, and
slew every male. . . . And the people of Israel took captive the
women of Midian and their little ones; and they took as booty all their
cattle, their flocks, and all their goods. All their cities in the places
where they dwelt, and all their encampments, they burned with fire,
and took all the spoil and all the booty, both of man and of beast. Then
they brought the captives and the booty and the spoil to Moses. . . .
And Moses was angry with the officers of the army, the commanders
of thousands and the commanders of hundreds, who had come from
service in the war. Moses said to them, "Have you let all the women
live? Behold, these caused the people of Israel, by the counsel of
Balaam, to act treacherously against the LORD in the matter of
390 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
Peor,x and so the plague came among the congregation of the LORD.
Now therefore, kill every male among the little ones, and kill every
woman who has known man by lying with him. But all the young girls
who have not known man by lying with him, keep alive for your-
selves. . . . Now the booty remaining of the spoil that the men of war
took was: six hundred and seventy-five thousand sheep, seventy-two
thousand cattle, sixty-one thousand asses, and thirty-two thousand
persons in all, women who had not known man by lying with him.
(Bible)
On the assumption that "keep alive for yourselves" did not
mean invitations to the tent for a lovely cup of tea with manna,
and the additional assumption that these women were not in the
mood to have sex with men they had watched slaughter their
brothers, fathers, sisters, and mothers, we have "reason to infer
that our male evolutionary ancestors sometimes enhanced their
reproductive success through rape" 2 and that this was more
likely to occur when the probability of punishment was low.
Indeed, this behavior was codified into general policy by the
god of the Israelites, and in a manner that would greatly enhance
the probability of reproductive payoff:
When you goforthto war against your enemies, and the LORD your
GOD gives them into your hands, and you take them captive, and see
among the captives a beautiful woman, and you have desire for her
and would take herforyourself as wife, then you shall bring her home
to your house, and she shall shave her head and pare her nails. And
she shall put offher captive's garb, and shall remain in your house and
bewail her father and her mother a full month; after that you may go in
to her, and be her husband, and she shall be your wife. Then, if you
have no delight in her, you shall let her go where she will; but you
shall not sell her for money, you shall not treat her as a slave, since
you have humiliated her. (Bible)
It might be objected to here that the intention to marry
obviated the circumstance of rape, but the Israelites did not see
things that way when one of their women was the victim,
suggesting that attitudes toward rape varied greatly according to
perceptions of cost and benefit:
Now Dinah the daughter of Leah, whom she had borne to Jacob,
went out to visit the women of the land; and when Shechem the son of
Hamor the Hivite, the prince of the land, saw her, he seized her and
lay with her and humbled her. And his soul was drawn to Dinah the
daughter of Jacob; he loved the maiden and spoke tenderly to her. So
Shechem spoke to his father Hamor, saying, "Get me this maiden for
my wife." . . . The sons of Jacob came in from the field when they
heard of it; and the men were indignant and very angry, because he
[Shechem] had wrought folly in Israel by lying with Jacob's daughter,
for such a thing ought not to be done. But Hamor spoke with them,
saying, "The soul of my son Shechem longs for your daughter; I pray
you, give her to him in marriage. Make marriages with us; give your
daughters to us, and take our daughters for yourselves. You shall
dwell with us; and the land shall be open to you; dwell and trade in it,
and get property in it." Shechem also said to her father [Jacob, aka
Israel] and to her brothers, "Let me find favor in your eyes, and
whatever you say to me I will give. Ask of me ever so much as
marriage present and gift, and I will give according as you say to me;
only give me the maiden to be my wife." The sons of Jacob answered
Shechem and his father Hamor deceitfully, because he had defiled
their sister Dinah. They said to them, "We cannot do this thing, to
give our sister to one who is uncircumcised, for that would be a
disgrace to us. Only on this condition will we consent to you: that you
will become as we are and every male of you be circumcised. Then we
will give our daughters to you, and we will take your daughters to
ourselves, and we will dwell with you and become one people. . . .
So Hamor and his son Shechem came to the gate of their city and
spoke to the men of their city, saying, "These men are friendly with
us; let them dwell in the land and trade in it, for behold, the land is
large enoughforthem; let us take their daughters in marriage, and let
us give them our daughters. Only on this condition will the men agree
to dwell with us, to become one people: that every male among us be
circumcised as they are circumcised. . . . And all who went out of the
gate of his city hearkened to Hamor and his son Shechem; and every
 Commentary/Thornhill
male was circumcised, all who went out of the gate of his city. On the
third day, when they were sore, two of the sons of Jacob, Simeon and
Levi, 3 Dinah's brothers, took their swords and came upon the city
unawares, and killed all the males. They slew Hamor and his son
Shechem with the sword, and took Dinah out of Shechem's house,
and went away. And the sons of Jacob came upon the slain, and
plundered the city, because their sister had been defiled; they took
their flocks and their herds, their asses, and whatever was in the city
and in the field; all their wealth, all their little ones and their wives, all
that was in the houses, they captured and made their prey. (Bible)
Under very different cost/benefit circumstances, specifically
under the threat of homosexual rape, it was considered un-
abashedly reasonable to offer one's mate or daughter as a
substitute:
In those days, when there was no king in Israel, a certain Levite was
sojourning in the remote parts of the hill country of Ephraim, who
took to himself a concubine from Bethlehem in Judah. . . . So they
passed on and went their way; and the sun went down on them near
Gibeah, which belongs to Benjamin, and they turned aside there, to
go in and spend the night at Gibeah. And he went in and sat down in
the open square of the city; for no man took them into his house to
spend the night. And behold, an old man was coming from his work in
the field at evening; the man was from the hill country of Ephraim,
and he was sojourning in Gibeah; the men of the place were Ben-
jaminites. And he lifted up his eyes, and saw the wayfarer in the open
square of the city; and the old man said, "Where are you going? and
whence do you come?" And he said to him, "We are passing from
Bethlehem in Judah to the remote parts of the hill country of
Ephraim, from which I come. I went to Bethlehem in Judah; and I am
going to my home; and nobody takes me into his house. We have
straw and provender for our asses, with bread and wine for me and
your maidservant and the young man with your servants; there is no
lack of anything." And the old man said, "Peace be to you; I will care
for all your wants; only, do not spend the night in the square." So he
brought him into his house, and gave the asses provender; and they
washed their feet, and ate and drank. As they were making their
hearts merry, behold, the men of the city, base fellows, beset the
house round about, beating on the door; and they said to the old man,
the master of the house, "Bring out the man who came into your
house, that we may know him." And the man, the master of the
house, went out to them and said to them, "No, my brethren, do not
act so wickedly; seeing that this man has come into my house, do not
do this vile thing. Behold, here are my virgin daughter and his
concubine; let me bring them out now. Ravish them and do with
them what seems good to you; but against this man do not do so vile a
thing." But the men would not listen to him. So the man seized his
concubine, and put her out to them; and they knew her, and abused
her all night until the morning. And as the dawn began to break, they
let her go. And as morning appeared, the woman came and fell down
at the door of the man's house where her master was, till it was light.
And her master rose up in the morning, and when he opened the
doors of the house and went out to go on his way, behold, there was
his concubine lying at the door of the house, with her hands on the
threshold.. He said to her, "Get up, let us be going." But there was no
answer. Then he put her upon the ass; and the man rose up and went
away to his home. And when he entered his house, he took a knife,
and laying hold of his concubine he divided her, limb by limb, into
twelve pieces, and sent her throughout all the territory of Israel. 4
(Bible)
Learning hypotheses of rape are especially difficult to hold
onto when one considers the sociobiological compatibility of the
intricacies of law that applied among Israelites in the postcon-
quest era. Those laws, ramifications of the following passage
from the Torah, were expanded in the Talmud and consolidated
by Maimonides:
If there is a betrothed virgin, and a man meets her in the city and lies
with her, then you shall bring them both out to the gate of that city,
and you shall stone them to death with stones, the young woman
because she did not cry for help though she was in the city, and the
man because he violated his neighbor's wife; so you shall purge the
& Thornhill: Evolution of sexual coercion
evil from the midst of you. But if in the open country a man meets a
young woman who is betrothed, and the man seizes her and lies with
her, then only the man who lay with her shall die. But to the young
woman you shall do nothing; in the young woman there is no offense
punishable by death, for this case is like that of a man attacking and
murdering his neighbor; because he came upon her in the open
country, and though the betrothed young woman cried for help there
was no one to rescue her. If a man meets a virgin who is not
betrothed, and seizes her and lies with her, and they are found, then
the man who lay with her shall give to the father of the young woman
fifty shekels of silver, and she shall be his wife, because he has
violated her; he may not put her away all his days. (Bible)
The above is rather straightforward - in short, if an Israelite
rapes an engaged Israelite virgin, he dies, if he rapes an
unattached virgin he pays a fine and marries her - but what
about nonvirgins, non-Israelites, and so forth? Consonant with
the rape-specific hypothesis, it depended upon the social and
financial status of the rapist and his victim (fines being fungible
and the stipulation about marriage being controvertible accord-
ing to prior law):
Who is a seducer, and who is a violator? A seducer acts with the
victim's consent; a violator has intercourse with her against her
will. . . . If a violated woman refuses to marry her violator, or if her
father refuses to give her in marriage to him, they may do so, and the
violator may then pay the fine and go his way. . . . A High Priest who
violates a virgin may not consummate the marriage, because he is
commanded to marry a virgin, whereas this one is no longer a virgin at
the time when he is about to marry her. If he does consummate the
marriage, he must dismiss her with a get [legal document]. . . . The
violator is not liable to a fine unless he has intercourse with the girl in
a normal manner, before witnesses. . . . At what age does a girl make
a man liable to a fine? From the age of three full years until she comes
of age. If she has intercourse within these three years [age 0 to 3] it is
not considered intercourse. If he has intercourse with her after she
has come of age [12.5 years plus one day], she is not entitled to the
fine, for it is said, a damsel that is a virgin (Deut. 22:28), not one who
has come of age. . . . The following are not entitled to the fine: a
woman who is of age, a girl who has exercised her right of refusal [has
previously refused to marry], a barren woman, an imbecile, a deaf-
mute, a woman known since her childhood to be of ill repute, with
two witnesses testifying that, she had requested them to engage in
harlotry with her, and a woman who, though divorced after marriage,
is still in fact a virgin maiden. If, however, a woman divorced after
espousal is violated, the violator is liable to a fine, and the fine accrues
to herself. . . . A female proselyte [one seeking to become Jewish], a
woman held in captivity, or a freed-woman [ex-slave], if proselytized,
redeemed [purchased from slavery], or emancipated [freed from
slavery without purchase] when three years old or less, is entitled to
the fine. If she is three years and one day old when proselytized,
redeemed, or emancipated, she is not entitled to the fine, for
inasmuch as intercourse with such girls is valid, they are accounted
nonvirgins [it is not clear whether these women were considered
nonvirgins because they were not Israelites or because they had been
raped (as per the logic that applied to rapists who were High Priests -
see above)]. . . . If the man has intercourse with his victim, and she
then dies, he is exempt from the fine, for it is said, then the man that
lay with her shall give unto the damsel's father fifty shekels (Deut.
22:29) - not "unto the dead damsel's father" - provided that she dies
before she appears in court. . . . The fine of fifty silver shekels
constitutes payment for the enjoyment of the intercourse alone. The
seducer must also pay, in addition to this fine prescribed by the
Torah, compensation for the humiliation and the blemish. The
violator must pay, in addition to all these, compensation for the pain, 5
for a woman who submits to intercourse willingly suffers no pain,
whereas if she is violated she does suffer pain. Hence it is said of the
violated, because he has pained her (Deut. 22:29). . . .
The fine [for the enjoyment] is the same for all women: whether a
man has intercourse with the daughter of a High Priest, or with the
daughter of a proselyte, or with the daughter of a bastard, the fine for
her is fifty silver shekels. On the other hand, compensation for the
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 391
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
humiliation, for the blemish, and for the pain is not the same for all
women, and requires assessment. How is compensation for the
humiliation to be assessed? It all depends on the status of him who has
inflicted the humiliation and of her who has been subjected to it, for
there is no comparison between him who humiliates an esteemed
maiden of distinguished family and him who humiliates a girl of poor
and humble family. Nor is there any comparison between one who is
humiliated by an important and eminent man and one who is
humiliated by an ignoble and utterly worthless man. 6 The judges
must therefore consider the status of the man and of the girl involved,
and assess the amount the girl's father and her family would have paid
to prevent such a thing from happening to them at the hand of this
man, who must then pay the equivalent of such a sum. Compensation
for the blemish is assessed according to the girl's beauty. The judges
must therefore consider her as if she were a bondswoman being sold
in the market place, and must estimate her value as a nonvirgin as
against her value as a virgin. . . . The judges must thus determine the
amount of her deterioration in value, and the offender must pay
accordingly. . . . And if he violates an imbecile or a deaf-mute he
must pay only for the pain. (Maimonides)
Contrasting the concerns of two thousand years ago with the
concerns of today, it seems safe to conclude that moral evolution
has occurred. Good thing, too, because it is virtually certain that
no substantively relevant biological evolution has occurred -
which is important to remember when considering the merits of
Thornhill & Thornhill's hypothesis versus the side-effect hy-
pothesis. If one views rape today as the coincident manifestation
of a general tendency toward aggression with a general tendency
to seek sexual gratification - a coincidence made more probable
by current socialization practices - then one is likely to advocate
change in socialization practices (perhaps more playing with
dolls and less playing with guns, or more boys playing with dolls
and girls playing with guns), which may produce some interest-
ing results, but is less likely to reduce the frequency of rape than
would turning up the down side of potential rapists' cost/benefit
equations.
Begging the reader's indulgence for a personal case-study
analysis, I recall that rape is conspicuously infrequent (parochi-
ally considered to be virtually absent) in Ethiopia. It is also the
case that male relatives of a rape victim cannot function as
respected individuals until they catch and kill the perpetrator.
After living near the much feared Shankilla of the Dedessa
Valley for two years, I got an opportunity to visit them. My
introducer warned me not to stare at any women. I inquired
why. "Because they will kill you," he replied, "just last spring I
took police down there to look into a murder. The man was very
straightforward, he said that the victim had stared at his
wife . . . as if we would all say 'Oh' and leave." I would not be
surprised to learn that there is literally no rape among the
Shankilla.
The much more feared Danikil of far eastern Ethiopia are a
special breed. All married women wear a necklace, presented to
them by the men who became their husbands, on which has
been strung the dried testicles of potential or actual prior
suitors. I lived among these people for three months. The
women are extraordinarily beautiful by most accounts, includ-
ing mine. Perhaps it was the malaria, but I never had sexual
fantasies about them.
The point here is that Ockham's razor cuts clean. Parsimony is
on the side of those who would decrease the frequency of rape
by increasing the cost. If those who think that rape is not a sexual
act, those who would address the problem by molding the
nature of men through proper socialization, were only a danger
to themselves, we could let them play. But that is not the case,
and rape is not a game.
NOTES
1. Moses appears to have gotten his enemies mixed up here - it was
actually the Moabites, an entirely distinct people, who, according to the
account, hired Balaam, and whose women engaged in sexual intercourse
392 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
with Israelite men, thus facilitating their worship of the Moabite god,
Peor. This error is particularly vexing because two of Moses's wives
were Medianites and the Medianites gave Moses refuge when he first
fled from Egypt.
2. The historical validity of such stories is independent of the attitude
that they reveal.
3. Two of the tribes of Israel, each tribe having its own retinue of
warrior-herdsmen.
4. Another example of this strategy was provided by Lot, Abraham's
nephew, in an extraordinary display of obliging hospitality: "My lords,
turn aside, I pray you, to your servant's house and spend the night, and
wash your feet; then you may rise up early and go on your way." They
said, "No; we will spend the night in the street." But he urged them
strongly; so they turned aside to him and entered his house; and he
made them a feast, and baked unleavened bread, and they ate. But
before they lay down, the men of the city, the men of Sodom, both
young and old, all the people to the last man, surrounded the house; and
they called to Lot, "Where are the men who came to you tonight? Bring
them out to us, that we may know them." Lot went out of the door to the
man, shut the door after him, and said, "I beg you, my brothers, do not
act so wickedly. Behold, I have two daughters who have not known man;
let me bring them out to you, and do to them as you please; only do
nothing to these men, for they have come under the shelter of my roof."
5. Maimonides, the most respected scholar of the Talmud ever, is not
usually so resolute about matters that were left unresolved by The
Sages. Both Rabbi Simeon son of Judah and Rabbi Simeon son of
Menasya held that no supplemental fine was due for pain: "Even in the
case of Rape [as distinct from a case of seduction] no compensation is
made for Pain, as the female would in any case have subsequently to
undergo the same pain through her husband" (Talmud).
6. The fact that modern Western law contradicts this premise in
principle should not obscure the fact that vestiges of ancient law still
apply in practice, of late, for example, when the case involves a
prominent preacher ("High Priest"?), a prize fighter, or the descendant
of a cultural hero.
issociations
Philip Kitcher
Department of Philosophy, University of California at San Diego, La Jolla,
CA 92093
Electronic mail: pkitcher@uscd.edu
Human sociobiology makes some progress. Few are now in-
clined to advance claims about the genetic determination of
behavioral traits on the basis of premises about adaptive advan-
tage, and as the target article reveals, the naive behaviorism of
much early human sociobiology has given way to an appreciation
of the need to think in terms of the adaptiveness of human
psychological mechanisms. Some things do not change, how-
ever. Thornhill & Thornhill (T & T) are still enamored with the
project of advancing casual, qualitative speculations about the
reproductive advantages accruing to certain behavioral strat-
egies and tying them loosely to current research about sexual
behavior. The dreary errors of early work in human sociobiology
- what I have sometimes (unkindly?) called "pop sociobiology"
(Kitcher 1985) - have not been overcome.
According to the Thornhills, there is such a thing as "the"
male human reproductive strategy and we should expect it to
have been shaped by selection so as to favor those males who
obtained as many copulations as possible. Men should thus be
disposed to be more eager to copulate than women. In the
ancestral hominid environment, the use of force to secure
copulations with unwilling females would have been advan-
tageous (despite acknowledged costs of engaging in coercion).
Selection has thus bequeathed to contemporary men a propen-
sity for coercing women into copulation, and the effects of this
propensity are allegedly seen in the studies of sexual behavior
that the Thornhills review.
Against this story, several points should be made:
(1) As has been shown on a number of occasions, when the
mathematical modelling is done rigorously, it is not clear that, in
 Commentary/Thornhill
species with parental care, males who desert their mates to
copulate with others are favored by selection (Dawkins 1976,
pp. 162-65; Grafen & Sibly 1978; litcher 1985, pp. 170-71;
Maynard Smith 1982, pp. 127-28; Schuster & Sigmund 1981).
The effects of any reproductive strategy on reproductive suc-
cess, construed as number of offspring reared to sexual matu-
rity, have to be carefully measured by offering precise models
that display the costs and benefits. Utterly lacking in the target
article is any serious attention to how coercive copulation might
have affected reproductive success for primitive hominids living
in relatively small groups. Until the Thornhills show exactly how
rape is supposed to yield a reproductive advantage - and just
what assumptions about the hominid environment and the
available behavioral strategies of the agents are required to
generate their results - the only possible verdict on their claims
is that they are pieces of evolutionary guesswork.
(2) Analyses of other interactions between the sexes (e.g., my
own account of the desertion situation in Kitcher 1985, pp. 170-
71) show that the optimal strategy for an organism is likely to
depend on the behavior of other organisms in the population. If
this is approximately correct, then the idea of identifying a
single "male reproductive strategy" is a myth, and the loose,
general talk about psychological mechanisms in "men" is quite
unwarranted. Similar conclusions are suggested by the variety
of studies of nonhuman primates revealing the variability of the
sexual strategies used by males (Smuts 1987; Strum 1988).
(3) How exactly does the hypothetical psychological adapta-
tion relate to current male psychology? In the (quite reasonable)
sense of adaptation used by the Thornhills, an adaptation is a
trait (in the present instance a psychological mechanism) pro-
duced by selection in the ancestral environment. More pedan-
tically, genes that, in the ancestral environment, directed the
formation of the mechanism would have been favored by selec-
tion. What can we conclude about psychological mechanisms in
contemporary men? Not much. Suppose that the hypothetical
favored genes are fixed (just in males?) in Homo sapiens today.
That does not imply that, in the developmental environments we
inhabit, those genes direct the formation of the same mecha-
nisms. Behind the Thornhills' claim is a mass of unarticulated
speculation about developmental psychology.
(4) Supposing that there are genes fixed in contemporary
Homo sapiens that direct the formation, in the present develop-
mental environments, of some psychological propensity for
sexual coercion, how can we derive predictions from this about
measurements of penile tumescence under artificial conditions?
We can do so only by specifying very clearly the precise
character of the mechanism and by making assumptions about
what other psychological mechanisms might interfere with or
reinforce it. Because they make up their psychology in small
unacknowledged pieces, ad hoc, the Thornhills' claim to predict
the results of laboratory studies is completely unjustified. The
"predictions" have no logical link to the favored hypothesis, nor
could they - unless the authors were prepared to advance clear
psychological principles about the complex factors that underlie
sexual behavior in contemporary human society. But candid
admissions that they were guessing about psychological princi-
ples would be preferable to the loose association of behavioral
studies with an incomplete and imprecise hypothesis.
I agree with the Thornhills that it is important to identify the
kinds of social contexts that discourage coercive sex. Hence
there are serious questions about the proximate mechanisms of
sexual behavior and the ontogenesis of those mechanisms.
Unfortunately, what we need in order to answer those questions
is the type of detailed psychology that is entirely lacking from
the target article. Deciding whether men have a psychological
adaptation for rape - that is, whether genes that, in ancestral
environments, disposed our hominid ancestors to go in for
coercive sex were favored, in those ancestral environments, by
natural selection — is profoundly irrelevant. Evolutionary analy-
sis, or more precisely the type of evolutionary analysis practiced
& Thornhill: Evolution of sexual coercion
by the Thornhills, even if it were well done, could only illumi-
nate issues of function. It sheds no light on how to understand -
or how to reshape - ontogeny. The target article therefore
succeeds neither as a contribution to theoretical science nor
even as a hamfisted attack on an urgent social problem.
What exactly does it achieve? Men sometimes engage in
sexual coercion, and from this the Thornhills infer an underlying
tendency. Reviewing other people's behavioral studies, they
leap from all those reports of erections to the conclusion that the
underlying tendency is pervasive. (As I have remarked, a mass
of psychological speculation is taken for granted here.) Add the
commonplace that copulation sometimes leads to pregnancy,
and the stage is set for hypothesizing selective advantage for an
ancestral tendency for rape. (Once again, I have noted the
guesswork involved). The article dissolves into some banalities,
a review of some behavioral studies, and a mass of loose
associations.
Empirical criteria for ewaluating rape as an
ewolutionarf phenomenon
Travis Langley
Department of Psychology, Tulane University, New Orleans, LA 70118
Electronic mail: ps3gawg@music.tcs.tulane.edu
Thornhill & Thornhill (T & T) present an empirically useful
approach to examining coercion and deceit in human mating
strategies. Unfortunately, they weaken their case by discussing
rape as a psychological adaptation in a way that makes some of
their underlying hypotheses untestable; and they reject perti-
nent empirical sources. It is a central problem with T & Ts
approach that they dismiss as irrelevant data on contemporary
offspring production by victims of sexual coercion.
If men's use of sexual coercion varied not with genetic
differences but with environmental conditions (sect. 3.4), the
environment in which the behavior evolved would be irrelevant
- unless T & T mean that rape is presently not heritable but was
at some point in the past. The environment in which a character
may have evolved was probably no more static than the current
environment, and many similarities will exist across time. We
must speculate on the relevant environmental conditions that
would drive coercion as a strategy, then we must see whether
they occurred, and finally examine what they are today. Until
we know that the present is not related to the past, the be-
havior's consequences in the existing environment must be
considered; and the behavior is maladaptive if costs outweigh
benefits.
Costs (see examples in target article, sect. 5) could well
outweigh benefits because rape is highly unlikely to produce
offspring (Baron 1985; Harding 1985). Rather than reject con-
temporary reproductive outcomes altogether, one might con-
sider how they indicate the conditions under which a behavior
would be maladaptive - which in turn might provide some
indication as to conditions under which it might be adaptive.
In rejecting data that might be related to determining
whether or not rape is an adaptation, the Thornhills write (sect.
3.3): "The only way to provide evidence about the selection
pressures that designed humans is to identify and characterize
human adaptations." The model is therefore used to validate
itself. A behavior or character is first identified as evolved
simply because it exists; then it is used to infer how evolution
could have produced it - much as Creationists first assume
supernatural creation without evolution and then infer how the
same things fit their viewpoint. Likewise, although Carl Jung
spent a good part of his life identifying and characterizing
archetypes as evolved psychological mechanisms, his ability to
do so proved nothing regarding whether they even existed.
As used in the target article, the term rape psychology is
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 393
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
ambiguous because psychology is a science. Psychological pro-
cesses, however, cannot be revealed except through inferences
based on observations of behavior. In fact, every "prediction" in
the target article Involves behavior. It Is unreasonable to reject
those behaviors that relate most directly to the adaptive value of
rape.
The reasoning behind the predictions In the target article is
not always apparent. For example, ejaculate competition does
not provide a sufficient explanation for why a pair-bonded male
will be most likely to use sexual coercion with a mate when he
suspects Infidelity. If rape were related to adaptive motivations,
one could well argue that the man would be least likely to
engage in any sexual act with the unfaithful mate because he
would not want to risk raising another's offspring as his own. The
different predictions might be compared by examining whether
the conditions under which a man rapes his mate are more
compatible with the ejaculate competition hypothesis. For
example, If recent intercourse between the pair-bonded man
and woman prior to the discovery or suspicion of cuckoldry
already makes paternity uncertain, it would be to his advantage
to engage In ejaculate competition to increase the likelihood that
any offspring would be his own. If, on the other hand, the
bonded pair had not had sexual Intercourse recently, he should
avoid additional Intercourse because It would allow him to be
fairly certain that any possible offspring were not his.
Alternative interpretations can be offered for other empirical
findings supporting T & T's predictions. For example, men's
sexual arousal in response to bondage of women (Quinsey et al.
1984) may reflect a reaction to novelty In a sexual situation (e.g.,
Garrett et al. 1989; Kelley & Muslalowski 1986). Furthermore,
Chrlstensen (1990, p. 79) notes that the majority of men In
paraphllic "bondage and discipline" situations desire the sub-
missive role (see also Soble 1986). It may be unnecessary,
however, to address the data Indicating that nonrapists are not
aroused by rape depictions, because patterns of sexual stimula-
tion (particularly as measured In the cited studies) in humans
have not been shown to relate directly to reproductive strategy.
Consider men who may experience difficulty achieving turaes-
cence when highly motivated to perform, or specifically to
procreate (Masters & Johnson 1970).
T & T's Interpretation of some research findings may have led
them to overstate their case at times, as may be seen under the
reasons advanced for considering that coercion may have been
advantageous to our ancestors (sect. 5.1). Contemplating the
traditional psychological literature In terms of evolutionary
Ideas should Indeed be encouraged, but the Thornhills write
that there is "strong evidence" that various traits (all measured
In terms of behavior) are evolved - without describing any such
evidence. The studies cited (e.g., Abbey 1982) simply indicate
that the traits appear to exist; they do not provide evidence
regarding evolution.
Not only are some of T & T's interpretations of past empirical
findings questionable, but contradictory evidence exists as well.
Harding (1985; also Sanday 1981; Schwendinger & Schwen-
dlnger 1985) reports that rape occurs rarely or never In more
egalitarian societies or tribes and was virtually unknown to some
until learned from outsiders who sought to "civilize" them.
Cultural anthropology can provide clues to understanding the
natural history of male coercive sexual behavior. Natural history
seems to be Inferred In the target article, leaving us to wonder
what available Information might support or contradict the
underlying assumptions. Percentages, frequencies, observa-
tions of Individuals over long terms or under different environ-
mental conditions - In short, the kinds of data and detail that
make up a natural history of any other well-studied population -
are lacking here.
In conclusion, Thornhill & ThornhiU's approach Is thought-
provoking and generates Important empirical questions. We
disagree as to which empirical data are appropriate or inap-
propriate to explore in pursuing the answers. Whether hypoth-
394 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
esizing that human rape could be a reproductive strategy or that
it could not, Investigators must consider conditions that would
drive coercion as a strategy, contemporary costs and benefits,
and evidence from cultural anthropology. Even Information
Indicating that rape should be maladaptive may suggest condi-
tions under which it could have been adaptive. Patterns of
sexual arousal, especially as measured in laboratory studies,
may not be pertinent to understanding reproductive strategies.
ACKNOWLEDGMENT
I wish to thank Terry Chrlstensen for evaluation of a draft of this
manuscript, Edgar O'Neal for helpful discussions, and Jacqueline White
for bringing the target article to my attention.
Solution and 1; arch on m
ejcuai arousal: ita show
can we ex
Neil M. Malamuth
Departments of Communication and Psychology, University of Michigan,
Ann Arbor, Ml 48109
Electronic mail: Im1v@umichum.bitnet
The Thornhills present an intriguing analysis of evolutionary
factors that may have contributed to men's sexual arousal pat-
terns, with particular relevance to rape. An Important part of
their analysis pertains to laboratory research on men's sexual
arousal. As one of the contributors to that literature, I believe
there is a need to clarify further some of the findings reviewed
by the Thornhills. On the basis of such a clarification and on
some theoretical inconsistencies, I question the current version
of the "adaptation to rape" hypothesis. In addition, I suggest
some building blocks that I believe need to be incorporated Into
a revised Information processing model. Finally, I point to the
need to account for individual-differences data, which have
revealed contrasting sexual arousal patterns among substantial
portions of the male population.
The findings of laboratory studies on men's sexual arousal. In
the laboratory studies discussed by the Thornhills and in addi-
tional research in this area, there have been three primary
dimensions manipulated in sexual portrayals. They are (1) in-
dications (usually verbal) of the female's consent versus noncon-
sent, (2) her physiological reactions of sexual arousal versus
disgust, and (3) the degree of violence (i.e., potentially injurious
behaviors such as hitting, cutting with a knife, etc.), which has
ranged from none at all to severe. Some studies varied these
dimensions orthogonally, although unfortunately no study has
systematically varied all three of them. Although space limita-
tions preclude a detailed review of this literature, I am confident
of the following conclusions, which differ In some instances from
the impression created by the Thornhills' discussion.
Most studies have covarled these three dimensions In de-
scriptions of rape versus consenting depictions, so the former
Included verbal nonconsent, disgust reactions, and violence
while the latter included consent, woman's arousal, and no
violence. It Is clear that when such comparisons are made, most
men show considerably more sexual arousal to the latter por-
trayals than to the former ones (e.g., Barbaree et al. 1979).
When the "main" effects of these dimensions have been assessed
separately, a high degree of violence was found to Inhibit most
men's sexual arousal relative to lower levels of violence or no
violence at all (e.g., Quinsey et al. 1984), although there does
not appear to be research that has included the critical test of the
effects of high violence In the context of a depiction of consen-
sual sex (e.g., a consenting sadomasochistic Interaction In which
the man brutalizes and bloodies the woman). Research that
might be considered an exception to the generalization about
the Inhibitory effects of violence (e.g., Briddell et al. 1978) Is
 Commentary/Thornhill & Thornhill: Evolution of sexual coercion
discussed below. Descriptions of women's disgust also appear to
have an inhibitory effect (e.g., Malamuth & Check 1980b). In
contrast, verbal nonconsent alone does not appear to inhibit
sexual arousal, although it is obviously difficult to manipulate
this dimension without also including some information about
the means by which the male coerces the female into sex once
she has indicated her nonconsent. Research that has attempted
to systematically manipulate the consent dimension without
confounding it with other variables has generally shown that it
does not affect men's sexual arousal, however (e.g., Malamuth &
Check 1980b).
The Thornhills' "adaptation to rape" hypothesis predicts in-
teraction effects between the dimension of consent versus
nonconsent and various other dimensions that are relevant to
rape-specific risks. Such evidence would be consistent with
"regulation specific to sexual coercion" (sect. 12), suggesting
that men should be differentially sensitive to certain rape-
specific risks. Although the Thornhills' target article itself will
hopefully stimulate additional studies that systematically assess
such interactions, there already appear to be some relevant
findings. For example, the Thornhills' emphasize that the "will-
ingness to use force will be constrained by the countervailing
desire of displaying sexual 'morality'" (sect. 6). This should lead
to an interaction between the consent versus nonconsent di-
mension and the conditions of assessment (e.g., permissive
instructions or being required to report arousal). As noted later,
such an interaction does not appear to have been found. Yet an
interaction has been found between the consent dimension and
the woman's physiological responsiveness to the sexual ad-
vances (e.g., Malamuth & Check 1980b). As elaborated upon
later, however, the Thornhills and I differ in our views of the
woman's responses dimension.
In their review of the laboratory findings, the Thornhills make
an important distinction between the role of violence and
aggressive control of the sexual partner, but their discussion
should be clarified somewhat. Although they conclude that
"violence per se will not be sexually stimulating for most men"
(sect. 9), literature actually demonstrates that the use of vio-
lence inhibits most men's sexual arousal (e.g., Quinsey et al.
1984). This may be relevant to the Thornhills' "rape adaptation"
hypothesis, because if the use of violence could facilitate coer-
cive sex, why would inhibition of arousal be found? In this same
section of their article the Thornhills also refer to the Quinsey et
al. (1984) study as showing that "men find the theme of noncon-
senting bondage . . . sexually arousing despite the 'absence' of
sexual content" (sect. 9). But the descriptions used by Quinsey
et al. in these bondage stories included references to elements
that could be considered "sexual," such as necking, having the
woman with a slender, beautiful body undress and get "her ass
high in the air." In addition, while the Thornhills consider the
data from this part of the study consistent with the "adaptation to
rape" hypothesis, the fact that these investigators did not find
any significant differences in arousal to the consenting versus
nonconsenting versions of these bondage depictions would
seem instead to contradict this hypothesis.
Disinhlbiting manipulations* The Thornhills point to four main
studies that they contend show that inhibition of men's sexual
arousal in response to rape can be eliminated by altering the
conditions of assessment. Unfortunately, in this part of their
discussion they do not distinguish clearly enough between the
nonconsent aspect of rape and the violence that frequently
accompanies it. If we consider the former, then the Thornhills
and I agree that men generally have been shown not to be
inhibited by nonconsent alone. As I attempt to show below,
however, I do not believe that these four studies (and related
research) have demonstrated the elimination of the inhibitory
effects of violence on sexual arousal.
First, Briddell et al. (1978) reported that when undergradu-
ates believed they had ingested alcohol (even though they
actually had not) they were as sexually aroused by a violent rape
depiction as by a mutually consenting depiction (both read by a
female). This study has not been successfully replicated, how-
ever (Barbaree et al. 1983), and research has shown that alcohol
expectancy effects of the type used by Briddell et al. (1978) may
be highly confounded with the demand characteristics of the
experiment (e.g., Knight et al. 1986).
Second, Quinsey et al. (1981) focused on the effect of telling
men that sexual responsiveness to unusual themes was expected
in the testing situation. The Thornhills believe that this study
indicates that the "community men who had received permis-
sive instructions showed a significantly greater response to rape
narratives than the community men with regular instructions."
Also, the former community men "did not differ significantly
from the rapists in their response to the rape narratives" (sect.
8.2). This community sample did show more sexual arousal to
both rape and consenting depictions than those without permis-
sive instructions, however. There was no differential effect of
the instructions on responsiveness to rape themes. These data
appear to be inconsistent with the "adaptation to rape" interac-
tion effects described earlier in the target article. Moreover,
even for these community subjects with permissive instruc-
tions, the overall pattern indicated that they were less aroused
in response to the rape depictions than the mutually consenting
ones. Given the relatively small size of this sample (n = 10), it is
not surprising that some of the comparisons only approached
conventional levels of statistical significance. Nevertheless,
using sensitive indices such as the "rape index" (sexual arousal to
rape depictions divided by arousal to consenting depictions) it
was found that rapists had higher rape indices than the com-
munity subjects with permissive instructions. In addition, as the
Thornhills note, other studies (e.g., Blader & Marshall 1984;
Malamuth 1981b) have not found that normative versus nonnor-
mative instructions significantly alter men's arousal to rape
depictions.
Third, as the Thornhills correctly note, in the first phase of a
study by Malamuth (1981b), males did not show significant
differences in arousal to a rape versus a mutually consenting
version of a slide-audio show narrated by a male voice. These
two versions were created using a story from a popular erotica
magazine, however, with little violence in either version and
with the suggestion in the rape version that "she is enjoying the
experience" (Malamuth 1981b, p. 37). In the second phase of
this study, subjects were relatively highly aroused while listen-
ing to a rape portrayal depicting victim abhorrence, read by a
female. It is, however, difficult to determine what the influence
was of having been exposed only a few moments earlier to highly
arousing visual images, despite the fact that time was allowed
between exposures for arousal to return to baseline. The com-
bination of hearing the female's description of sexual as well as
violent content in the second story, particularly if her voice had
subtly seductive qualities, might have led subjects to focus more
on the sexual than on the violent content. To be fair to the
Thornhills, I myself had speculated in this 1981 article that the
use of the female voice might have "disinhibited" arousal. What
I am pointing out now is that the use of a woman's voice may not
necessarily allow for subjects' "true" arousal to be shown, but it
may be a stimulant in the context of her describing sexual
content. Systematic manipulations are needed in the relevant
variables to assess these different interpretations.
Fourth, Blader and Marshall (1984) found that when men
were required to report their arousal (via a lever) during assess-
ment, they were less aroused by a violent rape depiction than
when they were not required to report arousal; however, no
such effect of reporting was evident for the rape-restraint
depiction. The fact that this effect occurred as a function of the
violence rather than the nonconsent-consent dimension seems
inconsistent with the rape adaptation hypothesis. It is also
important to note that under all testing conditions, the rape
depictions, particularly the violent version, elicited less arousal
than the mutually consenting portrayals.
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 895
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
A continuum wersus "adaptation to rape" In attempting to
provide some of the ingredients for a model that could account
lor the laboratory findings on men's sexual arousal, I begin by
noting what I believe is an inconsistency between the Thorn-
hills' description of a continuum perspective of mating tactics (in
contrast to a dichotomy of forced vs. unforced copulations) and
their development of an "adaptation to rape" hypothesis. If rape
cannot easily be distinguished from varying gradations of other
tactics because there are "only arbitrary boundaries between
them" (sect. 7.1), how can there be a specific adaptation to rape?
It would seem more consistent to hypothesize that information
processing mechanisms affecting men's sexual arousal adapted
to the costs versus the benefits of using different "mixes" of
tactics (honest advertisement, deceptive advertisement, and
coercion) under various conditions. Selection would have fa-
vored the inhibition of males' sexual arousal if the information
processed suggested higher costs than benefits associated with
sexual persistence in the face of female opposition to the male's
advances.
Communicating sexual interest. The three dimensions manip-
ulated in the laboratory may be viewed as multiple informa-
tional sources communicating the potential costs and benefits to
males of pursuing sexual advances. When signals from different
"channels" conflicted, such as when the woman verbally denied
her interest but became sexually aroused physiologically, men
were confronted with the need to evaluate to which source they
should give more credence.
When female resistance was anticipated or actually occurred,
the means by which such resistance could be overcome would
be likely to have different costs associated with them. Here it
may be important to distinguish between nonviolent means of
overcoming resistance (e.g., trickery or mere restraint) and
violent means, by which I mean those that are likely to inflict
physical injury (e.g., hitting, cutting with a knife, etc.). All else
being equal, selection would have favored men who were
sexually inhibited by the use of violence that was likely to
seriously injure themselves or the woman (Malarnuth et al.
1977). Males who were not sexually inhibited by actions causing
serious injury, particularly death, to those with whom they
copulated would have been less likely to bear offspring and
would therefore be selected against. This does not mean I am
suggesting that sex and all forms of aggression are necessarily
incompatible. On the contrary, as also noted by Malamuth et al.
(1977), selection may have favored males who sometimes used
force instrumentally for sexual purposes and may have shaped
the linking of male sexuality with status seeking and dominance
tendencies. This analysis suggests that it may be critical to
distinguish between men who use force or even violence instru-
mentally versus those for whom violence is sexually arousing in
and of itself. The latter group would be considered relatively
deviant from the arousal pattern of most men. This view is
consistent with the observation that among most primates there
is little injury producing aggression during copulation (Smuts
1992), particularly since human social evolution is uniquely
characterized by "pair bonding" (i.e., long-term, relatively
exclusive mating relationships; Alexander & Noonan 1979).
In keeping with the Thornhills' information processing ap-
proach, we might speculate about the evolutionarily based
informational utility of the three dimensions used in the labora-
tory research. The most powerful signal indicating a high likeli-
hood of injury and therefore high costs might be derived from
the actual description of destructive, violent (i.e., potentially
injury producing) behaviors. I believe that it is this dimension,
rather than the consent dimension, that is the most important to
focus on in developing "adaptational" hypotheses. The second
most relevant dimension may be female repulsion or disgust,
which could be a prelude to the onset of additional resistance or
even violence on her part. We can speculate further, in light of
the findings reported above, that the least informative dimen-
396 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
sion regarding "costs" might be verbal indications of nqnconsent
alone. As suggested earlier, however, the processing of informa-
tion from these dimensions is likely to occur simultaneously,
and the effects of certain combinations are likely to be very
different from those of others. For example, if one were to pair
somewhat ambiguous indications of male violence with the
sexual arousal of a woman, the potential costs would not be
perceived the same as if she responded with high revulsion and
anger.
Women's arousal wersus disgust The ideas I have discussed
provide a rather different interpretation from that given by the
Thornhills to findings based on story manipulations of the
woman's sexual arousal versus disgust. The Thornhills argue
that this is just one of many variables (e.g., perceived alcohol
intoxication, permissive instructions, etc.) that can be used to
remove situational inhibitions and to reveal the "true" arousal
patterns of men. I suggest that within the evolutionary context
this is one of the crucial signals to which men became sensitized
in order to determine the woman's likely responses (and the
potential costs vs. benefits to them).
The Thornhills believe that a part of the Barbaree et al. (1979)
study shows that manipulations of the woman's reactions are not
crucial to stimulating men's arousal. They report that this study
did not find variations in men's arousal as a function of different
levels of a woman's arousal. I disagree with the implications they
draw from this study. The three stories varying the woman's
arousal were all mutually consenting depictions and varied only
in whether the woman was relatively passive or highly aroused.
None showed negative reactions such as abhorrence or disgust.
The importance of portraying a woman as sexually aroused
versus disgusted may depend on contextual information from
other cues. As suggested earlier, if she has indicated her
disinterest verbally, her nonverbal responsiveness might be
interpreted as "discounting" her verbal indications while disgust
in that context would reinforce the message. There is no reason
to suggest however, that a varying degree of arousal in the
context of consenting depictions, as in the Barbaree et al. (1979)
study, would similarly affect men's perceptions of the "danger"
signals and thereby influence their sexual arousal.
Indiwlduai differences. Before closing, it is also important to
note that there are large individual differences that moderate
the effect of manipulations in the depictions that may be masked
in studies that do not take such differences into consideration.
Using a variety of theoretically related variables (e.g., reported
likelihood of raping, sex-role stereotyping, psychoticism, etc.)
we have shown in a series of studies (Barnes et al. 1984b; Check
& Malamuth 1983; Malamuth 1989; Malamuth & Check 1983;
Malamuth et al. 1986) that there are powerful interactions
between such individual-difference dimensions and manipula-
tions in the type and degree of coercion depicted. The result is
that different groups of men sometimes show opposite patterns
of sexual arousal. In an evolutionary model such differential
patterns might be accommodated by variations in the likely
evolutionary costs and benefits to different men of using various
"mixes" of copulation tactics, depending on variables such as
their learning histories, access to resources, anticipation of
female opposition to male's sexual advances, and so on. Al-
though a full discussion of these is beyond the scope of this
commentary, any model that attempts to explain laboratory
findings in this area, including that of the Thornhills, should also
account for these individual differences in arousal patterns.
ACKNOWLEDGMENT
I would like to thank David Buss and Christopher Heavey for providing
valuable feedback on an earlier draft of this paper.
 Commentary/Thornhill
The evolutionary psychologf of rape and
Allan Mazur
Maxwell School, Syracuse University, Syracuse, NY 13244
Electronic mail: u1503@suvm.hitnet
Hunger is as natural a motive for eating as libido is for coitus.
Both appetites increase with the duration of time since the last
consummation and both are stimulated by the presence of
attractive objects of consumption. Normally, both appetites are
satisfied without coercion, but either may be satisfied forcibly,
by rape or by robbery of food (or of money to buy food). The
sexes differ more in sexuality than in eating, but still, it is
primarily males who use coercive means in both cases.
Paraphrasing the Thornhills' opening sentence, one may ask:
Is robbing food just a side-effect of hunger, or does it arise
directly from an evolutionary adaptation to food robbery itself?
The logic of their ensuing target article is as applicable to food as
sex, including the six analogous predictions.
The Thornhills' first prediction, that men are aroused and
perform sexually in both coercive and noncoercive mating, has
its equivalence for food: Hungry men are aroused to eat and will
do so when presented with food, whether it is obtained legally or
by force. Predictions 2-5 also have food equivalents, which
assert that a successful food robbery is arousing to the robber
("stolen sweets"); robberies are committed more by younger
than older men, and more by men of low than high socioeco-
nomic status; and men are more likely to rob food when their
social reputations will be unharmed. Probably all these claims
are correct (Katz 1988).
Prediction 6 does not have a clear equivalent for food, but I
think its spirit is captured by the prediction that food robbery is
especially likely to occur when a hungry robber learns that his
mate, whom he trusted, has been deceitfully withholding food
from him. Intuitively, we expect that the robber, seeking
retribution, will wrest thefoodaway, rather like the Thornhills'
cuckold will force sex from his deceitful mate.
If one accepts the Thornhills' argument that men have psy-
chological traits designed for the specific purpose of rape, one
should by the same reasoning, accept that men have psychologi-
cal traits designed specifically for food robbery as well. Rape is a
special strategy to satisfy sexual desire. Food robbery is a special
strategy to satisfy hunger. In both instances, the case is made
that our psychology is evolutionarily designed for that particular
special strategy.
We can make the argument more specific yet. All six predic-
tions made for food robbery apply - and are probably true - for
any specific kind of food robbery, say, for robbing cheese
Danish. It would then follow that evolution has given us a
psychology especially designed to rob cheese Danish. At this
level of specificity, if not before, the argument becomes
unconvincing.
Where are we left? The Thornhills have given us a set of
assertions that ought to be true if men are especially designed to
rape, but they acknowledge that these assertions also ought to
be true if men are not so specially designed. What, then, is the
validity of using this evidence as an argument for special design?
For me it is sufficient to assume that evolution gave us
motives for sex and food. Whether these are obtained forcibly or
not depends on how much our appetites are whetted, on the
availability of objects of consummation, and on the efficacy of
societal constraints against coercion.
& Thornhill: Evolution of sexual coercion
Alternative adaptive models of rape
Linda Mealey
Psychology Department, College of St. Benedict, St. Joseph, MN 56374
Electronic mail: lmealey@csbsju.bitnet
Donald Symons, one of the most outspoken proponents of the
ubiquity of domain-specific, evolved psychological mecha-
nisms, suggests that selected, nonarbitrary traits are often
intuitively perceived as such, but that when intuition fails, those
using Darwinian thinking are more likely than others to "carve
nature at a joint" in their study of the human psyche (Symons
1987a). Using Darwinian logic is thus thought to enhance the
ability to formulate new hypotheses as well as to rule out non-
Darwinian explanations (see also Buss 1991a). After using Dar-
winian reasoning to postulate an evolved mechanism, the scien-
tist must then provide evidence that the behavior "is achieved
with a sufficient degree of precision, economy, efficiency, and
complexity to rule out the operation of unspecialized mecha-
nisms and/or chance" (Symons 1987a, p. 140). [See also Buss:
"Sex Differences in Human Mate Preferences" BBS 12(1) 1989.]
Unfortunately, the authors of the target article have failed to
carve nature at a joint in their definition of sexual coercion, and
consequently, they have been unable to formulate and test
hypotheses that suggest a specialized mechanism. Rather, avail-
able data suggest that sexual coercion is better explained as part
of a spectrum of sexual strategies generated by less specific, yet
still adaptive rules.
Carwing nature at a joint: The complexity of sexual interactions*
There is virtually no disagreement among Darwinians that
males, on the whole, will be selected to be more promiscuous
and more sexually aggressive than females. Females, on the
other hand, will be coyer and more likely to attempt to extract
various forms of investment (e.g., resources or emotional com-
mitment) from their partners in exchange for sexual access. This
differential approach to sexual encounters inevitably sets up
conditions of male-female conflict (e.g. Buss 1989b; 1991b). As a
result, males use a variety of strategies to try to influence their
partners, and, as Thornhill & Thornhill (T & T) point out, these
strategies range from "honest advertisement" to a continuum of
deceptive and coercive tactics.
Females also use a variety of tactics to influence their partners
(e.g., O'Sullivan & Byers 1990). That females are selected to be
coy will mean that sometimes saying "no" really does mean "try a
little harder." In fact, women who have been "date raped" are
more likely to continue dating the "perpetrator" of the rape than
are females who were sexually assaulted by their partner but
who thwarted the rape attempt (Wilson & Durrenberger 1982).
Several explanations have been put forth to explain this phe-
nomenon (see Ellis 1989), but I have not yet seen anyone
suggest that soliciting rape might be a female strategy: Analo-
gous to the "sexy-son" model of sexual selection, women who
selected as mates those males who persisted in copulation
attempts would in turn have sons who were more successful at
copulating than females who did not select mates based on this
criterion.
The result of this complex situation is that sexual "persua-
sion," "coercion," and "rape" are hard to delineate, just as
"persuasion" and "coercion" are in other social settings. This
intergradation of strategies and contexts suggests that one or
more general mechanisms (such as emotion) come into play
during male-female conflict, rather than that a specialized,
context-specific mechanism has evolved that leads to rape.
The critical prediction. T & T note that of their six predictions,
only one is helpful for distinguishing between the rape-specific-
mechanism hypothesis and the "byproduct" hypothesis - the
prediction that: "Achieving sexual control over a woman by force
will be sexually arousing to men." Although they claim to have
found support for that prediction, a review of the literature they
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 397
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
cite (Blader & Marshall 1984; Malamuth 1981b; Malamuth &
Check 1980a; 1980b; 1983; Malamuth et al. 1980) shows clearly
that male sexual arousal decreases as female resistance, pain,
and disgust increase. In addition, since the "rape" depictions
viewed in these studies are intended to be similar to por-
nographic depictions of rape rather than to "real" rape, the
prediction that most males would be aroused in response to
truly forceful stimuli or conditions remains unsupported. Exist-
ing data in fact indicate that contrary to T & T's prediction,
extraversion, which is positively correlated with reported likeli-
hood of using force to attain sex and with reported actual use of
force, is negatively correlated with enjoyment of such episodes
(Barnes etal. 1984a). Psychoticism, on the other hand (more like
psychopathy), is positively correlated with enjoyment, suggest-
ing differential reinforcement of coercive behavior in different
phenotypes (see Ellis 1991).
Indiwidual differences and heritability* That even relatively
homogeneous individuals (college students) differ significantly
in their propensity to rape and their response to rape also argues
against the psychological mechanism model. T & T say they
expect zero heritability in the propensity to rape, because, as a
highly selected, species-wide strategy, genetic variance would
have been selected out, leaving only environmental variables
explaining differences in strategy (see Crawford & Anderson
1989). Yet even in studies of nonrapists, Malamuth and Check
(1983) and Barnes et al. (1984a) found differences in a "likelihood
of raping" scale, which correlated with personality differences
known to be substantially heritable - extraversion and psychot-
icism on the Eysenck Personality Questionnaire (Eysenck 1990;
Neale et al. 1986).
T & T acknowledge that some personality features that may
"contribute" to rape may be heritable, but they go on to write
that this fact is irrelevant to whether or not there is a specific
rape adaptation; clearly it is not irrelevant. Extraverts are more
likely to participate in all kinds of sexual activities than intro-
verts, including the use of coercive tactics (Barnes et al. 1984a;
Eysenck 1976); this fact, too suggests that "coercive strategies"
are simply part of a continuum of more general sexual strategies
that vary as a function of both heritable and nonheritable factors.
Other models. In an earlier paper (Thornhill & Thornhill
1987), the authors of the target article noted that population
variation in the use of adaptive strategies can be achieved in
several ways. They included in their list: genetic differences
between individuals; the use of mixed strategies by individuals;
and the use of species-wide, condition-dependent strategies
(such as the ones they are positing here). They left out: heritable
variations in behavioral thresholds, and developmentally con-
tingent, species-wide strategies (see Buss 1991a); both of these
models seem to be more plausible than the mechanism T & T
favor. [Interested readers should consult Palmer (1991a) regard-
ing the philosophy behind these approaches and Ellis (1991) and
Rowe et al. (1989) for the models themselves.]
inns wdirsus teoiiwior:
Craig T. Palmer
Institute of Social and Economic Research, Memorial University of
Newfoundland, St. John's, Newfoundland, Canada A1C 5S7
The target article is another example of Randy and Nancy
Thornhills' laudable ability to generate many diverse predic-
tions from alternative hypotheses. The main difference between
this paper and their earlier attempts to provide an evolutionary
explanation of human rape (Thornhill 1980; Thornhill & Thorn-
hill 1983; 1987) is that adaptation is now being sought in
psychological mechanisms instead of behavior. This difference
398 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
is presented as a fundamental and necessary theoretical shift
that greatly alters the types of predictions that can be used to
decide between the "adaptation" and "side-effect" hypotheses. I
suggest that the switch from behavior to psychological mecha-
nisms has more to do with terminology than substance, how-
ever, and that exaggerating the significance of this shift hinders
attempts to determine the evolutionary basis of human rape.
The Thornhills follow other proponents of evolutionary psy-
chology in distancing themselves from earlier approaches by
dismissing the role of behavior in their analysis: "An empirical
focus on rape behavior rather than on rape psychology is not
appropriate, . . . There is nothing about rape behavior it-
self . . . that could indicate whether or not there is rape-specific
design" (sect. 3, para. 3, emphasis added; see also Cosmides &
Tooby 1987; Symons 1987a; 1989; 1990; Tooby & Cosmides
1989; 1990a). Although the fact that the Thornhills proceed to
analyze behavior throughout the rest of the target article indi-
cates that these statements are not to be taken literally, this
extreme rhetoric is used to justify excluding types of data
discussed in their earlier work. For example, the cross-species
comparative method, which was once claimed to be "a major
strength" of the Thornhills' explanation of rape as an adaptation
(Thornhill & Thornhill 1987, p. 286), is now claimed to be
irrelevant to the new psychological approach (sect. 3.3, para. 1).
The Thornhills also follow the position of other evolutionary
psychologists when they claim that data on offspring production
are "irrelevant to the rape-adaptation hypothesis" (sect. 3.2,
para. 1). Although caution must be used in interpreting all types
of data on rape (see Crawford & Galdikas 1986; Palmer 1988;
1989a; 1989b; 1991a), the complete dismissal of these bodies of
data as irrelevant requires greater justification than is provided
in this target article (see Irons 1990; Palmer 1991b).
Taking the place of the excluded types of evidence are many
new predictions. The most prominent of these concern experi-
ments on the sexual arousal of males exposed to audio and visual
depictions of different types of sexual encounters. Because the
interpretation of these data is cast in terms of psychological
mechanisms instead of correlations between behavior and en-
vironmental variables, the emphasis evolutionary psychologists
have placed on these mechanisms being "domain-specific" be-
comes crucial (see Cosmides & Tooby 1987; Symons 1987a;
1989; 1990; Tooby & Cosmides 1989; 1990a). Although the
existence oionly a few general-purpose learning mechanisms is
unlikely, some psychological learning mechanisms are almost
certainly more general-purpose than others (see Alexander
1990). The danger in overemphasizing the domain-specificity of
psychological mechanisms is that it can lead to the assumption
that a given correlation between behavior and environmental
variables is produced by a mechanism designed specifically for
that particular situation. This appears to have taken place in the
analysis of data showing that men become aroused by depictions
of the nonconsenting bondage of women. In reply to the Thorn-
hills' question, "why should this scenario be interpreted as
sexual at all, unless there is adaptation to rape?", I would suggest
that the males' arousal results from their anticipation of sexual
acts to follow the bondage, and that this anticipation results from
more general-purpose psychological mechanisms with which
they have simply learned that the bondage of women is often
followed by sexual acts. Hence, the arousal of males in response
to images of females being tied up only implies that males are
aware that rapes occur, not that males have been designed by
natural selection to rape.
My final comment concerns the claim that there is a "con-
tinuum" of human matings and that all divisions such as
rape/nonrape, forced/unforced, or coerced/uncoerced are
based on "only arbitrary boundaries" (sect. 7.1, para. 1). If all
such boundaries were really arbitrary and nonobjective then the
entire question of whether or not there is adaptation to rape (or
to coerce sex) would become meaningless because there would
be no way to test any of the predictions generated by the
 Commentary /Thornhill & Thornhill: Evolution of sexual coercion
Thornhills. For example, how could it be determined that
young men "exhibit more coercive sex than older men" if there
were no way to objectively distinguish an act of coerced sex from
an act of uncoerced sex? The problem becomes compounded
when one tries to "examine the sexual responses of nonrapist
[and rapist] men to sexually coercive and noncoercive stimuli in
the lab." Not only would it be impossible to decide which slides
or audio tapes to label as coercive or noncoercive, but it would
also be impossible to assign men to the rapist and nonrapist
categories in anything but an "arbitrary" way. The fact that talk
about "rape" and "coercive sex" does "make sense" implies that
the meaning of these terms is not arbitrary (see Palmer 1991a;
1991b).
In conclusion, the adoption of evolutionary psychology may
have helped the Thornhills generate new predictions, but the
overemphasis on the "newness" of the approach has hindered
their diligent efforts to determine the evolutionary basis of
human rape. This, despite the fact that the Thornhills point out
that hypotheses about psychological mechanisms were "im-
plicit" in their earlier works. In fact, all of the predictions in
Thornhill and Thornhill (1983) can be so easily reworded in
terms of psychological mechanisms without changing their
meaning that Paul Turke has called this 1983 paper an example
of evolutionary psychologists applying their theory (Turke 1990,
p. 306). The similarity between evolutionary psychology and
earlier approaches results from the fact that what evolutionary
psychologists call discovering psychological mechanisms in-
volves essentially the same steps that have always been involved
in determining whether the precision and efficiency of the
behavior indicates that the behavior has been "designed" by
natural selection (see Williams 1966). Contrary to some of the
extreme claims by evolutionary psychologists, both approaches
simply involve the observation of patterns of behavior in relation
to environmental variables. An objective consideration of all
relevant data, based on a realization that the difference between
studying behavior and studying psychological mechanisms does
not really make a difference may be the key to finally determin-
ing whether or not human rape is an adaptation.
Specific versus general adaptations: Another
unnecessary dichotomy?
Daniel Perusse
Department of Human Genetics, Medical College of Virginia, Richmond, VA
23298-003
Electronic mail: dperusse@gems.vcu.edu
Thornhill & Thornhill (T & T) have provided a valuable review of
the natural history and experimental studies on human rape. In
doing so, they have enriched our understanding of men's and
women's divergent sexual psychologies. Specifically, their view
of coercive and noncoercive male sexuality as a continuum, in
which male coercion appears much more significant than pre-
viously appreciated, is original and compelling. They also pro-
vide novel insights about how male and female psychosex-
ualities might interplay, with men having evolved, for example,
to find women's sexual arousal exciting because it would have
been interpreted as a cue to sexual exclusivity and high proba-
bility of paternity. Implicit in this hypothesis is the equally
interesting notion that females have evolved to experience
sexual arousal mainly in stable pair bonds that were the ones
most likely to have led to male parental investment. In the
tradition of Symons (1979), Daly and Wilson (1983), and others,
these reflections point to the existence of complexly asymmetric
yet intricately interlocking male-female sexual psychologies, for
which conventional studies lacking an evolutionary framework
have failed to provide alternative explanations.
Less convincing, however, is the casting of the article in the
specific versus general-purpose adaptation dichotomy. To de-
cide between these two adaptive scenarios, T & T examine six
predictions against a wealth of empirical data on human rape.
They are themselves the first to recognize, however, that in the
end this exercise fails to falsify either hypothesis. One is there-
fore left to wonder whether these are truly competing hypoth-
eses and if so, what tests would have really succeeded in
discriminating between them; and if not, whether the dichot-
omy on which they are predicated might not itself be ill-
conceived from the start. To try to answer these questions, I will
briefly examine the authors' main conclusions.
T & T devote the largest part of their review to laboratory
studies of men's sexual arousal in response to rape scenarios.
They conclude that all other current data, with the possible
exception of these, are compatible with both hypotheses (sect.
12). I will therefore limit my comment to the discussion of the
laboratory evidence.
The studies reported use a variety of experimental, designs,
which makes their comparison difficult. Some studies find that
nonrapists show little arousal in response to rape depictions,
others that they do, but only if the victim becomes aroused, and
still others that these findings can be reversed by factors such as
the narrator's gender or the subjects' feelings of normality or
security in being aroused. If there is a common thread running
through these diversefindings,the general adaptation explana-
tion seems more likely to reveal it than the specific one,
provided that a third component is appended to the general
framework: In addition to a species-wide male desire for sex and
a willingness to use coercion to attain goals, it is likely that there
also exists a general tendency on the part of individuals to make
cost-benefit assessments of the outcome of their potential ac-
tions. If so, the perception of low risk that is engendered by
being made to feel secure in the experiments (not having to
report one's arousal, being told that responding to unusual
stimuli is normal, being under the (placebo) influence of alcohol,
etc.) - not to mention the low risks intrinsic to an experimental
situation itself- might well explain the above results as part of a
favorable cost-benefit calculation without having to invoke a
specific adaptation to rape. T & T allude to this third possibility
(sect. 10), but they do not interpret the experimental evidence
in relation to it.
Hence rape (or fantasies thereof) would sometimes occur as
the result of the particular interplay of three general-purpose
adaptations that would also function in a large variety of other
situations not specifically related to rape. This seems a more
parsimonious and powerful explanation because it would ac-
count for both widespread behaviors (e.g., courtship) and par-
ticular ones (rape). Specific adaptations might exist, to be sure,
but to contrast the power of the two approaches, let us consider
deviations other than rape. T & T mention bondage and spank-
ing, implying that an absence of arousal to tied-up men on the
part of women would reinforce the hypothesis of a specific male
adaptation to rape. I suggest: (a) Women would indeed show
such a lack of arousal, but it would be entirely compatible with
the general gender dimorphism in sexual interest predicted by
sexual selection theory, (b) Numerous men would find being
tied-up very arousing, as suggested by the vast amount of
pornographic material devoted to "female domination." In fact,
the Diagnostic and Statistical M.anual of Mental Disorders
(1987) estimates that sexual masochism (a fantasy often involving
being raped) is 20 times more prevalent in males than in
females. The other DSM-III-R paraphilias are diagnosed in
males only and include such fantasies as fetishism (the sexual
involvement with inanimate objects, usually female undergar-
ments) and voyeurism, which is often catered to by the por-
nographic industry by providing male clients with depictions of
women in the act of lesbianism.
I suppose one could provide specific adaptive hypotheses for
all these deviations (e.g., men might be aroused by the sight of
sexually interacting women as this might be a clue, in our
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 399
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
evolutionary past of male polygyny, of an especially good en-
tente between co-wives, which would favor husband reproduc-
tive success). I would also suspect that such fantasies are fairly
common among males (consumers of pornography being the
"average man"), and that laboratory studies would find ample
evidence for specific male arousal from graphic depictions of
such situations. All these findings, however, seem more eco-
nomically interpretable as the results of a general adaptive and
obsessive male interest in all things sexual, of which the specific
actualizations are probably better pursued in individual on-
togenies than in phylogeny.
iediwidua! differences in
rape
Vernon L. Quinsey
Forensic/Correctional Studies, Psychology Department, Queen's University,
Kingston, Ontario, Canada K7L 3N6
Electronic mail: quinseyv@qucdn.queensu.ca
The target article argues that all males are equally sexually
aroused by consenting and nonconsenting sexual interactions.
Most of the evidence that is cited in support of this assertion
comes from phallometric studies of sexual interest in audiotaped
scenarios depicting consenting intercourse and rape. These
laboratory studies show, as Thornhill & Thornhill (T & T) assert,
that both incarcerated rapists and nonrapists (or, at least, un-
identified rapists) show sexual interest in depictions of both
consenting intercourse and rape.
The observation that males show some sexual interest in
depictions of both consenting intercourse and rape, however,
does not preclude substantial individual differences among
them in the degree of their relative interest in these stimuli.
Phallometric studies frequently find that incarcerated rapists
show an equal or greater amount of sexual arousal to depictions
of violent brutal rapes than to descriptions of consenting sexual
activity, whereas nonrapists are more aroused by consenting sex
than by rape descriptions. The degree of relative interest in rape
compared to consenting sexual activity, expressed as difference
scores or ratios, discriminates rapists from nonrapists (Barbaree
et al. 1979; Harris et al. submitted; Quinsey & Chaplin 1982;
1984; Quinsey et al. 1981; Quinsey et al. 1984). Incarcerated
rapists show more relative interest in rape cues than nonrapists
even in studies where both groups show less sexual arousal to
rape than to consenting sex (Baxter et al. 1986, p. 518). These
findings are rather remarkable, because rapists who are assessed
in a forensic context have exceptionally strong incentives to
show low sexual arousal to rape stimuli. Because phallometric
technology is not immune to dissimulation (Quinsey & Chaplin
1988), it is very likely that the literature substantially underesti-
mates the amount of sexual arousal these stimuli occasion among
rapists.
Rapists are differentiable from nonrapists in phallometric
assessments even when the nonrapists are instructed that sexual
arousal to unusual stimuli is both common and expected
(Quinsey et al. 1981). Although it is true, as T & T assert, that
nonrapists increase their responses to rape cues under these
"permissive" instructions, in the Quinsey et al. (1981) study they
simultaneously increased their responses to consenting cues,
too, so that their relative arousal to consenting and nonconsent-
ing stimuli still differentiated them from rapists.
From the viewpoint of individual differences in male sexual
preferences, the comparison of sexual arousal to rape and
consenting sex cues is more important than the absolute magni-
tude of the response to either. Relative sexual arousal in re-
sponse to brutal rape cues in comparison to descriptions of
consenting sex not only better discriminates rapists from nonra-
pists than absolute levels of responding to either rape or con-
400 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
senting cues (Harris et al. submitted), it also better predicts
future sexual recidivism among rapists (Rice et al. 1990).
It is important to note that high levels of brutality in the
depictions of rape facilitate the differentiation of rapists from
nonrapists in phallometric assessments. For example, the bond-
age and spanking stimuli in the Quinsey et al. (1984) study
probably did not differentiate between rapists and nonrapists
because they were too "soft." Incidentally, these stimuli, con-
trary to their description in the target article, were unam-
biguously sexual in nature.
Thus, although the phallometric literature supports the view
that males show sexual arousal to both consenting sex and to
rape, there are large and consistent differences shown in the
amount of their relative interest in these depictions. Not sur-
prisingly, men who stalk and rape women in a repetitive and
predatory manner report greater interest in coercive and violent
sex and the use of more violent imagery in masturbation, and
they show greater arousal to brutal rapes (in comparison to
consenting activity) than date rapists or nonrapists.
Although under normal testing conditions brutal rapes elicit
much less sexual arousal among nonrapists than consenting sex
(e.g., Quinsey & Chaplin 1984), it is nevertheless clear that men
in general show at least some interest in coercive sex and that
the degree of this interest can be facilitated in a variety of ways.
This does not falsify the idea that an interest in coercive sex may
be an adaptation, however. In the case oiPanorpa scorpionflies,
where males have evolved notal organs and abdominal claspers
to facilitate coercive copulations, males still appear to prefer
unforced matings to rape (Thornhill 1980).
Of course, there is, as T & T assert, much more to rape
behavior than sexual interests or a specific adaptation. A variety
of cultural factors clearly influence rape rates (for a review see
Quinsey 1984). In addition, there are other individual differ-
ences among men that affect their propensity to rape. In line
with one of the side-effect predictions of the target article, for
example, men who are insensitive to social sanctions are more
likely to rape than other men: Among rapists, those who have
high scores on the Psychopathy Checklist are much more likely
to rape again than those with low scores (Quinsey et al.,
submitted; Rice et al. 1990).
Although I accept T & T's contention that a rape-specific
adaptation might not be reflected behaviorally in a straightfor-
ward manner, the lack of such a correspondence makes it
extremely difficult to test empirically. Phallometric assessment
data offer one of the few ways of deciding between rape specific
adaptation and the side effect hypothesis. As discussed above,
phallometric data have both discriminant and predictive valid-
ity. Laboratory studies indicate that nonrapists respond less to
rape cues, the more violent and brutal these stimuli are; rapists
respond relatively more to these stimuli; heterosexual males,
whether rapists or not, do not respond to nonsexual violence
involving males but some rapists become aroused to nonsexual
violence directed toward females. Rapists who respond to such
violence are more likely to have injured their victims in the past
(Quinsey & Chaplin 1982; Quinsey et al. 1984). Rapists who
show greater relative arousal to depictions of rape are more
likely to rape again (Quinsey et al. submitted; Rice et al. 1990).
In my view, these findings support a modified view of the
rape-specific adaptation hypothesis. If we suppose that natural
selection may have made connections between dominance,
aggression, and sexual arousal easy for men to learn (Quinsey
1984), one can explain individual differences among them in
their sexual interest in rape and one can derive testable predic-
tions concerning the relative learnability of particular associa-
tions between biologically relevant stimuli and sexual arousal.
This evolutionary approach, together with a specification of the
proximal environmental determinants of rape and other person-
ality characteristics and beliefs that are associated with a pro-
pensity to rape, could provide a complete account of rape
behavior.
 Commentary/Thornhi\\ & Thornhill: Evolution of sexual coercion
Psychological adaptation: Alternatives and
implications
P. A. Russell
Psychology Department, King's College, University of Aberdeen, Old
Aberdeen AB9 2UB, Scotland
It is fascinating to see evolutionary biologists tackling aspects of
human behaviour that were once the virtually exclusive pre-
serve of psychologists and social scientists operating in the
context of purely sociocultural theories. Most of the target
article is devoted to the question of whether men's coercive
sexuality has been specifically selected for or is a side-effect of
other adaptations. The authors conclude that the available
evidence is largely equivocal, so the value of the paper lies
ultimately in its synthetic review of the data and in its demon-
stration of the potential of an evolutionary approach. My com-
ments concern two issues that the Thornhills touch on without
elaborating. Both have implications beyond the evolutionary
analysis of coercive sexuality, for the evolutionary analysis of
human behaviour generally.
One issue relates to the alternatives to an evolutionary the-
ory. Sociocultural. theories of coercive sexuality exist and the
Thornhills are clearly aware of them. The possible roles of social
learning and socialisation are mentioned briefly in section 3.1,
where it is correctly pointed out that such mechanisms are not
incompatible with the assumption that coercive sexuality has
been selected for, whether directly or indirectly through selec-
tion for other behaviours such as aggression. What the authors
do not consider, however, is the possibility ofapurely sociocul-
tural theory. No doubt this is because, as evolutionary biolo-
gists, they can scarcely countenance such a notion. They take it
as read that sexual coercion is the result of psychological adapta-
tion: This follows from the central evolutionary axiom that
"Psychological adaptation must somehow causally underlie all
human feelings, emotion, learning, and behavior" (sect. 3, para.
2, emphasis added). I happen to agree with them, but it is worth
noting that although this is axiomatic to evolutionary biologists,
it is certainly not axiomatic to many social scientists. The latter
will require convincing that there is a need to base theories on
the concept of psychological adaptation: In the long run this will
probably involve the demonstration that such theories provide a
better data fit than purely sociocultural theories. For some of
the aspects of human behaviour recently analysed by evolution-
ary biologists this demonstration is lacking and is often difficult
to achieve (see, e.g., Russell 1991; Wallen 1989). In the present
case, however, it seems unlikely that a purely sociocultural
theory can be made to generate the various specific predictions
derived from the evolutionary hypothesis (section 6 and else-
where). To the extent that these predictions are confirmed by
the data, the evolutionary theory does offer a better fit. One
might argue, however, that this is simply because existing
sociocultural theories are not stated with sufficient precision.
Perhaps theories of this sort can be formulated in such a way as
to generate these predictions. As an example of the kind of
argument that can be adduced, consider Prediction 4, that low
socioeconomic status men will be more inclined to rape. This
prediction can be generated socioculturally in various ways: For
example, it can be argued that low socioeconomic status families
socialise their sons in such a way as to increase their tendency to
engage in coercive sex. Whether all the predictions could be
generated by sociocultural theory, however, is a moot point and
certainly one suspects that, even if they could, the end result
might be a motley collection of ad hoc arguments. A great
attraction of the evolutionary theory here is its ability to syn-
thesise the data through the use of a single unifying principle,
psychological adaptation based upon evolved reproduction in-
terest. According to this theory, if low socioeconomic status
families do socialise their sons differently, this is not arbitrary.
Rather, it is ultimately linked to the fact that the reproductive
payoff of rape is likely to be greater for low socioeconomic status
men.
My second point concerns the ontological implications of the
psychological adaptation theory. As the Thornhills write, this
theory implies "that men have psychological traits whose under-
lying genes are virtually fixed or invariant in the human gene
pool" (sect. 3.4, para. 2). This is equally true whether coercive
sexuality is the result of specific selection or of selection for other
traits, such as male aggression. Both imply that men possess
genes promoting rape. Such genes could operate through
various mechanisms, including a disposition or preparedness to
learn, through social experience, to use coercion in sexual
encounters. This idea stands in contradistinction to a purely
sociocultural explanation, according to which rape is the out-
come of arbitrary cultural values and socialisation processes
relating to sexual conduct (at least these are assumed to be
arbitrary in the sense that they are not genetically constrained.
They may not be literally arbitrary, if, for example, they are seen
as serving some social function such as maintaining the power of
men over women).
The psychological adaptation theory and the purely sociocul-
tural theory may have very different implicationsforhow one is
to reduce the incidence of rape in societies. If, as sociocultural
theorists maintain, rape is promulgated as a result of the way
men are socialized (i.e., to be aggressive in sexual encounters
and to minimise the rights of women to choose whether or not to
have sex) then appropriate changes in socialisation processes
and in the cultural values on which these processes are based
should lead to a reduction in rape. Indeed, on this basis it is
possible to conceive of a society in which rape is absent, except
perhaps as a genuine pathology. The implications of the psycho-
logical adaptation theory, however, are less clearcut. Does this
theory predict that changes in cultural values and socialisation
processes would have less effect on the incidence of rape,
because the rape-promoting genes would place limits on the
responsiveness of men to such changes? Or is it that the effects of
these genes are specific to the kind of social environment in
which they were originally selected, that is, one in which
socialisation favours rape, so that their effects in other social
environments, including an environment in which socialisation
does not favour rape, are unpredictable? Or yet again, is it
possible that the genes themselves tend to constrain the nature
of cultural values and socialisation processes, rendering it im-
probable that a society will develop values and processes that
are out of line with the male genotype? These appear to be
difficult questions but a psychological adaptation theory of rape
can scarcely avoid confronting them. Moreover, questions of
this sort assume a more general importance in the context of
attempts to understand the relations between the human geno-
type and the nature of human societies' and culture.
Psychological adaptations, development and
individual differences
Barbara Smuts
Evolution and Human Behavior Program, Center for Human Growth and
Development, Departments of Anthropology and Psychology, University of
Michigan, Ann Arbor, Ml 48109-1346
Electronic mall: barbarasmuts@um.cc.umich.edu
Male sexual coercion of women is an important topic for evolu-
tionary analysis, but the focus by Thornhill & Thornhill (T & T)
on whether or not there exists a specific "rape-adaptation" is not
productive. I will indicate four weaknesses of their approach.
First, the meaning of the terms "rape" and "sexual coercion,"
which T & T explicitly equate, shifts in very confusing ways. The
authors point out that male sexual coercion of women occurs on a
continuum so that one cannot clearly distinguish forced versus
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 401
Commentary/Thornhill & Thornhill: Evolution of sexual coercion
unforced copulations (an idea first proposed by feminists such as
Dworkin [1987], not acknowledged by T & T). According to T &
T, the acts that constitute sexual coercion include an enor-
mously varied set of behaviors, ranging all the way from the use
of violence to the use of "implicit threats" of "withdrawal
of . . . emotional involvement." T & T ignore the significance of
this diverse behavioral repertoire, however, when they discuss
the hypothesis that men have a "specific rape-adaptation." They
give no explicit definition of "rape" in this context, yet all the
laboratory studies they cite involve depictions of the use of
physical force (or directly threatening the use of physical force)
to obtain sex with unwilling women. They cite no laboratory
evidence, or other kinds of experimental evidence, relevant to
the other kinds of coercive behavior they describe earlier.
In addition, T & T claim that the "rape-adaptation" hypothesis
can be tested most directly by finding out whether "gaining
physical control over an unwilling sexual partner by force should
be sexually arousing to men." The laboratory evidence that they
include and the emphasis on a link between the use of physical
force and sexual arousal strongly suggest that when T & T
discuss "rape-adaptation" they are referring to acts that fall at the
extreme end of the continuum of sexual coercion they pre-
viously identified. Failure to acknowledge this shift from a focus
on a wide range of coercive behaviors to a focus on forced
copulation allows T & T to gloss over a major inconsistency in
their target article. If, as they first claim, sexual coercion
involves a wide range of diverse behaviors, then the existence of
one specific underlying "psychological adaptation" seems un-
likely. Furthermore, if such an adaptation were hypothesized, it
could be tested only by referring to evidence concerning the full
range of sexually coercive behaviors. On the other hand, if, as
they later imply, the "rape-adaptation" hypothesis refers only to
physically forced copulations, then no rationale has been pre-
sented for isolating the extreme end of the continuum of sexual
coercion as a meaningful unit for evolutionary analysis.
T & T justify their search for an underlying psychological
rape-adaptation by arguing that "it seems reasonable to assume
that ancestral human males increased their reproductive suc-
cess by rape." However, whether this assumption seems reason-
able depends critically on the definition of rape one uses. Are
the authors referring here to the full range of sexually coercive
behaviors, or only to physically forced copulations? Probably
few would disagree that men have long used their ability to
withhold or withdraw social and economic resources to obtain
matings they would not have obtained otherwise, but this is a far
cry from the argument that rape per se - using force to obtain
copulations - increased male reproductive success in the past. If
it did not, there is no justification for proposing a psychological
adaptation specifically designed to achieve it.
A second problem concerns the authors' failure to consider
how development affects psychological adaptations. Although
they rightly emphasize that expressions of sexually coercive
behavior vary depending on environmental conditions, they
completely ignore a second major source of variation: individual
differences in response to the same conditions. Nearly all of the
laboratory studies T & T cite reveal a considerable range of
responses to experimental stimuli, even among subjects who
belonged to the "same group" (e.g., "nonrapists"). A focus on
"average" responses ignores this variability. T & T implicitly
deny the existence of individual differences by repeatedly using
phrases like "men are sexually aroused by physical control of an
unwilling mate through force" (when, in fact, the studies they
describe indicate that some men find such aggressive control
arousing and some men do not. Attempts to characterize the
psychological adaptations underlying sexual coercion must focus
on the developmental processes by which such important indi-
vidual differences come about. Because T & T completely
ignore individual differences and their developmental precur-
sors, their analysis comes dangerously close to biological
determinism.
402 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
A third, related problem involves the authors' assumption
that the responses of mostly young, mainly middle-class white
college men exposed to pornographic stimuli in the laboratory
necessarily reveal the nature of the psychological adaptations
favored by natural selection in ancestral men. T & T dismiss the
relevance of some kinds of modern evidence, such as current
relationships between the use of sexual coercion and male
reproductive success. They claim that because the modern
environment is so different from the one in which we evolved
such relationships tell us little about the adaptive significance of
sexual coercion in the past. However, similar reasoning can be
used to argue that because modern white male college students
have grown up under conditions very different from those in
which our ancestors developed significant discrepancies are
likely to exist between their psychology and the psychology of
ancestral men. Suppose, for example, that men possess evolved
psychological mechanisms that help them identify the likely
costs and benefits of various behavioral tactics based in part on
observations of the behavior of other men and its consequences,
as suggested by Draper and Harpending (1988). In our society,
boys and young men are bombarded with media images telling
them that sexual coercion of women is both sexually stimulating
and socially acceptable (i.e., low cost). Such developmental
experiences could play an important role in producing the male
psychology that T & T characterize on the basis of laboratory
studies. To deny that development in different environments
can result in very different psychologies, as T & T implicitly do,
is tantamount to claiming that psychological adaptations are
genetically determined, rather than viewing them as products of
a complex interaction between genetic makeup and individual
experience.
Afinalproblem concerns the failure to use the comparative
method to analyze evolutionary phenomena. T & T seem to
think that because the hypothesized psychological adaptations
underlying sexual coercion evolved in ancestral humans infor-
mation about sexual coercion in other animals and in various
human societies today is entirely relevant to investigating the
nature of these adaptations. Such information is critical, how-
ever, to developing hypotheses about both the adaptive prob-
lems our ancestors faced and the environmental conditions that
influence the reproductive costs and benefits associated with
particular forms of sexual coercion (Smuts 1991a; 1992b). In
addition, obtaining such information would be the most direct
approach to the authors' stated goal to "identify the kinds of
social contexts that discourage coercive sex." If we fail to use
information from other species, and particularly from other
cultures, proposed characterizations of human adaptations are
likely to be little more than "just so" stories that rationalize the
particular manifestations of human nature that develop in twen-
tieth century America.
Selection for rape or selection for sexual
opportunism?
Eckart Voland
Institut fuer Anthropologie, Universitat Goetiingen, W-3400 Goettingen,
Germany
Electronic mail: ewoland@dgogwdg1 .bitnet
In my view, three points especially contradict the rape-specific
adaptation hypothesis:
(1) The reproductive success of rapists is marginal (and was
probably marginal as well in the Pleistocene era). If the rape-
specific adaptation hypothesis were correct, then Pleistocene
males would have had to increase their personal reproductive
success to a significant degree through sexual violence against
females. This appears to me to be improbable because rapes
rarely lead to the birth of a viable child, as women react to
personal stress experiences with various forms of reproductive
 Commentary/Thorehill
dysfunction (Ellison 1990; Wasser & Isenberg 1986). I cannot
think of any reason why these evolved mechanisms, which
protect women against unfavorable pregnancies, should not also
have worked in the Pleistocene era. Rape is a high-risk/low-
benefit strategy, which is not worth the gain and is therefore not
subject to any great selection pressure.
(2) Human sperm competition is obviously a female-driven
phenomenon. Female sexual disinterest in their partners can
have innumerable causes, including a general loss of libido. This
would be more likely to reduce the probability of extra-pair
copulations, contrary to the hypothesis presented in the target
article. If men were to assess the sexual disinterest of their wives
as a sign that their wives were maintaining extramarital relations
(section 1.1), they would, without a doubt, be wrong more often
than they would be right. On the contrary, they should be
particularly concerned about their paternity when their part-
ners show above-average sexual demands, because, as Bellis
and Baker (1990) have been able to demonstrate, female sexual
unfaithfulness correlates positively with in-pair copulations.
Human sperm competition is obviously a female-driven phe-
nomenon and not a male-driven one.
The converse conclusion also appears to be problematic. Men
do not perceive female sexual arousal as stimulating because
they can interpret this as an expression of exclusive pair rela-
tions (sect. 6, predictions 2-5) but because they react to sexual
opportunities. How else would prostitutes be able to stimulate
business, except by sending signals of willingness to engage in
sex? No client thinks of an exclusive-partner relationship with a
prostitute - not even unconsciously - or does he?
(3) Men are sexually aroused and competent only or primarily
when they perceive that a woman is interested or not resisting
coitus. In my view, prediction 1 is more likely to be falsified than
verified. The Thornhills themselves formulate the decisive
criterion: "In contrast, the rape-adaptation hypothesis would be
falsified by evidence that men are sexually aroused and compe-
tent only or primarily when they perceive that a woman is
interested or not resisting coitus" (sect. 6). But it is exactly this
that corresponds to daily experience. Men are regularly sur-
rounded by women who act sexually neutral without being
constantly aroused. Only signs of sexual interest and willingness
trigger prompt sexual arousal. It may very well be that these
signs work subtly, however, and it is probably not always
possible to decode them clearly, which is why sexual com-
munication is disturbed and can be extremely conflict-laden.
This does not change the fact, however, that male sexual arousal
is primarily a function of signaled (real or presumed) oppor-
tunities. In my opinion, all of the empirical evidence of phal-
lometry research reviewed by the Thornhills can also be inter-
preted in terms of a general psychological adaptation: Sexual
opportunities arouse men; the more consent there is between
the partners, the more they do so.
My three points cannot disprove the rape-adaptation hypoth-
esis, of course. Pursuant to the role of "Ockham's razor,"
according to which the most economic explanation is to be
accepted in case of doubt, the side-effect hypothesis should
function as a null hypothesis for future research, because in
contrast to the rape-adaptation hypothesis, it does not require
specific additional assumptions.
What about the ewoiutionary psyecology of
coerciweness?
Margo Wilson and Martin Daly
Department of Psychology, McMaster University, Hamilton, Ontario,
Canada L8S 4K1
Electronic mail: daly@mcmaster.ca
The target article is concerned with one of the most fundamental
issues for those interested in developing an evolutionarily so-
& Thornhill: Evolution of sexual coercion
phisticated psychology, namely, the problem of recognizing and
characterizing psychological adaptations. Specifically, the au-
thors grapple with the question of how one decides whether (a)
human male minds manifest "design" features that have evolved
for purposes of sexual coercion, or (b) such coercion is better
understood as an incidental consequence of mental adaptations
whose functions reside elsewhere. We expect that many will
reject this framework, and yet the framing of this question (rape-
specific adaptation versus rape as a byproduct of other adapta-
tions) should not be controversial. The only available alterna-
tives to evolutionary adaptationism are creationism or the denial
that adaptive function is a scientifically useful concept; both
stances are sterile. More appropriate subjects for debate are
Thornhill & ThornhilFs (T & T's) specific arguments about the
predictions that may be derived from these alternatives and
their interpretations of the evidence.
As T & T argue, present-day occurrences of sexual coercion
must somehow be the outcome of a masculine psychology that is
sensitive to cues or other informational tokens of situational
predictors of statistically expectable fitness consequences in the
past. There is an obvious adaptive rationale for persistent male
sexual interest in the face of lesser female interest, and there is
an obvious adaptive rationale for agonistic capability and the
effective communication of threat when others' interests are in
conflict with one's own. There is surely sexual adaptation and
there is surely coercive adaptation, but it does not follow that
there is psychological adaptation for specifically sexual coercion.
In trying to decide whether such adaptation exists, T & T pay
considerable attention to comparing coercive and noncoercive
sexuality, but they pay no attention to a comparison that may be
equally relevant: sexual versus nonsexual coerciveness.
T & T argue that the historical incidence of female refusals has
been an adaptive problem for males that has shaped male sexual
psychology to include coercion. And yet, by the same reasoning,
coercive tactics must have had fitness consequences in other
areas of interpersonal social conflicts. So how specific should the
claims be about sexual coercion as an adaptation? We venture
that a review of the literature on robbery or other personal
property offences, on slavery, on warfare, and even on politics,
would reveal that conflict in these domains has been a poten-
tially potent selective context for the evolution of carefully
modulated coercive capabilities. If this is the case, then the
interesting task for selection-minded psychologists is to unpack
the phenomenon of the use of coercion in different contexts and
then evaluate what (if any) are the distinct features of specifically
sexual coercion.
According to T & T, male scorpionflies have a clamp that is
designed for retraining unwilling females. A (the?) major reason
for believing this to be a rape adaptation is that it is evidently
used in no other context. (But even if males also used their
clamps to grapple with rivals or prey, it might be possible, in
principle, to ask whether the detailed topography and function-
ing of the clamp implied that it was specifically adapted for
restraining mates.) Is there anything analogous to the scor-
pionfly's clamp in the evolved structure of human male behav-
ioral control mechanisms? Although T & T argue for the plau-
sibility of rape-specific mental adaptation - and although they
show that the idea survives an important challenge in that male
sexual arousal is often surprisingly little affected by female
unwillingness (which may not be recognized as remarkable by
those lacking a comparative perspective on the importance of
female choice in controlling insemination) - it is only in their
concluding paragraphs that they begin to formulate specific
hypotheses about the form(s) that such "mental clamp(s)" might
take. Arguably, some of these concluding hypotheses - sugges-
tions about sensitivity to the perceived threat from a victim's kin
group, for example - can still be accommodated to a notion of
generalized rather than sex-specific coercion adaptation.
Whether there are mental adaptations for sexual coercion
remains an open question, but at least it has now been posed.
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 403
Response/Thorahill & Thornhill: Evolution of sexual coercion
We laud the efforts of Thornhill & Thornhill to address - at a
basic and hitherto largely ignored level - the question of why
men use violence in the context of heterosexual intercourse.
sexuality: What
Randy ThornhSII3 and Nancy Wilmsen ThornhiIIb
^Department of Biology and ^Departments of Biology and Anthropology,
University of New Mexico, Albuquerque, NM 87131-1091
Electronic mail: nthorn@unmvm.hitnet
R1. The undisputabSe
Some commentators found certain ideas in the target
article incredible that are not legitimately disputable. We
briefly treat these comments here in order to move to
more appropriate areas of debate.
Contrary to the opinions of Akins & Windham,
Brownmiller & Mehrhof (para. 4) and apparently Bayer
& Steele (para. 2), rape really does stem from psychologi-
cal adaptation (sect. 3). All psychological change (emo-
tion, learning, etc.) and all behavior are the products of
psychological mechanisms processing environmental in-
formation, and environmental information processing re-
quires psychological structure/adaptation. Our current
understanding of the evolution of life strongly suggests
that adaptations are ultimately the products of past selec-
tion (e.g., Dawkins 1986). As Wilson & Daly point out
(para. 1), the only alternative ultimate theory of phe-
notypic design is divine creation, and that is a myth, not a
scientific deduction. For those who doubt that rape is a
consequence of evolved structure in men's brains, we
suggest the references in section 3 (second paragraph).
Thus, we did not treat the "sociocultural" theory of rape as
a valid, ultimate alternative, as Russell (para. 2) and
Archer (para. 1) suggest we should have, because it is not
a legitimate alternative.
Ecclesiastical celibacy unquestionably reflects psycho-
logical adaptation, as pointed out by Freyd & Johesoe9
but there is no psychological adaptation to the evolu-
tionarily novel circumstance of institutionalized celibacy.
Celibacy is a very complex and interesting issue that has
received only limited attention from adaptation-minded
researchers (e.g., Alexander 1979a; Dickemann 1981;
Hager, in press).
As we explain in section 3.2, data on the current
reproductive consequences of rap.e are irrelevant to the
hypothesis that there is a psychological adaptation to rape
per se. Current adaptiveness is not a legitimate theoreti-
cal expectation of any adaptation. Moreover, byproducts
of adaptations may or may not be adaptive now. Despite
all this, Futterman & Zirkel (para. 2) claim that in order to
demonstrate that male coercive sexuality is the result of
selection it must be shown that individual differences in
the tendency to rape lead to differential reproductive
404 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
success. Langley (para. 1) and Palmer (para. 2) apparently
also feel that rape researchers should be collecting data on
the current adaptiveness of rape and that our ignoring it is
a flaw in the approach we propose. We suggest the
references cited in section 3.2 plus Grafen 1988 and
Symons (in press).
Futtermae & Zirkel (para. 1-3) claim we must demon-
strate that rape is "determined by the genes" in order to
apply adaptationism to rape. All behavioral and non-
behavioral traits of individuals have an ontogeny and thus
they are caused by gene-environment interaction during
development. Of course, genes alone cannot cause any
phenotypic trait of an individual and in this sense the
phrase "determined by the genes" is misleading. Futter-
man & Zirkel (para. 1, 6, and 8) also consider it necessary
to show that rape is heritable (as defined in the target
article, sect. 3.4) in order to analyze rape from an evolu-
tionary perspective. That the evolutionary perspective
can be applied only to heritable traits is an erroneous but
not uncommon view (e.g., Lewontin et al. 1984).
Whether or not rape is heritable is an open question, as
pointed out in the target article (sect. 3 and 4) and by
Mealey (para. 5 and 6), Bixler (para. 6), and Ghiselin
(para. 2), but in order for any adaptation to evolve, there
must in the evolutionary past have been genetic variation
(heritability) in the trait, because the evolution of adapta-
tion involves changes in gene frequencies. We can be
certain that regardless of the nature of the psychological
adaptation underlying rape, there was heritable variation
in it in the human evolutionary line. There is no "skirting
the trap of heritability" (Brownmiller & Mehrhof, para. 1)
in the target article because there is no trap (for a full
discussion, see Crawford & Anderson 1989; R. Thornhill
1990; Tooby & Cosmides 1990b).
evolutionary psychoSogy appropriate for
s of rape?
Although Dupre (para. 1) and Gavey & Gray (para. 2)
apparently believe that rape reflects psychological adap-
tation, they feel the idea is trivial or naive. This idea must
be the starting point of any meaningful analysis of men's
sexuality, however, for two reasons: (1) It can counter the
erroneous view that the adaptationist and sociocultural
perspectives on rape are mutually exclusive. Adaptation
does not imply an absence of environmental influences -
in fact, it implies many necessary environmental influ-
ences (see below). (2) The idea that rape reflects psycho-
logical adaptation allows the immediate formulation of the
two alternative ultimate hypotheses to explain rape,
thereby identifying a basic empirical challenge for re-
search on men's sexuality: to determine whether there is
psychological adaptation for sexual coercion itself (i.e.,
adaptation specifically designed for generating rape in
human evolutionary history) or whether rape is an inci-
dental byproduct of noncoercive sexual psychological
adaptations together with coercive nonsexual psychologi-
cal adaptations. Determining which of these alternatives
is correct is not a trivial problem. We believe it is
important and urgent to examine empirically the evolu-
tionary design of the psychology underlying sexual coer-
cion. A focus on the design features of men's sexual
psychology will be required in order to understand rape
 Response/Thornhill & Thornhill: Evolution of sexual coercion
whether or not there is specific phenotypic design for
rape. Human psychology is based on the most fundamen-
tal information-processing machinery in the living world.
We need to study its evolved cue-processing properties to
understand behavioral variation within and between indi-
viduals: Which cues, at the various stages in a man's life,
shape and influence his attitudes toward rape.
Kitcher points to our poor understanding of the "pre-
cise character of the mechanism and . . . other psycho-
logical mechanisms [that] might interfere with or rein-
force it." He further points out that "detailed psychology
is entirely lacking from the target article." He feels that
because of this lack we can do nothing but provide the
reader with "banalities, a review of some behavioral
studies, and a mass of loose associations." Gavey & Gray
have a similar view. It would seem to follow from this that
all scientific endeavors are commonplace and unworthy of
attention because no study ever begins with a complete
understanding of the phenomenon under investigation. If
one ever did, it would be banal by definition. The un-
known is what scientists strive to discover. And, yes, we
are poorly informed about the exact psychological ma-
chinery involved in promoting sexual coercion. That is
why we wrote the target article, to begin to become better
informed. If no one ever asked the question, no one
would ever get the answer. Gavey & Gray and Kitcher's
position here is anii-scientific.
We do not mean to imply in the target article that an
analysis of psychological design is an alternative to behav-
ioral analysis, as inferred by Palmer (para. 2 and 5) an
Langley (para. 3). Instead, we are suggesting that pheno-
typic variation in the behaviors, feelings, arousal, learn-
ing, and so forth surrounding sexual coercion provides a
window on the design of the underlying species-typical,
sexual psychological structure. Thus, we are arguing for a
change in emphasis and goals in the analysis of sexual
coercion. The diversity in the forms of sexual coercion,
coupled with sexual arousal to laboratory stimuli, a behav-
ioral, emotional, and motivational measure, could iden-
tify specific environmental cues processed by the under-
lying psychological machinery. Rape behavior per se is
too variable to serve as the ultimate descriptive goal of an
adaptational analysis of sexual coercion. An adaptation
exhibits phenotypic invariance and is a species-typical
(often sex- or age-specific) phenotypic feature.
Smuts (para. 3) and Fetterman & ZirteT(para. 5) argue
that a single psychological adaptation for rape is inconsis-
tent with the wide diversity of behavior comprising sexual
coercion. A wide range of phenotypic variation is not
inconsistent with a single psychological control mecha-
nism. For example, a single psychological mechanism
seems to control the variable behavioral phenomenon of
incest avoidance in human beings (N. Thornhill 1990; in
press; see also R. Thornhill 1991 for a discussion of the
evolutionary psychology of the high intraspecific vari-
ability of male display). Even so, specificity of design does
not rule out multiplicity. Many pieces of psychological
machinery could be specific to rape (see sect. 1, para. 1).
R3» Development and indiwiduai differences
Several commentators question our view about the devel-
opment of the psychology underlying sexual coercion.
Dupre (para. 1 and 2), Futterman & Zirkel (para. 9),
Browemiller & Mehrhof (para. 9), and Gavey & Gray
(para. 7-9) feel that our perspective denies any possible
role for social factors, or, in the case of Gavey & Gray,
environmental factors in general, in the ontogeny of the
psychological mechanism underlying rape. The view of
ontogeny we discuss briefly in section 3 of the target
article is the epigenetic evolutionary one universally
adopted by biologists (e.g., see excellent treatments in
Ch. 10, Daly & Wilson 1983; Oyama 1985). The misun-
derstandings about development embedded in the social
science/feminist theory of rape have led many investiga-
tors to assume an incorrect view of ontogeny - one based
on capricious environmental determinism. The assump-
tion is that the only thing necessary for human sexual
behavior to develop - the behavior's sole or only signifi-
cant cause, or the only one worth mentioning - is arbi-
trary social learning. Because of this assumption, the
social scientific theory of rape is inconsistent with what is
known about ontogeny, including knowledge about the
role of learning experiences during ontogeny and about
how adaptations guide developmental trajectories. Three
insurmountable problems with the social scientific theory
of rape are immediately apparent: First, the sexual be-
havior of an individual human develops as the product of
the intimate and inseparable interaction of both genes
and environment. Second, social learning is only one
category of a vast multitude of necessary environmental
influences (experiences) during the development of the
sexuality of a woman or a man. And third, the environ-
mental influences on the ontogeny of human sexuality are
the product of evolutionary history and are thus specific
and nonarbitrary.
(1) For any feature, adaptive, nonadaptive or other-
wise, of an individual human (or other organism) current
understanding of development rules out a dichotomy of
genetic versus environmental factors, an opposition be-
tween subsets of the environment such as learned or
cultural ones. First, the functional expression of genes
during development is always environmentally triggered.
Second, each adaptation guiding ontogeny has its own
specific ontogeny. These, like all adaptations and all
individual traits that are incidental effects of adaptations,
are the outcomes of innumerable causal interactions be-
tween genes and environment. Finally, the functional co-
relationship of these interactions is too intimate to sepa-
rate into genes and environment; it is meaningless to
suggest that any trait of an individual is environmentally
or genetically determined. It is also inaccurate to use
terms such as "genetic leash" (Futterman & Zirkel, para.
2), or "genes . . . direct the formation" of psychological
adaptation (Kitcher, para. 6). Applying these terms to any
trait in an individual implies that genes have a primary
influence. To regard a trait as influenced only by the
environment is also to misunderstand development.
Dupre (para. 2), for example, feels that learning is a
sufficiently complete explanation for most human be-
havior
(2) With respect to ontogeny, the environment means
everything that is essential for development that is exter-
nal to genes: the immediate environment of the genes
within an individual (other genes and their products); the
environment within/the developing individual, including
adaptations that are generated by ontogeny; and the
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 405
Response/Thoinhill & Thornhill: Evolution of sexual coercion
environment - social and nonsocial - external to the
individual.
(3) It is always the phenotypic effect of the interaction of
genes and environment that selection evaluates, and
when a given interaction promotes individual reproduc-
tion more than an alternative interaction, the genetic
underpinning of the interaction increases in frequency in
the population. When selection acts in a directional way
over long periods of evolutionary time, adaptations,
which are outcomes of the retention by selection of the
most reproductively successful gene-environment inter-
actions, are created. Adaptations then, as Tooby and
Cosmides (1990) have emphasized, are manifestations of
evolved gene-environment interactions. The environ-
mental (experiential) and the genetic influences during
development are not only equally important and totally
inseparable, they are specific and nonarbitrary because
they both reflect evolutionary history, and equally so.
Points (1) to (3) together, we suggest, reveal the errors
in Dupre's view in causally dichotomizing psychological
adaptation and social conditioning (para. 5) and in Gavey
& Gray's view contrasting social versus biological (para.
9). According to this epigenetic view of ontogeny, each
psychological adaptation, including those that affect hu-
man sexuality, is designed by selection during its evolu-
tion to process specific, now-arbitrary information in the
environment. Such design is expected whether a psycho-
logical adaptation uses learning experiences or is influ-
enced only by experiences during ontogeny that do not fit
standard definitions of learning. Individuals whose psy-
chological traits did not guide behavior, learning, feel-
ings, development, hormone release, and so on, adapt-
ively in human evolutionary history are no one's
evolutionary ancestors. Psychological traits of individuals
that were incorporated into psychological adaptations as
design features by selection during human evolution had
one essential property that made them and not alternative
traits successful under the scrutiny of selection: This was
their greater ability to solve the environmental problem
affecting selection. The perception and processing of
arbitrary environmental information by psychological
features can only lead to psychological changes and be-
havioral effects that do not provide solutions to an en-
vironmental challenge causing selection. Thus each psy-
chological adaptation (as well as each nonpsychological
one) has evolved because of a precise, specific, nonarbitr-
ary relationship between phenotype and environment.
Furthermore, psychological adaptations, whose on-
togeny depends in part on social learning, evolved in
humans in a complex social environment of changing
conflicting and confluent interests among individuals. It
is highly likely, therefore, that these adaptations are
designed to be sensitive to specific, nonarbitrary social
information.
Contrary to Gavey & Gray's (para. 8) claim, we do not
suggest in the target article that learning or sex-specific
socialization is unimportant in the ontogeny of the psy-
chology responsible for rape. We noted that if socializa-
tion and learning are involved (and we believe they are)
then the cues are not arbitrary; they are specific, evolved
cues. We also mentioned that the usual view of learning is
unsatisfactory for explaining any human behavior (also see
Cosmides & Tooby 1989). Learning is merely an experi-
ence that subsequently affects action; all ontogenetic
406 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
environmental influences fit this definition. Social learn-
ing or socialization resembles learning as an explanation;
it merely means that the relevant experience is with
conspecifics. We feel it will be necessary to modify the
current concepts of learning and socialization in the social
and behavioral sciences and include the modern perspec-
tive that clarifies differences between categories of on-
togenetic/environmental experiences.
Contrary to the comment by BrownmiSSer & Mehrhof
(para. 5), we did not claim in the target article that
Brownmiller (1975) put forth the arbitrary socialization
theory of rape in Against our with We cited other sources
of this theory.
Contrary to what Smuts (para. 5) and Quinsey (para. 1)
infer, we do not deny individual differences nor do we
believe they are trivial, as suggested by Malamuth. We
instead point out in the target article that, according to
the rape-specific hypothesis, the differences are expected
to reflect different evolved environmental experiences
during the development of men's sexuality (sect. 3.4,
para. 3). We hope it is clear that individual variation in no
way argues against the hypothesis that there is rape-
specific psychological adaptation, as suggested by Mealey
(para. 5 and 6).
Certain commentators argue that we do not analyze in
enough detail individual differences in the propensity to
use sexual coercion and be aroused by rape stimuli in the
laboratory. This is a fair criticism. More research is
needed here. Malauiuth and Quinsey focus on this im-
portant issue. Malamuth mentions that the differences
between men in response to laboratory depictions corre-
late in the expected way with variables such as self-
reported likelihood of raping and sex-role stereotyping. If
the rape-specific hypothesis is correct, the individual
differences must be explicable by variables that reflect
the benefits and costs of using sexual coercion (e.g., age,
male resources, men's perceptions of the social conse-
quences of raping [punishment vs. impunity] etc.). Cer-
tain research that. Malamuth and colleagues have already
done on individual differences seems to provide signifi-
cant support for the rape-specific hypothesis. For exam-
ple, Malamuth et al. (1986) found that 71% (of 367) male
college students indicated that they would find some or
considerable sexual excitement in forcing a woman to
have sex. In the second part of the study subjects' arousal
to rape and consensual sexual laboratory depictions was
measured; 118 of the 367 volunteered. There was no
statistically significant difference between the subjects'
penile responses to rape and consensual depictions. Yet
the relative responses of the volunteers to rape depictions
correlated positively and significantly with the degree to
which they justified male aggression against and domi-
nance over women. We believe that different individual
ontogenies — that is, different evolved experiences —
affected the differences in ideology and arousal in this
study. Analyzing and understanding the causal ontogene-
tic experiences, however, will require further research.
In this study, a conclusion that the results are explained
by learning is misleading. It was not demonstrated that
learning, social or otherwise, applies to these results, or
even to the general domain of interest here. Indeed, none
of the ontogeeetic experiences playing a causal role were
identified.
Confusion surrounding how individual differences re-
 Response/Thornhill & Thornhill: Evolution of sexual coercion
late to our perspective may stem in part from the er-
roneous but common view that development ends when
an individual reaches adulthood. (It is also erroneously
assumed by some that socialization begins at birth [see
Dupre, para. 1]; it begins at conception.) Development
continues throughout adult life. Thus, to determine the
ontogeny of individual differences in attitude about sexual
coercion or arousal to it, one must keep in mind all
developmental events up to the time of the subjects'
assessment.
Thus, we believe that the differences between incar-
cerated rapists and normal men and the variations among
normal men have developmental explanations based on
how different experiential backgrounds affect the proba-
bility of rape. This will be true whether rape reflects
adaptation to rape or is an incidental effect of adaptation to
domains other than rape. It will also be true if rape
proclivities are heritable. Behavioral variation is never
independent of ontogeny and environmental information
processing by psychological mechanisms. Even if rape
tendencies were highly heritable it would still be neces-
sary, in order to understand rape and predict its occur-
rence, to determine how differences in the environment
affect the genetically varied sexual psychologies of men
related to their propensity to rape.
During development, environmental cues that were
the original evolved cues (see point 3 above) or are
currently correlated with them will shape sexual psychol-
ogy. Etcher (para. 5) and Smuts (para. 6) make the
interesting suggestion that the novelty of Western human
environments may result in the ontogeny of a sexual
psychology that is significantly different from the one that
arose in the ancestral evolutionary environment. This is,
of course, a possibility, but Smut's (as well as Bupre's)
candidate novel ontogenetic experiences of TV and movie
images that endorse sexual coercion (para. 6) are not
convincing. Sexual coercion in the context of warfare has
been common throughout history in traditional societies
that have no contact with Western media (see examples
discussed by Bartueg; also see Brownmiller's (1975)
review of rape in warfare). Even if aspects of men's sexual
psychology are novel evolutionarily, the psychology is
still responsible for rape and hence the necessity of the
evolutionary psychological approach we propose. Only
when the causal cues are understood will people be able
to determine whether it is feasible to create environments
that will prevent rape.
ist fallacy
In the target article we felt it unnecessary to remind
readers of the naturalist fallacy because it is treated
extensively in the literature of human evolution and the
literature of philosophy. Even so, Dupre (last paragraph)
calls our target paper "harmful" science because we por-
tray rape as a natural phenomenon, which could imply
that it is inevitable, even moral or good. He goes on to say
that since our claim is not true, its propagation should be
strongly resisted. Our claim is true: Rape is a product of
nature, a natural phenomenon. As we discussed above,
rape reflects psychological adaptation and thus the natu-
ral processes of historical selection and ontogeny. It does
not follow that rape is inevitable or good. We suggested in
the target article that rape may be controllable if the
design of its underlying psychology is understood. Nature
simply is, and what ought to be is inferred by people.
Brownmiller & M-ehrhof (last sentence) succumb to the
naturalist fallacy when they suggest that we dignify rape
by discussing it in terms of adaptation. Aklns & Windfaam
endorse the naturalist fallacy in their point that the target
article treats rape without making moral judgments (see
also Bayer & Steele, last paragraph; Dupre, para. 2 and
6). Our response is not the forum for discussing our own
ideologies about rape. But given that we suggest in the
target article that our approach may help identify en-
vironments that could reduce rape, it follows that we
would like to see rape eliminated. Dupre (para. 3) won-
ders why we don't discuss the evolution of the psychologi-
cal machinery underlying moral sentiment. This topic has
been treated in detail by Alexander (1987b) and Irons
(1991).
Gowaty introduces her commentary with the claim that
the target article is sexist because it proposes that men
and women have sex-specific, species-typical sexual psy-
chologies, and thereby creates sexual asymmetric stereo-
types: What if within-sex variation in sexuality is greater
than between-sex variation? If so, Gowaty would appar-
ently conclude that there are no evolved sex differences in
mental traits affecting behavior. First, there is tremen-
dous variability within the sexes in all aspects of sexual
behavior. This we do not deny. Second, high phenotypic
variance within a sex does not imply that no psychological
machinery occurs in all members of that sex and that sex
alone. Women whose lifetime numbers of unique sexual
partners exceed those of most men are not acting in the
absence of sex-specific psychological adaptations, the
expressions of which might be predictable. Variance in
athletic ability might be similar in men and women, for
example, but this does not negate sex-specific, species-
typical psychological mechanisms that result In higher
average athletic ability in men because of hormone titers,
muscle development, and so forth. We emphasize that a
belief in sexual differences is not the same as a belief in
sexual inequality. The sexes either are or are not different
in a certain domain. To suggest that a conclusion of
difference Implies that one sex is superior to the other is
oneformof the naturalist fallacy. Dupre (para. 6) commits
thisformof the fallacy in his argument that the whole area
of sociobiology has a "deeply political nature."
The target article seems sexist to Bayer & Steele too.
They argue that we provide only "phallocentric focus" and
perpetuate "male rape culture" by not discussing
women's experiences of being raped. Until this paper,
our work on rape tested our hypothesis for the evolution-
ary significance of psychological pain and focused almost
entirely on the psychological trauma experienced by rape
victims of all ages (R. Thornhill & N. W. Thornhill 1989).
The work on victims is published separately (N. W.
Thornhill & R. Thornhill 1990a; 1990b; 1990c; 1991); it
addresses rape victims exclusively. The target article is
about men's sexual psychology and is thus primarily
focused on men.
R5. Adaptationism
The great value of the study of adaptation is widely
recognized in biology. Adaptations are the products of
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 407
Response/Thornhill & Thornhill: Evolution of sexual coercion
past directional selective pressures. Thus, the under-
standing of the functional design of an adaptation is in
itself a demonstration of the nature of the directional
selection that shaped the adaptation. Adaptations are
biologists' sole source of information about the forces of
directional selection that were actually effective in de-
signing phenotypes during the evolutionary history of
life.
BMer (para. 2) criticizes the target article for using
teleological terms to describe adaptation. Evolution by
cumulative directional selection is not a purposive pro-
cess, but it Incorporates, by gradual, persistent effects,
purpose into its products, that is, its adaptations. This is
why many evolutionists refer to adaptations as goal-
directed or purposeful (e.g., Dawkins 1986; Symons
1989; Williams 1966; and many others; see R. Thornhill
1990 for a review).
Dupre (para. 4) and Gavey & Gray (para. 2) criticize
the adaptationist approach of the target article. Dupre is
astonished that we propose a perspective of "universal
adaptationism." His preference is proudly to join the
ranks of those (e.g., Gavey & Gray) who blithely agree
that psychological machinery or neurological mechanisms
underlie all human behavior and then suggest that inves-
tigating the exact character of these pieces of the human
mind Is so trival (apparently because knowledge of the
truth of psychological design is so trival) as to render it
nonsclentific. Our claim that psychological adaptation
underlies all human behavior is, again, not a reasonably
disputable one, and this realization is the necessary initial
intellectual step toward identifying the domain of alterna-
tive ultimate hypotheses that may explain human rape
(see also commentary by Wilson & Ba!y5 para 1). Dupre
would apparently find no investigation of complex prob-
lems preferable to a complete investigation of them. He
also criticizes us for ignoring the "devastating critique" of
adaptationism by Gould and Lewontin (1979). Certain
conceptual weaknesses underlying these critiques have
already been addressed (see Alexander 1987b; Bateson
1985; Dawkins 1985; 1986; Konner 1984; Thornhill &
Alcock 1983). Gould and Lewontin's central point - that
biologists should be conservative in applying the concept
of adaptation — is valid and worthwhile, but it is neither
devastating nor especially original. Williams (1966) made
It earlier and better. In the target article, we heeded
Willlams's suggestion to apply adaptation only when
necessary. In our analysis of men's sexuality, we sought
the level of phenotypic analysis that describes species-
typical phenotypic design and we use the levels of high
variance (coercive sexual behavior) in an attempt to dis-
cover that level. Gowaty's suggestion that we should
hypothesize adaptive value for almost every sexual vari-
ant used by men and women is the antithesis of our own
and Williams's approach. Dupre also points out that we
ignore Kitcher's (1985; see also BBS multiple book review
of'Vaulting Ambition, BBS 10(1) 1987) critique of adapta-
tionism and the various feminist critiques of adaptation-
ism as applied to sexual coercion, including the feminist
critiques of our ideas on rape. We have the opportunity to
respond to Kitcher in this Response.
The two major feminist critiques of our ideas and those
of others concerning the evolutionary significance of rape
were Fausto-Sterling's (1985) and Sunday and Tobach's
(1985). This Response is not the place for a point-by-point
408 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
critique of these two books. However, the target article
and this Response contain the essence of our answers to
all the major criticisms in the two books: The criticisms
range from the claim that the adaptationist perspective on
rape is racist and genetically deterministic (Tobach &
Rosoff 1985) to the claim that (1) adaptationism does not
apply because rape probably rarely results in offspring
(i.e., it is not currently adaptive), (2) rape has not been
shown to be heritable, and (3) rapists are not motivated by
a desire to increase their reproductive fitness (see espe-
cially Harding 1985 in Sunday & Tobach, and Fausto-
Sterlieg 1985, Ch. 6). We have dealt with all of these
points in writing for BBS. We hope the inaccuracy of
certain feminists' claims about racial issues is obvious to
readers. Race per se is irrelevant in our analysis of rape.
Our argument is for species-typicality in the adaptation
underlying rape and thus for the psychic identity of all
men.
Gavey & Gray suggest that the proper way to study
adaptation involves a combination of optimality and com-
parative analysis. These are two useful ways to study
adaptation, but they are not the only approaches of value
(see R. Thornhill 1990 for a discussion of the diversity of
methods for analyzing evolutionary purpose/phenotypic
design). The target article adopts the most widely used
approach in the study of adaptation (see R. Thornhill
1990). Phenotypic design is observed or Inferred and then
the investigator asks, "What is its evolutionary function -
that is, its functional design?
Archer suggests that an ESS (evolutionarily stable
strategy) game-theoretic approach should be the first step
in the analysis of rape and that if the theoretical outcome
is favorable, then and only then is it appropriate to
examine the data on men's sexuality in the literature. ESS
is only one of many useful approaches in the study of
adaptation and has no special priority (R. Thornhill 1990).
It has In fact been used to analyze male sexual coercive-
ness; the results suggest that selection for coercive male
adaptation should often have been effective and male rape
adaptation may therefore be common across animal spe-
cies (Hammerstein & Parker 1987; Parker 1979).
Futterman & Zirkel's list of criteria for demonstrating
that a history of selection is responsible for a trait exhibits
a misunderstanding about the difference between the
study of evolutionary history and the study of current
microevolution and selection (see R. Thornhill 1990).
Gavey & Gray (para. 6) stumble on the same ground.
They miss the important distinction between evolution-
ary design and the present-day effects of adaptation on
fitness when they claim that there is "considerable evi-
dence" that men's sexual psychology is poorly designed
and then cite various current maladaptive effects of this
psychology. There is no reason to believe that any adapta-
tion will yield positive fitness consequences currently.
Furthermore, the conclusion that men's sexual psychol-
ogy is poorly designed is obviously premature because so
little is known about its actual evolutionary design.
AMes & Wiedfaam raise the question of what sort of
explanation is useful for understanding human sexuality,
given its complexity. They do not answer the question,
but feel that evolutionary adaptationism is not adequate
to handle the complexity of the topic. Life is the most
complex phenomenon in the known universe and its
diversity is felt by biologists to be explicable in ultimate
 Response/Thornhill & Thornhill: Evolution of sexual coercion
terms by evolution. Evolutionary hypotheses have led to
deep insight into and a satisfactory understanding of very
complex natural phenomena - For example, sex ratio
adaptations, the diversity of sexual differences and sim-
ilarities across species, and, more to the point of Akins &
Windham's concern, human behaviors such as homicide
(Daly & Wilson 1988). Complexity in living systems
implies the need for evolutionary thinking and not an
exemption from it. We should also mention that the
sexuality of snails is extremely complex - arguably more
so than that of people.
Langley (para. 2) thinks we believe that because a trait
exists it must be an adaptation or product of historical
selection. Given that we suggest in the target article that
rape behavior is not an adaptation, it is difficult to under-
stand how Langley arrived at his interpretation. On the
contrary, we would argue that most individual traits,
human and nonhuman, are not adaptations (see Symons
1987a; R. Thornhill 1990; Williams 1966).
Langley feels that there is "evidence" against the rape-
specific adaptation hypothesis. He cites Sanday's (1981)
claim that there are some societies in the ethnographic
record in which rape is absent or rare. It should be
obvious that this is not necessarily contrary evidence. The
rape-specific hypothesis proposes that rape is a condi-
tional tactic. In addition, it should be kept in mind that
Sanday's cross-cultural study uses inappropriate com-
parative analysis (see Shields et al. 1982).
R6- General wersus special-]
Chiselin (para. 3-5) questions the utility of the concept of
special-purpose adaptation in general and in the particu-
lar case of rape. He uses two examples to support his view
that, under some cases, general adaptations are optimal
evolutionary solutions. In the target article we mention
that the human heart is specifically designed to pump
blood in a human body. Ghiselin claims this is not true; he
says that the heart evolved over half a billion years and
humans must make do with what nonhuman primate
ancestors passed down. We certainly believe that selec-
tion will always act on what came before, but it is not the
case that the human heart shows no design specific to the
human body plan and life style. Just as the human eye is
designed to cope with many human-specific problems - it
is part of the machinery involved in assessing a human
mate's reproductive potential rather than the reproduc-
tive potential of a bovine mate, for example - the human
heart is a phenotypic solution to numerous species-
specific problems (e.g., pumping blood in a bipedal
organism of human size). Phylogeny and adaptive design
are different causal matters. Darwin argued that descent
(phylogeny) with modification by selection accounts for
the diversity of species-specific adaptations.
Ghiselin again fails to distinguish phylogeny and adap-
tation in his second example, the expression of emotion.
Darwin demonstrated that facial and postural expressions
of human emotions and the underlying psychological
mechanisms have a phylogenetic history by showing that
the expressions have homologues in certain nonhuman
mammals - that is, he demonstrated the descent of
emotions and their associated psychology. The profound
differences between the emotional expressions of humans
and our closest nonhuman primate relatives demonstrate
modifications in morphology and psychology influencing
emotional expression as a result of selection ot tactors that
were specific to the human evolutionary line (Thornhill &
Thornhill 1989). Phylogenetic analysis describes the
movement of a phenotypic trait in a phylogenetic tree and
can identify the original state (or origin) of an adaptation
that was modified by selection to meet species-specific
problems. The functional design of an adaptation and the
original state of an adaptation prior to its subsequent
modification by selection are different causal problems
(see R. Thornhill 1990). Eibl-Eibesfeldt sketches a sce-
nario in which the phylogeny of male sexuality and taxon-
specific male sexual adaptation are separated.
Thinking of adaptations as general purpose rather than
special purpose is widespread in the biological literature.
This general-adaptation concept stems in large part from
superficial thinking about functional level. If one con-
cludes that the human heart's evolutionary function is to
pump blood, it is not difficult to imagine that any general
pump design will work. But if the adaptive problem to be
solved by selection is framed more specifically and appro-
priately as the need for a pump that will work in a human
body (not a general primate or a baboon body), the
general adaptation hypotheses is unreasonable. Mealey
(para. 3) implies that a general mechanism such as emo-
tion underlies rape. But there are numerous emotions,
and the underlying psychological machinery of each emo-
tion is almost certainly designed to cope with a specific
problem.
Symons (1987a; 1987b) has pointed out that environ-
mental problems affecting fitness are specific and not
general problems. Thus, it is expected that phenotypic
solutions to environmental problems - that is, adapta-
tions - will be special purpose in design, because a
general-purpose mechanism cannot solve a specific prob-
lem (see also Tooby & Cosmides 1989). Beyond this
theoretical advance in the concept of adaptation is the vast
body of empirical data suggesting that the best-
understood adaptations in plants and animals are indeed
special purpose in their functional design. In part, these
considerations led us to our hypothesis of rape-specific
adaptation. Also, because rape involves factors unique
within the scope of human sexuality, we reasoned that
rape may reflect rape-specific information processing
mechanisms. Perusse (para. 4 and 5) suggests that rape
reflects three general-purpose psychological adaptations
rather than the two we identify in the target article as
components of the side-effect hypothesis. He argues that
viewing rape as a side-effect of multiple adaptations
provides a more parsimonious and powerful explanation
than rape-specific adaptation. A similar argument based
on parsimony is offered by Gladue (para. 2), Figueredo
(para. 3), and Voland. Smuts rejects the rape-specific
adaptation hypothesis without discussing her alternative,
presumably because she believes that general-purpose
adaptation is sufficient. However, for the reasons men-
tioned above (because environmental problems are spe-
cific and not general, and because known adaptations are
special purpose in design) the rape-specific hypothesis is
actually more parsimonious than the general-purpose
one.
Mazur points out that the same two alternative hypoth-
eses for robbing sexual access could be applied to robbing
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 409
Response/Thornhill & Thornhill: Evolution of sexual coercion
food. Obviously, food robbing reflects either adaptation
to food robbing or it is a side-effect of other adaptation.
Although it is possible that adaptation to food robbing has
evolved, Mazur's cheese-Danish-robbing adaptation
clearly could not have evolved, because, unlike food,
cheese Danish was not a feature of the evolutionary
environment of human adaptedness. Also, regarding the
similarity of predictions from rape-specific and food-
specific robbery discussed by Mazur, people do not
become hungrier when they successfully wrestle a Big
Mac to the floor and gain physical control over it (see
prediction 2, target article).
R7. Alteraatiwe models
Several commentators suggested alternative models of
sexual coerciveness. Mealey (para. 3 and 7) proposes
three. One is that coercion evolved because females
prefer coercive males, or simply males that persist in
copulation attempts. She has in mind a form of the sexy-
son model of sexual selection. This model is not easily
applied to sexual coercion in humans, however, because
it seems to predict that women should test the persistence
or forceful skills of men in all copulations (see Thornhill &
Sauer 1991); apparently, this is not the case. Mealey also
suggests that rape might reflect a "developmentally con-
tingent, species-wide" strategy. She see this as distinct
from a species-wide, condition-dependent strategy,
which we argue is the proper description of men's mating
strategy, with rape as one of its tactics. We fail to see the
difference, because the evolved cue conditions that de-
termine the adoption of the tactics of a condition-
dependent strategy are developmental ones.
Quinsey suggests that a "modified view of the rape-
specific adaptation hypothesis" is needed, one that is
based on a male adaptation that promotes learning to
connect aggression and sexual arousal Eibl-EIbesfeldt
and Makmuth suggest a similar hypothesis but without
invoking learning as a mechanism affecting the connec-
tion. It is possible that an evolved connection between
aggression and sexual arousal occurs during the ontogeny
of the psychological adaptation underlying rape; the con-
nection need not involve learning. This hypothesis must
be framed differently, however; it lacks the specificity to
account for sexual coercion, because most male aggres-
sion is directed at other men, and without accompanying
sexual arousal.
Mealey suggests that variation in rape behavior may
reflect heritable variation in behavioral thresholds. The
same idea is suggested by Bixler (para. 4) and a similar
suggestion is made by Ghiselin (para. 1-3). The heritable
threshold model is not an alternative to our proposal. We
tentatively suggest a threshold as a possibility in regard to
human population differences in rape (sect. 3.4, para. 4),
but we emphasize in the target article that the rape-
specific hypothesis predicts that the basic psychology
underlying rape is virtually fixed genetically - it has a
heritability at or near zero. Small heritable differences in
thresholds for rape activation may exist within popula-
tions as a result of mixing between populations with
genetically different thresholds. Also, some heritability in
rape adaptation per se may arise as an incidental effect of
the heritability of resistance to disease organisms. The
410 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
continual revolutionary race between hosts and their
parasites and pathogens probably accounts for most of the
genetic variation in plant and animal populations (see
reviews in Howard 1991; Tooby & Cosmides 1990).
Ghiselin (para. 2) apparently feels that human popula-
tion differences in nonpsychological traits imply impor-
tant population differences in sexual-psychological de-
sign. Such inferences across categories of phenotypic
traits do not seem logical. Furthermore, it is becoming
increasingly apparent that in terms of evolved mental
structure, the people of the world are fundamentally the
same. Of course, there are also age- and sex-specific,
species-typical psychological adaptations in Homo sa-
piens (e.g., see Brown 1991 on human universals for an
excellent treatment).
Kitcher (para 4) proposes that men have multiple
evolved strategies. In the evolutionary literature dealing
with types of alternative phenotypic traits within popula-
tions, the concept of multiple strategies refers to genet-
ically distinct alternatives (Dawkins 1980; Gross, in press;
Maynard Smith, 1982; Thornhill & Alcock 1983; West-
Eberhard 1979), but there have been very few convincing
demonstrations of alternative male mating strategies de-
spite much research on the topic in the last 15 years. And,
contrary to Kitcher's claim, genetically distinct alterna-
tive male mating strategies have not been demonstrated
in nonhuman primates. The general conclusion is that
almost all male within-species alternatives are conditional
tactics of single strategies (Gross, in press; Thornhill &
Alcock 1983; West-Eberhard 1979). Kitcher's view that
men have genetically distinct mating strategies is also
difficult to accept, on both theoretical and empirical
grounds: Theoretically, the ability to switch alternatives
will usually be favored by selection in situations of social
competition (e.g., Gross, in press; West-Eberhard 1979).
Empirically, men do switch sexual behavior in a
condition-dependent manner (see sect. 7).
Archer constructs a weak hierarchy of conditional strat-
egies. He suggests (para. 3) that there are three kinds of
conditional strategies but that we present only one of the
three possible models: We view rape as a conditional
tactic used In relation to evolutionary cues of cost and
benefit. Surely, conditional tactics adopted, as Archer
puts it, because "the main tactic is unavailable" involve an
assessment of costs and benefits. If individuals adopt a
conditional tactic because they are poorly fitted to under-
take the main tactic - another of Archer's models - then
the main tactic Is unavailable because of costs (see Thorn-
hill & Thornhill 1983 for a complete discussion).
The target article does not propose that sexual coercion
is a "sex-linked" adaptation, as suggested by Futterman &
Zirkel. It Is sex-limited, not sex-linked. Both sexes have
the relevant genes, but they influence development only
in males.
Gavey & Gray (para. 2) suggest that we argue in the
target article (sect. 5) that evidence of selection for rape is
evidence for rape-specific adaptation. In section 5, we
discussed evidence that ancestral human males some-
times increased their reproductive success by rape. We
mentioned in the target article that this is not evidence for
the existence of rape-specific adaptation (sect. 12, para.
1). It is important to distinguish between the occurrence
of selection and selection that yields adaptation. Most
selection is mere noise in long-term evolution and does
 Response/Thornhill & Thornhill: Evolution of sexual coercion
not yield phenotypic design; this is one important reason
why the evolutionarily significant selective forces can
only be inferred from their consequences - that is, from
the design of adaptations (see R. Thornhill 1990; Williams
1966).
Kitcher feels we are wrong to ignore how coercive
copulation might have affected male fitness for "primitive
hominids living in relatively small groups." There is no
reason to believe that all human psychological adaptations
evolved in social contexts of small groups of related
interactants. There is also no evidence that the relevant
social environment was small groups of close relatives.
The social organization of certain extant hunter-gatherers
and other "referential models" (Tooby & Devore 1987),
for example, nonhuman species, especially certain pri-
mates, does not provide a mirror of the environments in
which human nature evolved. To use these as heuristic
models in assessing assumptions or predictions from the-
ory is scientifically unsound. The leading hypothesis for
the evolution of the human psyche is that the mental
capabilities uniquely expressed in H. sapiens arose from
selection in the context of very complex social competi-
tion (Alexander 1989; Humphrey 1976). The social en-
vironment of small family groups does not provide the
social complexity needed for the evolution of the human
mind. It could only have been produced in an environ-
ment of effectively large groups with the concomitant
complexity of confluences and conflicts of interests (see
especially Alexander 1979b; 1989). The nature of evolu-
tionarily significant past environments can only be in-
ferred from the adaptations that arise to meet these
environments. Only the designs of human adaptations
contain data about the selectively significant features of
the environment of human evolution.
Hartung's commentary identifies war as a potentially
important context for the positive selection of rape. We
suggest that war was the most common context for selec-
tion for rape during human evolution, but not the only
important context. Circumstances of low cost/high bene-
fit rape may have been a frequent aspect of the within-
group social environment of H. sapiens, as we discuss in
section 5.
The prevalence of rape in war (see Hartueg; also
Brownmiller 1975) demonstrates that many men (of all
ethnic groups, which is relevant to our claim of species
typicality of rape adaptation) are sexually aroused and
competent in the total absence of female sexual interest
and willingness. As Hartung points out, it is reasonable to
assume that the women involved were not in the mood to
copulate with men they had watched slaughter their
relatives. Thus, perhaps even more than the data in the
target article, rape in war calls into serious question
Voland's suggestion that signs of sexual interest are neces-
sary to arouse men.
Eibl-Eibesfeldt believes that sexual coercion in war is
primarily an act of male dominance. Although aggression
is involved in rape and aggressive control of victims is one
motive behind rape, the motivation to have a sexual
experience is also central. Rape in war is focused on the
sexual preferences of men in general - women of re-
productive ages - not on men or females of nonreproduc-
tive ages (see Hartung's commentary; also Brownmiller
1975).
Archer and Votaed suggest that further considerations
about sperm competition might clarify the rape-specific
hypothesis by better identifying the conditions under
which rape was adaptive for males during human evolu-
tion. Archer and Voland believe that females benefit from
sperm competition. This benefit, Archer feels, bears on
the cost of rape to females and explains why females of
many pair-bonding species (birds and H. sapiens) fre-
quently engage in extra-pair copulations (EPCs). We do
not suggest in the target article (nor do we believe) that
EPCs have evolved only in the context of rape, but it is
premature to conclude that any aspect of EPCs by females
has evolved in any species to promote sperm competition.
Bellis and Baker's (1990) results on EPCs of women are
consistent with this sperm competition interpretation,
but there is an alternative explanation. Benshoof and
Thornhill (1979) suggested that noncoerced extra-pair
copulation by women is a means by which females paired
to males of low genetic quality upgrade the genetic
quality of future offspring (also Hamilton 1990; M0ller
1988). Double-mating, coupled with women's differential
acceptance of the sperm of men (e.g., by pres-
ence/absence of orgasm; see Archer's discussion of Baker
& Bellis 1989c) or differential use of sperm of men could
lead to effective sire choice and could minimize the pair-
bond male's detection of infidelity. Subtle, postmating
forms of female choice of sires are receiving increased
attention and empirical support (Eberhard 1985; R.
Thornhill 1983; Watson 1991).
The above suggests that there is as yet no reason to
conclude that ejaculate competition is female driven, as
Archer and Volaed (point 2) suggest. Sperm competition
stemming from rape in war is clearly not female driven,
and there is no unequivocal evidence that EPC behavior
by females reflects a tactic to increase sperm competition
to produce sexy sons (in this case, sons who are good at
sperm competition).
There are reasons to believe that high levels of sexual
selection in the form of sperm competition would often
have accompanied rape during human evolution. Rape in
war probably often led to significant sperm competition
because of temporal overlap of ejaculates of different
rapists in the victims, in addition to any overlap between
the pair-bond mate's ejaculates and those of the rapists.
Also, as we mention in sections 5 and 11 of the target
article, rape of pair-bond mates in the context of sexual
jealousy and infidelity would lead to sperm competition,
and in general, a woman's resistance to rape by a non-
pair-bond male may often stem from her sexual interest in
a pair-bond male, a condition promoting sperm competi-
tion in rape. Thus, rape ejaculates that are designed for
high sperm competition (sperm numbers and/or spe-
cialized morphs) might have evolved in men. The study of
ejaculate design may be possible using laboratory audio
and video depictions of consensual and coercive sex.
R8- Distinguishing the rape-specific and side-
effect hypotheses and predictions
Several reviewers focused on how the two hypotheses
might be distinguished in future work, offered alternative
interpretations of the data discussed in the target article,
or suggested additional predictions from the hypotheses.
Figueredo calls for more specific differential predic-
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 411
Response/Thornhill & Thornhill: Evolution of sexual coercion
tions from the two hypotheses. However, we emphasized
in section 12 that future work should focus on determin-
ing whether there is one (or several) psychological mecha-
nism^) in men that processes information that is specific
to the benefits and costs of rape. That is, the empirical
Issue here is whether sexual coercion Is regulated by cues
specific to the benefits and costs of rape. If so, rape-
specific psychological machinery is implied.
We suggest In the target article that carefully designed
laboratory studies of men's sexual arousal in response to
depictions of rape and consensual sex are an important
technique for addressing the question of rape-specific
adaptation. Blip re (para. 4) feels that this approach is
"shady," "bizarre," and "worthless." We briefly discussed
the validity of the laboratory approach in the target article
(sect. 8.3, para. 2). Qiiimsey5 one of the pioneers in this
Important area of sexual research, discusses in some detail
the discriminant and predictive validity of phallometric
assessment and stresses our point that phallometry is one
of the best ways of distinguishing the rape-specific and
side-effect hypotheses. Despite Dupre's strong words, he
offers no significant support for his criticism. The one
point he makes Is that men's excitement while watching
televised sports does not imply that men are designed to
play sports. But men do not show sexual excitement when
they watch televised intergroup competition. The Inter-
est that people show in social scenarios depicted In the
media, and the age and sex differences in the scenarios of
Interest, are neither trivial nor well-understood phenom-
ena; as Alexander (1989) has emphasized, such patterns
probably contain important information about the design
of human psychological adaptations.
Figueredo (para. 4) suggests that the rape-specific
hypothesis requires that coercive sexual laboratory stim-
uli be more arousing to men that noncoerclve sexual
stimuli. This is not a valid prediction. We emphasized
that certain scenarios - for example, wartime versus
peacetime - are predicted to affect the relative magni-
tude of men's response to coercive and noncoercive
scenarios (sect. 12).
Langley argues that instead of raping an unfaithful
mate, it would be more in the reproductive interests of a
man to refrain from sex with an unfaithful mate in order to
avoid cuckoldry. Langley's prediction assumes that pair-
bonded men have many alternative mating opportunities.
Otherwise, the waiting male might fail to reproduce.
Alternative Interpretations of certain data In the target
article were entertained by several commentators.
Gladue (para. 4), Laegley (para. 5), Palmer (para. 4), and
Perusse (para. 5) offer other Interesting interpretations of
the bondage and spanking study discussed in the target
article (sect. 9). Quinsey and Malamiith feel that the
bondage and spanking transcripts In the study by Quinsey
et al. (1984) were sexual in nature. Our interpretation that
they were without sexual content was based on the fact
that the sample bondage and spanking transcripts, unlike
others that were used in the study, did not include
mention of sexual intercourse. We believe that laboratory
studies that refine this original study may provide a
greater understanding of men's sexuality (see sect. 9,
para. 7).
Malamuth, another pioneer In the study of men's
sexual arousal to laboratory depictions, offers what he
feels are significant clarifications of some of the findings
412 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
from this literature that we review. Initially, he makes
three claims. First, that most subjects show more sexual
arousal to depictions that combine consent, woman's
arousal, and no violence than to those that combine verbal
nonconsent, disgust, and violence. Second, (a) a high
degree of sexual violence against a victim appears to
"inhibit" men's sexual arousal relative to low/no violence
and (b) the victim's disgust, in contrast to her arousal, has
an inhibiting effect. Third, consent alone does not affect
men's arousal (e.g., Malamuth & Check 1983). We agree
with his interpretations of the literature bearing on the
first and third points. We disagree with his use of the
word "inhibit" In the second points. Mealey apparently
has the same view of the second points as Malamuth. The
data show that the victim's disgust and gratuitous violence
against the victim result in less sexual arousal than vic-
tim's arousal and limited violence (see sects. 8.2, 8.4).
The data do not show that disgust and gratuitous violence
prevent - that is, Inhibit - men's sexual arousal: For
example, In the Blader and Marshall (1984) study dis-
cussed in the target article, subjects in general show
significant sexual arousal to depictions in which the victim
shows disgust and In which gratuitous violence is used.
Malamuth goes on to suggest that the rape-specific
hypothesis would seem to predict an interaction between
the verbal consent/nonconsent and other variables that
might be perceived by men as influencing their general
social reputations. In the experimental scenarios he dis-
cusses, this would involve permissive/normal Instruc-
tions (Quinsey et al. 1981). As Malamuth points out, no
such interaction has been found, despite the rather well
established interaction between the consent dimension
and the dimension of victim arousal/disgust. However,
the increasing evidence that men's willingness to use
force in mating is constrained by the influence of rape on
their social reputation (see discussion below of rape in
warfare and of Quinsey and colleagues' recent work)
suggests that further experiments are In order. As Mal-
amuth emphasizes, how men weigh the information
about verbal consent/nonconsent compared to other in-
formation about the costs of rape (presence or absence of
arousal) Is not understood. This is certainly a central Issue
and could probably be clarified by laboratory experi-
ments. The interaction between consent and arousal that
Malamuth discusses extends the empirical Implications of
the rape-specific hypothesis.
Malamuth seems to realize the Importance of the
distinction between violence toward victims and physical
control over the sexual partner that we discuss in the
target article. Yet he asks why violence against women
should inhibit men's arousal. It Is physical control of an
unwilling woman that the rape-specific hypothesis pre-
dicts will be sexually arousing to men because it implies
that rape can be successfully accomplished. High degrees
of violence toward a victim may be perceived as an
absence of physical control of the victim, and this may be
why violence against women In laboratory depictions
lowers men's sexual arousal. Again, Malamuth's term
"inhibition" is misleading here (see discussion of Blader
and Marshall [1984] in the target article; sect. 8.2, para.
14). Thus, it is not necessary for us to rely only on Briddell
et al. 's (1978) study, which was not replicated successfully
by Barbaree et al. (1983), to conclude that true inhibition
of men's arousal in the context of viewing sexual violence
 Response/Thornhill & Thornhill: Evolution of sexual coercion
against women has not been demonstrated. Blader and
Marshall's (1984) study shows significant sexual arousal in
response to limited violence as well as to gratuitous
violence against women in rape depictions. The rape-
specific hypothesis predicts that cues of violence that are
most closely associated with achieving physical control
over victims will be the ones that arouse'men. Inciden-
tally, we agree with Malamuth that males who were
primarily sexually aroused by violence that would se-
riously injure themselves or their rape victims were not
the reproductively successful males in the human evolu-
tionary line.
Malamuth and Quinsey suggest that the rape-specific
hypothesis requires that Quinsey et al.'s (1981) com-
munity men who received permissive instructions should
show a greater response to rape than to consenting depic-
tions relative to community men with regular instruc-
tions. This was not found: The men with permissive
instructions showed similar increases in response to both
rape and consenting depictions. The permissive instruc-
tions probably produced greater arousal to consenting
cues merely because testing community men in the
environment of a maximum security mental health center
may lower their responsiveness to otherwise arousing
cues (see sect. 8.2, para. 1). What is remarkable about this
study is that only three sentences of permissive dialogue
gave rise to the statistically significant pattern we discuss*
in the target article, despite the small sample size. These
results suggest the need for other studies of permissive
instructions.
Malamuth questions his earlier interpretation of an
aspect of the research reported in Malamuth (1981b). He
suggests that the audio rape depiction containing victim
abhorrence read by a female assistant may have had
subtle seductive qualities and therefore caused the sub-
jects to focus only on the sexual content and not on the
violent content. He points out the need for better experi-
ments to determine whether a woman's voice per se is a
cue that increases men's arousal to rape. We agree, of
course, and suggest that a young woman's voice reading
sexually explicit material, whether it includes rape or not,
will be sexually interesting and often arousing to many
men.
AHgeier & Wiederman suggest that further research is
needed to clarify the cues that arouse men viewing
coercive scenarios in the laboratory. We agree. This is a
major point of the target article (sect. 12).
Wilson & Daly point out that our prediction about how
the power of a rape victim's family affects men's arousal to
rape (sect. 1.2, para. 6), if met, could be explained by the
side-effect hypothesis. We acknowledge the ambiguity of
this prediction. Remarkably little is known about the
actual cues that affect men's arousal in coercive settings;
we accordingly need the laboratory experiments manip-
ulating all the cues discussed in sect. 12, para. 6.
We also agree with Wilson & Daly that it will be
important in future research to pay attention to sexual and
nonsexual coerciveness and not just coercive and non-
coercive sexuality. This would be useful for distinguishing
coercive cues that are specific to sex from coercive cues
that are important in nonsexual contexts.
AHgeier & Wiederman argue that the support we claim
for the rape-specific hypothesis is based on selective
inclusion of data and misinterpretations of findings. They
point to a study by Abbey (1987), not cited by us, as
contradictory. However, all studies, the three cited by us
(Abbey 1982; Abbey & Melby 1986; Saal et al. 1989), as
well as Abbey (1987), show the same general pattern:
Men are more likely than women to infer sexual interest
in a potential partner when no such interest exists. We fail
to see the contradiction. Also, certain studies suggest that
pornographic literature commonly includes sexually
coercive cues. The studies mentioned by AHgeier &
Wiederman do not show this. There is an inconsistency
between studies. New studies that include rigorous inter-
rater reliability checks should be conducted to answer
this question. The literature on sexual fantasies that
AHgeier & Wiederman discuss indicates a consistent sex
difference in fantasies of forcing sex on a partner, with
men having more; thus, the pornography industry, which
competes for men's sexual arousal, is likely to have ex-
plored this sex difference.
AHgeier & Wiederman cite a study by Groth and
Burgess (1977) of incarcerated rapists showing that these
men are frequently sexually incompetent in coercive
settings. Gavey & Gray (para. 4) cite Harding (1985) to
make this same point. We do not claim that men will show
equal sexual competence in coercive and noncoercive
sexual settings. Rape often occurs under circumstances
that are dangerous to the perpetrator; this could lead to
lower sexual performance (see above discussion of the
role of victim's physical control). However, we have
questioned the validity of the comparisons of sexual
performance in rapists described in the literature (see
Thornhill & Thornhill 1983; also Palmer 1989). Further-
more, the widespread use of sexual coercion by men
discussed in the target article implies that men are often
sexually competent during rape (sect. 8.3, para. 3).
AHgeier & Wiedermae criticize our interpretation of
Buss (1989b). Buss did find statistically significant sex
differences in the incidence of anger about sexual rejec-
tion (target article, sect. 7.1, para. 4). This as an interest-
ing and relevant result. AHgeier & Wiederman rightly
point out that although the gender difference was statis-
tically reliable, it was small. We re-emphasize that fur-
ther studies of this sexual conflict might be revealing of
subtle forms of male coercion.
AHgeier & Wiederman (para. 8) apparently believe
that sexual coercion is rare, in part because a study by
Byers and Lewis, not referred to in the target article,
showed that date rape is infrequent. We contend that
men rape conditionally; an analysis of the circumstances
surrounding different frequencies of date rape would be
interesting and illuminating. We stand by our conclusion
that there is an abundance of sexual coercion in dating and
other similar situations. The position of AHgeier &
Wiederman on the rarity of sexual coercion is called into
question by certain research results: In studies of self-
reported likelihood to rape, roughly one-third of men
reported a likelihood, and, under certain conditions, no-
likelihood reporters showed sexual arousal to laboratory
depictions that included coercion (sect. 8.2, para. 5); their
position is also challenged by Malamuth et al. (1986)
discussed above.
AHgeier & Wiederman and Gavey & Gray (para. 3)
question the quality of the pattern indicating that men of
lower social status use more sexual coercion than other
males. There are limited data that suggest that the pattern
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 413
Response/Thornhill & Thornhill: Evolution of sexual coercion
is not strong, and possibly nonexistent. This inconsis-
tency underlies the need to bring the issue into the
laboratory and to examine the arousal of men of different
socioeconomic status to coercive and consensual stimuli;
we also hope this will be done to examine the effect of
men's age (see sect. 10, para. 1).
Allgeier & Wiederman (last paragraph) make the inter-
esting point that women sometimes sexually coerce men.
The approach in the target article does not deny this. It
predicts that sexual coercion by women is an incidental
effect of adaptation to conditions other than coercing
sexual access.
Browemiller & Mehrhof (para. 6) suggest that rape is
random. There are now considerable data on rape victim-
ization revealing that men do not randomly select victims
in terms of age and that the occurrence of penile-vaginal
intercourse and ejaculation are not random during rape.
Young women are overrepresented as rape victims (Fel-
son & Krohn 1990; Russell 1982; Thornhill & Thornhill
1983). Also, recent research on a large sample of victims of
all ages indicates that the probability of penile insertion,
ejaculation, and multiple copulations during sexual as-
sault depends on the victim's age, and that women of
reproductive ages are more likely to experience these
events than nonreproductive aged victims (N. W. Thorn-
hill & R. Thornhill 1991).
Akins & Wiedham and Gavey & Gray (also Grauerholz
& Koralewski 1991; Harding 1985) suggest that the adap-
tationist view implies that rape results from motivation to
reproduce or increase fitness. This is not one of its
predictions. No organism is motivated to enhance fitness.
Motivation refers to proximate rewards and punishments,
and not ultimate goals. Thus we have proposed that one
motive behind rape is the desire for a sexual experience
and that another is to physically control the victim (see R.
Thornhill & N. W. Thornhill 1991).
Contrary to the suggestion by Gavey & Gray9 the
prediction that sexual coercion will be sensitive to the
probability of detection and negative social consequences
(prediction 5, sect. 6) does not imply that stranger rape
will be more common than acquaintance rape. Acquain-
tance rape is expected to be far more common (and
apparently is) because of greater opportunity.
Prediction 5 is further supported by data from recent
studies by Quinsey and colleagues cited in Quinsey's
commentary. Also, Hartraig's case study analysis of rape
in war, like the material on rape in warfare that we discuss
in the target article (sect. 10, para. 2), indicates that men
increasingly rape when the probability of punishment is
low.
Blxler (para. 5) and Allgeier & Wiederman (para. 5)
discuss the interesting issue of women's arousal to labora-
tory rape depictions. The perspective in the target article
predicts that women will not be found to have a psycho-
logical adaptation that processes cues that are specific to
women raping. As mentioned above, women's arousal to
the depictions probably reflects responses to cues that are
not specific to rape. This hypothesis is supported by
Stock's (1988) results on women's arousal to the realistic
versus "distorted" rape depictions discussed by Allgeier
& Wiederman. It is also supported by Malamuth et al.'s
(1980) study of male and female college students' self-
reported responses to rape depictions with and without
victim pain and arousal. They conclude that, "Sur-
414 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
prisingly, . . . although female subjects were most
aroused when the rape victim was portrayed as experienc-
ing an orgasm and no pain, males were most aroused
when the victim experienced an orgasm and pain" (p.
399). These two studies point to the value of an empirical
evaluation of the cues that arouse men and women in
laboratory rape depictions; future work on these sex
differences could be important in distinguishing the two
ultimate hypotheses for rape.
We do expect women, however, to have mechanisms
designed specifically for assessing the probability of their
rape in social settings. Also, evidence is accumulating that
women have a psychological adaptation to assess and
reduce the costs of rape to them (N. W. Thornhill & R.
Thornhill 1990a; 1990b; 1990c; 1991).
Archer and Voland (point 1) discuss other possible
female counteradaptations to rape, specifically mecha-
nisms that prevent conception or lead to early termina-
tion of pregnancies due to rape. Despite the logic behind
the idea that such adaptation should exist, there is no
evidence that it does.
R9B The continuum and! definition
Palmer (para. 4) appears to misunderstand our position
about the continuum of sexual coercion/noncoercion
(sect. 7.1). We did not write, nor do we believe, that all
human copulations are forced to some degree. Instead,
we pointed out in the target article that the dichotomy of
forced versus unforced copulation is too simplistic; these
two tactics often grade into each other, with only arbitrary
bounds between them. Thus, contrary to Palmer's claim,
our view does not mean that researchers cannot possibly
devise coercive and noncoercive laboratory stimuli.
Part of Palmer's discussion of the continuum between
coercive and noncoercive sex confuses two issues. One is
our claim that forced and unforced copulation often grade
into each other, which we address Immediately above.
The second is the matter of how people use terms such as
rape versus consensual sex. These two terms "make
sense" to many people, and many people find them
useful, but this is not evidence that the terms describe a
real dichotomy. The major reason words are viewed as
useful or nonuseful is that they are or are not consistent
with the user's interests. There are reasons why people
would find their interests served by these dichotomous
terms; in this sense, the terms are not arbitrary. For
example, men often view themselves as engaging only in
consensual sex - the sexually coercive ones are the other
guys. This dichotomy may be comparable in utility to the
distinction made by many heterosexual men between a
homosexual and a heterosexual sexual preference, two
preferences that grade into one another, as shown by
Kinsey and his colleagues. The utility of the coer-
cion/noncoercion dichotomy for women may be that it is
used to identify and present socially the significant males
in their lives (husbands, brothers, sons, etc.) as distinct
from certain others. Also, the dichotomy may be useful to
many women because their interests may be served by
remaining in a sexually coercive relationship that has not
reached the point of costs exceeding benefits; such rela-
tionships would not involve rape, according to the
women. Arguments similar to these may explain the
 Response/Thornhill & Thornhill: Evolution of sexual coercion
utility of the common distinction between rape and sexual
coercion.
If sexual coercion cannot easily be distinguished from
noncoercion, Malamuth asks, how can there be specific
adaptation to rape? First, as we pointed out immediately
above, many human copulations are not coerced. Also,
consensual and coercive sex present different problems to
males and may thus affect different adaptive solutions.
Sexual coercion has many costs, some obvious, others
subtle. An example of the latter is a man's nonviolent
threat of withdrawal of financial support in order to obtain
a copulation with a pair-bond mate that would be other-
wise unavailable. Such a threat might be very costly,
leading to the mate's general suspicion and distrust. We
suggest that men often use subtle hints in this regard that
imply withdrawal of emotional or other resources without
an explicit statement of withdrawal. Such hints make for
ambiguity, but the ambiguity itself can be an important
threat. In human evolutionary history, males obtaining
matings from unwilling and sometimes resistant mates
under these circumstances would seem to require infor-
mation processing specific to the benefits and costs of
sexual coercion. For this reason and others we have
discussed, we do not believe that the problems of adapt-
ive sexual coercion can be solved by the kind of general-
purpose adaptation Malamuth proposes.
Gowaty emphasizes that there are exceptions to some
of our statements about human sex differences and sug-
gests that her more "robust framework" be adopted in the
study of human sexuality. One aspect of her program is a
description of all circumstances under which vaginas and
penises come together. The second aspect is imagining
hypotheses about the adaptive significance of each of the
sexual variants. As we mentioned above, phenotypic
variation is not an adaptation or even numerous adapta-
tions but the manifestation of the underlying invariant
species-typical mechanisms that are the sexual adapta-
tions. The continuum we proposed incorporates variation
in sexuality by definition and is the relevant variation for
identifying species-typical rape adaptation. The diversity
of ways human vaginas and penises come into contact is
not the limit of the relevant variation in this case. Not
infrequently, menforcecopulation with men and nonhu-
raan species. These variations also imply the existence of
species-typical male sexual psychology that does not re-
quire the partner's sexual interest or willingness (see R.
Thornhill & N. W. Thornhill 1991 for a discussion).
Moreover, we believe that forms of sexual harassment
such as those mentioned by Brownmiller & Mehrhof —
frottage and genital exhibitionism - are appropriately
viewed as manifestations of the same male sexual psychol-
ogy that causes rape. Contra Brownmiller & Mehrhof,
these are not phenomena that we have failed to consider
(see R. Thornhill & N. W. Thornhill 1991). Such assaults
are further evidence that women's sexual arousal is often
irrelevant to men's sexual arousal.
Smuts (para. 4) argues that there is a qualitative distinc-
tion between rape, which she defines as "using force to
obtain copulation," and men threatening to withhold or
withdraw social or economic resources to obtain matings
they would not obtain otherwise. The continuum we
propose has, of course, two ends which are more distinct
from each other thanfromother points in the continuum.
And, of course, copulations that are physically forced are
distinct from copulations that resultfromsome amount of
"bargaining," but elements of coercion are automatically
involved in situations in which men threaten to withhold
or withdraw social or economic resources to achieve
matings (sect. 7.1, para. 3). We recognize that a signifi-
cant portion of the interaction between men and women
surrounding copulation is appropriately viewed as re-
ciprocity (noncoercion), but this form of reciprocity and
coercion will often grade into each other.
Futterman & Zirkel (para. 5) and Gowaty find the
definition of rape in the target article inappropriate be-
cause it does not include the diversity of definitions across
cultures, nor the multiple definitions in Western culture.
Elsewhere we discuss cross-cultural patterns in rape and
rape attitudes; that analysis suggests that the cross-
cultural diversity follows evolutionary logic (see Thornhill
& Thornhill 1983). Our definition in the target article is
appropriate for our evolutionary view that male sexual
coercion circumvents mate choice and is on a continuum
with noncoercive sex.
Smuts (para. 2) notes that the continuum idea was first
proposed by feminists. We cited the early feminist litera-
ture that contained this idea (sect. 7.1, para. 1) and we
thank Smuts for providing another recent reference.
>e in nonhuman animals
A few commentators challenged our view of how rape in
nonhuman animals bears on our approach (sect. 3.3).
Bixler (para. 3) points out that the comparative method
might help illuminate rape. We agree. We emphasize
that we do not think the comparative method of biology is
an inappropriate tool for studying adaptation. It is one of
two legitimate general approaches to identifying and
characterizing phenotypic design - the other is used in
the target article. Both of us use the comparative method
often and have written extensively about its validity (N.
W. Thornhill 1990; 1991; Thornhill 1984a; Thornhill &
Alcock 1983; Thornhill & Thornhill 1983). There are, how-
ever, appropriate and inappropriate uses of comparisons.
The successful application of the comparative method
to sexual coercion would uncover the ecological factor(s)
associated with the presence and absence of rape-specific
male and female (counter-rape) adaptation across species.
Such an application might even lead to successful predic-
tion of different kinds of rape adaptations among species.
For example, in certain scorpionfly species males force
initial sexual access; in another, males do not force access
but instead force copulation to continue after it is ob-
tained without coercion (see Thornhill & Sauer 1991).
Examples of inappropriate comparisons are those sug-
gested by Brownmiller & Mehrhof (para. 3) and Kitcher
(para. 4). Brownmiller & Mehrhof claim that the or-
angutan is the only nonhuman primate with forced
copulation. They also feel it relevant to add that rape in,
the orangutan has not been observed to result in preg-
nancy. This information is used by Brownmiller &
Mehrhof to support their opinion that rape did not yield
offspring during human evolution. Orangutans and peo-
ple are the end products of different evolutionary lines. If
one wishes to understand the selective forces that de-
signed orangutans, one must study the functional design
of orangutan adaptations. But if one wishes to illuminate
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 415
References/Thornhill & Thornhill: Evolution of sexual coercion
the selective forces that created human adaptations, only
the designs of human adaptations contain the relevant
information. Orangutans do not give insights into the
selective forces that were acting during the evolution of
If. sapiens. The opinion that primates give us deep
insights about the environment in which humans evolved
is not uncommon (e.g., Hrdy 1981), but It is fraught with
difficulties (see also Symons 1982; Tooby & Devore 1987).
KItcher (para. 4) makes this error when he discusses his
opinion (see above) that men have multiple sexual strat-
egies that are genetically distinct. He draws this conclu-
sion, in part, based on his interpretation of studies of
nonhuman primates.
Thornhill (1980) and Crawford and Galdikas (1986)
discuss the difficulty of demonstrating rape in nonhuman
species; for example, in many species male courtship
contains elements of aggressive behavior (also see Eibl-
Eibesfeldt's commentary). It is because of these difficul-
ties that rape has been demonstrated in only a handful of
species, and actual rape-specific adaptation in only scor-
pionflies, waterstriders, and possibly women (see Thorn-
hill & Sauer 1991 and references therein). Male scor-
plonflies and waterstriders have morphological rape
adaptations. Women may have psychological machinery
specifically designed to deal with social problems that
result from rape (see above). It is an open question
whether men have rape-specific adaptation. This ques-
tion will only be answered by understanding the design of
men's sexual psychology. Even If future studies show that
rape-specific adaptation has evolved in males of many
species of primates, it does not follow that men have It
too.
We emphasize that cross-cultural comparative infor-
mation about sexual coercion is very relevant to the study
of men's sexual psychology. Smuts (last paragraph) claims
we think that sexual coercion in cross-cultural perspective
"is entirely irrelevant." This is puzzling, given that in the
target article we attempt to analyze a species-typical
adaptation and in an earlier paper we have analyzed cross-
cultural attitudes about rape (see above).
The sQciocuituraS perspectiwe on rape: Ss it
We mentioned above that the social science /academic
feminist's view of rape (i.e., the sociocultural one: see
Russell) Is not a valid, alternative ultimate explanation to
the evolutionary perspective. The only true alternatives
are mythological beliefs. The socioculturalists are not
proposing that rape Is a product of supernatural creation.
Thus, both the sociocultural and the adaptationist views
have the following beliefs in common: Rape reflects
psychological adaptation that is a product of selection, the
relevant adaptation has an ontogeny, and men's use of
coercion Is the result of environmental information pro-
cessing by psychology. The difference between the two
perspectives is in regard to the nature of the ontogeny
involved. The adaptationist maintains that during its
ontogeny, men's sexual psychology is generated by
evolved and thus nonarbitrary environmental events.
Socioculturalists typically make two interrelated claims:
(1) Rape Is primarily or only caused by arbitrary learning
or socialization; (2) learning is a robust and sufficient
416 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
explanation of rape. For reasons we discussed in the
target article and above, the first claim is very likely to be
wrong, and the second Is definitely wrong. The sociocul-
tural view distracts us from the most significant issue -
the study of the salient evolved ontogenetic events during
the development of men's sexual coerciveness and the
evolved cues that affect the use of rape by adult men. The
sociocultural view does seem to offer hope and a simple
remedy in that it implies that we need only fix the way
boys are socialized and rape will disappear. This naive
illusion is widespread. Browemlller & Mehrhof suggest
that social policy would be useful to correct the minds of
confused young men; Futtermae & Zirkel offer a similar
suggestion. Dupre feels that learning is a sufficient expla-
nation for human behavior and that the most Important
issue or the only one is the study of the social factors
involved in rape, such as the way women are depicted in
the media. The fact is that we simply do not know which
factors, social or nonsocial, to fix at this time - that is, the
design features of rape psychology are unknown. It Is
certainly in the interests of proponents of the sociocul-
tural theory to have their colleagues and the public
believe that they have identified a remedy for rape and
are working toward Implementing It. But this can only be
an effective social tactic in an environment of ignorance.
As Harteeg points out, those who feel that the social
problem of rape can be solved by changing the nature of
men through naive and arbitrary social adjustments
should "get real about rape" because their perspective is a
danger to us all.
ACKNOWLEDGMENTS
Thanks to Martin Daly and Patrick Thornhill for comments on
the Response. Thanks also to Anne Rice for typing the manu-
script, and to Andi Causeyformiscellaneous assistance. Finally,
we thank the Harry F. Guggenheim Foundation for support of
our research on sexual coercion.
References
Letters a and r appearing before authors' initials refer to target article and
response respectively.
Abbey, A. (1982) Sex differences in attributions for friendly behavior: Do
males misperceive female friendliness? Journal of Personality and Social
Psychology 42:830-38. [arRT, ERA, TL]
(1987) Misperceptions of friendly behavior as sexual interest: A survey of
naturally occurring incidence. Psychology of Women Quarterly 11:173-
94. [rRT, ERA]
Abbey, A., & Melby, C. (1986) The effects of non-verbal cues on gender
differences in perception of sexual intent. Sex Roles 15:283-98. [arRT]
Abel, G. G., Barlow, D. H., Blanchard, E. & Guild, D. (1977) The
components of rapists' sexual arousal. Archives of General Psychiatry
34:395-403. [aRT, ERA]
Abel, G. G., Blanchard, E., Becker, J. V. & Djenderedjian, A. (1978)
Differentiating sexual aggressiveness with penile measures. Criminal
Justice and Behavior 5:315-32. [aRT]
Administrative Office of the U. S. Courts (1988) In U. S. Bureau of Census,
Statistical Abstract of the United States, Table 297. U. S. Government
Printing Office. [AF]
Ageton, S. S. (1983) Sexual assault among adolescents. Lexington
Books. [ERA]
Alexander, R. D. (1979a) Sexuality and sociality in humans and other
primates. In: Human sexuality: A comparative and developmental
perspective, ed. H. Katchadourian. University of California
Press. [arRT]
(1979b) Darwinism and human affairs. University of Washington
Press. [arRT]
 References/Thornhill & Thornhill: Evolution of sexual coercion
(1987a) The biology of moral systems. Aldine de Gruyter. [aRT]
(1987b) The evolutionary approach to human behavior: What does the
future hold? In: Human reproductive behaviour: A Darwinian
perspective, ed. L. L. Betzig, M. Borgerhoff Mulder & P. W. Turke.
Cambridge University Press (Cambridge). [rRT]
(1989) The evolution of the human psyche. In: The human revolution, ed.
C. Stringer & P. Mellars. University of Edinburgh Press. [arRT]
(1990) Epigenetic rules and Darwinian algorithms: The adaptive study of
learning and development. Ethology and Sociobiology 11:241-
303. TaRT, CTPl
Alexander, R. D. & Noonan, K. M. (1979) Concealment of ovulation, parental
care, and human social evolution. In: Evolutionary biology and human
social behavior: An anthropological perspective, ed. N. A. Chagnon &
W. Irons. Duxbury Press. [aRT, AF, NMM]
Allgeier, E. R. & Wiederman, M. W. (1991) Love and mates selection in the
1990s. Free Inquiry ll(3):25-27. [ERA]
Bailey, K. (1987) Human paleopsychology. Applications to aggression and
pathological processes. LEA Publishers. [IE-E]
Baker, R. R. & Bellis, M. A. (1988) "Kamikaze" sperm in mammals? Animal
Behaviour 36:936-39. [JA]
(1989a) Elaboration of the Kamikaze sperm hypothesis: A reply to Harcourt.
Animal Behaviour 37:865-67. [JA]
(1989b) Number of sperm in human ejaculates varies in accordance with
sperm competition theory. Animal Behaviour 37:867-69. [aRT, JA]
(1989c) Sperm competition in rats and humans: An experimental evaluation
of the Kamikaze sperm hypothesis. Paper presented at Meeting of the
Association for the Study of Animal Behaviour, Zoological Society of
London, December. [rRT, JA]
Baker, R. R., Bellis, M; A. & Hudson, G. (1989) The orgasm: You've
redefined it. Company Magazine September: 60-62. [JA]
Barbaree, H. E., Marshall, W. L. & Lanther, R. D. (1979) Deviant sexual
arousal in rapists. Behavior, Research and Therapy 17:215-22. [aRT,
NMM, ERA, VLQ]
Barbaree, H. E., Marshall, W. L., Yates, E. & Lightfoot, L. O. (1983)
Alcohol intoxication and deviant sexual arousal in male social drinkers.
Behaviour Research and Therapy 4:365-73. [rRT]
Barnes, G. E., Malamuth, N. M. & Check, J. V. P. (1984a) Personality and
sexuality. Personality and Individual Differences 5:159-72. [LM]
(1984b) Psychotocism and sexual arousal to rape depictions. Personality and
Individual Differences 5:273-79. [NMM]
Baron, L. (1985) Does rape contribute to reproductive success? Evaluation of
sociobiological views of rape. International Journal of Women's Studies
8:266-77. [TL]
Bateson, P. (1983) Genes, environment and the development of behaviour. In:
Animal behaviour, vol. 3. Genes, development and learning, ed. T. R.
Halliday-& P. J. B. Slater. Blackwell Scientific Publications. [NJG]
(1985) Review: Not in our genes by Le won tin et al. (1984). New Scientist
105:58-59. [rRT]
Baxter, D. J., Barbaree, H. E. & Marshall, W. L. (1986) Sexual responses to
consenting and forced sex in a large sample of rapists and nonrapists.
Behaviour Research and Therapy 24:513-20. [VLQ]
Belknap, J. (1989) The sexual victimization of unmarried women by non-
relative acquaintances. In: Violence in dating relationships: Emerging
social issues, ed. M. A. Pirog-Good & J. E. Stets. Praeger
Publishers. [aRT]
Bellis, M. A. & Baker, R. R. (1990) Do females promote sperm competition?
Data for humans. Animal Behaviour 40:997-99. [rRT, JA, EV]
Benshoof, L. & Thornhill, R. (1979) The evolution of monogamy and
concealed ovulation in humans. Journal of Social and Biological
Structures 2:95-106. [rRT]
Betzig, L. L. (1985) Despotism and differential reproduction: A Darwinian
view of history. Aldine de Gruyter. [aRT]
Betzig, L., Borgerhoff Mulder, M. & Turke, P., eds. (1988) Human
reproductive behavior: A Darwinian perspective. Cambridge University
Press. [aRT]
Bible, The Oxford Annotated Revised Standard Version (1965) Numbers 31:7,
9-12, 14-18, 32-35; Deuteronomy 21:10-14, 22:23-29; Genesis 34:1-4,
7-16, 20-22, 24-29, 19:2-8; Judges 19:1, 14-29. [JH]
Bixler, R. (1980) Nature versus nurture: The timeless anachronism. Merrill-
Palmer Quarterly 26:153-59. [RHB]
(1986) Of apes and men (including females!). Journal of Sex Research
22:255-67. [RHB]
(1992a; submitted) Why littermates don't: The avoidance of inbreeding
depression. Annual Review of Sex Research. [RHB]
(1992b) Do we/should we behave like animals? In: The sexual abuse of
children: Theory, research and therapy, ed. W. O'Donohue & J. Geer.
Erlbaum. [RHB]
Blader, J. C. & Marshall, W. L. (1984) The relationship between cognitive
and erectile measures of sexual arousal in non-rapist males as a function
of depicted aggression. Behaviour Research and Therapy 22:623-
30. [arRT, LM, NMM]
Blakemore, C. & Cooper, G. F. (1970) Development of the brain depends on
the visual environment. Nature 228:477-78. [NJG]
Briddell, D. W., Rimm, D. C , Caddy, G. R., Krawitz, G., Sholis, D. &
Wunderlin, R. J. (1978) Effects of alcohol and cognitive set on sexual
arousal to deviant stimuli. Journal of Abnormal Psychology 87:418-
30. [arRT, NMM]
Brown, D. (1991) Human universals. McGraw-Hill. [rRT]
Brownmiller, S. (1975) Against our will: Men, women, and rape. Simon &
Schuster; Bantam Books. [arRT, SB]
Brundage, J. A. (1987) Law, sex, and Christian society in medieval Europe.
University of Chicago Press. [JJF]
Bumiller, K. (1991) Fallen angels: The representation of violence against
women in legal culture. In: At the boundaries of law: Feminism and legal
theory, ed. M. A. Fineman & N. S. Thomadsen. Routledge. [BMB]
Burian, R. M. (1983) Adaptation. In: Dimensions of Darwinism: Themes and
counter themes in 20th century evolutionary theory, ed. M. Grene.
Cambridge University Press. [aRT]
Buss, D. M. (1987) Sex differences in human mate selection criteria: An
evolutionary perspective. In: Sociobiology and psychology: Ideas, issues,
and applications, ed. C. Crawford, M. Smith & D. Krebs.
Erlbaum. [aRT]
(1988) From vigilance to violence: Tactics of mate retention in American
undergraduates. Ethology and Sociobiology 9:291-317. [aRT, JJF]
(1989a) Sex differences in human mate preferences: Evolutionary
hypotheses tested in 37 cultures. Behavioral and Brain Sciences 12:1-
14. [aRT, ERA]
(1989b) Conflict between the sexes: Strategic interference and the evocation
of anger and upset. Journal of Personality and Social Psychology 56:735-
47. [arRT, ERA,'LM]
(1991a) Evolutionary personality psychology. Annual Review of Psychology
42:459-91, [LM]
(1991b) Anger, sadness, jealousy and abandonment: Events producing
conflict between the sexes. Presented at the Human Behavior and
Evolution Society, Hamilton, Ontario, August. [LM]
Byers, S. E. & Lewis, K. (1988) Dating couples' disagreements over the
desired level of sexual intimacy. Journal of Sex Research 24:15-
29. [rRT, ERA]
Cantor, N. (in press) Social psychology and sociobiology: What can we leave
to evolution? Motivation and Emotion. [AF]
Check, J. V. P. & Malamuth, N. M. (1983) Sex role stereotyping and
reactions to depictions of stranger versus acquaintance rape. Journal of
Personality and Social Psychology 45:344-56. [NMM]
(1985) An empirical assessment of some feminist hypotheses about rape.
International Journal of Women's Studies 8:414-23. [NJG]
Christensen, F. M. (1990) Pornography: The other side. Praeger
Publishers. [TL]
Clark, L. M. G. & Lewis, D. J. (1977) Rape: The price of coercive sexuality.
Women's Education Press. [aRT]
Cosmides, L. & Tooby, J. (1987) From evolution to behavior: Evolutionary
psychology as the missing link. In: The latest on the best: Essays on
evolution and optimality, ed. J. Dupre. MIT Press. [aRT, CTP]
(1989) Evolutionary psychology and the generation of culture, part II: Case
study: A computational theory of social exchange. Ethology and
Sociobiology 10:51-98. [arRT]
Crawford, C. & Anderson, J. (1989) Sociobiology: An environmentalist
discipline? American Psychologist 44:1449-59. [arRT, LM]
Crawford, C. & Galdikas, B. M. F. (1986) Rape in nonhuman animals: An
evolutionary perspective. Canadian Psychology 27:215-30. [rRT, CTP]
Daly, M. & Wilson, M. (1983) Sex, evolution and behavior, 2nd ed. Duxbury
Press. [arRT, RHB, DP]
(1987) Evolutionary psychology and family violence. In: Sociobiology and
psychology: Ideas, issues and applications, ed. C. Crawford, M. Smith &
D. Krebs. Erlbaum. [aRT]
(1988) Homicide. Aldine de Gruyter. [arRT, RHB]
Daly, M., Wilson, M. & Weghorst, S. J. (1982) Male sexual jealousy.
Ethology and Sociobiology 3:11-27. [aRT]
Darrach, B. & Norris, J. (1984) An American Tragedy. Life 7:58-74. [aRT]
Darwin, C. (1859) The origin of species (Reprinted.) Penguin. [aRT]
(1872) The expression of the emotions in man and animals. John
Murray. [MTG]
Dawkins, R. (1976) The selfish gene. Oxford University Press (Oxford).
[PK]
(1980) Good strategy or evolutionarily stable strategy? In: Sociobiology:
Beyond nature/nurture? ed. G. W. Barlow & J. Silverberg. Westview
Press. [rRT]
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 417
References/Thornhill & Thornhill: Evolution of sexual coercion
(1985) Review: Not in our genes by Lewontin et al. (1984). New Scientist
105:59-60. [rRT]
(1986) The blind watchmaker. W. W. Norton. [arRT]
Diagnostic and statistical manual of mental disorders (DSM-III-R) (1987)
American Psychiatric Association. [DP]
Dickemann, M. (1979a) Female infanticide, reproductive strategies, and social
stratification: A preliminary model. In: Evolutionary biology and human
social behavior: An anthropological perspective, ed. N. A. Chagnon &
W. G. Irons. Duxbury Press. [aRT]
(1979b) The ecology of mating systems in hypergynous dowry societies.
Biology and Social Life 18:163-95. [aRT]
(1981a) Paternal confidence and dowry competition: A biocultural analysis of
purdah. In: Natural selection and social behavior, ed. R. D. Alexander &
D. W. Tinkle. Chiron Press. [aRT]
(1981b) Female infanticide, reproductive strategies and social stratification:
A preliminary model. In: Evolutionary biology and human social
behavior, ed. N. Chagnon & W. Irons. Duxbury. [rRT]
Dietz, P. E., & Evans, B. (1982) Pornographic imagery and prevalence of
paraphilia. American Journal of Psychiatry 139:14-39. [aRT]
Dominey, W. J. (1984) Alternative mating tactics and evolutionarily stable
strategies. American Zoologist 24:385-96. [JA]
Draper, P. & Harpending, H. (1982) Father absence and reproductive
strategy: An evolutionary perspective. Journal of Anthropological
Research 38:255-73. [JA]
(1988) A sociobiological perspective on the development of human
reproductive strategies. In: Sociobiological perspectives on human
development, ed. K. B. MacDonald. Springer-Verlag. [BBS]
Dupre, J., ed. (1987) The latest on the best. Bradford Books/MIT
Press. [JD]
Dworkin, A. (1987) Intercourse. Free Press. [BBS]
Earls, C. M. & Marshall, W. L. (1983) The current state of technology in the
laboratory assessment of sexual arousal patterns. In: The sexual aggressor:
Current perspectives on treatment, ed. J. G. Greer & I. R. Stuart. Van
Nostrand Reinhold. [aRT]
Eberhard, W. G. (1985) Sexual selection and animal genitalia. Harvard
University Press. [rRT]
Eibl-Eibesfeldt, I. (1970) Liebe und Hah. Zur naturgeschichte elementarer
verhaltensweisen. Piper-Verlag. [IE-E]
(1971) Love and hate. Rinehart and Winston. [IE-E]
(1989) Human ethology. Aldine de Gruyter. [IE-E]
(1990) Dominance, submission, and love: Sexual pathologies from the
perspective of ethology. In: Pedophilia. Biosocial dimensions, ed. J. R.
Feierman. Springer-Verlag. [IE-E]
Eibl-Eibesfeldt, I. & Siiterlin, C. (1992) Im Banne der Angst. Zur Natur- und
Kunstgeschichte menschlicher Abwehrsymbolik. Piper-Verlag. (in
press) [IE-E]
Ellis, L. (1989) Theories of rape: Inquiries into the causes of sexual
aggression. Hemisphere Publishers. [aRT, LM]
(1991) A synthesized (biosocial) theory of rape. Journal of Consulting and
Clinical Psychology 59(5):631-42. [LM]
Ellison, P. T. (1990) Human ovarian function and reproductive ecology: New
hypothesis. American Anthropologist 92:933-52. [EV]
Eysenck, H. J. (1976) Sex and personality. Open Books. [LM]
(1990) Genetic and environmental contributions to individual differences:
The three major dimensions of personality. Journal of Personality 58:245-
61. [LM]
Fagen, R. (1981) Animal play behavior. Oxford University Press. [RHB]
Falconer, D. S. (1981) Introduction to quantitative genetics.
Longman. [aRT]
Farkas, G. M. (1979) Trait and state determinants of male sexual arousal to
descriptions of coercive sexuality. Doctoral dissertation, University of
Hawaii. [aRT]
Farkas, G. M., Sine, L. F., & Evans, I. M. (1978) Personality, sexuality and
demographic differences between volunteers and non-volunteers for a
laboratory study of male sexual behavior. Archives of Sexual Behavior
7:513-20. [aRT]
Fausto-Sterling, A. (1985) Myths of gender. Basic Books. [rRT, JD]
Felson, R. B. & Krohn, M. (1990) Motives for rape. Journal of Research in
Crime and Delinquency 27:222-42. [arRT]
Fine, M. (1989) Coping with rape: Critical perspectives on consciousness. In:
Representations: Social constructions of gender, ed. R. K. Unger.
Baywood. [BMB]
Finkelhor, D. (1984) Child sexual abuse: New theory and research. Free
Press. [aRT]
Finkelhor, D. & Yllo, K. (1983) Rape in marriage: A sociological view. In: The
dark side of families, ed. D. Finkelhor, R. J. Gelles, G. T. Hotaling &
M. A. Straus. Sage. [aRT]
418 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
(1985) License to rape: Sexual abuse of wives. Holt, Rinehart &
Winston. [aRT]
Flinn, M. V. (1988) Mate guarding in a Caribbean village. Ethology and
Sociobiology 9:1-28. [aRT]
Foucault, M. (1981) The history of sexuality, vol. 1: An introduction. (Trans-
lated by R. Hurley; Original work published in 1976). Penguin. [NJG]
Freedman, D. (1979) Human sociobiology. Free Press. [RHB]
Freud, S. (1936/1964) A disturbance of memory on the Acropolis. In: The
standard edition of the complete psychological works of Sigmund Freud,
vol. XXII), ed. J. Strachey. Hogarth Press. [BMB]
Frieze, I. (1980) Causes and consequences of marital rape. Paper presented at
the American Psychological Association Meeting, Montreal, Canada,
September. [aRT]
Gager, N. & Schurr, C. (1976) Sexual assault: Confronting rape in America.
Grosset & Dunlap. [BMB]
Gagnon, J. H. (1977) Human sexualities. Scott, Foresman. [aRT]
Garrett, J. C , Harrison, D. W. & Kelly, P. L. (1989) Pupillometric
assessment of arousal to sexual stimuli: Novelty effects of preference?
Archives of Sexual Behavior 18:191-201. [TL]
Gavey, N. (1991a) Sexual victimization prevalence among New Zealand
University students. Journal of Consulting and Clinical Psychology
59:464-66. [NJG]
(1991b; submitted) Rape and sexual coercion within heterosexual
relationships: A review of research and feminist perspectives. [NJG]
Gordon, M. T. & Riger, S. (1989) The female fear. Free Press. [BMB]
Gould, S. J. & Lewontin, R. C. (1979) The spandrels of San Marco and the
Panglossian paradigm: A critique of the adaptationist programme.
Proceedings of the Royal Society of London 205:581-98. [JD]
Gowaty, P. A. (in press) Evolutionary biology and feminism. Human
Nature. [NJG]
Grafen, A. (1988) On the uses of data on lifetime reproductive success. In:
Reproductive success: Studies of individual variation in contrasting
breeding systems, ed. T. H. Clutton-Brock. University of Chicago
Press. [rRT]
Grafen, A. & Sibly, R. (1978) A model of mate desertion. Animal Behaviour
26:645-52. [PK]
Grammer, K. (1989) Human courtship behavior: Biological basis and cognitive
processing. In: The sociobiology of sexual and reproductive strategies, ed.
A. E. Rasa, C. Vogel & E. Voland. Chapman Hall. [aRT]
Grauerholz, E. & Koralewski, M. (1991) What is known and not known about
sexual coercion. In: Sexual coercion: A sourcebook on its nature, causes
and prevention, ed. E. Grauerholz & M. Koralewski. D.C. Heath &
Co. [rRT]
Gray, R. D. (1987) Faith and foraging: A critique of the "paradigm argument
from design." In: Foraging behaviour, ed. A. C. Kamil, J. R. Krebs & H.
R. Pulliam. Plenum Press. [NJG]
(1988) Metaphors and methods: Behavioural ecology, panbiogeography and
the evolving synthesis. In: Evolutionary processes and metaphors, ed.
M.-W. Ho & S. W. Fox. John Wiley. [NJG]
Greer, J. G. & Stuart, I. R., ed. (1983) The sexual aggressor: Current
perspectives on treatment. Van Nostrand Reinhold. [aRT]
Griffin, S. (1979) Rape: The power of consciousness. Harper & Row. [BMB]
Gross, M. (1985) Disruptive selection for alternative life histories in salmon.
Nature 313:47-48. [JA]
(in press) Evolutionary alternative reproductive tactics and strategies.
Behavioral Ecology. [rRT]
Groth, A. N. (1979) Men who rape: The psychology of the offender.
Plenum. [aRT]
Groth, A. N. & Burgess, A. W. (1977) Sexual dysfunction during rape. The
New England Journal of Medicine 297:764-66. [rRT, ERA]
Hager, B. (in press) Get thee to a nunnery: Female religious claustration in
Medieval Europe. Ethology and Sociobiology. [rRT]
Hall, J. D. (1983) The mind that burns in each body: Women, rape, and racial
violence. In: Powers of desire: The politics of sexuality, ed. A. Snitow, C.
Stansell & S. Thompson. Monthly Review Press. [BMB]
Hamilton, W. D. (1990) Mate choice near or far. American Zoologist
30(2):341-52. [rRT]
Hammerstein, P. & Parker, G. A. (1987) Sexual selection: Games between the
sexes. In: Sexual selection: Testing the alternatives, ed. J. W. Bradbury &
M. B. Andersson. John Wiley. [rRT]
Harding, C. F. (1985) Sociobiological hypotheses about rape: A critical look at
the data behind the hypotheses. In: Violence against women: A critique
of the sociobiology of rape, ed. S. R. Sunday & E. Tobach. Gordian
Press. [rRT, NJG, TL]
Harris, G. T., Rice, M. E., Quinsey, V. L., Chaplin, T. C. & Earls, C.
(submitted) Maximizing the discriminant validity of phallometric
data. [VLQ]
 References/Thornhill & Thornhill: Evolution of sexual coercion
Harvey, P. & Pagel, M. (1991) The comparative method in evolutionary
biology. Oxford University Press. [NJGJ
Hausfater, G. & Hrdy, S. (1984) Infanticide: Comparative and evolutionary
perspectives. Aldine. [RHB]
Horn, G. (1991) Cerebral function and behaviour investigated through a study
of filial imprinting. In: The development and integration of behaviour, ed.
P. Bateson. Cambridge University Press. [NJG]
Howard, J. C. (1991) Disease and evolution. Nature 352:565-67. [rRT]
Howells, K. (1981) Adult sexual interest in children: Considerations relevant
to theories of aetiology. In: Adult sexual interest in children, ed. M. Cook
& K. Howells. Academic Press. [aRT]
Hrdy, S. (1981) The woman that never evolved. Harvard University
Press. [rRT, RHB]
Humphrey, N. K. (1976) The social function of intellect. In: Growing points
of ethology, ed. P. P. G. Bateson & R. A. Hinde. Cambridge University
Press (Cambridge). [rRT]
Hunt, M. (1974) Sexual behavior in the 1970s. Playboy Press. [ERA]
Irons, W. (1990) Let's make our perspective broader rather than narrower.
Ethology and Sociobiology 11:361-74. [CTP]
(1991) How did morality evolve? Zygon 26(l):49-89. [rRT]
Johnston, T. D. (1987) The persistence of dichotomies in the study of
behavioral development. Developmental Review 7:149-82. [NJG]
Katz, J. (1988) Seductions of crime: Moral and sensual attractions in doing
evil. Basic Books. [AM]
Kelley, K. & Musialowski, D. (1986) Repeated exposure to sexually explicit
stimuli: Novelty, sex, and sexual attitudes. Archives of Sexual Behavior
15:487-98. [TL]
Kinsey, A. C , Pomeroy, W. B. & Martin, C. E. (1948) Sexual behavior in the
human male. W. B. Saunders. [aRT]
Kinsey, A. C , Pomeroy, W. B., Martin, C. E. & Gebhard, P. H.
(1953) Sexual behavior in the human female. W. B. Saunders. [aRT]
Kirkendall, L. A. (1961) Premarital intercourse and interpersonal
relationships. Julian Press. [aRT]
Kitcher, P. (1985) Vaulting ambition: Sociobiology and the quest for human
nature. MIT Press. [JD, PK]
Knight, L. J., Barbaree, H. E. & Boland, F. J. (1896) Alcohol and the
balanced-placebo design: The role of experimenter demands in
expectancy. Journal of Abnormal Psychology 95:335-40. [NMM]
Konner, M. (1984) The politics of biology. Natural History 93:66-69. [rRT]
Koss, M. P., Gidycz, C. A. & Wisnieski, N. (1987) The scope of rape:
Incidence and prevalence of sexual aggression and victimization in a
national sample of higher education students. Journal of Consulting and
Clinical Psychology 55:162-70. [ERA, NJG]
Lakatos, I. (1970) Falsification and the methodology of scientific research
programmes. In: Criticism and the growth of knowledge, ed. I. Lakatos
& A. Musgrave. Cambridge University Press. [AJF]
Langevin, R. (1983) Sexual strands: Understanding and treating sexual
anomalies in men. Erlbaum. [aRT]
Laws, D. R. & Osborn, C. A. (1983) How to build and operate a behavioral
laboratory to evaluate and treat sexual deviance. In: The sexual aggressor:
Current perspectives on treatment, ed. J. G. Greer & I. R. Stuart. Van
Nostrand Reinhold. [aRT]
Lewontin, R. C. (1983) The organism as the subject and object of evolution.
Science 118:65-82. [NJG]
(1984) Sociobiology: Another biological determinism. In: Biology as destiny:
Scientific fact or social bias, ed. Science for the People Study Group.
Science for the People. [AF]
Lewontin, R. C , Rose, S. & Kamin, L. J. (1984) Not in our genes: Biology,
ideology and human nature. Pantheon Books. [rRT]
Leyton, E. (1986) Compulsive killers: The story of modern multiple murders.
Washington News Books. [aRT]
Lickliter, R. & Berry T. D. (1990) The phylogeny fallacy: Developmental
psychology's misapplication of evolutionary theory. Developmental Review
10:348-64. [NJG]
Lockwood, D. (1980) Prison sexual violence. Elsevier. [aRT]
Loh, W. D. (1981) What has reform of rape legislation wrought? The Journal
of Social Issues 37(4): 28-52. [AF]
Low, B. (1989) Cross-cultural patterns in the training of children: An
evolutionary perspective. Journal of Comparative Psychology 103:311-
19. [aRT]
Maimonides, Moses (1135-1204) The hook of women (1972) Yale University
Press. [JH]
Malamuth, N. (1981a) Rape proclivity among males. Journal of Social Issues
37:138-57. [aRT]
(1981b) Rape fantasies as a function of exposure to violent-sexual stimuli.
Archives of Sexual Behavior 10:33-47. [arRT, SB, ERA, LM, NMM]
(1984) Aggression against women: Cultural and individual causes. In:
Pornography and sexual aggression, ed. N. M. Malamuth & E.
Donnerstein. Academic Press. [aRT]
(1986) Predictors of naturalistic sexual aggression. Journal of Personality and
Social Psychology 56:953-62. [aRT]
(1989) The attraction to sexual aggression scale: Part two. Journal of Sex
Research 26:324-54. [NMM]
Malamuth, N. M. & Briere, J. (1986) Sexual violence in the media: Indirect
effects on aggression against women. Journal of Social Issues 42:75-
92. [aRT]
Malamuth, N. M. & Check, J. (1980a) Penile tumescence and perceptual
responses to rape as a function of victim's perceived reactions. Journal of
Applied Social Psychology 10:528-47. [aRT, SB, LM]
(1980b) Sexual arousal to rape and consenting depictions: The importance of
the woman's arousal. Journal of Abnormal Psychology 89:763-66. [aRT,
NMM]
(1983) Sexual arousal to rape depictions: Individual differences. Journal of
Abnormal Psychology 92:55-67. [arRT, LM]
Malamuth, N. M., Check, J. & Briere, J. (1986) Sexual arousal in response to
aggression: Ideological, aggressive and sexual correlates. Journal of
Personality and Social Psychology 50:330-40. [arRT, SB, NMM]
Malamuth, N. & Donnerstein, E. (1984) Pornography and sexual aggression.
Academic Press. [ERA]
Malamuth, N. M., Feshbach, S. & Jaffe, Y. (1977) Sexual arousal and
aggression: Recent experiments and theoretical issues. Journal of Social
Issues 33:110-33. [NMM]
Malamuth, N. M., Haber, S. & Feshbach, S. (1980) Testing hypotheses
regarding rape: Exposure to sexual violence, sex differences, and the
"normality" of rape. Journal of Research in Personality 14:121-37. [aRT]
Malamuth, N. M., Heim, M. & Feshbach, S. (1980) The sexual
responsiveness of college students to rape depictions: Inhibitory and
disinhibitory effects. Journal of Personality and Social Psychology 38:399-
408. [arRT]
Masters, W. H. & Johnson, V. E. (1970) Human sexual inadequacy. Little,
Brown. [TL]
Maynard Smith, J. (1982) Evolution and the theory of games. Cambridge
University Press (Cambridge). [rRT, JA, PK]
Mayr, E. (1982) The growth of biological thought. Harvard University
Press. [RHB]
Mazur, A. & Lamb, T. A. (1980) Testosterone, status, and mood in human
males. Hormones and Behavior 14:236-46. [IE-E]
McDermott, M. J. (1979) Rape victimization in 26 American cities. National
crime survey victimization and attitude data. U. S. Government Printing
Office. [aRT, SB]
Meaney, M. J., Aitken, D. H., Berkel, C. V., Bhatnagar, S. & Sapolsky, R.
M. (1988) Effect of neonatal handling on age-related impairments
associated with the hippocampus. Science 239:766-68. [NJG]
Medea, A. & Thompson, K. (1974) Against rape. Farrar, Straus and
Giroux. [aRT]
Mitani, J. (1985) Mating behavior of male orangutans in the Kutai game
reserve, Indonesia. Animal Behavior 33:392-402. [SB]
Mock, D. W. & Fujioka, M. (1990) Monogamy and long-term pair bonding in
vertebrates. Trends in Ecology and Evolution 5:39-43. [JA]
Muehlenhard, C. L. & Cook, S. W. (1988) Men's self-reports of unwanted
sexual activity. Journal of Sex Research 24:58-72. [ERA]
Muehlenhard, C. L., Friedman, D. E. & Thomas, C. M. (1985) Is date rape
justifiable? The effects of dating activity, who initiated, who paid, and
men's attitudes toward women. Psychology of Women Quarterly 9:297-
310. [AF]
Muehlenhard, C. L. & Linton, M. A. (1987) Date rape and sexual
aggressiveness in dating situations: Incidence and risk factors. Journal of
Counseling Psychology 34:186-96. [AF]
M0ller, A. P. (1988) Female choice selects for male sexual tail ornaments in
the monogamous swallow. Nature 332:640-42. [rRT]
Neale, M. C , Rushton, J. P. & Fulker, D. W. (1986) Heritability of item
responses on the Eysenck personality questionnaire. Personality and
Individual Differences 7:771-79. [LM]
O'Sullivan, L. & Byers, S. (1990) Female use of sexual influence strategies in
dating relationships. Presented at the Society for the Scientific Study of
Sex. Minneapolis, MN, November (unpublished). [LM]
Oyama, S. (1985) The ontogeny of information: Developmental systems and
evolution. Cambridge University Press (Cambridge). [rRT, NJG]
(1989) Ontogeny and the central dogma: Do we need the concept of genetic
programming in order to have an evolutionary perspective? In: Systems
and development, ed. M. R. Gunnar & E. Thelen. Erlbaum. [NJG]
Palmer, C. T. (1988) Twelve reasons why rape is not sexually motivated: A
skeptical examination. Journal of Sex Research 25(4):512—30. [aRT,
CTP]
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 419
References/Thornhill & Thornhill: Evolution of sexual coercion
(1989a) Is rape a cultural universal? A re-examination of the ethnographic
evidence. Ethnology 28:1-16. [CTP]
(1989b) Rape in nonhuman animal species: Definitions, evidence, and
implications. Journal of Sex Research 26(3)-.355-74. [CTP]
(1991a) Demonstrating adaptiveness in Darwinian anthropology and
evolutionary psychology: The example of human rape. Presented at the
Human Behavior and Evolution Society, Hamilton, Ontario, Canada,
August 25. [LM, CTP]
(1991b) Human rape: Adaptation or by-product? Journal of Sex Research
28:365-86. [ERA, RHB, BAG, CTP]
Parker, G. A. (1970) Sperm competition and its evolutionary consequences in
the insects. Biological Reviews 45:525-67. [JA]
(1974) Courtship persistence and female-guarding as male time-investment
strategies. Behaviour 48:157-84. [JA]
(1979) Sexual selection and sexual conflict. In: Sexual selection and
reproductive competition in insects, ed. M. S. Blum & N. A. Blum.
Academic Press. [rRT]
(1984) Sperm competition and the evolution of animal mating strategies. In:
Sperm competition and the evolution of animal mating systems, ed. R. L.
Smith. Academic Press. [JA]
Priog-Good, M. A. & Stets, J. A., eds. (1989) Violence in dating relationships:
Emerging social issues. Praeger. [aRT]
Polk, K., Adler, C , Bazzmore, G., et al. (1981) Becoming adult: An analysis
of maturational development from age 16 to 30 of a cohort of young men.
Final report of the Marion County Youth Survey. University of
Oregon. [ERA]
President's Commission on Obscenity and Pornography (1970) The report of
the president's commission on obscenity and pornography. Bantam
Books. [aRT]
Quinsey, V. L. (1977) The assessment and treatment of child molesters: A
review. Canadian Psychological Review 18:204-20. [aRT]
(1984) Sexual aggression: Studies of offenders against women. In: Law and
mental health: International perspectives, 1, ed. D. Weisstub. Pergamon
Press. [VLQ]
Quinsey, V. L. & Chaplin, T. C. (1982) Penile responses to non-sexual
violence among rapists. Criminal Justice and Behavior 9:312-24. [aRT,
VLQ]
(1984) Stimulus control of rapists' and non-sex offenders' sexual arousal.
Behavioral Assessment 6:169-76. [aRT, VLQ]
(1988) Preventing faking in phallometric assessments of sexual preference.
In: Human sexual aggression: Current perspectives, ed. R. A. Prentky &
V. L. Quinsey. Annals of the New York Academy of Sciences,
528. [VLQ]
Quinsey, V. L., Chaplin, T. C. & Upfold, D. (1984) Sexual arousal to non-
sexual violence and masochistic themes among rapists and non-sex
offenders. Journal of Consulting and Clinical Psychology 52:651-
57. [arRT, ERA, TL, NMM, CTP, VLQ]
Quinsey, V. L., Chaplin, T. C. & Varney, G. (1981) A comparison of rapists'
and non-sex offenders' sexual preferences for mutually consenting sex,
rape, and physical abuse of women. Behavioral Assessment 3:127-
35. [arRT, VLQ]
Quinsey, V. L., Rice, M. E. & Harris, G. T. (submitted) Psychopathy, sexual
deviance, and recidivism among sex offenders released from a maximum
security psychiatric institution. [VLQ]
Reiss, I. A. (1990) An end to shame. Prometheus Books. [ERA]
Report of the U. S. Merit Systems Protection Board (1981) Sexual harassment
in the federal workplace: Is it a problem? U. S. Government Printing
Office. [aRT]
Rice, M. E., Harris, G. T. & Quinsey, V. L. (1990) A follow-up of rapists
assessed in a maximum security psychiatric facility. Journal of
Interpersonal Violence 5:435-48. [VLQ]
Rimmer, R. H. (1984) The X-rated videotape guide. Arlington House. [ERA]
Rosen, D. E. (1982) Teleostean interrelationships, morphological function and
evolutionary inference. American Zoologist 22:261-73. [NJG]
Rowe, D. C , Rodgers, J. L., Meseck-Bushey, S. & St. John, C. (1989) Sexual
behavior and non-sexual deviance: A sibling study of their relationship.
Developmental Psychology 25:61-69. [LM]
Russell, D. E. H. (1982) Rape in marriage. Macmillan. [arRT]
(1984) Sexual exploitation: Rape, child sexual abuse, and workplace
harassment. Sage. [aRT, NJG]
Russell, P. A. (1991) Evolutionary theories must fit the data better than other
theories. Behavioral and Brain Sciences 14:277-78. [PAR]
Saal, F. E., Johnson, C. B. & Walker, N. (1989) Friendly or sexy? It may
depend on whom you ask. Psychology of Women Quarterly 13:263-
76. [arRT, ERA]
Sanday, P. R. (1981) The socio-cultural context of rape: A cross-cultural study.
Journal of Social Issues 37(4):5-27. [rRT, JJF, AF, TL]
420 BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2
Scacco, A. M., Jr. (1975) Rape in prison. Charles C. Thomas. [aRT]
Schmidt, G. (1975) Male-female differences in sexual arousal and behavior.
Archives of Sexual Behavior 4:353-64. [aRT]
Schuster, P. & Sigmund, K. (1981) Coyness, philandering, and stable
strategies. Animal Behaviour 29:186-92. [PK]
Schwendinger, J. S. & Schwendinger, H. (1985) Homo economicus as the
rapist in sociobiology. In: Violence against women: A critique of the
sociobiology of rape, ed. S. R. Sunday & E. Tobach. Gordian
Press. [TL]
Shields, N. M. & Hanneke, C. R. (1983) Battered wives' reactions to marital
rape. In: The dark side of families, ed. D. Finkelhor, R. J. Gelles, G. T.
Hotaling & M. A. Straus. Sage. [aRT]
Shields, W. M. & Shields, L. M. (1983) Forcible rape: An evolutionary
perspective. Ethology and Sociobiology 4:115-36. [aRT, BAG]
Shields, W., Shields, L., Thornhill, R. & Thornhill, N. W. (1982) What
causes rape? A dissenting view. Science 82 December: 16. [rRT]
Smith, J. (1989) Misogynies. Faber & Faber. [BMB]
Smith, P. (1984). Play in animals and humans. Basil Blackwell. [RHB]
Smith, R. L. (1984) Human sperm competition. In: Sperm competition and
the evolution of animal mating systems, ed. R. L. Smith. Academic
Press. [aRT, JA]
Smith, V. (1990) Split affinities: The case of interraciaLrape. In: Conflicts in
feminism, ed. M. Hirsch & E. F. Keller. Routledge. [BMB]
Smithyman, S. D. (1979) Characteristics of "undetected" rapists. In:
Perspectives on victimology, ed. W. H. Parsonage. Sage. [ERA]
Smuts, B. (1987) Sexual competition and mate choice. In: Primate societies,
ed., D. Cheney, R. Seyfarth, R. Wrangham & T. T. Strahsake.
University of Chicago Press. [PK]
(1991; submitted) Male aggression against female primates: Evidence and
theoretical implications. In: Advances in the study of behavior, ed. P. J.
B. Slater, J. S. Rosenblatt, M. Milinski & C. T. Snowdon. Academic
Press. [BBS]
(1992) Male aggression against women: An evolutionary perspective. Human
Nature 3:1-44. [NMM, BS]
Soble, A. (1986) Pornography: Marxism, feminism, and the future of sexuality.
Yale University Press. [TL]
Stearns, S. C. (in press) Comparative and experimental approaches to the
evolutionary ecology of development. In: Evolution et development, ed.
R. Chaline & B. David. [NJG]
Stephens, D. & Krebs, J. R. (1986) Foraging theory. Princeton University
Press. [NJG]
Stock, W. (1982) The effect of violent pornography on women. Cited in
Allgeier, E. & Allgeier, A. (1984) Sexual interactions. D.C.
Heath. [RHB]
(1983) The effect of violent pornography on women. Paper presented at the
national meeting of The Society for the Scientific Study of Sex, Chicago,
IL, November. [rRT, ERA]
Stoltenberg, J. (1989) Refusing to be a man: Essays on sex and justice.
Breitenbush Books. [BMB]
Struckman-Johnson, C. (1988) Forced sex on dates: It happens to men, too.
Journal of Sex Research 24:234-41. [ERA]
Strum, S. (1988) Almost human. Random House. [PK]
Studd, M. V. & Gattiker, U. E. (in press) The evolutionary psychology of
sexual harrassment in organizations. Ethology and Sociobiology. [aRT]
Sue, D. (1979) Erotic fantasies of college students during coitus. Journal of
Sex Research 15:299-305. [ERA]
Sunday, S. R. (1985) Sociobiology theory and the violent abuse of women. In:
Violence against women: A critique of the sociobiology of rape, ed. S. R.
Sunday & E. Tobach. Gordian Press. [TL]
Sunday, S. R. & Tobach, E., eds. (1985) Violence against women: A critique
of the sociobiology of rape. Gordian Press. [rRT]
Symons, D. (1978) Play and aggression: A study of rhesus monkeys. Columbia
University Press. [RHB]
(1979) The evolution of human sexuality. Oxford University Press
(Oxford). [aRT, DP]
(1982) The woman that never existed. Quarterly Review of Biology 57:297-
300. [arRT]
(1987a) If we're all Darwinians, what's the fuss about? In: Sociobiology and
psychology: Ideas, issues and applications, ed. C. Crawford, M. Smitfe &
D. Krebs. Erlbaum. [arRT, LH, CTP]
(1987b) An evolutionary approach: Can Darwin's view of life shed light on
human sexuality? In: Theories of human sexuality, ed. J. H. Greer & W.
T. O'Donahue. Plenum Press. [arRT]
(1989) A critique of Darwinian anthropology. Ethology and Sociobiology
10:131-44. [arRT, CTP]
(1990) Adaptiveness and adaptation. Ethology and Sociobiology 11:427-
44. [CTP]
 References/Thornhill & Thornhill: Evolution of sexual coercion.
(in press) On the use and misuse of Darwinism in the study of human
behavior: In: The adapted mind: Evolutionary psychology and the
generation of culture, ed. J. Barkow, L. Cosmides & J. Tooby. Oxford
University Press (Oxford). [arRT]
Symons, D. & Ellis, B. (1989) Human male-female differences in sexual
desire. In: The sociobiology of sexual and reproductive strategies, ed. A.
E. Rasa, C. Vogel & E. Voland. Chapman Hall. [aRT]
Talmud, The Babylonian (1935) Nezikin, vol. 1, Baba Kamma 59a. Soncino
Press. [JH]
Thiessen, D. (1989) Hormone correlates of sexual aggression. In: Biosocial
criminology, ed. L. Ellis & H. Hoffman. Praeger. [aRT]
Thornhill, N. W. (1990) The comparative method of biology in the study of
societies of history. International Journal of Comparative Sociology 27:7-
28. [rRT]
(1991) An evolutionary analysis of rules regulating human inbreeding and
marriage. Behavioral and Brain Sciences 14:247—93. [rRT]
(in press) Human inbreeding. In: The natural history of inbreeding and
outbreeding: Theoretical and empirical perspectives, ed. N. W. Thornhill.
University of Chicago Press. [arRT]
Thornhill, N. W. & Thornhill, R. (1990a) Evolutionary analysis of
psychological pain of rape victims I: The effects of victim's age and
marital status. Ethology and Sociobiology 11:155-76. [rRT]
(1990b) Evolutionary analysis of psychological pain following rape II: The
effects of stranger, friend and family member offenders. Ethology and
Sociobiology 11:177-93. [rRT]
(1990c) Evolutionary analysis of psychological pain following rape III: The
effects of force and violence. Aggressive Behavior 16:297—320. [rRT]
(1991) Evolutionary analysis of psychological pain following rape IV: The
effect of the nature of the sex act. Journal of Comparative Psychology
105:243-52. [rRT]
Thornhill, R. (1980) Rape in Panorpa scorpionflies and a general rape
hypothesis. Animal Behavior 28:52-59. [rRT, CTP, VLQ]
(1983) Cryptic female choice and its implications in the soorpionfly
Harpobittacus nicriceps. American Naturalist 122(6):765—87. [rRT]
(1984a) Scientific methodology in entomology. Florida Entomologist 67:74-
96. [arRT]
(1984b) Alternative hypotheses for traits believed to have evolved by sperm
competition. In: Sperm competition and the evolution of animal mating
systems, ed. R. L. Smith. Academic Press. [aRT]
(1986) Relative parental contribution of the sexes to their offspring and the
operation of sexual selection. In: Evolution of animal behavior:
Palentological and field approaches, ed. M. Nitecki & J. Kitchell. Oxford
University Press (Oxford). [aRT]
(1990) The study of adaptation. In: Interpretation and explanation in the
study of animal behavior, vol. 2, ed. M. Bekoff & D. Jamieson.
West view Press. [arRT]
(1991, in press) Teleonomy and the study of sexual selection. Proceedings of
the 20th International Ornithological Congress. [rRT]
Thornhill, R. & Alcock, J. (1983) The evolution of insect mating systems.
. Harvard University Press. [arRT, RHB]
Thornhill, R. & Sauer, K. P. (1991) The notal organ of the scorpionfly
(Panorpa vulgaris): An adaptation to coerce mating duration. Behavioral
Ecology 2:156-64. [arRT]
Thornhill, R. & Thornhill, N. W. (1983) Human rape: An evolutionary
analysis. Ethology and Sociobiology 4:137-73. [aRT, BAG, CTP]
(1987) Human rape: The strengths of the evolutionary perspective. In:
Sociobiology and psychology: Ideas, issues and applications, ed. C.
Crawford, M. Smith & D. Krebs. Erlbaum. [aRT, LM, CTP]
(1989) The evolution of psychological pain. In: Sociobiology and the social
sciences, ed. R. Bell. Texas Technical University Press. [rRT, NJG]
(1991) Coercive sexuality of men: Is there psychological adaptation to rape?
In: Sexual coercion: A sourcebook on its nature, causes and prevention,
ed. E. Grauerholz & M. A. Koralewski. Lexington Books. [rRT]
Tieger, T. (1981) Self-rated likelihood of raping and the social perception of
rape. Journal of Research in Personality 15:147-58. [aRT]
Tinbergen, N. (1959) Einige Gedanken iiber "Beschwichtigungsgebarden."
Zeitschrift fur Tierpsychologie 16:651-65. [IE-E]
Tobach, E. & Rosoff, B. (1985). Preface. Violence against women: A critique
of the sociobiology of rape. Gordian Press. [rRT]
Tooby, J. & Cosmides, L. (1989) Evolutionary psychology and the generation
of culture, Part 1: Theoretical considerations. Ethology and Sociobiology
10(l-3):29-51. [arRT, CTP]
(1990a) The past explains the present: Emotional adaptations and the
structure of ancestral environments. Ethology and Sociobiology 11:375-
424. [CTP]
(1990b) On the universality of human nature and the uniqueness of the
individual: The role of genetics and adaptation. Journal of Personality
58:17-67. [arRT]
Tooby, J. & Devore, I. (1987) The reconstruction of hominid behavioral
evolution through strategic modeling. In: The evolution of human
behavior: Primate models, ed. W. G. Kinzey. SUNY Press. [arRT]
Townsend, J. M. (1987) Sex difference in sexuality among medical students:
Effects of increasing socio-economic status. Archives of Sexual Behavior
16:415-44. [aRT]
(1989) Male selection criteria: A pilot study. Ethology and Sociobiology
10:241-53. [aRT]
Townsend, J. M. & Levy, G. D. (1990) Effects of potential partners' physical
attractiveness and socioeconomic status on sexuality and partner selection.
Archives of Sexual Behavior 19:149-64. [ERA]
Trivers, R. L. (1972) Parental investment and sexual selection. In: Sexual
selection and the descent of man, 1871-1971, ed. B. Campbell.
Aldine. [aRT]
(1985) Social evolution. Benjamin/Cummings. [aRT]
Turke, P. W. (1990) Which humans behave adaptively, and why does it
matter? Ethology and Sociobiology 11:305-39. [CTP]
Van Rhijn, J. G. (1974) Behavioural dimorphism in male ruffs Philomachus
pugnax (L.). Behaviour 47:153-229. [JA]
Voland, E. & Engel, C. (1990) Female choice in humans: A conditional mate
choice strategy of the Krammhorn women (Germany 1720-1874).
Ethology 84:114-54. [aRT]
Wallen, K. (1989) Mate selection: Economics and affection. Behavioral and
Brain Sciences 12:37-38. [PAR]
Wasser, S. K. & Isenberg, D. Y. (1986) Reproductive failure among women:
Pathology or adaptation? Journal of Psychosomatic Obstetrics and
Gynaecology 5:153-75. [EV]
Watson, P. (1991) Multiple paternity as genetic bet-hedging in female sierra
dome spiders, Linyphia litigiosa (Linyphiidae). Animal Behaviour 41:343-
60. [rRT]
West-Eberhard, M. J. (1979) Sexual selection, social competition and
evolution. Proceedings of the American Philosophical Society 123:222-
34. [rRT]
Wickler, W. (1967) Socio-sexual signals and their intraspecific imitation among
primates. In: Primate ethology, ed. D. Morris. Weidenfeld &
Nicolson. [IE-E]
Williams, G. C. (1966) Adaptation and natural selection. Princeton University
Press. [arRT, CTP]
Wilson, M. & Daly, M. (1985) Competitiveness, risk taking, and violence:
The young male syndrome. Ethology and Sociobiology 6:59-73. [aRT]
Wilson, W. & Durrenberger, R. (1982) Comparison of rape and attempted
rape victims. Psychological Reports 50:198-99. [LM]
Zuckerman, M. (1971) Physiological measures of sexual arousal in the human.
Psychological Bulletin 75:297-329. [aRT]
BEHAVIORAL AND BRAIN SCIENCES (1992) 15:2 421