A Mind of Her Own: The Evolutionary Psychology of Women

  • 20 767 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up

A Mind of Her Own: The Evolutionary Psychology of Women

Mhoa01.fm Page i Monday, December 17, 2001 9:25 AM A mind of her own The evolutionary psychology of women Mhoa01.fm P

3,357 1,120 14MB

Pages 402 Page size 348 x 500 pts Year 2011

Report DMCA / Copyright


Recommend Papers

File loading please wait...
Citation preview

Mhoa01.fm Page i Monday, December 17, 2001 9:25 AM

A mind of her own The evolutionary psychology of women

Mhoa01.fm Page ii Monday, December 17, 2001 9:25 AM

This page intentionally left blank

Mhoa01.fm Page iii Monday, December 17, 2001 9:25 AM

A mind of her own The evolutionary psychology of women

  Psychology Department Durham University

Mhoa01.fm Page iv Monday, December 17, 2001 9:25 AM

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Bangkok Buenos Aires Cape Town Chennai Dar es Salaam Delhi Hong Kong Istanbul Karachi Kolkata Kuala Lumpur Madrid Melbourne Mexico City Mumbai Nairobi São Paulo Shanghai Taipei Tokyo Toronto and an associated company in Berlin Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © Oxford University Press, 2002 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2002 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose this same condition on any acquirer A catalogue record for this title is available from the British Library. Library of Congress Cataloging in Publication Data (Data available). ISBN 0 19 850498 5 (Hbk)

10 9









Typeset by Integra Software Services Pvt. Ltd., Pondicherry, India www.integra-india.com Printed in Great Britain on acid-free paper by Biddles Ltd. Guildford & King’s Lynn.

Mhoa01.fm Page v Monday, December 17, 2001 9:25 AM

To the Blackwell Ox crew—for keeping my feet on the ground and for reminding me that, although the unexamined life may not be worth living, the unlived life is not worth examining.

Mhoa01.fm Page vi Monday, December 17, 2001 9:25 AM

Acknowledgements I want to thank all those who have taken time away from their own work schedule to offer me debate, advice, instruction and correction on the issues in this book. These include: John Archer, Paul Brain, Jill Becker, Dean Hamer, Simon Hampton, Eric Keverne, Hilary Little, Dennis McGuire, Robert Plomin and Richard Trelfa. Needless to say, I have not always followed their advice to the letter and so any remaining errors are my fault and not theirs. Many exciting ideas and fruitful references also came from my third-year undergraduate students at Durham University. Most of all, thanks to Steven Muncer for cheerfully tolerating the unilateral and progressive narrowing of our conversational repertoire over the past two years and to my son Jamie who tried to compensate by widening my horizons to encompass mountain bikes and rugby.

Mhoa01.fm Page vii Monday, December 17, 2001 9:25 AM

Contents  The essential woman: Biophobia and the study of sex differences 1  Mothers matter most: Women and parental investment 34  High stakes and low risks: Women and aggression 64  Who does she think she is? Women and status 101  Like a sister: Women and friendship 137  But she that filches from me my good name: Women and competition 169  In the underworld: Women and crime 208  A coincidence of interests: Women and marriage 233  Counting the ways: The unique woman 269 References 311 Index 383

Mhoa01.fm Page viii Monday, December 17, 2001 9:25 AM

This page intentionally left blank

Mhoc01.fm Page 1 Friday, December 14, 2001 10:14 AM

 


The essential woman: Biophobia and the study of sex differences

In the past twenty years over 110,000 studies of women, gender and sex differences have appeared in academic journals. The questions that researchers have posed, the methods they have used and the recommendations they have made have been profoundly guided by the zeitgeist of the post-war years. In the West, incomes rose, educational opportunities increased and women began to discern their very unequal standing in the world of work, public achievement and recognition. These forces informed an implicit belief that society was perfectible and that we should aim to equalize the standing of women and men. Quite right too. But something else was going on—the political ideology that drove this laudable quest for social equality began to drive psychological theories too. The only acceptable account of sex differences was one which explicitly acknowledged the socially constructed, arbitrary and malleable nature of sex differences. Women’s studies became steeped in a politically-driven rejection of essentialism (the idea that the sexes differ at a fundamental psychological level) and committed on the one hand to social constructionism (there is no objective truth ‘out there’, only negotiable subjective representations) and on the other to extreme environmentalism (all sex differences result from factors external to the person). Neither road has taken us very far towards an accurate understanding of why men and women differ. Social constructionists effectively remove gender from the human mind and instead allow it to float freely in an insubstantial ether as a ‘social construction’ or an ‘emergent property’ or an ‘interpretative repertoire’. This is why it is possible to read statements such as the following written in all seriousness: ‘Gender distinctions as dichotomous categories are perpetrated and maintained by social mechanisms and are socially constructed’ (Epstein 1997). The prevailing dogma is that the distinction between men and women is a collective and tyrannical fiction. There are no real biological or psychological differences other than ones that we (arbitrarily?) construct through discourse. For these writers, the question of the causes of sex differences never rears its head because positivistic science (with its traditional obsession with causality) is disparaged as simply another rhetoric among many—and an outdated

Mhoc01.fm Page 2 Friday, December 14, 2001 10:14 AM

      one at that (Woolgar 1996). Humans are the sole focus of interest and any comparison between our behaviour and that of lower animals is unjustified, demeaning and reductionist. This is because humans have language, language enables discourse and it is through discourse that social reality, including gender, is constructed. (This is especially true of educated, middle-class Western humans judging by the disproportionate attention they receive.) The study of discourse is the study of implicit meaning and all meaning is subjective so there can be no one authoritative or ‘privileged’ reading of a text. Although social constructionists recognize the implications of this observation for their own analyses, they nevertheless ‘deconstruct’ (often in dense literary and psychoanalytic terms) the ways in which gender is created in social talk. To give a flavour of their approach to gender differences, I quote from one of the most frequently cited writers of this genre (Hollway 1984, pp. 227–8) Hence recurrent day-to-day practices and meanings through which they acquire their effectivity may contribute to the maintenance of gender difference (reproduction without a hyphen) or to its modification (the production of modified meanings of gender leading to changed practices) . . . I am interested in theorising the practices and meanings which re-produce gendered subjectivity (what psychologists would call gender identity). . . . Gender differentiated meanings (and thus the positions differentially available in discourse) account for the content of gender difference.

In this article, Hollway goes on to explain how different discourses about sexuality locate women and men in different positions relative to one another. She writes of the ‘discourse’ of the stronger male sex drive, the ‘discourse’ of the Madonna–whore distinction and the permissive ‘discourse’ which appeared to (but did not) liberate women’s sexuality. Now each of these topics is of some considerable interest to evolutionary psychology, as we shall see, but in that discipline rather than locating them as discursive fictions they are taken as answerable empirical hypotheses about which evolutionary theory makes clear predictions. Men’s sex drive should be stronger—and it is (Oliver and Hyde 1993). Women should experience a reputational cost if they gain a reputation for promiscuity—and they do (Cashdan 1996). Women should find casual sexual liaisons less satisfactory than men—and they do (Townsend et al. 1995). Nor are these findings exclusive to a single culture or language community—they exist independent of so-called ‘constitutive’ discourse. For social constructionists the key question of the origins of these discourses is strenuously avoided: But to assume the mechanical reproduction of discourse requires asking how it got to be like that in the first place. And that question is in danger of throwing theory back into answers according to the terms of biological, Oedipal or social and economic determinisms (Hollway 1984, pp. 238–9).

In short, better not to ask the question if you think you may not like the answer. But elsewhere in the social sciences some academics were indeed resorting to ‘social determinism’. Sex differences come from outside the child. Babies are not

Mhoc01.fm Page 3 Friday, December 14, 2001 10:14 AM

        born wanting to play football or dress dolls. These preferences are imposed by parents and by the media, and then encoded into children’s cognitive frameworks, magnifying and reifying the differences between masculine and feminine behaviour. Socialization explanations of sex differences are built on the foundation of the tabula rasa infant shaped, rewarded and punished until it conforms to societal demands for sex-appropriate behaviour. They took shape in the era of behaviourism and learning theory. The account was a simple one. Parents treat boys and girls differently, reinforcing the correct behaviour in each. Boys are encouraged to fight, climb trees and play football. Girls are forced to wear dresses, play with dolls and share. Despite the fall from grace of radical behaviourism, nobody seriously doubts that reinforcement can shape behaviour. The question was whether it was strong enough to account for the worldwide patterns of sex difference that we see. The ‘Baby X’ paradigm was hailed as conclusive evidence of socialization differences (e.g. Will et al. 1976). A six-month-old baby was wrapped in a blue or a pink blanket, identified as a boy or a girl and then handed to a woman who was asked to look after it for a few minutes. When told it was a girl, the women more often offered the infant a doll in preference to other toys. Surely this showed that parents treat infants differently as a function of their sex? But there was a problem. Despite many attempts to replicate the effect, it seemed even weaker than it had on first sight appeared (and recall the effect was found only for toy selection—there were no differences in social behaviour to the infant). It was certainly not strong enough to support the whole edifice of sex differences (Stern and Karraker 1989). And even if parents gave their children different toys, such a finding would be trivial unless it could be shown that the toys changed the child’s subsequent behaviour. But the real crunch came when Lytton and Romney (1991) collected 172 studies from around the world which had examined the way in which parents treat their sons and daughters. Considering them all together, the evidence for differential treatment was virtually nil. Parents did not differ in the amount of interaction with the child, the warmth they showed, their tendency to encourage either dependency or achievement, their restrictiveness, their use of discipline, their tendency to reason with the child or the amount of aggression that they tolerated. There was one area that showed a difference. Parents tended to give their children sex-appropriate toys. But sex-differentiated preferences for toys have been found in infants from nine months of age (Campbell et al. 2000). Children play more with sex-appropriate toys even when their parents do not specifically encourage them to do so (Caldera et al. 1989). It is quite likely that parents are not using toys to turn their children into gender conformists but are simply responding to the child’s own preferences. Anyway, if parents’ behaviour towards their children was being guided by their desire for them to conform to traditional gender stereotypes then we would expect to find that the most sex-typed adults have the most sex-typed children. Yet studies find that there is no relationship between traditional household division of labour, parents’ attitudes to sex-typing, their sex-typical activities and their reactions to

Mhoc01.fm Page 4 Friday, December 14, 2001 10:14 AM

      children’s behaviour on the one hand and children’s degree of sex-typing on the other (Maccoby 1998). Following Skinnerian views came social learning theory which emphasized a hitherto neglected (but altogether central primate) capacity—imitation. No-trial learning. We can acquire a piece of behaviour merely by watching it performed by others. But the trick was to co-opt this observation into an explanation of the acquisition of sex differences. This was done by proposing that children selectively imitate their samesex parent. Laboratory studies were done in which children were exposed to adult ‘models’ performing a variety of novel behaviours. If social learning theorists were right, then the statistical analysis would show a significant interaction between sexof-model and sex-of-child—girls would imitate women and boys would imitate men. Dozens of such studies failed to find such an effect (Huston 1983; Maccoby and Jacklin 1974). Undeterred, Perry and Bussey (1979) devised a cunning experiment that avoided the pitfalls of the previous studies where children had a one-off exposure to an adult model. They showed children a film of eight adults selecting a preferred fruit. In one condition all four men made one choice (e.g. orange) while all four women made another (e.g. apple). In another condition, three men and one woman chose an orange while three women and one man chose an apple. In another condition half the men chose oranges and half the women chose apples. They found that the extent to which children copied an adult preference depended upon the proportion of their sex that made that choice. In the first condition, there was a high degree of same-sex imitation, in the second a much smaller amount, and in the third condition, there was no significant difference between the girls and boys in their choices. What this meant was that children were not slavishly imitating a same-sex adult but rather judging the appropriateness of a particular (in this case wholly arbitrary) preference on the basis of the proportion of male or female adults who made it. These results helped to make sense of previous work which had already shown that children tended to imitate activities that they already knew to be sex-typed regardless of the sex of the model who was currently engaged in it (Barkley et al. 1977). What was important was the child’s internal working model of gender and behaviour. Until then, the differential treatment and selective imitation views had painted a thoroughly passive view of the child. There he or she sat, being slowly filled with sex-contingent (or, as it turned out, gender neutral) reinforcement and exposure to adult models. Some developmentalists rebelled. They knew that children are active participants in their own development, Piaget had shown this already. Now Perry and Bussey had put the child’s own understanding of gender centre stage. Martin and Halverson (1981) argued that children have a natural tendency to think categorically. They form categories about all sorts of things from animals to sports and it would be surprising if they did not, very early in life, form categories of male and female. Once these categories are formed, all incoming information that is gendertypical gets shunted into the correct binary slot and over time a stereotype is built up about what males and females look like, do and enjoy. It is this internal model or

Mhoc01.fm Page 5 Friday, December 14, 2001 10:14 AM

        schema, not the surveillance of parents, that drives the child towards sex-appropriate behaviour. It was clear that what Perry and Bussey had done, in their search for the mechanisms of imitation, was to lay bare the process of creating gender schema. At the very same time that this proposal was being offered for child development, Bem (1981) was proposing an identical scheme to explain adult differences in sex-typing. The degree to which we ‘type’ information as gender-relevant is an individual difference variable. Women who strongly sex-type information become more stereotypically feminine than women who are less inclined to tag information with gender labels. The cognitive revolution had come to sex differences—it was not a matter of behavioural training, it was a matter of mental categorizing, organizing and recalling. But the cracks soon began to appear. Children show sex-typed behaviour before they are able to label the sex of other children (Ruble and Martin 1998). Toy choice, play styles, activity levels, and aggression are found as early as two years of age (Brooks and Lewis 1974; Fagot 1991; Freedman 1974; Howes 1988; Kohnstamm 1989; O’Brien and Huston 1985; Roopnarine 1986) but children are not able to correctly sort pictures of boys and girls into piles until their third year (Weinraub et al. 1984). Although children can point to pictures of boys and girls when instructed to do so somewhat earlier at about 30 months (Etaugh et al. 1989; Fagot and Leinbach 1989), for a gender schema to operate spontaneously and successfully, children should be able to categorize without specific verbal cueing to do so. Children prefer sex-congruent toys before they are able to say whether the toy is more appropriate for a boy or a girl (Blakemore et al. 1979). They prefer to interact with members of their own sex and show sex differences in social behaviour before they can label either toys or behaviours as being more common among boys or girls (Serbin et al. 1994; Smetana and Letourneau 1984). Having gender labels at the age of two does not predict sex-typing either at the same age or one year later (Campbell et al. submitted). Even where a cross-sectional study does find a behaviour difference between children who can label and children who cannot, it is found for some behaviours not others or for one sex but not the other (Fagot et al. 1986). Children seem to need neither the ability to discriminate the sexes nor an understanding of gender stereotypic behaviour to show sex differences. Even in later years, as children’s gender stereotypes become more crystallized and peak at about 7 years of age, there is no relationship between a child’s gender knowledge and how sex-stereotypic their own behaviour is (Martin 1994; Powlishta 1995; Serbin et al. 1994). As Carol Martin (1993) ruefully concluded after twenty years of immersion in the field: ‘Seldom are individual differences in behavior and thinking explained by differing levels of gender stereotype knowledge.’ But perhaps children really recognize gender at a much earlier age than experimenters’ artificial request to point to pictures of boys and girls reveal. Perhaps they lack the verbal or cognitive skills to execute such a task until they are three. After all, animals seem to make no mistakes about the sex of their conspecifics and they lack the sophisticated cognitive machinery that we possess. Researchers turned to infants

Mhoc01.fm Page 6 Friday, December 14, 2001 10:14 AM

    

using an ingenious method of uncovering their ability to categorize the world. Infants, like adults, get bored when they are exposed for too long to the same thing and they turn away—a phenomenon called habituation. Leinbach and Fagot (1993) showed a group of infants aged between 9 and 12 months a series of photographs of different men’s or women’s faces. Every now and again a face of the opposite sex would be shown. The infants would show a sudden recovery of interest when this unexpected face appeared. This suggested that infants had an implicit category of male and female for, if they did not, how could they detect the category shift when the ‘unusual’ face appeared? This seemed to solve the cognitive problem—infants understand sex much earlier than we thought. But wait—the same type of study has also been performed in the laboratory using different categories such as animal species, rising and falling tones, numbers, colours and patterns (Bhatt and RoveeCollier 1996; Wagner et al. 1981; Xu and Carey 1996; Younger and Cohen 1983). All these studies show that infants can habituate and recover from habituation. Would we want to conclude, therefore, that 6 month-old infants brought to the laboratory with them an acquired understanding of the difference between a zebra and a kangaroo? Given their limited exposure to such novel stimuli, how could they? Rather, we infer that in the laboratory, infants develop (rather than reveal) categories for dividing up the world. So while infants can be experimentally primed to make a male–female distinction, this is no evidence that they have brought it with them from the outside world. Even if they did have it, it would be of no use to them unless they knew to which sex they themselves belonged. Gender schema can only guide behaviour when sex-of-self is incorporated into the schema. Children sort pictures of themselves correctly into the boy or girl pile at about the same time they sort pictures of others— around 36 months (Thompson 1975; Weinraub et al. 1984). Indeed they do not seem to even recognize themselves in mirrors until about 20 months (Amsterdam 1972). (This is tested by surreptitiously placing a blob of rouge onto the child’s nose and allowing them to view themselves in a mirror. If they try to wipe their own nose they have self-recognition. If they try to wipe the nose of the child in the mirror, they do not.) In any case, self-recognition is a necessary but far from sufficient condition for knowing one’s gender. Gender schema theory was all too cognitive and the cognitive data would not fit the behavioural time course. And there was the question too obvious to be asked— why do children choose to socialize themselves to behave sex-typically? Is the process one of simple social conformity? If so, is gender a special case or do children also categorize themselves in other ways (swats, athletes) and strive to conform to these categories as much as they do to boy or girl? There lurks beneath cognitive theories a pervasive feeling that there is a something different and special about gender and that the engine that drives categorization and conformity is an innate propulsion—perhaps not to be aggressive or nurturing—but at least to realize oneself as a male or female.

Mhoc01.fm Page 7 Friday, December 14, 2001 10:14 AM

      

At the heart of gender schema theory lay stereotypes. Initially very crude (boys like trucks, girls like dolls), they become increasingly complex with age (men are more competitive, women more cooperative). We construct them from the bits and pieces of observation that we can—from the media, from watching others, from gossip and myths. And it is stereotypes that form the foundation for another explanation of sex differences—social role theory. According to this formulation the division of labour in society, rather than the child’s natural tendency to form categories, is the starting point for sex differences. Men occupy roles that require competitiveness, autonomy and aggression. Women occupy roles that require nurturance, caring and cooperation. These roles draw out of their occupants the commensurate qualities and skills. These in turn set up stereotypes that embody beliefs in the appropriateness of expected characteristics. ‘Expectancy-confirming behavior should be especially common when expectancies are broadly shared in a society, as is the case for the expectancies about women and men’ (Eagly 1987, p. 15). These expectancies are internalized, resulting in sex differences in both behaviour and self-perception. During the past twenty years there has been a significant change in the nature of women’s labour as women move into many arenas traditionally occupied by men. We might therefore expect to see a shift in both stereotypes and self-perceptions by men and women. No such shift has occurred (Helmreich et al. 1982; Lewin and Tragos 1987; Lueptow 1985). Furthermore, we would expect to see a fair degree of cultural specificity with ‘traditional’ societies showing more marked stereotypes than more egalitarian ones. We do not (Williams and Best 1982). Social role theory supposes that sex differences are responsive to stereotypes and hence that stereotypes should be more extreme and polarized than actual sex differences. They are not (Swim 1994). We are left with the alternative suggestion that stereotypes are reasonably accurate assessments of the typical differences between men and women and that rather than stereotypes causing sex differences, the reverse is the case. If this is true then we at least have a means of explaining the typical division of labour between the sexes (women elect to spend more time than men do in parenting activities). As it stands, the cause of universal differences in child care remains completely opaque and Eagly (1987, p. 31) acknowledges—literally as a footnote—that there may be some biological factors involved. Nobody can seriously doubt that environmental factors modify the expression of sex differences. The foci of environmental theories—reinforcement, imitation, cognitive schema, conformity—all modulate our actions. The pleasure of social approval, the ability to learn through observation, and the desire to be like others are part of human psychology everywhere. The question is whether these processes alone can explain the origins of the cross-cultural differences between male and female. Altering reinforcement contingencies for sex-typical behaviour can change it: boys and girls will show cross-sex play where the environment is manipulated to encourage it and where social approval is contingent on it. But when that intervention is removed,

Mhoc01.fm Page 8 Friday, December 14, 2001 10:14 AM

      children revert to the same-sex preference that characterizes children everywhere (Serbin et al. 1977; Theokas et al. 1993). Cultural icons, especially teenage ones, can subtly and not so subtly alter our prevailing image of femininity. Demeanour and language that used to be frowned on in young women as ‘masculine’ is now unremarkable. But there is no link between girls’ approval of these new female behaviours and their level of aggression (Muncer et al. 2001) and as yet we have not seen any change in the universal tendency for men to be more violent than women. As new opportunities open to women, they eagerly accept them. Women’s performance in hitherto masculine areas of academic achievement (such as science and mathematics) and business entrepreneurship has been remarkable. Yet, for the majority of women occupational choice still rests as heavily on the social as the monetary rewards and the extent to which the work can be effectively combined with motherhood (Browne 1995; Geary 1998). When we can open up new opportunities for expression, enjoyment and achievement for women, we should do it because it is morally right. But that is very different from saying that gender has no biological basis and that the nature of men and women is wholly constructed by society. The problem with such a position is that it fails to address the issue of why sex differences take the particular form that they do. If gender differences are arbitrary, it is a curious coincidence that they follow such a similar pattern around the world. Even if sex differences were driven by differential parental treatment, we would still want to ask why a trait is considered more desirable for one sex than another. If they were driven by selective imitation, we would still want to ask why children might show an untutored interest in their own sex. If driven by gender schema, we would need to ask why sex-specific conformity is so attractive to children. If driven by the division of labour, we still need to explain the preference of men and women for agentic and expressive occupational roles. Social constructionist and environmental theories explain the transmission of the status quo—but without asking where it came from.

Evolutionary psychology Evolutionary theory addresses this very question. And the Darwinian algorithm is so elegant that it can be stated in five words: random genetic variation, non-random selection. Evolutionary psychology is the application of evolutionary principles to the study of the evolution of mind (Tooby and Cosmides 1992). Natural and sexual selection pressures which shaped species-typical aspects of our anatomy (bipedalism, cranial capacity, gestation length) are assumed to have orchestrated the architecture of the human mind which in turn drives behaviour. Evolutionary psychology holds that psychological attributes that conferred significant benefits in terms of survival and reproduction upon their bearer (relative to others who did not possess such attributes) are present today in the form of evolved modules designed to solve such specific ancestral problems as detecting social cheaters (Cosmides and Tooby 1992),

Mhoc01.fm Page 9 Friday, December 14, 2001 10:14 AM

      

enhancing paternal certainty (Wilson and Daly 1992), optimizing mate selection (Buss 1989), speedily acquiring language (Pinker 1994), comprehending the mental state of others (Baron-Cohen 1997) and weighting the costs of risky encounters (Campbell 1999). The distinguishing features of evolutionary psychology are fourfold. First, it is ultimately concerned with mechanisms of mind and not simply behaviour. This distinguishes it from sociobiology where comparisons are made between animal and human behaviours and implications are drawn about a common evolutionary pathway or about convergent evolution under similar selection pressures. Primate behaviour is often described and discussed by evolutionary psychologists (and I will be doing this too) because many human adaptations are shared with other species and emerged prior to human speciation (Foley 1996). Such behavioural comparisons are a starting point for attempting to locate the mental mechanisms which produce it. A good starting point for such an analysis is with a description of function—what does this behaviour achieve? To answer this we need a description of the circumstances under which the behaviour appears and whether or not it solves an adaptive problem. But evolutionary psychology also asks about the relevant inputs to the mental device and the range of outputs that can appear. This is important in understanding flexibility of action—how the life-stage and competencies of the organism together with perception of the past and current environment affect the strategy that is implemented. The same mechanisms can give rise to different manifest behaviours. Competition for resources, for example, can lead to combat, the formation of advantageous alliances, and to dispersion to new niches. The same mechanism can produce different manifest behaviours given different inputs; babies raised in China speak a different language from infants raised in England but that does not invalidate the existence of a universal mental device for acquiring the ability to produce the language heard in the community. We are searching for the deep structure not only of language but of other universal human actions including kin recognition, mate selection and sexual jealousy despite the fact that their behavioural expression may vary. Secondly, evolutionary psychology conceives of the mind as modular. We presume that the environment of adaptation presented similar classes of problem again and again, resulting in selection of those specific mental abilities that were advantageous in solving it. The presence of a predator produces activity in the fear-centre of the amygdala at a pre-conscious level that triggers alertness and evasion even before we have consciously registered exactly what the threat is. (The path to the sensory cortex is slower and more roundabout than the direct pathway to the amygdala.) The ability to detect fast-approaching objects on a collision course represented a sufficient threat to us in our evolutionary past that infants today will fall backwards when an object is made to ‘loom’ (by simply increasing its size) on a screen in front of them. This reflex was sufficiently useful as an adaptation that it is now hardwired. The mind is a collection of modules (many requiring environmental input,

Mhoc01.fm Page 10 Friday, December 14, 2001 10:14 AM

      like language) that reliably develop under a wide latitude of environments. Some are simple reflexes but many more are not. The ability to detect social cheaters requires conscious consideration of the information available but the mechanism that governs it works far more speedily and efficiently than it could by the application of formal logic (Cosmides 1989). Humans have a tendency to commit various cognitive ‘errors’ that have been successful rules of thumb in our evolutionary past. One is the availability heuristic—we judge the likelihood of events in terms of the ease with which we can recall instances of their occurrence. When asked whether accidents or cardiovascular disease accounts for more deaths in the United States, most people reply that accidents do. In fact, accidents account for 5 per cent of deaths annually compared to 50 per cent from heart attacks and strokes. Accidents are more vivid and memorable and their prominence in our memory misleads us (Tversky and Kahneman 1974). In ancestral communities only about half of infants born survived to adulthood and many of these deaths would have been traumatic. The ability to attend to and recall such lethal threats (and consequently to overestimate their frequency) had advantages. Our ability to attribute, with much greater than chance accuracy, beliefs and desires to other people’s actions conferred a special advantage when the main selection pressure was the behaviour of other people (Baron-Cohen 1997). Thirdly, evolutionary psychology does not conceive of the mind as a fitness-maximizing device—and it is here that it parts company with Darwinian anthropology (Tooby and Cosmides 1990). Anthropologists who share a Darwinian view focus on the way in which current human communities show evidence of optimality in their interaction with others and their environment. For example, optimal foraging theory is concerned with the net gain or loss in calories that are contingent on different organization of foraging. Their measure of fitness of a strategy is the extent to which it is the single most efficient means of calorie replenishment. (Notice also that the focus is on behaviour rather than on the mental module that organizes it.) The assumption is that humans do whatever they can to maximize their adaptation to the current environment and hence the usually unspoken presumption is that the mind is an all-purpose fitness maximizer. Somehow it reads the environmental problem accurately (food located at minimum two kilometres away), generates a number of possible solutions (the net utilities of various permutations of travelling alone, carrying the baby, leaving child at camp, travelling early before the sun gets hot, relying on left-overs from relatives) and adopts the most successful strategy. As I have noted, evolutionary psychology focuses on mechanisms not behaviour and on domain-specific adaptations not on an all-purpose adaptation generator. Just as important, evolutionary psychologists argue that present behaviour is a function of the past adaptive success of genetically-encoded mental modules. For them, there is no point in speaking of currently adaptive behaviour. Adaptiveness is a property of the past for that is where selection occurred. To know if a current behaviour is adaptive we would have to return in several hundred thousand years, find what

Mhoc01.fm Page 11 Friday, December 14, 2001 10:14 AM

      


traits had gone to fixation and trace the reproductive success of humans who had the rudimentary adaptations compared to those who did not. What we do know is that where we find a species-typical module (such as language), this mechanism was selected at some point in human history. Because the current environment differs from the one in which we evolved, it is quite possible that an adaptation is not currently adaptive. Our preference for fat and sugar was useful at a time when meat and berries were nutritious and rare. They are currently responsible for obesity and heart disease in an environment where sources are too plentiful. Indeed, our appetite for sugar is so strong that rather than simply refusing it, we go to extraordinary lengths to develop chemicals that mimic the taste while removing the calories. The question for any putative adaptation is ‘What did it do for us back then?’ Although we can and do surmise on the apparent mismatches between evolved adaptations and current environments (Crawford 1998), we cannot meaningfully speak of adaptations-in-the-making until the unknown future environment has a chance to make its genetic selection. Lastly, evolutionary psychology is chiefly concerned with species-typical adaptations. It seeks to explain the emotions, algorithms and strategies that are common and central to all human experience (even though their surface manifestation may vary from one culture to the next and they may only be activated given appropriate environmental input). This sets it apart from behaviour genetics. Behavioural geneticists are engaged in a statistical attempt to partition the variance between people with respect to a given psychological trait. Using adoption and twin studies, they attempt to fit mathematical models that distribute the differences between people to environmental and genetic sources. The whole enterprise depends essentially on the presence of variance. But for species-typical traits no genetic variance exists. Because we have all evolved to have one heart and two lungs then (genetic abnormality aside) there is no variability on this attribute. The trait has gone to fixation and falls out of the purview of behaviour genetics. The very existence of a heritable component for any trait tells us that it has not reached fixation and is not possessed uniformly by every human being. Evolutionary psychologists are not uninterested in variability (and in my final chapter I shall have more to say on this) but to see the long view of evolution we must dwell not on the noise but on the signal—those traits that were acted upon by selection to the point that they came to characterize the whole species. We need a crucial caveat, however, when we talk of universality. Humans come in two distinct morphs—women and men—differentiated by the size of the gametes that they contribute to sexual reproduction. The bulk of selection pressures— disease, predators, famine—affected both sexes equally and no sex differences are expected in psychological mechanisms that allow us to cope with these threats. The majority of traits that were advantageous were passed on by sexual recombination to both daughters and sons regardless of whether they were contributed by the mother or the father. (Later in the book this statement will have to be complicated by the discussion of genomic imprinting—a process by which the expression of some traits depends upon the parent that contributed them.) Where the sexes differ most

Mhoc01.fm Page 12 Friday, December 14, 2001 10:14 AM

      clearly it is the result of sexual, not natural, selection. The strategies that enhanced reproductive success for females were not identical to those that enhanced it in men. Through sex-linkage and sex-limitation evolution has coupled genetically encoded adaptive strategies to the sex of the individual receiving them. What evolutionary psychology offers is the hope of integration in the understanding of human behaviour. Using the most powerful theoretical development of the past one hundred years, we are finally able to address the ‘Why?’ question and to ask it together with other disciplines that have long ago accepted the premise of adaptation. Psychologists must work with and depend on other disciplines if the enterprise is to be successful. We need primatology to help us understand the common and unique paths of adaptation in the anthropoid line. We need paleoanthropology to track the evolving size and shape of the brain. We need archaeologists to describe the man-made tools and art that were part of the early emergence of Homo sapiens. We need anthropology to describe and document the varieties of human solutions to ecological and social problems. We need geneticists to map the genome and tie complex traits to their even more complex interacting genetic loci. We need biologists to identify the mutual paths between genes, hormones and environment. We need developmentalists to document the trajectory and constraints on the successful emergence of human capabilities. We need neuroscientists to identify the evolution and modification of structures that govern specific human emotions and actions. We need pharmacologists to help us understand the actions of neurotransmitters and their relationship to categories of experience. As psychologists, our contribution is to identify the characteristics and the parameters of the mind mechanisms that drive behaviour. It will be a long and cooperative undertaking if it is to be finally successful. The meteoric rise of evolutionary psychology over the past two decades has been impressive—but it has not gone unchallenged.

Bad politics? Sociobiology functions as a political theory and program. (Bleier 1984, p. 46) Evolutionary psychology is not only a new science, it is a vision of morality and social order, a guide to moral behaviour and policy agendas. (Nelkin 2000, p. 20) . . . the biological accounts of male–female difference and male dominance that have emerged since the mid-nineteenth century have merely used the language of science, rather than the language of religion, to rationalise and legitimise the sexual status quo. (Bem 1993, p. 6) Nevertheless, the social construction of the categories ‘woman’ and ‘man’ has been historically justified by reference to biological differences, and the modern tendency to provide essentialist and reductionist explanations which include the effects of genes and hormones can be viewed as a contemporary manifestation of this long-standing tradition. (Muldoon and Reilly 1998, p. 63) Sociobiologists are . . . constructing a framework of ideas about what is natural and what is not. Women who enter professions that are typical of men are therefore seen as unnatural

Mhoc01.fm Page 13 Friday, December 14, 2001 10:14 AM

      


and going against their biology; so too are men who take up professions using abilities considered typical of women. These ‘unnatural’ women and men are considered to threaten the fabric of society, as seen and maintained by those (scientists, politicians, business leaders and the general public) who see genes as paramount in causing sex differences in behaviour. (Rogers 1999, pp. 48–49) Ought science to be seen as truth-telling, or as politics by other means, or can it be both things at the same time? (Fausto-Sterling 1997, p. 58)

As the above quotes show, one line of attack has come from those who are more concerned with the political implications of evolutionary psychology than its truth value. As they see it, any attempt to identify a universal human nature and to posit a biological basis for it is equivalent to abandoning all hope of amelioration. For them, evolutionary psychology is about the maintenance of the status quo and the rejection of liberal progress. Many of the most vehement objections come from feminists who have been particularly offended by the proposal that universal sex differences may have a biological basis. But let’s unpack their arguments and take them one at a time. Charge 1: Evolutionary theory is biological determinism Evolutionary theory is certainly biological. It argues from the premise that the genes associated with phenotypic characteristics that increase survival and reproductive success will increase over generations. In rejecting Lamarck, it rejects the idea that individuals are capable of passing on by genetic means skills or abilities that they themselves have acquired through cultural transmission or trial-and-error learning in the course of their lifetime. For feminists, the real issue is whether there are other biological differences between males and females aside from the bodily changes that are triggered by the twenty-third chromosome pair. In short, are some genetically influenced traits sex-linked (carried on one of the sex-determining chromosomes) or sex-limited (carried on autosomes and triggered by the presence of male or female hormones)? It is widely agreed that there are two sexes possessing different reproductive organs and that the two sexes may also differ with regard to body form such as average differences in height, strength and fat distribution (Lewontin 1994). However, some writers do not even concede these facts—Muldoon and Reilly (1998, p. 55) believe that ‘the objectivity of ‘hard science’ in this area can be questioned, so much so that the biological definition of sex itself becomes untenable’. They suggest that there is no biological basis for our belief in male and female as ‘dichotomous, mutually exclusive categories’ (see also Bem 1993). Notwithstanding these authors’ uncertainty, most feminists broadly agree that there are two discriminable sexes. Indeed most feminists are willing to go as far as acknowledging that biological differences are the result of evolution—provided that biology stops at the neck (Bem 1993). Even though the brain is the most expensive organ in the human body in terms of calorie consumption, even though feminists accept that hominid brain size itself was a

Mhoc01.fm Page 14 Friday, December 14, 2001 10:14 AM

      result of natural selection and even though the production of the very hormones that orchestrate bodily differences originate in the brain, feminists reject the notion that evolution could have had an impact on the minds of the two sexes. Though successful reproduction is the reason for our existence today and though the sexes play vital and different roles in that process, feminists reject any notion that their minds may have been sculpted by thousands of years of evolution to set different goals or pursue different strategies. Most 5-year-old children also agree that people are either male or female. They will also tell you that boys are rougher and fight more than girls do. They are correct. The sex difference in physical aggression is evident in naturalistic studies of playgrounds, in experimental studies of undergraduates, in psychometric inventories and in criminal justice statistics (Eagly 1987; Hyde 1986; Kruttschnitt 1994). This sex difference is cross-cultural and trans-historical (Daly and Wilson 1988). There are no human societies in which women commit more violent crime than men. We also know that in animal species like our own, in which females provide the bulk of parental investment, the same sex differences exist (Geary 2000). We have seen too that children show sex differences in aggression before the age at which they can correctly label the sex of others or sort photographs correctly by sex. Together these facts strongly suggest a very fundamental difference between the sexes in aggression and one that, however biologically mediated, may be traced back ultimately to differential evolutionary pressures on the two sexes. It is important to bear in mind that evolutionary theory predicts sex differences in a few psychological domains only—those that are relevant to male and female roles in sexual reproduction. There would be no evolutionary reason to suppose that men and women should differ on average with regard to sociability, intelligence, sense of humour or openness to experience (to name but a few) and they do not. There is every reason to think that they should differ in nurturance, hostility and assertiveness, and they do (Feingold 1994). The ‘determinism issue’ relates to the erroneous belief that genes alone direct development and behaviour. Patently, this in not the case—although it is straw man that is frequently used to bait evolutionists. Take three examples culled from many dozen similar pronouncements: . . . sociobiologists argue that these strategies are given by biology and thus imply that they are eternally fixed features of human sexual relations. (Sayers 1982, p. 60) By reducing human behavior and complex social phenomena to genes and to inherited and programmed mechanisms of neuronal functioning, the message of the new Wilsonian Sociobiology becomes rapidly clear: we had best resign ourselves to the fact that the more unsavory aspects of human behavior, like wars, racism, and class struggle, are inevitable results of evolutionary adaptations based in our genes. (Bleier 1984, p. 15) [Genes] are therefore seen as the source of human behaviour, including sex differences in monogamy and polygamy, aggression and the perception of beauty. This is clearly a reductionist’s position. Of course, genes have a part in the development of sex

Mhoc01.fm Page 15 Friday, December 14, 2001 10:14 AM

      


differences and other behaviours, but it need not be any more important than other influences from both inside and outside the body, and the part played by genes might not be separable from these other influences. (Rogers 1999, p. 44)

Where are these evolutionary psychologists who allege that genes operate without reference to hormones, experience or environment? Wherever they are, I have not been able to locate them. Rather it is environmentalists that continue to set up mythical distinctions between nature and nurture in order to maintain a clear line between the politically correct and incorrect. As Margo Wilson and her co-workers (1997, p. 437) explain: ‘“Biology” is the study of the attributes of living things, and only living things can be “social”. So whence this idea of antithesis? . . . The irony is that developmentally, experientially and circumstantially contingent variation is precisely what evolution-minded theories of social phenomenon . . . are all about.’ Surely it is evident that ultimate success of a given individual in evolutionary terms depends upon manifest behaviour which in turn derives from a particular gene–environment complex. Smuts (1995) provides a good example of this. Cooper and Zubek (1958) took strains of maze-bright and maze-dull rats that had been so successfully bred that there was no overlap in their maze performance. They then reared a new generation under normal, enriched or impoverished conditions. All of the rats who were raised in the enriched environment performed as well as the maze-bright rats who were reared normally. All the rats raised under impoverished conditions performed as poorly as maze-dull rats reared normally. Behaviour depends upon the confluence of genetic disposition and environmental influences. The environment interacts with genetic predispositions in a variety of ways (Buss et al. 1998). Some environmental parameters are necessary for the emergence of adaptations; absence of contact with a language-using community can severely disrupt the development of language and early close contact with a reliable care-taker seems to be important for later social and emotional functioning. Early developmental events may channel individuals into different pathways by setting different expectations about the environment. Father-absent children are more inclined to pursue short-term mating patterns with the expectation that paternal investment is a statistically rare event. Families and communities characterized by high levels of competition and hostility set children on a more aggressive life strategy than those whose early experiences of others are more stable and cooperative. Other genetic tendencies are only expressed if the environment provides the necessary trigger— nicotine addiction has a high genetic component but whether it is activated depends upon environmental (and sometimes chance) factors such as exposure to peers who smoke. Environmental experiences can alter the expression of adaptations—sexual jealousy seems to be activated only when people experience a deep attachment to an exclusive romantic and sexual partner. The choice of one strategy rather than another depends upon a variety of contemporaneous factors such as the life-stage of the individual, the prevailing sex ratio, the number of alternative avenues of action

Mhoc01.fm Page 16 Friday, December 14, 2001 10:14 AM

      and her behavioural and psychological competencies. Humans are characterized by facultative responses to the demands of their environment. Differences between societies and between individuals are seen by evolutionary psychologists as environmentally induced variation in the expression of similar genotypes. In evolutionary theory, genetic predispositions orchestrate universal trends in human psychology and behaviour and thus create what is often called human nature. It also directs particular male and female natures in some psychological modalities. However, superimposed upon these, we must take into account the impact of the environment in both development (nutrition, education, opportunities congenial to the development of particular abilities) and facultative adaptation (the particular environmental constraints that inform decisions about behavioural strategies). Charge 2: Evolutionary theory is simplistic and reductionistic Evolution is the process of selection which causes differential survival and reproduction of individuals as a function of their performance in a particular ecological niche. I can think of few other theories that can be expressed so succinctly. Yet pejorative accusations of ‘simplistic’ notwithstanding, it is apparently not simple enough to escape misunderstanding. Fausto-Sterling (1992), a biologist and vocal critic of evolutionary psychology, correctly explains the evolutionary premise that the sex which makes the lower parental investment (typically males) tends to display greater promiscuity in its mating habits. She then points to female promiscuity in phalaropes (sea snipes), exclaiming, ‘You name your animal species and make your political point.’ The non-political point which she completely misses is that in the phalarope it is the male not the female that makes the greater parental investment. Hence this example is entirely consistent with evolutionary theory and merely demonstrates what the theory has always argued—it is parental investment not sex per se that drives mating strategy. The truly remarkable thing about evolution is that, although the theory itself is simple, it leads to highly varied and often counter-intuitive hypotheses. An evolutionary analysis of incest developed by Westermark argued that people develop an aversion to later sexual contact with those with whom that they spent their infancy and childhood years (normally siblings). One counter-intuitive prediction was that children raised in kibbutzim should avoid marriage with their kindergarten peers despite their non-relatedness. This prediction turns out to be true (Sheper 1983). A different strand of evolutionary thought has been concerned with homogamy (the tendency for like to mate with like). Together these two pieces of work can explain the recent reports that siblings separated at birth and then reunited in adulthood tend (to their distress) to find one another sexually attractive. This seems to me to be simplicity at its best—a simple theory that is able to explain apparently unrelated and unexpected findings in the real world. There was a time when simplicity used to be called elegance and constituted one of the criteria for quality—if the data equally support two theories then the simpler one was the better one.

Mhoc01.fm Page 17 Friday, December 14, 2001 10:14 AM

      


The charge of reductionism comes in two forms. The first objects to the ‘reduction’ of complex human behaviour to the action of genes and we have already discussed the fact that no serious evolutionary psychologist believes that genes can operate independently of the environment (although many feminists apparently fear that they may). The second highlights a failure to include the full range of variables that are needed to account for a given behaviour. No scientist really wants this contrived simplicity, any more than their critics do. We would all love to offer complete theories that fully account for the range and diversity of human behaviour. But reductionism is a necessary evil. It is a stepping-stone that allows us to work towards the truth by first decomposing the explanation into its constituent elements. If (as many feminists prescribe) we reject reductionism and take on the full complexity of a phenomenon as it appears in the real world, we are faced with an insurmountable problem. We can offer only a description and nothing more. We cannot generalize beyond the historical moment and the actors involved. Feminists, like Gestalt psychologists, argue that the whole is more than the sum of the parts. If we wanted to remove or introduce variables to observe their effect, we would have ‘committed reductionism’ in accepting the potential decomposability of the event. Now many feminists are happy to accept these limitations. But in so doing they become historians (not psychologists) describing (but not explaining) non-generalizable and unique events by the use of a subjective interpretation (that is itself the product of a particular moment in history, geography and culture). Charge 3: There are no human universals and hence no such thing as human nature The denial of human universals is central to the liberal agenda because of critics’ erroneous acceptance of the naturalistic fallacy and their mistaken belief that biology is destiny. If something is universal it may reflect fundamental human nature and if such a thing exists at a biological level then attempts to ameliorate the status quo are doomed. This shaky reasoning underpins the enormous kudos given to anthropologists who return with reports of novel and bizarre behaviour in exotic locations. Obvious hoaxes such as Carlos Castenada’s dissertation on Don Juan or the discovery of the Tasaday (who were inventions of the Marcos government) were eagerly and, for many years, uncritically embraced. An anthropological challenge to the traditional equation of masculinity with aggression and femininity with gentleness was also welcomed. Margaret Mead (1935) conveniently found, within a hundred mile area, three tribes in which these equations broke down; the Arapesh (both sexes gentle), Mundugumor (both sexes aggressive) and the Tschanbuli (sex role reversal). Since that time, her claims have been discredited by other researchers and these have been carefully reported by Freeman (1983). Her determination to demonstrate the existence of the three other logical permutations of sex and temperament lead to some strange interpretations of her observations. Though she argues that among the Mundugumor both sexes are violent, the men express it by murder, rape and

Mhoc01.fm Page 18 Friday, December 14, 2001 10:14 AM

      head-hunting raids while the women express it by serving tastier dishes to their husbands than their co-wives can. Among the allegedly gentle Arapesh, young men were not initiated into adulthood until they had committed homicide. Among the sex-role reversed Tschambuli, the make-up worn by men celebrates their killing of an enemy and the aggressive women were frequently beaten by their gentle husbands. Anthropology vocally encourages reports of cultural difference rather than cultural similarity (Geerz 1984) and this no doubt affects the way in which field workers interpret the behaviour that they witness. Critics often seem confused about just what is meant by human nature. Consider the following quote from Sandra Bem (1993, pp. 21–2): ‘As a biological species, human beings do not have wings, which once meant that it was part of universal human nature to be unable to fly. But now human beings have invented airplanes, which means that it is no longer part of universal human nature to be unable to fly.’ Now the idea that sitting in a seat a few thousand miles above the ground constitutes an alteration of ‘human nature’ is an odd distortion of the concept. The evolution of wings might have led to a radical alteration of human nature but we have not evolved them. Has air travel in any way led to an alteration in our physical morphology? Do we seriously suppose that aeroplane passengers show a different psychology or physiology to those who have not flown? It is equally hard to know what to make of Fausto-Sterling’s (1992, p. 199) claim that ‘there is no single undisputed claim about universal human behavior (sexual or otherwise)’. Presumably even the most ardent cultural relativist would accept that everywhere people live in societies, that they eat, sleep and make love and that women give birth and men do not. The problems seem to arise when we move from basic biological functions to behaviour. Though everywhere women are the principal care-takers of children, the fact that there may be variation in how that task is fulfilled leads some anthropologists to conclude that mothering is not universal. This is analogous to arguing that because people eat different food in different parts of the world, eating is not universal. Evolutionary psychologists do not argue for cultural invariance in the expression of evolved adaptations. As Tooby and Cosmides (1992, p. 45) put it: ‘manifest expressions may differ between individuals when different environmental inputs are operated on by the same procedures to produce different manifest outputs.’ At a behavioural level, the expression of the mechanism may vary but that does not question the universality of the generative mechanism itself. Fortunately, Donald Brown (1991), trained in the standard ethnographic tradition, has documented the extent of human universals. The list is astoundingly long but here is a taste of the hundreds that he finds: gossip, lying, verbal humour, storytelling, metaphor, distinction between mother and father, kinship categories, logical relations (not, same, equivalent, opposite), interpreting intention from behaviour and recognition of six basic emotions. Of special interest to the study of gender we find: binary distinctions between men and women, division of labour by sex, more child care by women, more aggression and violence by men, acknowledgement of differences

Mhoc01.fm Page 19 Friday, December 14, 2001 10:14 AM

      


between male and female natures and domination by men in the public political sphere. Now this latter observation is a nice example of the extreme reluctance of anthropologists to acknowledge universals. In 1973, Steven Goldberg wrote a book documenting the universality of patriarchy. He was inundated with letters informing him that he was wrong and pointing out counter-examples. (Other feminists were more willing to accept the premise, see Bem 1993; Millett 1969; Rich 1976.) Over the next twenty years he carefully examined the available ethnographic documentation for each example cited and in 1993 authored a second book in which he is emphatic that no society has yet been found that violates his rule. There are societies that are matrilineal and matrilocal and where women are accorded veneration and respect— but there are no societies which violate the universality of patriarchy defined as ‘a system of organisation . . . in which the overwhelming number of upper positions in hierarchies are occupied by males’ (Goldberg 1993, p. 14). Such a state of affairs is deplorable but mere denial of the facts will do nothing to alter it—women’s engagement in the political arena will. Charge 4: Evolutionary psychology is used to naturalize and legitimize the status quo Evolutionary psychology has considered a number of highly charged and socially relevant issues including infanticide, sociopathy, wife abuse and rape. In these areas, feminists (e.g. Fausto-Sterling 1992) have objected that such work opens the door to frivolous exonerating pleas in the criminal justice system (‘I beat my wife because it is man’s nature to experience extreme sexual jealousy because of internal fertilization’). The misuse of scientific research by defence attorneys is doubtless widespread but is certainly not confined to evolutionary theory. Contemporary work on neuroanatomy, hormones and clinical disorders can be misapplied in this way. Perhaps then we should abandon biologically based work which addresses behaviour that may be subject to criminal sanction? However, environmental factors are far more commonly misused. Parental abuse and neglect, inadequate educational opportunities and drug addiction have all been used to mitigate guilt and reduce sentences. The issue here is not evolutionary theory but the prevailing legal philosophy of responsibility and free will. This objection also takes the form that an evolutionary explanation of, for example, patriarchy will allow policy makers to view it as natural and therefore benign. It is deeply ironic that evolutionary psychologists have been the ones to argue most forcibly against the naturalistic fallacy—the belief that what is natural is morally right or desirable. This is not a post-hoc attempt to make their position socially acceptable but flows from the very nature of the theory itself. Natural selection operates as a sieve, allowing some variants in particular environments to pass through and others to die. This sieve knows nothing of good or bad, kind or cruel, desirable or unacceptable. It does not result in progress in the form of the survival of ‘better’ species or individuals. Humans have proliferated exceedingly successfully and one

Mhoc01.fm Page 20 Friday, December 14, 2001 10:14 AM

      feature which we possess is a large brain relative to our body size. That does not make high intelligence better than low intelligence, although it did make it a useful adaptation for Homo sapiens trying to survive in a particular ecology several hundred thousand years ago. If intelligence was ‘naturally’ good then ants and spiders would have developed enormous cortexes or would have become extinct—neither of which has occurred. It is vital to make a distinction between what we as humans hold to be morally desirable and what natural selection has retained as an adaptation. Malaria, tuberculosis and death in childbirth are natural but we hardly regard them as good. Indeed we have poured enormous (positivist) scientific effort into their eradication with considerable success. Our vision of a more perfect world cannot be found in evolutionary theory which can tell us only about our past, not our future. Charge 5: Culture and technology have so changed the environment that we are freed from natural selection Evolutionary psychologists hold that evolution has provided us with ‘fitness tokens’— desires, pleasures, aversions and goals that mould our preferences and habits. Teenagers do not attend discos in order to increase their share of the future gene pool but selecting the best possible mate is an activity that occurs across cultures and exerts an untutored fascination in most teenagers. A teenage couple who have sex after such an event are not seeking to increase their reproductive success—if they were, they would not employ contraception. Consensual sexual contact is inherently pleasurable (evolution has seen to that) and they are only enjoying that evolved pleasure. Natural selection worked to make sexual contact enjoyable because in ancestral environments contraception was not available and sex eventually led to pregnancy and increased reproductive success. The principal aim of evolutionary psychology is to understand the origins and the parameters of these mental adaptations. Such adaptations have not magically disappeared with the advent of new technology or social institutions in the past one hundred years. But, some argue, the world today looks nothing like the environment of evolutionary adaptation (EEA). Population density is greater. Social and political structures are more complex. We live in parts of the world that would have been uninhabitable then. International conflicts, air pollution and multinational companies dominate the world. At first glance, there is an enormous gulf between the hunter-gatherer societies in which humans spent 99 per cent of their existence and the contemporary societies we inhabit. But these changes have been wrought by the human mind and this mind was adapted in the EEA. As Crawford (1998, p. 291) points out, we have not created a world ‘designed for individuals who can fly, who choose mates indiscriminately, who have litters of offspring, who have fur to protect them from the cold, who make little investment in offspring once they are born, who do not mind being cheated in a social contract, who do not value close relatives, and so on’. At a social level, the size of intimate human groups has not substantially changed from the EEA (Dunbar 1993) and gossip is the major form of conversation with friends and relatives,

Mhoc01.fm Page 21 Friday, December 14, 2001 10:14 AM

      


as it is among hunter-gatherers. Even culture seems to support rather than undermine our evolved predispositions (Durham 1991). Cultures everywhere value altruism, reciprocity, fairness and in-group bias. Culture devalues selfishness, cheating, theft and in-group violence. Whole cultures which endorsed mass suicide or abstinence from sex would not last long. Cultural values are derived from adapted predispositions and because of this generally favour rather than oppose factors which in the past have enhanced survival. It is true, however, that adaptations now encounter a range of environments that are richer and more diverse that they were evolved to deal with. The stimulation of the dopaminergic reward pathways of the brain was a useful device for increasing appetitive behaviour but it can now be directly stimulated by illegal and addictive drugs. At the same time, some of our adaptations have been usefully recruited in the service of other unintended but beneficial outcomes. The pleasure that soccer brings to millions may derive from our natural fascination with ingroup–outgroup agonistic encounters, with male–male competition and with the demonstration of young male strength and agility. Our enjoyment of gossip is exploited in soap operas. Despite our knowledge that we are watching miniature representations of human beings (as pixels on an electronic machine) who are pretending to be people that they are not, we still find our emotions engaged in the drama of other people’s lives so artificially presented to us. The mind and body that we have today exist in their current form because we come from a long line of individuals who survived at least past puberty and successfully reproduced. Many did not survive. In Western society medical advances have led to the survival of severely premature infants and of individuals with diseases that would have been lethal in the EEA. We have seen significant increases in longevity in past decades. Infertility treatment has become widely available, as has contraception. It may well be the case that medicine, politics and personal choice have already begun to perform the roles of natural and sexual selection using criteria that are novel (the ability to pay) or no criteria at all (universal availability). Evolution is slow and we cannot know what the long-term effects will be.

Bad science? Another set of objections has come from those who question the scientific status of evolutionary psychology. We can begin with the most radical critiques from those who regard the whole scientific enterprise as an androcentric activity and those women who engage in it as ‘sleeping with the enemy’. Their wrath is not specifically directed at evolutionary psychology per se but at any use of the hypothetico-deductive method. Charge 6: Evolutionary theory, as part of traditional science, fails to recognize that there can be no objective truth Many feminists have objected that the very questions posed by scientists are laden with tacit political agendas and that the scientific method itself can never be value-free

Mhoc01.fm Page 22 Friday, December 14, 2001 10:14 AM

      (Fausto-Sterling 1992; Harding 1991; Hubbard 1990; Keller 1992). The solution they offer is for researchers to announce their politics at the same time as their results so that the reader may be aware of the possible bias of their data collection, analysis or interpretation. This has the side-effect of allowing the reader to pick and choose articles in terms of the author’s politics and to be prejudicially positive to articles that gel with their own agendas. Fausto-Sterling (1992, p. 212), for example, writes of the difficulty that she experiences in distinguishing between ‘science well done and science that is feminist’. She is also surprisingly honest about the double standard that she employs in evaluating data which are not congenial to her ideological position: ‘I impose the highest standards of proof, for example, on claims about biological inequality, my high standards stemming directly from my philosophical and political beliefs in equality’ (Fausto-Sterling 1992, pp. 11–12). Theories that are not consistent with a feminist viewpoint usually fail to achieve this higher standard. Feminists are keen to promote high-quality research—but this aim is made difficult by their inability to distinguish between feminist science and good science. Many feminist journals will refuse to publish data that are unacceptable to their ideological position. This state of affairs has already inhibited open debate among those who fear that they will incur feminist wrath and if it continues, it will seriously jeopardize academic freedom. There is a more liberal possibility. Traditional scientific method is a cyclical loop joining theory and data via hypotheses. It is a system that requires explicit and internally coherent reasoning at the level of theory-building and honesty and clarity about the method and results of data collection (but not about the ideological position of the authors). It also crucially depends upon the endeavours of many scientists working separately and together. The system allows freedom of politics to individuals, is self-correcting (unsupported hypotheses are modified or abandoned) and open (replications can be undertaken by detractors as well as supporters). Though hypotheses, or even whole theories, may be laden with tacit ideology, the data that they require for falsification are collected from ordinary people who have no axe to grind. Scientific method can be the antidote to hidden agendas. Charge 7: Evolutionary theory does not resonate with women’s experience Two strands of contemporary thinking have come together to create a new criterion for the acceptability of an explanation: it must reflect our lived experience and it must ‘feel’ like it is subjectively true. The first of these (which I wholeheartedly endorse) is the feminist demand that researchers should address women’s experiences, recognizing them to be potentially different from those of men. The second is the postmodern rejection of grand theory (feminist theory excepted) which emphasizes close qualitative description of experiences and discourse which are contextually and historically bound. This effectively replaces theory with subjectively interpreted description. Since there are multiple possible descriptions of any event and no objective criterion for deciding between them, the best one is the one that resonates with the feminist

Mhoc01.fm Page 23 Friday, December 14, 2001 10:14 AM

        reader’s own experience and intuition. (Parenthetically, it should be noted that the rejection of any objective set of truths necessarily excludes men’s oppression of women as a historical truth.) This emphasis upon personal resonance has been formalized and endorsed by writers who encourage women to transcend positivist empirical methods and to seek alternative ways of knowing that incorporate and are true to their own life experiences (Belenky et al. 1986). But this personalized and introverted approach to knowledge which places resonance with the reader’s experience above the truth, generality or coherence of the argument was taken to new heights by a book that recounts the biographical memories of five women. The authors proclaim that ‘The way forward is seen through the collective understanding opened to women by the collective method of memory-work. Such an approach allows us to evaluate our understanding in the inter-subjective realm, to explore resonances with the experience of other women.’ The other women of whom they speak are the five feminist psychologists who co-authored the book (Crawford et al. 1992, p. 14). Many women feel that evolutionary psychology does not resonate with their own experience. They do not ‘feel’ that parental investment impacts upon their emotions or behaviour. It does not ‘feel right’ to them that differences in gamete size could create differences between men and women’s behaviour or interests. However, intuition is not a reliable guide to the quality of an explanation. Children do not ‘feel’ as if their bodies are composed of 65 per cent water. People do not ‘feel’ that solid matter is made of atoms. ‘Feeling right’ is equally controversial with regard to psychological truth. Therapists can persuade clients that they are secretly in love with their father. Palm readers can persuade clients that they are destined for greatness. In an influential article, Nisbett and Wilson (1977) demonstrated that participants in psychology experiments perform no better than chance in guessing the causes of their behaviour when these are manipulated in experiments. Although we have conscious access to the products of lower-order processes (recalling our mother’s name or the ability to speak in grammatical sentences), we do not have access to the processes themselves (we do not know how our memory or language production systems work) except by the formal knowledge gleaned by psychological science. Evolutionary theory is concerned with how the mind was shaped in the environment of evolutionary adaptation approximately 100,000 years ago. We cannot have any ‘feel’ for what the selection processes were or even what it felt to be a proto-human at that time. What are available to us are the psychological products of evolution. We like the taste of sugar, feel sick in the first trimester of pregnancy, experience fear when we look out from a tall building, see three-dimensional images in stereograms, enjoy sexual contact and feel anger when our child is threatened. Evolutionary explanations have been offered for all these experiences. There are a number of feminists working in the evolutionary sciences. They have misgivings about the way in which women have been systematically excluded in their discipline both as objects of study and as collaborators (see Hager 1997). However,

Mhoc01.fm Page 24 Friday, December 14, 2001 10:14 AM

      few evolutionary feminists want to see the abandonment of empirical techniques, as do radical feminists in other branches of psychology. Instead, they seek to identify the evolutionary problems faced by women but ignored by male researchers, such as the evolutionary basis of patriarchy (Hrdy 1997; Smuts 1995) or male infanticide of infants (Hrdy 1981). Others highlight the selection pressures that operated specifically upon women, such as the care of slow-maturing, altricial young (Lancaster 1991) or the infant’s greater need for maternal rather than paternal survival (Campbell 1999). They believe that their work is important because it can illuminate the evolutionary basis of sexual inequality and in so doing complement mainstream feminist work and inform a more solidly grounded agenda for change. Evolutionary feminist research on wife abuse and male proprietary jealousy; on clitoridectomy, infibulation and claustration; and on polygyny and the gendered inheritance of wealth are a few examples of how evolutionary theory and feminism can work together to improve the status of women. Other critics of evolutionary theory, far from wishing to see the abandonment of scientific method, argue that evolutionary psychology is not scientific enough. Charge 8: Evolutionary theory is tautological In principal, this objection can be levelled at any functionalist explanation in the social or natural sciences. Some feminists argue that (male-dominated) governments allow the production of pornography in order to objectify women and to keep them in a state of fear about the possibility of rape. This is a functionalist explanation. It seeks to explain why the status quo is the way it is. The simple question ‘What is this for?’ is in fact the driving force of most scientific enquiry that goes beyond description or classification. When directed at evolutionary theory, the objection takes the form of allegations of the Just-So Story; the invention of plausible but unverifiable stories about how an adaptation occurred (Bleier 1984). In fact the methods of evolutionary psychology have been clearly explicated by Tooby and Cosmides (1992) and others (Halcomb 1998; Sherman and Reeve 1997). Broadly, evolutionary psychology proceeds by (1) identifying an adaptive problem that proto-humans would have faced in the environment of evolutionary adaptation, (2) developing a description of the module that is suggested to have evolved in response to this problem, including the range of inputs that would have activated it and the impact of its outputs in terms of differential survival and reproductive success, (3) formulating a description of the current environment and a map of correspondence between the ancestral and present conditions that allows specific hypotheses to be generated about the current activating inputs, and (4) undertaking tests of these hypotheses which, where appropriate, allow comparison with alternative (evolutionary or non-evolutionary) accounts. An adaptation is identifiable by evidence of special design such as complexity, economy, efficiency, reliability, precision and functionality (Williams 1966). The eye, for example, is so manifestly and complexly suited to the function of tracking the

Mhoc01.fm Page 25 Friday, December 14, 2001 10:14 AM

        environment and allowing the animal to orient itself that nobody could seriously doubt that it evolved to serve that function (Dawkins 1986). The ability to interpret the mental states of others also evidences special design—it is fast, accurate, automatic, follows complex rules, rarely makes mistakes and bestows obvious advantages on its bearers. Language also has these trademark features. But to make the techniques of determining an adaptation more concrete, let us take a different example. Let us posit for a moment that individuals prefer to avoid, rather than seek out, situations in which they are cheated or exploited. Universality is taken to suggest (but not prove) that the phenomenon may have been an evolutionary basis. The next step is to ask what purpose it might have served in the EEA about 300,000 years ago. The fossil record provides information about where we were living (on the savannah plains of sub-Saharan Africa) and roughly how (in small hunter-gatherer bands). We were highly social and survival depended as much on regulating social relationships as on ecological pressures. We can posit that it would have been important to detect social cheaters—those who failed to repay altruistic acts. If we could not do so, we would find our generosity exploited in the service of someone else’s reproductive success. Let us hypothesize that one module of the evolving mind might have been devoted to detecting cheats. Cosmides and Tooby (1992) made such an assumption and tested it on the Wason card task that goes like this. A person is presented with four cards; on one side is a number and on the other is a letter. The upturned cards read D F 3 7. They are then given the proposition ‘If a card has D on one side it must have 3 on the other’ and the person is asked which cards they need to turn over to determine whether this proposition is correct. This is a standard P not Q problem. They ought to turn over D and then a second card bearing a number that is not 3. Most people find this task quite difficult. But then Cosmides changed the task slightly. The subject is presented with a cheater detection problem. They are told that they are a barman and that they must ensure that nobody under the age of 18 is drinking alcohol. The four cards before them say Drinking Cola, Drinking Beer, Age 26 and Age 16 and they are again asked which cards they need to turn over. Most people rightly realize that if the person is 26 years old it does not matter what they are drinking and ignore this card. They also rightly realize that if a person is drinking cola, it also does not matter what age they are. Correctly, they turn two cards; Age 16 and Drinking Beer. The identical problem when translated into cheat detection becomes easy. And it is not just the fact that problem was made social rather than abstract. The cheater problem also seems to be activated only when a rule of exchange has been broken. If we change it to ‘If a person eats chilli peppers then they drink beer’, people seem as flummoxed as they were before because although the problem is still social, no rule violation is involved. As evolutionary theory builds an ever more complete picture of the underlying principles, these same principles can be used to generate further explanations and accompanying hypotheses. For example, Trivers’ (1972) key articulation of par-

Mhoc01.fm Page 26 Friday, December 14, 2001 10:14 AM


    

ental investment theory pointed to the differential investment made by males and females in offspring and the dangers posed to males by concealed ovulation and internal fertilization. In consequence, males should be especially sensitive to signs of sexual infidelity and females to the threat of lost resources resulting from male abandonment. Over a dozen studies in a variety of cultures (including sexually liberated ones) now show that this prediction is correct. There may be some other theory that can also explain this effect, but if this is the case it is incumbent on those who propose it to demonstrate that it does. Once a rival explanation is proposed, it becomes possible to generate decisive empirical tests between the competing theories. The EEA has become a particular source of concern to some. We have no time machine to travel to it and we must rely on an archaeological and paleoanthropological reconstruction. Because we believe that savannah living was the lifestyle that shaped our species, there is a tendency to caricature the EEA as something akin to an African hunter-gatherer community. This can be misleading if we fail to recall that evolution is deployed over vast stretches of time and that characteristics evolved at different time periods and conceivably in different ecologies. For example, humans’ tendency to live in kin groups can be traced back over 25 million years to a common ancestor shared with other anthropoids, meat-eating may have evolved about 2 million years ago while the evolution of language is traced to a mere 300,000 to 100,000 years ago (Foley 1996). There were a variety of adaptively relevant environments. But every adaptation comes from a statistical aggregate of selection pressures spread over time—the environment presented the same problems over and over again with sufficient regularity to make a problem-solving adaptation a distinct advantage. The uncertainty that must surround the details of life at that time of each adaptation does not preclude the generating and testing of adaptational hypothesis. If they evolved, they are part of human nature and will manifest themselves today over a wide latitude of cultures given the appropriate environmental inputs. Some have argued that the study of contemporary human adaptations will help to build a more accurate picture of the EEA (Tooby and Cosmides 1990). Even if it does not, the status of current adaptations is not dependent on a complete account of their genesis. It is dependent on their ability to offer elegant and empirically supported explanations of contemporary human behaviour. Adaptations exist as part of our nature—from the infant’s Moro reflex to our fear of heights—and a psychology that denies the available evidence in favour of a wishful world where humans are formless clay waiting to be moulded by cultural practices may serve politics but not truth. Tests of evolutionary predictions, however, will never satisfy critics if they insist on moving the evidentiary goalposts. The standards of proof that are required from evolutionary theory are far higher than those demanded for social environmental theories. The interpretative constructionism espoused by many feminists is subject to the single standard of ‘plausibility’. Ironically, this is the very standard that feminists

Mhoc01.fm Page 27 Friday, December 14, 2001 10:14 AM

      


reluctantly agree that evolutionary psychology meets. Fausto-Sterling (1992, p. 187) admits that ‘At one level it [evolutionary theory] all seems quite plausible’ and Bem (1993, p. 30) acknowledges ‘the consistent pattern of sexual difference and dominance that not only appears to exist across time and place but that a theory like sociobiology appears to so elegantly explain’.

Charge 9: Evolutionary theory is riddled with disagreements The ‘Modern Synthesis’ brought together Darwin’s nineteenth-century insights about natural and selection with our twentieth-century understanding of genetic transmission. Together they have inspired an ever-increasing body of empirical work that seeks to test and refine these ideas. In so doing, the modern synthesis has brought together biologists, anthropologists, geneticists, developmental and social psychologists and neuroscientists who, for the first time, have a common language with which to communicate. Notwithstanding this remarkable integration, Fausto-Sterling (1992, p. 169) asserts that ‘arguments continue to rage’ within evolutionary theory. The arguments to which she refers are those offered by Steven Jay Gould, an ardent Darwinian, which have been greeted with delight by many critics of evolutionary theory. His critiques are not rejections of adaptationist thinking but proposed modifications and admonishments of caution to his fellow Darwinians. First, a possible modification—Gould believes that the trajectory of evolution is discontinuous rather than smooth (see Eldredge and Gould 1972). He believes that the normal equilibrium of evolving life is punctuated by dramatic events that select for new macromutations before the pace slows again. His claim is far from radical— there is no reason why evolution should proceed at a constant pace and periods of stasis, when species were freed from dramatic environmental change, were probably commonplace (Dawkins 1986; Dennett 1995). Failure to find evidence of intermediate forms of life that evolved and were extinguished in a matter of a few hundred thousand years is not especially surprising. For a species to be identified as a unitary entity some degree of stasis has to be involved—we would not even entertain the idea of a species whether alive or extinct unless all of its members shared common attributes over some space of time. Whether the emergence of new life-forms happens ‘momentarily’ as he proposes largely depends upon one’s definition of a moment. The punctuated equilibrium argument depends upon the scale on which one draws evolutionary change. Gould himself appears to recognize this and agrees that ‘Our theory entails no new or violent mechanism, but only represents the proper scaling of ordinary events into the vastness of geological time’ (Gould 1992, p. 12). In this matter he hardly seems to part company from Darwin himself who wrote: ‘the long periods, during which species have undergone modification, though long as measured by years, have probably been short in comparison with the periods during which they retain the same form’ (see Dennett 1995, p. 290). So, for many evolutionists, Gould’s thesis hardly constitutes a major threat to Darwinian ideas.

Mhoc01.fm Page 28 Friday, December 14, 2001 10:14 AM

      Gould also argues that chance can be an important component of evolution—and again this is far from heretical. It is universally acknowledged that evolution (defined as a change over time in the relative frequency of genotypes) can be the result of things other than natural or sexual selection (Majerus et al. 1996). However, only these latter forces can produce adaptations—by which we mean a feature of design that becomes common by the differential success of phenotypic variants in previous generations. Mutation, migration and differential mortality due to chance events can all change the gene pool but not adaptively. Although random catastrophic events may cause the extinction of a whole species (for example, dinosaurs from a meteor strike) they do not cause systematic changes in the gene pool of an existing species. This is because such events, by their definition and nature, are random. Every year people die as a result of being struck by lightening. The people who die carry genes that are effectively wiped out (if they are childless). But the genes that are eliminated are a random selection of genes. Chance does not systematically retain and reject different variants. Some critics appear to believe that the mere existence of noise or chance is a serious challenge to the whole theory. For example, FaustoSterling offers a hypothetical example of an island of birds composed of blue and speckled variants in which the speckled variants are blown onto a neighbouring deserted island. She points out that this is an example of evolution by ‘a chance natural event, not natural selection’ (Fausto-Sterling 1992, p. 172). But the example is flawed because the probability of the wind carrying only one colour of bird by chance is vanishingly small. A chance event would be unselective about what colour birds it affected. When a natural event (wind) selectively affects one variant (speckled birds) while not affecting the other (blue birds) then it is most likely for a reason—perhaps the speckled birds weighed less than the blue ones. This made them more likely to be blown away and would constitute a clear case of natural selection. To give another example, imagine a community of people who vary genetically in their tendency to store fat. A famine occurs and those with low fat reserves perish before the rains come. The genes for low fat reserves will be selectively culled. Although the famine was a chance event, its effect was systematic selection. Chance events that do not distinguish their targets can have evolutionary effects, notably the extinction of whole communities and species, but only chance events that have systematic effects on different variants can generate adaptations. Gould and Lewontin (1979) coined the term ‘spandrel’ to describe epiphenomenal aspects of natural selection. They borrowed the term from architecture. In cathedrals and churches, spandrels are typically ornately decorated and give the appearance of having been put there specifically for this aesthetic purpose. But Gould and Lewontin argue that they are simply a by-product of the design—they had to exist as soon as the architect decided to join two rounded arches at right angles and thereby created a tapered triangular space. Their point was that we must not assume everything in nature to be a functional adaptation. Some apparent design features are merely side-effects of selection for something quite different. Bones are white and

Mhoc01.fm Page 29 Friday, December 14, 2001 10:14 AM

        we might be tempted to pose the question ‘What is the evolutionary advantage of white bones?’ if it were not for the fact that bones just happen to be made out of calcium which is white in colour. Blue or purple bones could have done the job just as well and the colour is of no evolutionary significance (although bones are). Because most genes are pleiotropic (they have multiple effects), some natural phenomena are simply side-effects that have survived because they have tagged on to an adaptive gene complex and, presumably, were not so detrimental as to outweigh the beneficial impact of its other phenotypic effects. Dennett (1995) argues that Gould is wrong—bones could not be made of just any old material. There were constraints on what material could be chosen (it had to be something that could be biologically manufactured, it had to be strong, it had to be attachable to tendons, etc.). Such constraints severely limit the material that could have been used—in fact they limit it to such an extent that calcium was the only one that would do. If we accept that evolution places constraints on possibilities, then white bones are in fact adaptations as much as anything else in the sense that calcium was the best available building material that the body had to hand. But if Gould’s fundamental message is simply that not everything is an adaptation then this hardly threatens evolutionary theory. Daisies float in water but no sensible person would ask what the adaptive significance of their buoyancy is. As Dennett puts it: The thesis that every property of every feature of everything in the living world is an adaptation is not a thesis anybody has ever taken seriously, or implied by what anybody has taken seriously, so far as I know. If I am wrong, there are some serious loonies out there, but Gould has never shown us one. (Dennett 1995, p. 276)

Gould (1991) has also introduced another term—the exaptation—to describe ‘any organ not evolved under natural selection for its current use—either because it performed a different function in ancestors (classical preadaptation) or because it represented a nonfunctional part available for later co-optation’. His example is bird feathers that originally evolved to conserve heat but were later exapted for use in flying. This example highlights the co-option of a functional feature from one role to another but his definition also includes the co-option of spandrels (non-functional features). In one sense it is misleading to coin a new term for this process because all current adaptations arose from previous states of the organism and genetic mutation can only use available components to engineer change whether they were functional in some other realm or not. Feathers happened to be available and, along with a host of other mutational changes, were exploited for flying purposes. But Gould’s writing has had the effect of generating confusion about the agency that is responsible for co-option (Buss et al. 1998; Pinker 1997). In places, he implies that natural selection does the work, while in others he seems to suggest that it is the human mind. If he means the former then there is nothing very new being added to the concept of adaptation. But if he means the latter, then we are talking about a different process. Reading and writing are human abilities that are too recent to have evolved by natural

Mhoc01.fm Page 30 Friday, December 14, 2001 10:14 AM

      selection and may well represent exaptations of existing human abilities (language, symbolic communication, manual dexterity, representational thought). The human ability to creatively exploit our mental and physical capacities in new ways deserves study in its own right but cannot (yet) be examined in terms of natural selection. There we have it. The pace of evolution, the role of chance and the status of spandrels and exaptations constitute the full extent of the ‘raging disagreements’ in evolutionary thought. Let us now turn to the unity of feminist theory. Though the term is widely used, it is hard to pin down a definition. The nearest I have been able to find is this: ‘Feminist theorists are concerned with how gender (which is the social construction of characteristics associated with sex) affects individuals’ access to control of their own and other people’s lives, power, and resources’ (Gowaty 1992, p. 218). This certainly defines the subject-matter of the discipline, but where is the theory? Theories are usually taken to be higher-level explanations from which local hypotheses can be drawn, but explanation is absent from this definition. Even a more direct statement of feminist theory such as ‘Women have been oppressed by men’ is essentially a statement of historical fact rather than an explanation. The absence of agreement about just what constitutes feminist theory is perhaps not surprising when its very proponents acknowledge that ‘There seem to be many kinds and varieties of feminism, as many kinds and varieties as there are individual feminists, with individual desires, notions and conceptions of what we are and want’ (Gowaty 1992, p. 225). Indeed we are chastized for expecting that there should be a commonly agreed-upon theoretical explanation: ‘Individuals unfamiliar with feminism or women’s studies often assume that feminist theory provides a singular and unified framework for analysis’ (Rosser 1997, p. 22). At least nine brands of feminist theory are currently available (Percy 1998; Rosser 1997). Liberal feminists argue for the advantages of psychological androgyny, the establishment of a gender-blind society with equal opportunities for men and women, and are unique among other feminists in continuing to accept traditional scientific method. Marxist feminists argue that gender oppression can be traced to capitalism as a means of production and to the power structures reproduced by class in capitalist societies. Freedom from gender role constraints can best be sought in a Marxist economy. Socialist feminism seek to give a more equal weighting to class and gender. Socialist feminists integrate material, social and unconscious processes in explaining how race, gender, class and sexuality produce power relations that disadvantage women. Afro-American feminists reject the Eurocentric approach to knowledge embodied in individualism and positivism. They maintain that race is the primary oppression and that gender is secondary to this. They are particularly critical of scientific work that under-emphasizes the impact of social and economic inequalities between the races. They deplore the failure of mainstream feminism to address the problems of women of colour. Radical feminists believe that men’s oppression of women is the most fundamental and widespread oppression in society. They urge

Mhoc01.fm Page 31 Friday, December 14, 2001 10:14 AM

      


women to reject all theories developed by men including Marxism, psychoanalysis, positivism and existentialism. Women can gain true knowledge only by using and reflecting on their own personal experiences and those of other women. Lesbian separatists allege that compulsory heterosexuality and engagement in patriarchal society makes it impossible for women to understand their own oppression. Women must refuse to collaborate with men in any way that oppresses women, lesbianism is the preferred sexuality and artificial insemination the preferred means of reproduction. Essentialist feminists argue that women by virtue of their biological and psychological qualities are equal to or superior to men. Although originally rejecting any implication of biological differences as ‘a tool for conservatives who wished to keep women in the home’, they have now rethought their position ‘with a recognition that biologically based differences between the sexes might imply superiority and power for women in some areas’ (Rosser 1997, p. 29). Psychoanalytic feminists use neo-Freudian theory to argue for the unconscious internalization of female powerlessness. Psychoanalytic feminists trace gender differences to the distinct ways of dealing with psychosexual development. Although they reject the ‘biological determinism’ of Freud, they trace male dominance to the fact that women are the chief care-takers of infants and children, resulting in boys distancing themselves from their mother and adopting independent and autonomous styles while girls become enmeshed in over-dependent relationships with their mothers. Existential feminists emphasize the ways in which women are raised to see maleness as the natural human state in which women form the objectified ‘other’. It is the importance that society accords to biological sex, rather than sex itself, that forces women into playing the role of the Other. Post-modern feminists reject the notion of a stable and unified self but rather see the self as a product of ideology, discourse and language. Equally, they reject the idea that women can speak with a unified voice. They argue against grand theoretical narratives and contend that gender, like the self, is neither real nor fixed but variously socially constructed in different contexts. In most of the examples above, it is clear that the political agenda of feminism or a putative description of the status quo has superseded ‘theory’ as it is normally understood. Be that as it may, it is a surprise to many evolutionary psychologists to find themselves accused of internal wrangling when faced with the bewildering disarray of feminist theory. While evolutionary psychologists argue over the fine-tuning of their favoured theory, feminists appear to be unable to agree on what their theory should be.

Three questions Are there inequities between women and men in society? Where did they come from? How shall we change them? The first is a straightforward empirical question that we can answer with respect to a variety of criteria including relative income, likelihood of promotion, leisure time, voting rights, participation in political life, public recognition of achievement and so

Mhoc01.fm Page 32 Friday, December 14, 2001 10:14 AM

      on (while holding all other variables apart from sex constant). Most people would agree that such differences do exist and, in the main, women fare less well than men. The second question is the subject-matter of this book. I will map some of the domains of women’s lives that are characteristically different from men’s and offer an interpretation of them from the viewpoint of evolutionary theory. I am addressing the distal causes of male–female difference stemming from disparate pressures on men and women several hundred thousand years ago. But these evolved differences can also set up a dynamic of their own. If fewer men than women excel in the field of interpersonal sensitivity and if fewer women than men excel at navigation, we can be misled into typologizing these activities as male or female. The differences between men and women are differences of degree not kind. The overlap on the distribution is great. Even for the most male-advantaged tasks (running a marathon, lifting weights) there are always some women who do better than the least able men (and vice versa). This is even more true of psychological characteristics. We must avoid restricting opportunities on the basis of crude stereotypes about what men and women are able to do. My concern with stereotypes is not so much that they drive people to conformity—I have already explained that stereotypes are more likely to be the product not the cause of sex differences—but that they may cause us, within the family or wider society, to debar entry on the basis of sex. For example, denying women the right to be fire-fighters or police officers is rationalized on the grounds of strength or endurance. But this is clearly wrong. The criterion should not be sex but the individual’s ability to perform the tasks that the job entails. Women should be given entry not in the belief that they will act as role models to other women but because to deny them the right to take up a job for which they are qualified and able is a basic human injustice. Whether or not we want to alter the status quo is not a matter for psychologists but for society at large. But the last half-century has shown that there is a public will to do so. But social engineering without a firm scientific understanding of sex differences is like a surgeon operating with a blindfold (Tooby and Cosmides 1992). In accepting the idea that womanhood is socially constructed and is without any psychological basis, we are already in danger of developing policies that are not in women’s best interests. Women’s natures, its seems, not only can but should be the same as men’s. Women are not ‘naturally’ maternal so there is nothing special about mothering. As a result single mothers have been forced to put their children in daycare and made to work. Their feelings of loss and guilt have been brushed away because they are ‘feminine’ feelings incompatible with effective performance in the workplace (Hrdy 1999). Women employees who do not show the ruthless drive of their male counterparts are disparaged as poor role models for other women and blamed for women’s failure to break the glass ceiling. Girls who resist the contemporary educational pressure to take science subjects are viewed as academic also-rans. If on the other hand we accept that women and men are different, we can think about a society that breaks down the barriers between children and work, that

Mhoc01.fm Page 33 Friday, December 14, 2001 10:14 AM

        allows women to see value in cooperation as well as competition and that allows women to capitalize on their linguistic advantages. If evolutionary theory is correct then we cannot design twenty-first century woman as if from scratch. Ideology, social policies, law and the media cannot in and of themselves make women into something they are not. What we can and should do is to give people choices that allow them the maximum freedom to be whatever they want. With that freedom, women’s nature can take its own course.

Mhoc02.fm Page 34 Friday, December 14, 2001 10:15 AM

 


Mothers matter most: Women and parental investment

Ask a child why mothers matter and there will be no shortage of answers: they make costumes for the school play, they organize birthday parties, they look after you when you have chicken pox, they do the school run, they make up the beds so friends can stay overnight, they take you to kids’ movies that bore grown-ups . . . Fathers do not lack the competence to do these things yet they do them less often. For many social scientists the reasons are to be sought in patriarchy, constricted gender roles, maternal guilt and family-hostile career structures. But evolutionary theory steps further back to look at the biological basis of reproduction because that is the wellspring from which these other more recent social and psychological causes spring.

Sexual, not natural, selection When we think of evolution we tend to think of natural selection—the competition for survival. But Darwin knew that there was more to the process of evolution than this. Survival without reproduction is a genetic dead-end. An animal that survives but does not reproduce leaves no genes behind. Our forebears may not necessarily have lived a long life, but we know that they successfully reproduced and ensured that their progeny, in turn, reached reproductive age. Darwin named this second strand of evolution sexual selection: ‘the advantage which certain individuals have over other individuals of the same sex and species, in exclusive relation to reproduction.’ So the prize is not necessarily to survive to an old age but to reproduce. After reproduction, natural selection is indifferent to us, even callous. Genes that enhance youthful reproduction will flourish even if as a side-effect they happen to cause earlier death. Cancer and heart disease are immune from natural selection, as long as they do not kill people too early in life. Death is the result of an absence of selection pressure on the diseases of old age because our death is of no consequence in the grand scheme of things. Sex differences derive from sexual rather than natural selection. If women were more vulnerable to death—if they made easier prey or succumbed more quickly to the effects of food shortage or had less efficient immune

Mhoc02.fm Page 35 Friday, December 14, 2001 10:15 AM

  :      systems—women (and the whole species with them) would have become extinct. In fact, if we look at the average age of death, women survive longer than men. If there is an imbalance it seems to work against men rather than women. For every hundred girls born, about 105 baby boys arrive. Because males take more risks, are more vulnerable to accidents and suicide, and experience more developmental difficulties such as attention deficit hyperactivity disorder, autism and dyslexia, we need to begin with a surplus of males in order to end up at maturity with an equal representation of both sexes. Inclusive fitness is the sum total of the genes that we leave behind in all of our blood relatives including children, grandchildren and great-grandchildren. This is probabilistically related to our own personal reproductive success. The more children we have, the more grandchildren we are likely to have. Each child will carry a half of our genes and each grandchild one-quarter. Any genetic trait that has the effect of increasing the number of children that we rear will be expressed in more future bodies than a genetic trait that puts us at a disadvantage in reproductive competition. But the traits that assist men and women in carrying their genes forward are not identical. What is a good strategy for a man may be counterproductive for a woman. Over evolutionary time we begin to see a sex difference appear. Baby girls receive those genes that selectively help females to become reproductively successful while baby boys receive a slightly different complement of genes that in the past have helped their fathers and grandfathers. There are a number of ways in which these genetic sex differences can be carried between generations. Some may be sex limited—here the genes are carried on autosomal chromosomes (ones that do not code for anatomical sex differences) but are activated in the presence of certain sex-specific hormones such as testosterone. They may be sex-linked, that is carried on the X or Y chromosome that determines the child’s sex. An added wrinkle to this pathway is the increasing number of genomically imprinted traits that are being found. In this process, whether or not a gene is expressed depends upon which parent donated the gene. For example, recent work suggests that women’s greater social intelligence is mediated by genes received from the father not from the mother. The X chromosome that the father contributes carries the critical genes. The homologous genes on the X chromosome that the mother contributes are silenced. Hence boys show lower social intelligence because they receive only the mother’s copy of the X chromosome (Skuse et al. 1997). Studies of rodents suggest that a similar process may operate for maternal behaviour—the mother’s tendency to retrieve straying pups and return them to the nest is mediated by the X chromosome of her own father. Whatever the pathway used, each sex receives the genetic instructions most useful for building a mind that will enhance the body’s reproductive success. So sex differences are expected only where they have a direct influence on sex-specific reproductive strategy. Biology can point to the sexual strategies of men and women that lie behind evolved psychological differences.

Mhoc02.fm Page 36 Friday, December 14, 2001 10:15 AM


    

Anisogamy: the start of parental inequity Sexual reproduction is not obligatory. There are plenty of ways to reproduce that do not require the fusion of gametes from two individuals. In parthenogenetic species, no mating is needed. North American whip tail lizards are all female and in the breeding season, they produce about ten unfertilized eggs that hatch carrying 100 per cent of their mother’s DNA (Fisher 1993). There are plenty of advantages to this strategy—these females do not waste time and energy on finding mates, they do not expose themselves to predators while copulating, they do not have to impress the opposite sex with their desirability, and most importantly they do not dilute their genetic legacy by 50 per cent. But it is this last factor that paradoxically becomes an advantage in mating. Sexual reproduction creates novel and unique individuals and this has a threefold implication. First, because offspring are different from their parents and their siblings, they can occupy a variety of different environmental niches and so create less competition with their kin while increasing the odds that at least some will survive in the face of unpredictable local hazards like sudden climatic changes. Secondly, each offspring’s unique genotype means that it has a unique immune system—when parasites take hold they may prove deadly to some of the brood but others will survive. Another advantage of sexual reproduction is genetic repair. Imagine a strand of DNA which is normally composed of sequences of the four nucleotides C, G, T and A. When an error occurs— CGX—the best way to find out what should really be there instead of the X is to examine the second copy of the sequence on the complementary strand of DNA. But it will do no good if the second strand has come from the same genetic line as the first because there is a good chance that it will carry the same error. In sexual reproduction, the enzymatic ‘proofreading’ machinery of the cell consults the corresponding sequence on the strand provided by the second, unrelated parent and fixes the error (Williams 1996). So two parents can be better than one. But why have an egg and a sperm? An egg is about one million times bigger than a sperm and far more costly to make. It carries not just its DNA message within the nucleus but the metabolic machinery and nutrients to supply the zygote until the tiny bundle of cells can attach itself to the uterus and set up its own supply line through the placenta. A sperm is little more than a DNA-containing nucleus with a tail to propel it in its search for a free genetic ride in an egg. Why not have more equality—two parents who each contribute equally to the zygote? Multicellular algae do this. Each alga releases cells into the water that fuse with cells from another alga creating second-generation compound cells. All that is required for basic sexual reproduction is organisms that are capable of releasing haploid gametes (ones that contain half the adult number of chromosomes) that can join with a gamete from another organism to form a viable zygote. The story of the evolution of sperm and egg begins with just such an egalitarian scenario (Parker et al.

Mhoc02.fm Page 37 Friday, December 14, 2001 10:15 AM

  :      1972). Each gamete carried half the nutrients needed by the about-to-be-formed zygote. But then came variability; some gametes carried less than their fair share and others more. Imagine that every individual made a thousand gametes for every milligram of material devoted to reproducing (Williams 1996). Each gamete was worth one microgram and when they fused they created a two-microgram zygote which made it strong enough to be viable. But now some cheating begins, thanks to the random process of mutation. Some mutants produce not 1000 but 1100 gametes even though each one is slightly smaller (only 0.9 microgram). If each one fuses with a normal, non-mutant gametes, the zygote will weigh a little less (only 1.9 micrograms) but that will be compensated for by the fact that there will be 10 per cent more of them. As long as the net disadvantage is less than 5 per cent, the mutant form will still be ahead of the game. When two mutant gametes meet the zygote drops to 1.8 micrograms so that the mutant forms lose their net advantage, preventing the whole species from adopting the cheating strategy. But the situation could stabilize, favouring ever larger and smaller gametes. Organisms producing small gametes can make a larger number of them and hence a larger number of offspring but they have the disadvantage of small size and lower viability. Organisms producing larger gametes make fewer of them but they are better equipped and more likely to survive. Gametes have to achieve two objectives—they must find a partner and create a well-resourced, nutrient-provided zygote (Low 2000). Small gametes do best at the former—sperm are cheap to make, light and motile. Big gametes do better at the latter—eggs carry the nourishment that the growing zygote needs to survive. Intermediate-sized gametes have neither advantage and so lose out to both. Once anisogamy (the size division between egg and sperm) had begun, the gulf between them could only get bigger. Sperm that were better designed to move fast and carry only the minimum fuel out-competed slower, fatter sperm. Eggs that carried the most nutrients and conserved their energy could survive longer and create stronger zygotes and so they became more numerous than thinner, more active eggs. So the process continued until the giant quiescent egg dwarfed the stripped-down searching sperm. The consequences of this gametic ‘cheating’ have been almost unimaginably far-reaching for the two sexes. The sperm will have to compete with one another for the limited number of eggs available, and will therefore retain the locomotor mechanisms (long propulsive tails in most species) needed in the race for fertilisation. The eggs can leave the work of finding a partner to the sperm. . . . When egg-producers reproduce, they must bear the entire nutritional burden of nurturing the offspring. By contrast, the sperm-makers reproduce for free. A sperm is not a contribution to the next generation; it is a claim on contributions put into an egg by another individual. Males of most species make no investments in the next generation, but merely compete with one another for the opportunity to exploit investments made by females. (Williams 1996, p. 118)

Thus were females first taken advantage of at a purely biological level. But the egg and sperm distinction did not end there. The greater gamete size and cost that

Mhoc02.fm Page 38 Friday, December 14, 2001 10:15 AM

      females incurred was only the beginning of their greater lifelong commitment to infant care. It was Robert Trivers (1972) who spelled out the key reproductive difference between the sexes. He called it parental investment and defined it as ‘any investment by a parent in an individual offspring that increases the offspring’s chance of surviving (and hence reproductive success) at the cost of the parent’s ability to invest in other offspring’. It is absolutely central to an evolutionary analysis of sex differences. The more time and effort an individual channels into any single offspring, the fewer offspring they will ultimately produce. This simple principle amounts to a distinction between offspring quantity and quality. Animals who have taken the quantity route are called r-selected species. Insects lay thousands of eggs but may invest very little time or energy in any of them. Instead they move swiftly on to the creation of a new batch. K-selected species, such as primates, give birth to one baby at a time and spend several years feeding and protecting it. Some animals, like cats, steer a middle course giving birth to multiple offspring but yet providing milk and care to them for a limited time. The point is that there is always a trade-off—the more time we devote to one child, the less time there is to produce more. The r–K distinction seems to hold for men and women also. The minimum biological costs of reproduction are greater in women than in men. Each month about twenty ovarian follicles prepare their oocytes for possible ejection, one is chosen and ripened, meiosis takes place and the ovum is released into the Fallopian tube. Meanwhile the endometrium of the uterus must be engorged and prepared to receive a fertilized egg (even if one fails to arrive which, as we will see, is usually the case). These twin processes are so time-consuming and costly in terms of calories that they take about 14 days (half a menstrual cycle) to achieve. If pregnancy occurs, the woman’s body is occupied for nine months with gestation. After the birth, lactation requires even more calories than were necessary to sustain the pregnancy (almost twice a woman’s normal daily calorie requirements) and, in the environment of our evolutionary adaptatedness, would have continued for up to four years. To every mother, every baby represented a very large investment of time and energy, to say nothing of emotion. Compare this with a man’s minimum investment. He can ejaculate several times a day. Admittedly, the number of sperm he produces falls with each intercourse from a high of 300 million on the first occasion to a low of a mere 30 million after a threehour interval. Nonetheless, pregnancy only requires one sperm in the right place at the right time. Once his sperm have been gallantly donated, any further contribution on his part is optional rather than biologically mandatory. It does not require a mathematical genius to realize that under optimal conditions a man could father half a dozen infants in a day while a women would take thirty years to do the same. A host of psychological differences between males and females, as we shall see, derive from the fact that women’s biological parental investment is greater than males’. But it is important to recall that the crucial factor that drives these differences is

Mhoc02.fm Page 39 Friday, December 14, 2001 10:15 AM

  :      parental investment, not maleness or femaleness per se. In our species it happens to be women who invest more heavily but in other species, like pipefish, sea horses and teleost fish, it is fathers not mothers who make the greater investment. And this seems to result from the mechanics of reproduction in these species (Dawkins and Carlisle 1976). Out of 49 species of fish where paternal care is found, 48 reproduce by external fertilization (Trivers 1985). Because gametes evaporate in the air, earthbound species copulate using intromission of sperm to keep them in a safe host environment. Among fish, however, the seawater protects them from evaporation. Because of this, in a number of species, a female lays her eggs and they are then fertilized externally by the male—he must wait until she spawns because his sperm are lighter than her eggs and are more likely to be carried away by passing currents. Because the female finishes her part of reproduction before the male, she can then desert knowing that the male will be forced to take responsibility for the brood. Of course, the male could desert also but there would be no selection pressures for abandoning viable offspring. A male who continued to do this would leave no progeny behind and so deserting last is not an adaptive choice. In any case, the real bonus of external fertilization for the male is that he is certain of his paternity and this means he can be equally certain that the effort he devotes to the egg batch is going to his own, rather than a rival male’s offspring (Trivers 1985). In humans and other mammals, however, fertilization occurs within the female’s body and the fertilized ovum remains there, developing, for several weeks or months. This gives the male ample time to desert, leaving the female holding the baby. Because deserting first offers advantages for further copulations with new and unencumbered partners, it would have been selected for and, because of internal fertilization in humans, this would lead to the evolution of a male strategy of seeking multiple sexual partners. Because the tendency to desert is related to future reproduction prospects, desertion will be less likely among females than males because their future prospects are always lower than males—that time is especially truncated in human females by the advent of menopause and because a human infant demands such a very long period of maternal care. Why has natural selection favoured this long period of parental care? After all, if infants were born ready to cope with the world by themselves, humans could produce several times the number of children that they currently do. Many mammals arrive in the world able, quite literally, to find their feet immediately and, within weeks, to feed themselves. The answer is our unusual intelligence. Compared to our hominid ancestors, Homo sapiens has an extraordinarily large brain—about three times the size of the Australopithecus brain (McHenry 1994). Brains are very expensive items, they consume more calories than any other organ of the body. The benefits of their steadily increasing size must have been consistently advantageous at every evolutionary step. They gave us a better memory (capable of storing massive amounts of information), representational thought (the ability to perform safer off-line ‘thought’ experiments rather than costly trial-and-error learning), consciousness (the ability

Mhoc02.fm Page 40 Friday, December 14, 2001 10:15 AM

      to represent ourselves in our internal model of the world), language (and through this speedy cultural transmission), a theory of mind (the ability to accurately impute mental states to others) and metacognition (the ability to reflect on what we know and how we know it). A brain capable of doing these things needs to be big. It was the increasing size of the brain that posed special problems for women (Lancaster and Lancaster 1983; Rosenberg and Trevathan 1996). Humans were walking upright and enjoying all the benefits that came with this ability. But bipedal locomotion required a small pelvis and a smaller pelvis meant problems with childbirth as the neonate’s head became bigger and bigger. Babies born any later than nine months were simply too big to get through the pelvic bones. So nine months became a compromise gestation period. Any longer and the baby could not be born, any shorter and the baby would be too immature to survive. But, as any women who has given birth knows, natural selection pushed gestation time to its upper limit. During childbirth, the four plates that compose the baby’s skull are forced to slide over one another leaving the newborn with a visible and vulnerable soft spot, the fontanel. Human infants are born too early in a maturational sense. They are helpless, totally dependent on parental care and have another thirteen years before they reach bodily maturity.

Men and the attraction of polygyny The optimal condition for male reproductive success is access to as many fertile women as possible. If a man remains with a single woman throughout his life and is completely faithful, he can only produce as many children as his wife can give birth to. Her intensive period of investment in each child places a ceiling on his success also. But a man with even two women can double his reproductive success. With a harem, he can increase it a hundredfold. In fact under such ‘perfect’ conditions, a man’s sexual performance would be even further enhanced because novelty increases sexual arousal. (This ‘Coolidge effect’ is named after the president’s famous remark on being shown around a chicken farm. His wife, informed that the cock copulated several times a day, asked her host to tell the president about it. When told, her husband asked if it was always the same hen. ‘No, Mr President, it is a different one every time.’ ‘Tell that to my wife’, he allegedly replied.) Concubinage was aimed at precisely this kind of maximization of male reproductive success. Ismail the Bloodthirsty of Morocco (1672–1727) is reputed to have had 500 concubines at any one time (they were recruited at puberty and ejected at the age of 30) and to have produced 888 children (McWhirter and McWhirter 1975). Most men are not so fortunate but that does not alter the basic fact that for several thousand years men’s reproductive success was directly related to their number of sexual partners. As we shall see, staying with one woman had advantages—men could offer protection and resources that increased the likelihood of each of their children surviving to maturity. But it had a definite cost too. Any man who took

Mhoc02.fm Page 41 Friday, December 14, 2001 10:15 AM

  :      advantage of the occasional opportunity to have sex with other women would have a definite reproductive edge (Geary 2000). There is much evidence that this strategy of at least mild polygyny was (and continues to be) a strong theme in human evolution. The evidence comes from comparing sex differences in humans to those that are evident in other polygynous species. Polygyny is strongly associated with differences in size and strength between the sexes. Because the ratio of males to females is about 1:1, any male who monopolizes or impregnates more than one female leaves another male without a mate. This means that males must compete with one another to gain access to extra female partners and this selects for ever larger and stronger males over the generations. (Big males have many children, half of whom will be sons who carry their fathers ‘big’ genes. Small males have fewer children and so fewer sons carrying ‘small’ genes. Over generations, male size increases.) The size difference between human males and females was very marked indeed in our hominid ancestors. Australopithecus males were between 50 and 100 per cent larger than females (Geary 2000). Though the size of the sex difference has diminished—probably as a result of less intense male competition resulting from an increase in paternal investment and a consequent decrease in polygynous matings—it has remained relatively constant over the past 300,000 years with males being about 20 per cent larger than females (Geary 1998; Hrdy 1999). At puberty boys’ height shoots up and most of men’s extra height is attributable to their longer leg length. They also develop larger skeletal muscles, a greater capacity for carrying oxygen in the blood and for neutralizing the chemical products of physical exercise. They can run faster, grip harder and jump higher than females (Tanner 1990). In polygynous animals, puberty occurs earlier in females than in males. At reproductive maturity, a young male must enter the ferocious male–male contest for mating opportunities. The bigger he is and the greater his experience of preparatory rough-and-tumble contests, the greater his chance of success. So there is an advantage for males in delaying puberty even though it means that they must wait longer for their first reproductive opportunity. In polygynous species, the female pubertal growth spurt starts earlier and peaks more quickly than in males (Leigh 1996). This is true of humans also and boys reach puberty at the age of about 13–14, approximately two years later than girls (Tanner 1990). Males in polygynous species also tend to die earlier than females. The Y chromosome instructs the formation of testes and so begins the generation of testosterone which energizes youthful sexual and competitive behaviour—but at a cost. In the long term it compromises the immune system leading to men’s earlier death relative to women (Folstad and Karter 1992). Another piece of biological evidence adds to the picture of polygyny—testes size (Harcourt et al. 1981). Animals that have no need to compete for females, like the solitary harem-holding gorilla, can manage with modest testes. But animals who mate with many females must produce copious quantities of sperm so that their sperm can successfully compete with those of other males who are also seeking extra-pair mating

Mhoc02.fm Page 42 Friday, December 14, 2001 10:15 AM

      opportunities. Admittedly, human testicles look rather undersized in comparison with chimpanzees but they are considerably more impressive than those of the gorilla. A final requirement of the successfully polygynous male is a strong sex drive. The male must be ready and eager to have sex at every possible opportunity and this seems to be marked feature of human males. In a review of 177 data sources (Oliver and Hyde 1993), men had a far more positive attitude towards casual sex than did women, they had intercourse more frequently and the biggest difference of all was in incidence of masturbation, an activity enjoyed more frequently by men—often as a substitute for sex with a partner. Men experience sexual fantasies and sexual arousal about once a day compared to about once a week in women and men more often fantasize about someone they have not yet had sex with while women tend to rerun past sexual encounters (Ellis and Symons 1990). Because of this disparity in motivation, the frequency of intercourse for couples depends upon the woman’s rather than the man’s willingness This is demonstrated by the very high rate of sexual contact among male (but not female) homosexuals where there is no female partner to gate or suppress high male sexual drive (Symons 1979). An ingenious study in the United States (Clarke and Hatfield 1989) documented men’s willingness to avail themselves of complementary sexual partners by using a male and a female confederate to approach students on campus and ask them if they wanted to come back to their apartment for sex. Predictably, 75 per cent of men but none of the women took the stranger up on their offer. One psychological mechanism that may give rise to this sex difference is in the standards that men and women apply to casual sexual partners. Men, more than women, are willing to relax their standards (especially in terms of intelligence, but also on personality and social status) very markedly for a short-term sexual partner compared to what they expect in a spouse (Kenrick et al. 1990). Though men may have paired with a single long-term partner, it seems as if nature fitted them to take advantage of sexual opportunities whenever they arose.

Why men invest at all Given the advantages of polygyny, it is reasonable to ask why males ever committed themselves to a single female and their mutual offspring. For some animals, the answer is simple—offspring die without the joint care of two adults. In some bird species two parents are needed to provide both protection and food. There is essentially no choice for a male in this situation. He could mate with dozens of females but if none of them is able to raise the brood alone then his reproductive success would be zero. This makes paternal care species-typical and obligate (Westneat and Sherman 1993). But in the majority of mammals, this is not the case. One parent can raise the offspring more or less successfully and this makes male desertion an attractive option. Why then do some males stay? An important factor is if and by how much paternal care improves the likelihood of infant survival. Does the presence of a father increase the chances that children

Mhoc02.fm Page 43 Friday, December 14, 2001 10:15 AM

  :      will survive to maturity? If so, does it make a sufficient amount of difference that, in terms of reproductive success, it outweighs the benefits of promiscuous matings and offspring abandonment? Suppose that a female has a 50 per cent chance of raising a child alone but a male’s presence doubles the probability of an offspring surviving. On average a male would have to mate with two females to equal the advantages of staying with one. Finding the data necessary to accurately specify the values needed to test such a hypothesis is hard. We can look, however, at human data on infant survival when fathers are absent (Geary 2000). In contemporary industrial societies, child mortality rates are very low and state provision is made to single mothers so that they are able to feed and clothe their children. We have to look at pre-state societies to observe the kind of fate that might have befallen fatherless children in our evolutionary past. Among the forest-dwelling Ache of Paraguay, the loss of a father before a child’s fifteenth birthday increases mortality rate from 20 per cent to 45 per cent (Hill and Hurtado 1996). Although the Ache share the fruits of their hunting across the whole community, children without fathers to contribute to the common pool fare less well nutritionally and may even be permanently removed. Children are thrown—sometimes alive, sometimes already killed—into the grave of a dead chieftain. The Ache say that the children are sacrificed as companions for the dead in the afterlife but the fact that the vast majority of the victims are orphans suggests that this is a justification for ridding themselves of extra mouths to feed. The loss of a father doubles the probability of a child being killed or sold into slavery. It triples the chance of a child dying from illness. In pre-industrial Europe, illegitimate children had a higher mortality rate than those born to a married couple and a United Nations study of contemporary developing countries similarly found an association between child mortality and father absence (Geary 2000). Although a father can provide a survival advantage for children (from increased resources and physical protection), nonetheless the majority of fatherless children can and do survive. And as we shall see later there is consensus that the loss of a mother is far more serious than the loss of a father. Across species, another powerful factor in paternal investment is paternal certainty. The trend is clear—the greater the doubt, the less willing is the male to invest time and energy in the offspring (Clutton-Brock 1991). Ancestral males who pursued such a selfless course of action would produce fewer of their own offspring while at the same time increasing the survival chances of their rivals’ children and this particular form of male altruism would be quickly removed from the gene pool. Among barn swallows, males who are the biological fathers of their brood provide 46 per cent of the nestlings’ food (nearly as much as the mother) while those who provision broods that contain some offspring sired by other males, provide only 34 per cent of the food intake (Moller and Tegelstrom 1997). In humans, both internal fertilization and concealed ovulation have worked against paternal certainty. When conception takes place inside the woman’s body, the male cannot be sure that it his sperm that have fertilized the egg. And because the period during which a woman can conceive

Mhoc02.fm Page 44 Friday, December 14, 2001 10:15 AM

      is not advertised, he cannot know whether he donated his semen at the ‘right’ time. A man’s suspicion may be well placed. When male parental investment is obligate, females are unlikely to risk their offspring by cuckoldry. But in humans, where women can raise the offspring alone, there is always the temptation to have it both ways—to find a male to provision her while obtaining the best genes from a higher quality but non-investing man. Studies of blood typing in newborns suggest that anywhere between 5 and 30 per cent of babies could not have been sired by the putative father (Bellis and Baker 1990). Men are more likely to invest in children when they feel relatively certain that they are their own. This certainty can be increased by mate guarding. If a man remains with one women over an entire cycle and ensures that she mates with nobody else, then he can be confident that he alone is responsible for a pregnancy. But given the fact that women may take several unprotected cycles to become pregnant and that pregnancy itself may not be detectable for several weeks, this commits him to an extensive period of monogamy. And this brings us to a third factor in paternal investment— the behaviour of females. Given the advantages of father-presence for women and the costs for men, we might expect that women would have evolved some strategies that made men more likely to invest exclusively in their offspring. They could do this by increasing the rewards of monogamy or by increasing the costs of polygyny. Women’s continuous sexual receptivity may have increased the rewards. Women’s reluctance (relative to men’s) to have casual sex would have made it harder for men to find willing but undemanding one-night-stand partners while at the same time reassuring prospective long-term mates that they were likely to be faithful in a monogamous relationship. Women’s intolerance of men’s additional sexual partners (and even prospective female rivals) would also make it harder for males to pursue polygyny. And monogamy and prolonged paternal investment might also have had advantages for those males who found themselves out in the cold under polygyny. Polygyny selectively benefits only the most desirable and well-resourced males. The less blessed face the possibility of zero reproductive success unless they are willing to make concessions. One concession might have been to abandon competition with other males for the desirable but elusive prize of many children and settle instead for the more modest, but altogether more certain, rewards of the fruits of a single woman’s body. Women and children benefit from paternal investment but the biological odds are stacked against equal investment by both partners. In a comprehensive review of the paternal investment, Geary reports that in all cultures so far studied, mothers are more available to and engaged with their children than are fathers. As he concludes: The pattern indicates that a reduction in paternal investment in favor of mating effort, that is, to pursue additional mates, is a viable reproductive strategy for some men but not for most women. One evolutionary result would be the maintenance of a greater focus of men, as a group, on mating effort than women and greater overall levels of maternal than paternal investment. (Geary 2000, p. 68)

Mhoc02.fm Page 45 Friday, December 14, 2001 10:15 AM

  :     

Women as choosy investors After conception, biologically speaking, men’s investment is completed and they are free to move on to pastures new. But women, unlike men, are quality not quantity specialists. Their investment is not limited to a few moments pleasure—they must live with the consequences for years and women produce only a limited number of offspring. In hunter-gatherer societies, birth spacing is about four years and a woman typically gives birth to four or five babies. Each one is precious and each one must be carefully nurtured to adulthood. So great is the commitment demanded by every child that women’s bodies and minds are exquisitely careful to invest only in high quality children. This fussiness is all the more remarkable since menopause drastically shorten a women’s reproductive career relative to a man’s, yet still she will not invest in an infant that has a poor chance of survival. First, she must ensure that her body has a reasonable chance of sustaining the pregnancy. Menstruation is the first casualty of a poorly resourced environment. Malnourishment signals that a pregnancy under current conditions would not be viable. Athletes and dancers, who put considerable physical demands upon their bodies while maintaining a very restricted diet, lay down little fat and so have later menarche than other girls (Angier 1999). Although anorexics deliberately choose to avoid food, the body does not distinguish between an active choice and an environmental scarcity and first ovulation, then menstruation, ceases. Women in Zaire lose weight and show a drop in progesterone levels and ovulation when food is scarce before the harvest. After the harvest, their fertility returns to normal (Ellison 1994, 1996). The same relationship between abundance and fertility can be detected by examining birth rates around the world in the wake of periods of famine or plenty. Nine months after a food shortage there is a significant drop in births (Betzig 1997). Stress is a warning that the body is under strain. That strain may result from resource shortage and competition in the external environment or from problems in the immune response inside the body. In many cases, the two are linked. High levels of stress can suppress ovulation, menstruation or both (Baker 1996; MacDonald 1997). The body shuts down the reproductive system and waits for a better time to start a pregnancy. Acute stress seems more likely to trigger ovulatory shutdown than does chronic stress. Among women who live in extremely poor and unrelentingly harsh environments, stress levels are consistently high and if the body responded by shutting down ovulation, women who hardly ever reproduce. In fact, in such circumstances of chronic stress, the woman’s body adapts and responds by adopting an r-strategy and producing as many offspring as possible in the hope that some at least will survive the adverse environmental circumstances. Stress is relative. A woman living in more favourable circumstances may inhibit ovulation in response to short-term stress such as her husband being made redundant—her body through cortisol release is signalling that she should wait for a more favourable moment to

Mhoc02.fm Page 46 Friday, December 14, 2001 10:15 AM


    

conceive. Yet her current circumstances might well seem like a life of luxury to a street beggar in India. Fertility is also affected by the season. Babies survive better when the climate is moderate and food supplies to the lactating mother are plentiful. In the northern hemisphere, more babies are born in early spring—February and March—than at any other time. The mothers time their baby’s arrival for the start of warmer weather and the blossoming time of berries and grain. In turn, ovulation is most likely to occur in May or June than at any other time. A woman should be fussy about which man she will allow to take a free genetic ride on her parental investment. Two ways that her body can do this are by concealing ovulation and by favouring conception by a male that is healthy and attractive. During a normal ovulatory cycle, a woman is capable of conception only for a very narrow band of time. The sperm must be delivered to her between 6 days before and 12 hours after ovulation with peak fertility occurring when sperm arrive precisely 48 hours before ovulation. The chief problem of course is that neither the man nor the woman knows when she is ovulating. It happens 14 days before her next menstruation but since she cannot see into the future she is left with retrospective estimations from the date of her last period. This is notoriously imprecise as cycles can vary by several days from one month to the next. The function of concealed ovulation has fascinated evolutionary thinkers because in many (but by no means all) primates the female’s fertile period or oestrus is signalled very clearly through genital swelling. It has been speculated that the point of concealment is to fool the woman herself. If early humans were capable of knowing when they were fertile and equally capable of observing the pain and death associated with childbirth, they might effectively use abstinence as a contraceptive measure. Women who recognized their own ovulation would have produced fewer children than their less sensitive peers and there would have been selection for ignorance (Burley 1979). Another useful function of concealed ovulation might have been to cut down the amount of aggression, between males and against the female. If everyone knew when a female was fertile, then all male attention and competition would be intensively and exclusively focused on that one female—the injury level would be overwhelming. But the more obvious consequence of concealed ovulation is that it creates uncertainty about who the father of a given child might be. And that leads directly to two apparently opposing hypotheses. One benefit of uncertain paternity may have been to encourage monogamy by effectively forcing the male to be vigilant in mate guarding over a long period in order to avoid cuckoldry. But concealed ovulation also allowed women with multiple mates to fool each man into believing they were the father and so encourage them all to supply food to and defend the infant—or at least to refrain from injuring it. Among the Ache of Paraguay, three types of fathers are recognized: the man to whom the pregnant woman is married, the man with whom she had sex just prior to her pregnancy and the man whom she believes is responsible for the pregnancy.

Mhoc02.fm Page 47 Friday, December 14, 2001 10:15 AM

  :      About 63 per cent of babies have multiple fathers or ‘partible paternity’ and the average number of fathers is just over two. Because the Ache sometimes dispose of a child whose father dies, children with multiple fathers are at an important advantage. Among the Bar of Venezuela and Columbia, a woman who claims multiple fathers for her unborn child runs a lower risk of miscarriage and of perinatal death presumably because all the fathers ensure that she is well supplied with food throughout the pregnancy. The child also has a greater chance of surviving until the age of 15 because it has the support of two fathers. These bizarre facts notwithstanding, not everyone accepts the partible paternity hypothesis and its plausibility depends upon exactly how strong human paternal certainty must be to ensure investment and whether that investment is graduated. If the certainty threshold is set high for investment (and men’s normally acute sensitivity to women’s infidelity and the frightening strength of their jealousy suggests it is) then trying to convince multiple men that they may be the child’s father ought to be doomed to failure because each of them would refuse to invest. This would argue strongly for the mate-at-home hypothesis of concealed ovulation. But it may be that, under some ecological conditions, a modicum of care from several men outweighs full investment from one. This depends upon paternal investment being graduated according to certainty rather than being an all-or-none affair. Men who harbour a hope of their possible paternity will make some efforts to protect the child as long as the cost is not too great. Even if they simply refrain from harming it, the child’s survival chances may be enhanced. Diamond (1997) suggests that both hypotheses are correct but for different points in our evolutionary history. By examining the family tree from which humans descended (Sillen-Tulberg and Moller 1993), it is possible to find at least twenty points at which changes in ovulatory advertisement occurred. If we look at those species that shifted from advertised oestrus to concealed ovulation, we find that the vast majority were either promiscuous or harem structured (where a newly established alpha male might well pose a threat to infants). This seems to suggest that concealed ovulation developed to confuse paternity rather than to enhance monogamy. But we can turn the question the other way round and ask what the prevailing oestrous conditions were in those species that went on to develop monogamy. Monogamy never evolved in species that had advertised ovulation but was largely confined to those species that had already developed concealed ovulation. This suggests that concealment was a female strategy to retain the attention of a single male. If we put the pieces of the puzzle together, the primate data suggest that concealed ovulation arose first in promiscuous or harem species in order to confuse paternity and prevent infanticide. Then once it had evolved, the species switched to a monogamous mating system in which concealed ovulation worked to keep a single male at home. Given the flexibility of human sexual strategies and our ability to fine-tune our behaviour to different circumstances, it may be that women exploit these two benefits under different circumstances (Jolly 1999) depending on the quantity and

Mhoc02.fm Page 48 Friday, December 14, 2001 10:15 AM

      quality of available men. What is clear is that concealed ovulation did not evolve as an advantage for men but for women. (An interesting question is why ancestral men did not evolve a way to counter this female strategy by detecting ovulation. Any man that could do so could very markedly raise his reproductive success.) When women are at mid-cycle and maximally fertile, they are more likely to have sexual fantasies, to dress provocatively, to initiate sex and to experience orgasm (Adams et al. 1978; Mateo and Rissman 1984). Married women are especially likely to show this timing in their infidelity and are less likely to use contraception when they are with a lover (Baker 1996). Whether or not conception takes place with a husband or a lover depends on the quality of his sperm and her own feelings about him. Only about one million of every 600 million sperm are capable of fertilizing an egg (Baker 1996). Why do men take the trouble to produce so many apparently useless sperm? They make them as an insurance against cuckoldry. The job of the 100 million blockers is to prevent any subsequent sperm from entering the uterus while the 500 million killers identify and destroy foreign sperm. When a woman has sex with two different men she sets in motion a ‘sperm war’ between them. This helps to ensure that only the most successful sperm win out and that her pregnancy will result from the male with the strongest army fighting on his behalf. And her orgasm can increases the likelihood of conception—she retains between 50 and 90 per cent of sperm as compared to between 0 and 50 per cent without orgasm. This is because during orgasm (in fact shortly after she actually psychologically experiences it) her cervix gapes open, widening the mucus channels, and dips down into the seminal fluid, both of which allow more sperm to pass through. The contraction of uterine and vaginal muscles create pressure changes which suck the semen more effectively through her cervix. The orgasm also expels old sperm from the uterus creating space for new sperm to enter. Women are more likely to experience orgasm with a lover than with a husband so in this way a woman can alter the probability of a pregnancy from an act of intercourse. Even after sexual intercourse, women have means of controlling whether or not conception will occur. Failure to conceive can result from a shortening of the normally consistent interval of 14 days between ovulation and menstruation which reduces the chance of a fertilized egg implanting. It is in this time window that a foetus can be ejected without even the woman’s knowledge. Only about 60 per cent of fertilized eggs succeed in implanting themselves in the uterus and of these, 60 per cent will not survive to the twelfth day of pregnancy and the woman experiences what she may believe to be a normal, if heavy, period. Of those that make it through the 12-day barrier, about 20 per cent will be miscarried in the first trimester (Baker 1996). The majority of these conception failures are a result of genetic abnormality in the foetus or of disastrous, high-stress maternal events. Miscarriage rates increase with the outbreak of war or with the death or infidelity of a woman’s partner.

Mhoc02.fm Page 49 Friday, December 14, 2001 10:15 AM

  :      But even when a pregnancy is established, women still exercise choice about continuing with it and even about continuing to care for a child after its birth. If a foetus or an infant is not viable then a woman is better off abandoning it as early as possible. Trivers (1972) argued that this was because the longer she continued to invest in a child the more time and effort she had wasted when it ultimately died. This was christened the Concorde fallacy by Richard Dawkins and Belinda Carlisle (1976). They drew an analogy with Britain’s sustained and economically disastrous commitment to the supersonic airliner which was driven by the logic that so much had already been spent that it was unthinkable to abandon it. Dawkins and Carlisle pointed out that this retrospective view of ‘investment to date’, however intuitively appealing, was incorrect. They argued, and Trivers (1985) later came to agree with them, that the true explanation of early abandonment lay in the amount of prospective investment that remained. A woman who abandons a non-viable foetus saves herself 15 years of future wasted investment while a woman who abandons a 12-year-old saves herself only a further 3 years of investment. These interpretative arguments notwithstanding, the evolutionary predictions are identical and are not in dispute. The cruel truth is that from the point of view of her own genetic success, a woman who spends time and effort in raising a child who is a poor reproductive prospect or who will be born into an environment too harsh to sustain normal development curtails her chance of investing in other more viable offspring. Natural selection should favour female bodies and minds that are profoundly selective about which infants should receive the mother’s scarce and valuable investment amounting to years of sustained feeding and protection. In many contemporary cultures, women have the option of terminating the pregnancy. We would expect that the decision to do so would be related to whether or not the mother would do better to wait for a more propitious time to bear the child and to the currently available resources for rearing the child. The decision to have a child now—versus waiting for a better set of emotional and material circumstances in which to make this enormous investment of time and energy—is closely related to the mother’s age. Younger women have a longer reproductive career ahead of them and can afford to wait for a more optimal time. Statistics confirm that the rate of voluntary abortion is high among young women aged 20 or less, with approximately three-quarters of 18–19-year-olds choosing to terminate their pregnancy (Hill and Low 1992). (Opting to terminate also rises after the age of 40 but for different reasons. Here the mother has finished her childbearing tasks and may feel that she is too old to provide the care needed by another child. As women approach menopause, they may do better to provide assistance to grandchildren (Euler and Weitzel 1996; Pope et al. 1993; Sear et al. 2000). In addition, the probability of bearing a child with genetic abnormalities also increases at this age.) Young women can afford to wait not only in terms of their biological clocks but also because they have a better future chance of finding a reliable male partner to assist in child-rearing. Younger women have a much higher probability of marriage and remarriage

Mhoc02.fm Page 50 Friday, December 14, 2001 10:15 AM

      than do older women and they are more likely to find a marriage partner if they are childless. Material, personal and interpersonal resources also play a role in the decision to abort. Fathers make a contribution to the economic and emotional welfare of the mother–child unit and statistics confirm that the best predictor of abortion is whether the father is willing to commit himself to the mother. In the United States, 65 per cent of pregnancies among unmarried women were terminated compared to 10 per cent among married women (Hill and Low 1992). A study of 300 middle-class American women asked them to estimate the probability of terminating a pregnancy. The estimated probability was 0.07 for married women but if the woman was unmarried at the time of the pregnancy that figure rose to 0.64. Fathers are less likely to contribute to the care of a child who is not their own and when women were asked about the likelihood of abortion, given uncertainty about who the father was, the probability rose to 0.57 in line with the figure for no supporting male (EssockVitale and McGuire 1988). Fathers are not the only source of possible support. State aid, while meagre, is better than the possibility of zero economic support. In a study of teenage girls who made different choices about their pregnancy, the provision of state aid decreased the abortion rate to 42 per cent compared to 71 per cent among those without such aid (Leibowitz et al. 1985). The woman’s own psychological resources are important too. Women suffering from psychological instability may lack the social, emotional and physical resources not only to bear and rear children but also to find and retain a long-term male partner. In a comparison of women suffering from anxiety and depression with a non-clinical control group, the clinical group was less likely to have ever married (70 versus 94 per cent) and far more likely to have divorced (87 versus 38 per cent). Clinic women had a 0.39 probability of abortion compared to 0.07 for the control women, although the two groups did not differ in their likelihood of ever having been pregnant or in the probability of miscarriage or stillbirth (Essock and McGuire 1989). Infanticide is an act that not only excites feelings of revulsion but one which at first sight seems wholly incompatible with evolutionary theory. Yet given the extreme parental investment demanded of human mothers, we should expect infanticide to be a line of last resort by a woman who believes that the investment in the current infant is not worth continuing. Although abortion has effectively removed the need for infanticide in many societies, in others it may be the only avenue for a woman who bears a child that she cannot support without compromising her existing family. In the United States, infanticide is mainly confined to young women who effectively deceive not only others but also themselves into believing that they are not pregnant at all. After giving birth alone and in secret, the young woman disposes of the child and may even convince herself that she has done nothing wrong. She believes that the child ‘belongs to her’ and as such she has a right to decide upon its life or death without state interference. In other cases, the woman convinces herself that the child was stillborn and that she has not committed any offence (Pearson 1998).

Mhoc02.fm Page 51 Friday, December 14, 2001 10:15 AM

  :      From the point of view of evolutionary theory, we would expect to see more frequent infanticide when the infant’s health is so poor as to cause doubt that it can survive to reproductive age. Daly and Wilson (1988) consulted cross-cultural evidence from 60 societies around the world, noting in each case the reasons that were given to account for infanticide. Twenty-one societies indicated that deformity at birth or serious perinatal illness would be grounds for infanticide. When the present economic circumstances are inadequate for child-rearing we would also expect infanticide to be more frequent. Fifty-six societies recognized infanticide as a response to the absence of male support for the child and economic hardship. Features of the infant also play a role in determining in its own life chances. Twins place an extra burden of pressure upon the mother and upon the economic resources available. Usually one rather than both of the twins is sacrificed and twins are less likely to be killed in societies where the mother has female relatives or friends to help her and is able to give up arduous work. An infant that is born too soon after a previous birth is also more likely to be killed. The mother must decide between the older child’s need for her milk and that of the new arrival—in line with evolutionary predictions, the child in whom she has already invested most is preferred. If the child is the product of a clandestine extramarital union, the chances of infanticide rise in line with the potential threat that the infant poses to the continued provisioning of older children by her husband. Contemporary data from Canada and France (Daly and Wilson 1988) support the idea that infanticide is related to the mother’s evaluation of her economic and social situation as well as her future likelihood of giving birth. Hagen (1999) has suggested that post-partum depression may be a psychological adaptation that disengages the mother from the infant, reduces attachment and may thus open the way to infanticide. He reviews a number of studies that confirm that post-partum depression is associated with inadequate emotional and material support from the father, family and friends. Another significant risk factor for depression is infant viability. Depressed mothers are more likely to have experienced difficult pregnancies and deliveries which are strong predictors of child morbidity and mortality. They are also more likely to have given birth to infants with health problems. From the woman’s reproductive condition even before the act of intercourse to the woman’s acceptance of her newly born infant, there are many points at which a less-than-optimum pregnancy can be knowingly or unconsciously halted. Once the baby is accepted, the women commits herself to nearly two decades of care and concern—small wonder that such an investment is carefully made.

Women as heavy investors For a woman birth is the beginning, not the end, of the story. The juvenile period is ‘a great selection funnel into which many enter but from which few emerge’ (Lancaster 1989, p. 65). Among carnivores about 10 per cent of animals born survive to

Mhoc02.fm Page 52 Friday, December 14, 2001 10:15 AM

      adulthood, the figure is about 25 per cent for non-human primates. Among human hunter-gatherers, half of the babies born survive to reproductive age. A critical factor in ensuring survival will be the mother’s ability not just to defend and feed her offspring but to ensure her own survival too. Without a mother, the life expectancy of infants is cut tragically short. As Mealey (2000, p. 341) succinctly puts it: ‘Desertion by one’s mother means almost certain death, whereas desertion by one’s father generally means only a reduction in resources.’ In the animal world, orphans do not do well alone. Young squirrel monkeys and rhesus macaques continue to have a strong emotional dependence on their mother even after they have been weaned. What happens when the mother disappears? Experimental studies of separation (Martin 1997) observe that initially both mother and infant appear extremely distressed and the infant gives repeated plaintive calls as if signalling its whereabouts. Within half an hour, both of them show rises in cortisol levels resulting from over-activation of the hypothalamic reaction to stress. Even when their behaviour begins to calm down, cortisol levels continue to rise and this has serious consequences for the immune system. They have fewer circulating lymphocytes (cells which attack antigens), lower levels of antibodies and a reduced antibody response. The relationship between losing a loved one and ill health may be all in the mind but it is far from imaginary. Animal behaviourists have observed the fate of orphaned primates in the wild. Jane Goodall (1986) followed the life of Flint who was born to an ageing female called Flo. He persisted with nursing until the age of 5 despite his mother’s attempts to wean him and he refused to give way even when a younger sister Flame was born. When Flame died, Flint became even closer to his mother, sharing her nest at night in the trees. When he was 8 years old, Flo died. Flint sat by his mother’s body for hours and became increasingly inert and depressed over the next few days. Despite attempts by his older siblings to coax him into activity, he developed gastric enteritis and peritonitis and three weeks later he was dead. A systematic study of Japanese macaques (Hasegawa and Hiraiwa 1980) followed the fate of a number of orphans. Seven out of nine who lost their mother in the first year of life died. The older orphans fared somewhat better. Of the 25 who lost their mother between one and 5 years of age, five disappeared or died. Others were adopted by female siblings or cousins and two infant females were even adopted by adult males. It seems, however, that getting adopted depends more on the determination of the orphan than the benevolence of the adoptive parent (Jolly 1985). Among hunaman langurs, adoptees seek out and cling on to alternative mothers without any sign of initial acceptance by the adoptive female. The ability of an infant to locate an adoptive parent depends on how much infant-sharing takes place in the group (and hence how willing the infant is to approach older group members), the independence and assertiveness of the orphan, the available number of foster mothers (including the number that are lactating and so can feed very young orphans) and the number of other juveniles in the group. These rival youngsters often try to exclude the infant from making

Mhoc02.fm Page 53 Friday, December 14, 2001 10:15 AM

  :      contact with their own biological mother and sometimes unintentionally exhaust the orphans with their unrelenting invitations to play. Fortunately, because of Western medicine, maternal death during childhood is a rare tragedy. When it does happen, the existence of monogamy, relative affluence and welfare provision mean that the father, other relatives or the state will assume the role of care-taker. But in the environment of evolutionary adaptation, such factors were not present. We can gain some idea of the fate of motherless children by examining the ethnographic and historical record. Ostfriesland is a north-western German coastal region. The births, marriages and deaths of the citizens who lived there between 1668 and 1879 were recorded in the church records. Voland (1988) focused on 870 families in which the last-born child either survived the death of one of its parents during the first year of life or was born alive after the father died. He examined how many would go on to survive until their first and fifteenth birthday. One-quarter died in the first year and 45 per cent died before the age of 15. In the first year of life, child mortality was twice as high when the mother died (33 per cent) compared to the father (18 per cent). Overall, 52 per cent of motherless children failed to survive until 15 years of age compared to 38 per cent of fatherless infants. When the Ache of Paraguay were still forest-living, on every occasion when a mother died in the first year of an infant’s life, the child died also. Even at later ages, the best predictor of child mortality was maternal mortality—motherless children were five times more likely to die than those whose mothers were alive at all ages and they were four and a half times as likely to be killed. On the reservation, far fewer adult deaths occurred. However, of the four children who lost their mothers, two also died before maturity while of the handful of fatherless children none died (Hill and Hurtado 1996). A study in rural Gambia (Sear et al. 2000) focused on child outcomes up to the age of 5. The death of the father had hardly any effect on the likelihood of his child’s survival but children who lost a mother were massively more at risk. If they were under the age of 2, these children were between 11 and 13 times more likely to die than those whose mothers were alive. The younger the child, the greater the danger posed by maternal death and it is not hard to see why. The Gambian toddlers who lost their mothers weighed significantly less than their peers. In our evolutionary past, mothers breast-fed for up to four years and her infant’s life literally depended upon her ability to sustain lactation. The mother’s supply of breast milk was, and in many places still is, the best predictor of whether an infant lives or dies (Hrdy 1999). To feed her baby, she required 600 to 700 additional calories every day creating a situation where she had to forage regularly and successfully to meet these needs, to say nothing of the extra calories she required to carry the growing baby who accompanied her on these expeditions. The composition of breast milk itself provides one of the strongest clues that she took her baby with her. Unlike the fat-rich solution of animals who must

Mhoc02.fm Page 54 Friday, December 14, 2001 10:15 AM

      leave their offspring for hours to forage for food or who have only a short time with the infant before climatic changes force weaning upon them, women’ s milk contains only 3 to 4 per cent fat and is 88 per cent water. To sustain an infant on such a weak mixture the mother must have carried it with her and fed it for several minutes every hour during the day. At night, she may have fed the baby (or more likely the baby latched on to the nipple and fed itself) three or four times. Beast-feeding is a positive feedback system—the more milk the baby takes, the more production is stepped up. In this way, the baby effectively calls the shots, ensuring the correct amount of nutrients for its weight and growth. Lactation carried benefits for both the mother and baby. For the mother, by suppressing ovulation, optimal birth spacing between children was ensured. Effective menstrual suppression depends upon many small feeds and those that take place during the night seem to be especially important. These trigger prolactin surges that are up to six times more intense than those that she shows during the day and it is prolactin that is responsible for the suppression of ovulation. The four-year gap between children in hunter-gatherer societies was doubtless a result of the mother’s continual breast-feeding. With agriculture and a more settled and predictable availability of food, mothers could grind grain and cook it to create an easily digestible food for infants which reduced the age of weaning to as early as six months. But this meant that mothers became pregnant again after shorter intervals and were placed in the evolutionary novel situation of caring for multiple dependent young at the same time. While we consider the low birth rate of Western populations as anomalous, Hrdy (1999) points out that a hundred thousand years ago, women produced as few as four children in a lifetime and this low birth rate seems to have been at the root of our very slow growth as a species. Anthropologists estimate that in the Pleistocene the human population was probably not more than about 100,000 people. It took 15,000 years for that population to double in size compared to our current doubling rate of 50 years (Hammel 1996). The investment costs of children have always been high and it is a wise woman who judges her timing well. It is the far more recent ‘excessive’ production of children that is the real historical anomaly. For the baby, breast milk brought the advantages not just of perfectly tailored and digestible nutrients but a second survival boost—ready-made immunological protection (Cunningham 1995). During pregnancy the foetus receives antibodies through the placenta from the mother. After the baby is born and before her milk arrives, the mother produces colostrum, a thick cream-coloured mixture that is rich in antibodies. By the fourth month of lactation, the baby receives as much as half a gram of antibodies each day. Because mother and infant, bonded together day and night by lactation, are exposed to the same pathogens, the mother is able to supply exactly the antibodies that the baby needs. Mother’s milk is especially geared to attack antigens that cause dysentery and diarrhoea, the most dangerous of potential infant killers.

Mhoc02.fm Page 55 Friday, December 14, 2001 10:15 AM

  :      For both, breast-feeding brought with it not only a prolonged but a pleasurable skin-to-skin contact. Suckling on the infant’s part stimulates the release of oxytocin on the mother’s. The chemical travels in the blood from the pituitary to the breasts, triggering the ‘let-down’ reflex—the contraction of muscles that ejects milk from the nipple. Oxytocin has effects on the brain too. It generates feelings of warmth, comfort and contentment (it is released during orgasm) and it may be that it is passed through the milk to the infant (Hrdy 1999). Both mother and child seem to be soothed by feeding and the experience of well-being may be as much to do with oxytocin as with the infant’s filling stomach and the mother’s relief at dispatching the heavy load of milk she has accumulated. This special kind of relationship with the infant was exclusively reserved for mothers. Though males of some species may deliver food to their young and some male birds produce crop milk (a mixture of partially digested food and mucous produced in the throat), males do not lactate. The female-specific nature of lactation, with the massive demands it makes on calorie intake and on time, was a natural consequence of maternal certainty and the already huge investment that the female had made. The bond between mother and infant is the primary social relationship (Fedigan 1982) and it extends far beyond milk provision. Nowhere is this more evident than among chimpanzees. Chimpanzees, like humans, are slow to mature. They are weaned at the age of 4 or 5, with the females reaching sexual maturity at 11 (though because of adolescent sterility the first pregnancy does not occur until about 13 years of age). The males may be able to produce sperm by the age of 9 but are not sexually mature until about 14 years of age. Childhood then is long and is spent in an intimate relationship with the mother. Despite the intense curiosity about the newborn shown by community members, the mother prevents any of them from touching the infant during the first six months of life. She provides not only milk but in later years she shares fruits, nuts and honey freely with her offspring. She fishes for termites and dips for ants, showing the offspring how to create and use the tools they need for this. She protects them from the volatile temper of adult males and is constantly watchful when adult males are close by. Although a band of chimpanzees may be composed of between 15 and 80 individuals, during waking hours they travel in smaller parties. Mothers usually travel alone with their offspring in matrifocal units, coming together at night to make their tree nests as a community. During the day, the mother seeks out areas that have enough food resources for herself and her infant and this means that adult females spend relatively little time together compared to the males. The mother–infant bond is strong and is broken only when the adolescent daughter must emigrate from the group to seek mates among neighbouring groups. Observations of mother–infant interactions among chimpanzees show a threephase development from infant dependence to juvenile autonomy (Marvin 1997). In the first stage, lasting until about 10 months of age, the infant has poor locomotor and communication skills and is unresponsive to adult attempts to communicate.

Mhoc02.fm Page 56 Friday, December 14, 2001 10:15 AM


    

Because of this, the mother maintains very close proximity, suckles the infant frequently and protects it from other troop members. This corresponds to the human age range from birth to about 15 months during which human infants are equally socially inept and mothers in a variety of cultures perform the same tasks. In the second phase, which for chimps extends from about 6 months to 4 years, the infant gains its milk teeth, begins to eat solid food in addition to suckling, becomes more adept at independent locomotion, masters communication skills that can be used over increasing long ranges to keep in contact with the mother and begins to communicate with other members of the troop. During this time, a secure attachment between the two becomes evident. The mother is used as a base for exploration. The infant leaves the mother but only for short excursions with frequent reunions, the infant (rather than the mother) becomes responsible for maintaining contact (although the mother maintains close visual monitoring of the infant), they reunite immediately when danger is perceived and the infant plays with other youngsters and practices communicative signals. This pattern shows strong parallels with human youngsters between one and five years. Bowlby (1969), in an explicitly ethological theory of human development, describes the development of attachment, the use of the mother as a safe base and the increasing ability of the child to internalize a working model of the absent mother as a source of security. At last, in the juvenile phase, the mother may have a new infant to care for and the youngster is independent of the mother in terms of feeding and locomotion. It possesses all the receptive and expressive communication skills required for integration into the social group. The juvenile travels with the mother but spends increasing time with other troop members and becomes more responsible for its own protection (though the mother may still intervene in an emergency). Again the parallels with human behaviour during middle childhood and adolescence are perhaps too obvious to note. Human mothers and infants form close bonds from the beginning of the infant’s life. Just 48 hours after birth, mothers can distinguish the cry of their own infant (Wiesenfeld and Klorman 1978). Babies reciprocate by distinguishing the sound of their mother’s voice (Field 1985) and the unique features of her face (Bushnell et al. 1989). During the first six months, mother and child engage in mutual imitation, smiling and vocal turn-taking. Between 2 and 4 months, studies using forehead temperature change indicate a positive reaction to the mother and a negative response to strangers (Mizukami et al. 1990). Between 7 and 11 months, the infant is able to display evidence of this specific attachment in its behaviour by responding with distress to the disappearance of the mother (Schaffer and Emerson 1964). For the majority of children, the mother is the primary attachment figure (Kotelchuck 1976) and, when upset, children are far more likely to seek comfort from their mother than their father (Cohen and Campos 1974). Like chimpanzees, human mothers also take the primary responsibility for childcare worldwide (Brown 1991; Ember 1981; Geary 2000). In hunter-gatherer societies,

Mhoc02.fm Page 57 Friday, December 14, 2001 10:15 AM

  :      like the ones in which humans evolved, women provide the majority of the calories consumed and, unlike men who share their meat throughout the group, they share their food only with their family. Most food comes from foraging and women undertake these expeditions two or three times a week, carrying their infant on their back. The infant is in close proximity to the mother at almost all times—they sleep, eat and travel together. Among the Kung San, fathers provide less than 7 per cent of childcare (West and Konners 1976). The greater involvement shown by the mother rather than the father is not limited to infants. Between the ages of 3 and 6, children spend between three and ten times as much time with their mother than with their father (Whiting and Whiting 1975) and by the ages of 4 to 10 years, they are with their mother two to four times longer than with the father (Whiting and Edwards 1988). Even in this liberal era, mothers do the majority of childcare. This has been found in observational studies of the home even when parents are visited at early evening periods—when many mothers have been at home all day and fathers have only just returned. Mothers perform routine care-taking chores (such as feeding and changing) between two and four times more often than fathers and spend between one-and-ahalf to twice as much time holding, touching or playing with their child (Belsky et al. 1984). At infant ages of 1, 3 and 9 months, mothers more than fathers respond to, stimulate, express positive affection towards and care for their infant. As Belsky et al. (1989, p. 142) concluded from their study: ‘The only thing that men did more of than women was engage in personal leisure activity (i.e. read/watch television)!’ This differential attention to children is not merely a function of the fact that mothers are more frequently in the home. In Israel, where both parents work and children are cared for during the day in nurseries by metapelets, mothers are still far more frequently involved in child-care duties and in spontaneous interaction with the child at the end of the day than fathers (Sagi et al. 1985). Lamb et al. (1982) took advantage of innovative Swedish legislation that allowed either parent to take nine months paid leave after the birth in order to examine child-rearing under the most liberal of gender role conditions. The leave could be divided between the parents as they wished. The take-up of this opportunity was revealing. Between the introduction of the legislation in 1974 and data collection in 1979, less than 15 per cent of fathers took leave and the majority of those that did took one month or less. Most fathers took this leave at the same time as their wife and the extent of their actual contribution to childcare is hard to ascertain. The researchers originally identified 52 couples, half of whom planned to take paternal leave for one month (the ‘nontraditionals’) and half of whom had no plans for paternal leave. At re-contact one year after the birth, only 17 of the original 26 ‘non-traditional’ fathers had in fact taken the month of leave. The researchers reclassified the families on this basis and analysed observational data that they collected in the home. In both traditional and non-traditional families, mothers displayed affectionate behaviour, vocalized, smiled, tended, held, disciplined and soothed the infant more than fathers. As the

Mhoc02.fm Page 58 Friday, December 14, 2001 10:15 AM

      authors concluded: ‘maternal and paternal interactional styles revealed in many previous studies may be biological in origin’ (Lamb et al. 1982, p. 134). After all, breast-feeding aside, men as just as capable as females of meeting children’s needs. Why they should do so less than females, particularly when paternity is virtually certain and when both parents work outside the home, remains unclear. Sarah Hrdy (1999) believes that the perceptible imbalance (women’s ‘extra shift’ as it has become known) results from an originally small, probably innate difference in parental responsiveness and solicitude. A team of researchers asked parents to listen to two recordings of a baby crying (Stallings et al. 1997). One was the sound of a hungry baby wanting to be fed, the other was the altogether more distressing pain cry of a baby being circumcized. They monitored the reactions and hormonal changes of mothers and fathers. While both responded equally and vigorously to the sound of pain, the mother showed greater responsiveness to the routine hunger cry. Hrdy suggests that this difference snowballs on a daily basis. The infant cries, the mother responds just a little quicker than her husband, the husband decides not to interfere as the baby is now settling again, the baby develops a slightly greater attachment than she had before to her mother. Soon the baby starts to show a distinct preference for the mother and the father is pushed (happily or unhappily) from the picture. But men may have another evolutionary force working against their involvement with young children. Human infants show greater fear of strange males than of strange females (Greenberg et al. 1977) and an examination of non-human primates suggests why this might be the case. A substantial threat to infants’ lives comes not only from predators and natural hazards but from strange males who enter the group. Newly-arrived males who achieve dominance in the group use infanticide as a means of dispatching the offspring of rival males, bringing the mother back into oestrus and fathering their own infants (Hausfater and Hrdy 1984). Sexual selection is not a kind process and ruthlessness about the fate of rivals’ genes can be advantageous to males. The female often resists (Palombit 1998) but if she cannot defend her infant, then her only option is to begin reproduction again. She will do this even with her infant’s killer because he has demonstrated his dominance and this acts as a guarantee that he will be equally ruthless in the defence of his own progeny. Pair-bonding may even have begun as a response to the threat of infanticide from strange males, with females trading sexual exclusivity for a male who could provide effective protection. Among chacma baboons and mountain gorillas 37 per cent of infants are lost to infanticidal males. This is a very high proportion of the female’s reproductive success and in these species females contribute disproportionately to maintaining their relationship with the resident dominant male—they perform 90 per cent of the grooming that takes place between them (Palombit 1998). The risk to infants from unrelated adult males has also been noted in humans. Despite years of research on child abuse, it was not until evolutionary theorists turned their attention to it that the extraordinary relationship between stepfathering

Mhoc02.fm Page 59 Friday, December 14, 2001 10:15 AM

  :      and abuse was established. Stepchildren are 65 times more likely to be murdered than are children living with their two natural parents (Daly and Wilson 1988). In Britain 52 per cent of babies killed in the home were murdered by a stepfather (Watson 1995). This is not to suggest that there is anything adaptive about men injuring or killing children. It does, however, point to the fact that the final limits of tolerance are reached more quickly by men who are not themselves the biological father of the child. Biological relatedness seems to act as a shield against the nerve-shattering, nightly crying of a difficult baby. In such situations, the baby’s well-being depends upon the strength of the mother’s bond to it and upon her ability to deflect or protect her infant. Even in intact biological families, the amount of care-taking and play between the father and the child depends upon the quality of the father’s relationship with the mother. The mother’s attachment to her child, however, is not affected by the quality of the couple’s relationship (Belsky et al. 1984; Parke 1995). Perhaps most critical for the importance of mothering is the nature of male reproductive strategy. Men’s quest for novelty and variety in their sexual life increases their chances of both death and desertion, leaving the mother firmly holding the baby. In animal species, polygyny is associated with earlier death among males. In part, this results from the dangers of male–male competition and from the generalized risky behaviour of young men. In England and Wales, the male-tofemale ratio for deaths from external causes is at its most extreme at ages 15 to 24 years when it reaches nearly four to one (Office of Population Censuses and Surveys 1995). However, even males who survive past their peak reproductive years still die earlier than do females. If a father dies, and he is more likely to do so than the mother, it is she who must shoulder the full burden of infant care. About 7 per cent of children living in single-parent homes in the United States do so as a result of the death of one parent (Weinraub and Gringlas 1995). But men’s polygynous inclinations more often increase the likelihood of desertion. Even in officially monogamous societies, men seek more pre- and extramarital affairs than do women (Daly and Wilson 1988; Fisher 1993). Men’s preference for youth and physical beauty in sexual partners means that as their wives age, younger women become increasingly attractive and divorce more likely. After divorce, men are more likely than women to remarry, to select younger partners and to produce further children in their second marriages (Buckle et al. 1996; Johanna et al. 1995; Kaar et al. 1998). When divorce happens, the burden of parental care is almost always taken up by the mother—in no country do mothers abandon their children at the rate which fathers do (Geary 2000). In the United States, custody is awarded to the mother in over 85 per cent of cases (Furstenberg et al. 1983). (Where fathers remain as the principal care-taker, it is the result of the mother’s incompetence or her desertion—it is rare for fathers to contest the custody of the child (Greif 1985.)) After divorce, only one in six fathers maintains regular contact with his child and about half of fathers who are not living with their child fail to pay any monetary support for their upkeep.

Mhoc02.fm Page 60 Friday, December 14, 2001 10:15 AM


    

Males do not always wait until after marriage to desert. Increasingly, they jump ship without bothering to marry at all. While the proportion of children in the United States living in single-parent families because of divorce declined from 73 per cent in 1970 to 62 per cent in 1990, the proportion of children born to unmarried women rose from 6.8 per cent to 30.6 per cent. For whites, premarital births rose from 9 per cent of all births to 22 per cent. Among blacks, the rise was from 42 per cent to 70 per cent. Sometimes males choose to desert and sometimes they are pushed because they can make no meaningful financial or emotional contribution to the household. Either way, the fact remains that unmarried fathers who seek custody of their children are rare relative to the number of men being pursued by government agencies to provide financial support for their illegitimate children. Mothers must take the parenting job on or find another woman (often their mothers) who will. For all these reasons, nature has crafted a special relationship between mother and child, mediated by mutual attachment and aimed at increasing the chance of an infant achieving independence and reproducing anew. Fathers contribute in their own way to a child’s safety and provisioning—a good father can improve his child’s social and economic future. (Although, because good fathers tend to marry good mothers, some of the apparent advantages accrued by their children may have more to do with his wise choice of mate than with his direct contribution to parenting, see Amato 1998; Geary 2000.) Mothers know that a good partner is worth having and, as we shall see, compete for these high-quality fathers. But through our evolutionary past, mothers, from an infant’s viewpoint, have always been essential while fathers have been an optional extra.

Pursuing half the point Sexual selection is at the very core of Darwinian theory. Understanding the difference between men and women in their parental investment was the key that opened the research doors (and academics poured through them) to the study of human reproductive psychology. The reproductive task can be broken into two stages and two kinds of problems: finding the right mate and raising offspring to maturity. Males direct a greater part of their energy to the former, females to the latter. The study of mating drew far more research attention than the study of parenting. It is tempting to think that this was the result of the predominance of male researchers whose interest was understandably in their own sex’s agenda and desires. And seeming to corroborate this, there was a disproportionate interest in what men, rather than women, desire in a mate. There were far more studies of women’s facial structure and body shape in terms of their desirability than of males’ (Etcoff 1999). This was remarkable because theory and data tell us that men are less choosy than women—if anything the emphasis should have been upon the criteria by which females make their mating choices. But I don’t want to dwell on this too heavily—the balance is redressing and we understand much more now about what women want. Besides,

Mhoc02.fm Page 61 Friday, December 14, 2001 10:15 AM

  :      the relative lack of attention to female preferences was hardly novel. It can be traced back well over one hundred years to the disbelief that greeted the critical role Darwin (1871) accorded to female choice. He pointed out that since males compete for indiscriminate copulations while females do the choosing, the logical gateway to the composition of the next generation was female preferences. But as Ridley (1993, pp. 131–2) points out: ‘Of all Darwin’s ideas, female choice proved the least persuasive. Naturalists were quite happy to accept that male weapons, such as antlers, could have arisen to help males in the battle for females, but they instinctively recoiled at the frivolous idea that a peacock’s tail should be there to seduce peahens. . . . Any theory, it seemed, was preferred to the idea of female preference for male beauty.’ More surprising was the relative neglect of the second component of sexual selection—parenting. With men less involved in the task than women, evolutionary researchers (with notable exceptions, e.g. Bowlby 1969; Geary 2000; Hrdy 1999) dropped the priority that they accorded to this altogether critical aspect of sexual selection. Yet, as Hrdy (1999, p. 81, original emphasis) puts it so aptly: ‘Unless mating results in production of offspring who themselves survive infancy and the juvenile years and position themselves so as to reproduce, sex is only so much sound and undulation signifying nothing.’ Raising offspring is as critical, if not more so, to reproductive success as wise mate choice yet women’s capacity to perform this feat—lasting over years, costing them dearly, constraining their own life choices, shaping the next generation—was taken as unproblematic. While the decisions involved in mate selection loomed large, the decisions about birth spacing, breastfeeding, weaning, alloparenting, the balance between infant protection and the fostering of juvenile independence took a back seat in most evolutionary psychology works. Mothering, until the arrival of Sarah Hrdy’s exceptional book Mother Nature, was taken for granted as an innate and straightforward exercise. Hrdy (1999, p. 69) was forced to underline again and again the critical importance of the mother as offspring dependency increases: For species such as primates, the mother is the environment, or at least the most important feature in it during the most perilous phase of any individual’s existence. Her luck plus how well she copes with her world—its scarcities, its predators, its pathogens, along with her conspecifics in it—are what determine whether or not a fertilization ever counts.

I think that an important reason for this neglect derived from the way parental investment was understood. Females invest heavily and are the prizes for which low-investing males must compete. The image is irresistible—a female sits on the sidelines, passively awaiting the arrival of a male who has demonstrated by his dominance over his rivals his undisputed right to claim her reproductive favours. It was male competition—risky, desperate, and tinged with sexual anticipation—that became the attention-grabber. The long hard road that lay ahead for the female— mundane, time-consuming, repetitive—seemed so much less exciting. And the

Mhoc02.fm Page 62 Friday, December 14, 2001 10:15 AM


    

focus on male winners and losers informed a belief that women, and the mothering that they did, was all much of a muchness. There was little to choose between them either in terms of strategy used or quality and quantity of children produced. The number of offspring that a male could sire ranged from a stunningly high in the hundreds to a catastrophic zero. Males were brilliant success stories or dismal failures, while women, we were told, took a moderate middle road, each producing as many offspring as her reproductive span would allow. Enter oestrus, copulate, get pregnant, lactate and begin again until menopause brings the whole repetitive cycle to a halt. If this were true, there would be no variability in female reproductive success and so little of interest to understand about differences between females in their choices and strategies. And if there were no differences, there could be no selection and no adaptation. The driving force of evolution, we were told, was male success or failure, with females relegated to convenient sperm receptacles and breeding machines. But we know there is considerably variability between females in their ability not just to produce children but to raise them. The variance may not be as great as between males but that is irrelevant because females are not in competition with males, they are in competition with other females. If you leave six children behind you instead of my one, you leave 300 per cent of your genes in the next generation as compared to my 50. Women were choosy investors alright but to opt out of the competition altogether was no more an adaptive option for a female than it was for a male. The difference was that females were forced to carry out their competition in a much lower key (for reasons I will discuss later) even though competition between them was just as important. The attraction of male competition derives from its dramatic life-or-death quality. The life or death of a male’s genes in future generations is at stake and for this it is worth a male risking his own death in the here and now. But, paradoxically, our preoccupation with the importance of male competition obscures a critical fact—males competed because they were less important than females in the reproductive process. The traditional evolutionary argument highlights the glittering reproductive prizes for which males competed and their risk-taking is seen as the route to the winner’s podium. But while males seem to be the stars of the show, we should bear in mind that they are essentially freeloading on women’s effort. Consider this: if we knew our planet was about to be struck by a meteor and only 100 people could be saved on an underground bunker, what proportion of men and women would you put down there? My suggestion would be about 10 men and 90 women. Ten should be able to do an adequate job of impregnating all the women and the fewer the men, the fewer the calories they would consume and the lower the competition between them would be. But imagine we had made the opposite choice—90 men and 10 women. What kind of hope would we have for the future of the human race? The fact is that the majority of men are, biologically speaking, dispensable but when the number of women drops too far, our future looks bleak. In fact by looking around us we can see that the sex ratio remains roughly equal and Fisher (1930) explained why. Imagine that the sex ratio shifts so that there are

Mhoc02.fm Page 63 Friday, December 14, 2001 10:15 AM

  :      not enough men around. Any boy that gets born now does very well indeed reproductively speaking and so those parents that are disposed to produce sons in this situation also produce more grandchildren. Genes that bias towards sons rather than daughters will start to spread but as the sex ratio reaches equality the reproductive advantages of sons begin to diminish. The same logic holds for a dearth of females and brings the system back into equilibrium. Eventually it stabilizes at a near equal number of sons and daughters. But in some species extraordinary sex ratios can occur (Hamilton 1967). They can be the result of ‘driving’ genes that seek to get into the next generation at the expense of others. If a driving Y chromosome emerges, consistently pushing maleness ahead of femaleness, it is estimated that in just 15 generations—a blink of the evolutionary eye—all females (and hence the species) will have been wiped out. But now imagine a driving X chromosome. Its disastrous effects are much slower. It spreads only about one-third as fast because X chromosomes spend two-thirds of their time in male bodies and the progress towards extinction is delayed even further under polygyny because so few males are needed to fertilize many females. In short, a loss of males is far less hazardous to a species’ health than a dearth of females. And when we look for species with imbalanced sex ratios, it comes as no surprise that they are female heavy. In some insects, there are 20 females born to every one male. In the next chapter I will try to convince you that men’s greater willingness to take risks and to live their lives in the fast lane is a testament not just to the reproductive gulf between male winners and losers but to men’s smaller contribution to parental investment. Men hold their lives less tightly in their hands because they have been more marginal than women in ensuring the successful survival of their offspring.

Mhoc03.fm Page 64 Friday, December 14, 2001 10:15 AM

 


High stakes and low risks: Women and aggression A sex difference in parental investment cuts two ways. It has implications for the evolutionary psychology of females as well as males. But it is men who have received the lion’s share of attention. Sex differences in aggression, coalition formation, status seeking, risk taking and self-esteem have all been explained by the fact that males needed to compete for mates while females did not. Females became the default sex—their psychology was determined by the fact that they were not males, and hence passive, cooperative, risk-avoidant and lower in self-esteem. In this chapter I want to redress the balance—to weigh the seesaw of theoretical implications a little more heavily on women’s side. Certainly male aggression was driven upwards by competition for mating opportunities but at the same time—and just as importantly—female competition was held in check by a mother’s need to ensure her own survival and, with it, her reproductive investment. Women had just as much to gain from competition as men but the crucial dependency of an infant on its mother meant that female competition had to take low-key and indirect forms. To correct the theoretical gender balance, we need to follow through (and question) the three main premises that have traditionally been used to explain the sex difference in aggression and risk taking.

Premise 1: Dominant males get the biggest prizes The first premise is that the desire to mate with multiple females causes intense male competition. If 10 per cent of men have a monopoly on 50 per cent of the female population, other men are faced with the possibility of going to their graves childless, unless they fight for their share of reproductive opportunity. Men are therefore always pitted against one another in an attempt to beat their neighbour’s share in the gene pool of the next generation and the route to that prize is to take control of as many mating opportunities as possible. The disparity between male winners and losers is great. Men who win, win very big. Those who lose, leave no descendants whatsoever. In short, the variance in reproductive success among men is greater than that among women. (Compare Ismail’s 888 children with the current female champion’s total of 69.) The most successful 5 per cent of men can father dozens of children so the pay-off for winning is large. Bigger prizes warrant greater risks.

Mhoc03.fm Page 65 Friday, December 14, 2001 10:15 AM

    :   


Males engage in risky physical aggression because the possible return offsets the risks. Remember that we are not speaking of men consciously and deliberately vying with each other to inseminate women. Male sexual competitiveness was forged long before the advent of our species when males who fought for females left behind more of their genes than their more restrained peers. In primate species, including our own, male competitiveness does not always seem to be about females (although a remarkable number of assaults between men are directly connected with romantic triangles) but they are about achieving dominance over other males, which is associated with greater reproductive success. In these days of contraception, some people find this whole notion preposterous. Victors of Saturday-night brawls, even when those fights are about women, rarely have impregnation in mind. Contraception has effectively cut the link between sex and reproduction and the fact that we have welcomed contraception so widely is evidence that human reproductive success was a by-product of our enthusiasm for sex, not for making babies. As long as sex naturally led to pregnancy, selection only needed to ensure a keen interest in sex. But contraception has been with us for half a century and this is insufficient time to undo hundreds of thousands of years of evolutionary pressure. The fact that men desire women (and lots of them) and are willing to fight for them is part of male evolved psychology. A glance at criminal statistics from any nation in the world confirms that men are indeed more likely to engage in criminal violence than are women (Kruttschnitt 1994; Simon and Baxter 1989). On average, 90 per cent of violence on our planet is committed by men. The same message comes from laboratory studies of aggression, despite the often artificial and contrived nature of these studies (Eagly and Steffen 1986; Hyde 1986). Men deliver higher voltages of electric shock to a rival than women do. Men are also higher than women on verbal insults and abuse which may not be dangerous in and of themselves but which are often the precursors to more serious physical assaults. On psychometric inventories, men rate their own aggression as more intense and more frequent (Buss and Perry 1992). Male competition can also account for differences in aggression between men in different sectors of society and different age groups (Daly and Wilson 1988). Among more affluent sectors of society dominance can be achieved by flaunting a high income, conspicuous material goods or membership of high status cliques—all of which form the basis of much male competition. But among the poor, males have fewer economic opportunities and the competition between males is more physical and more direct. Criminological studies confirm an association between violence and socio-economic status. Married men would be expected to show less involvement in male–male aggression because they have already succeeded in gaining control of at least one woman’s reproductive future. Again studies confirm that unmarried men have higher rates of criminal assault than comparably aged married men. The peak age for criminal assault among men is between the ages of 18 and 24 which

Mhoc03.fm Page 66 Friday, December 14, 2001 10:15 AM


    

corresponds to the time when male’s sexual drive and physical strength are at their peak. It is the time when young men seek to experiment with sexual partners and to select a long-term mate. It is also a time when they are most able to handle themselves in a fight. So the higher rate of male aggression compared to that of women derives from the greater variance in male reproductive success. Because the reproductive rewards of competition are so much greater for a male than a female, men are more willing to risk their lives in pursuit of them. This brings us to a very central question: just how big are the rewards? The majority of men throughout the world marry monogamously, either by virtue of legislation or because, even where polygamy is legal, very few men have sufficient resources to support multiple wives. This makes contemporary human data on the mating consequences of dominance problematic. Men (and probably women too) are also reluctant to allow DNA tests of paternity which are important in establishing the relationship between dominance and reproductive success. To make matters worse, the ‘demographic shift’ as it called has led to a novel situation where social status is in Western societies inversely correlated with number of children (but see Perusse 1993). For all these reasons and because male–male competition arose long before humans appeared, ethological data are important. Recent studies have begun to suggest that the advantages of dominance are not as great as we might have thought and it was female choice that threw something of a spanner in the works. For a male to risk his life, the benefits of success must be very high indeed and in many species they are. Studies of Northern elephant seals show that 5 per cent of the males can sire as many as 80 per cent of a year’s pups—a very big genetic prize (LeBoeuf and Peterson 1969). But among our closer relatives, the primates, the prizes for dominance may not be nearly as high. Until recently, male reproductive success was established by counting the number of copulations that they achieved. This was never a very satisfactory state of affairs (Martin et al. 1992) because some copulations do not end in ejaculation and even those that do are no guarantee of pregnancy. Primate males mate with females outside their period of receptivity and even with pregnant females. The fertilized egg may fail to implant or the male’s sperm may be beaten by those of another male who copulated with the female before or afterwards. In other words the male who achieves the greatest number of copulations may not be the one who fathers the most children. In the past twenty years it has become possible to establish paternity with much more certainty using biological techniques. Paternity exclusion analysis can rule out as a possible father any male who lacks a particular allele carried by an offspring but not its mother. DNA fingerprinting compares and matches bands of DNA appearing in the infant with both its mother and possible fathers. These techniques have produced some unexpected findings. A number of studies failed to find the expected relationship between dominance and paternity in the most popular species for study—Rhesus macaques (see Berard

Mhoc03.fm Page 67 Friday, December 14, 2001 10:15 AM

    :     et al. 1993; Inoue et al. 1993)—although some studies have reported positive results (de Ruiter et al. 1994; Melnick 1987; Pope 1990). A stronger relationship has been reported for Mandrills (Dixson et al. 1993) and Patas monkeys (Ohsawa et al. 1993). The data are far from conclusive and suggest that correlations are found for some years and not others, among some groups and not others. When they are found (see Smith 1994) they range in magnitude from very low (r = 0.06) to quite high (r = 0.77). But remember that even a correlation as large as 0.77 means that only about 50 per cent of the variance in reproductive success is attributable to rank. Bercovitch (1991) has argued that the size of the correlations may be artificially inflated by including sub-adult males. Since these males neither copulate nor have rank they automatically bias the data in favour of a positive relationship. But others have disputed this by showing that the correlation is even bigger if sub-adult males are included in the analysis (Smith 1993). In fact, turning the argument on its head, Smith points out that it is these younger males who seem to be particularly attractive to Macaque females and, far from achieving little mating success, these sub-adults are doing very nicely. In fact it is they who are largely responsible for the relationship between dominance and mating success. Smith goes beyond mere correlation to examine the temporal ordering of dominance and mating success over the 15 years he spent studying the macaque group NC7. His results invert the assumed causal relationship. Before achieving alpha status males No. 3 and No. 5 sired an average of 7.8 and 3.7 offspring per year, respectively; during their last five years at alpha rank, these two males sired on average only 1.2 and 1.4 offspring per year respectively. Thus, while rank and RS are correlated, the temporal changes in both suggest that if there is a simple causal relationship between the two, it is the latter which influences the former, not the reverse. Our results, therefore, do not suggest that high rank provides NC 7 males access to a larger number of estrus females. (Smith 1993, p. 478)

Among wild baboons also, the individual who achieved dominance and fathered 81 per cent of offspring during a 4-year period, sired 6 out of 14 infants born in the 4 years before he achieved dominance (Altmann et al. 1996). The young males that the females choose are those that subsequently rise in rank. So it seems there are two possibilities—either females sense those qualities that will later cause the male to become dominant and are placing their reproductive bets with greater than chance accuracy or the females are the real king-makers. And an experimental study that allows researchers to manipulate variables in order to establish a clearer cause–effect relationship suggests that a male’s relationships with females is vitally important in determining his rank. Raleigh et al. (1991) found that low status males had low levels of a neurotransmitter called serotonin. They used drugs to increase the monkeys’ serotonin levels and watched their subsequent behaviour. These monkeys rose in social status within the group and, most importantly, before they did so they showed an increase in their affiliative behaviour towards the females. In return they received the support of females in dominance interactions and began to rise up the

Mhoc03.fm Page 68 Friday, December 14, 2001 10:15 AM


    

social ladder. We shall have more to say about the powerful effects of serotonin on aggression shortly but for now we should note that its presence in the frontal cortex and the amygdala seems to be strongly implicated in prosocial behaviours such as grooming and affiliation. The result of this ‘act nice’ strategy is that females favour friendly males both as sexual partners and as group leaders. Dominance does not guarantee more offspring because a male’s success can be influenced by many factors such as the size of the group (a bigger group means more males have to be excluded from copulations and to make matters worse the females may cycle in synchrony, making it hard to guard more than one female at a time). In captive groups that are provisioned by humans, dominant males may be less able to control the behaviour of other group members and, through provisioning, young males may achieve sufficient body mass to make an earlier challenge to the dominant male. Males also diversify in their mating tactics. The strategy of dominant males is usually a direct one of mate guarding—they may tolerate other males mating with a female but not at the time of peak oestrus when she is most likely to conceive—made possible by their strength and strong coalitions with supporting males. Subordinate males, however, resort to sneak copulations with the full cooperation of the females who do not always show a preference for the dominant male. They may also use charm instead of stealth—after a period of grooming and food sharing, they escort the female to a quiet spot away from the troop for a period of hours or even weeks. Tutin (1979) found that only 2 per cent of chimpanzee matings took place during these consortships but 50 per cent of the offspring born were conceived during them. There is something about these matings (perhaps the absence of sperm competition from rival males) that makes them very effective forms of conception. And there is the matter of female choice. Females have their own preferences and they can thwart a male’s attempt at copulation and consortship by failing to cooperate or by protesting loudly. Other males will respond to a reluctant female by attempting to oust and replace her suitor. Female preference for novelty is also a problem for the dominant male. Among Rhesus macaques, genetic analysis revealed that 4 out of 11 infants born had been sired by males from other social groups (Berard et al. 1993). In a multimale chimpanzee group, group members fathered 7 out of 13 infants born with the remainder apparently sired by males from neighbouring groups with whom the researchers were not even aware that the female had mated (Gagneux et al. 1997). DNA fingerprinting was performed on a chimpanzee group that had only one resident male—a situation where effective mate guarding ought to be easy. But even here he fathered only three of the four infants born between 1985 and 1991 (Sugiyama et al. 1993). Again, the fourth infant seems to have been fathered by a male from a neighbouring group. So female preference based on youth, novelty, food-provision, grooming, protection for her and her infant or perhaps just plain whimsy does not always coincide with male dominance hierarchies. Even if a male achieves dominance, his tenure at the top may be brief and a snapshot correlation of dominance and reproductive success does not do justice to the

Mhoc03.fm Page 69 Friday, December 14, 2001 10:15 AM

    :   


complexity of the situation. Dominance can sometimes last only a year or less and, however much the male can capitalize upon his social kudos to inseminate females, this may have little overall impact on his lifetime success rate in relation to other group members. As Altmann et al. (1996, p. 5800) found in their study of baboons: ‘Lifetime reproductive success cannot in general be deduced from short-term measures of mating success. More males will contribute to the gene pool than predicted from highly skewed short-term reproductive success, resulting in greater effective population size.’ Another long-term factor that has thus far received little attention is the inclusive fitness achieved by a dominant male. Success depends not only on insemination and infant survival but also on the extent to which the progeny attain adulthood and are successful reproducers themselves. A male whose offspring produce no offspring themselves is a genetic cul-de-sac. And the task of ensuring offspring survival is a female one that can be aided by monogamy. Although polygamy is legal in the majority of human cultures, in fact only about 15 per cent of men in such societies are polygamously married (Murdock 1981). Women, as a limiting resource, are expensive. They have to be paid for with a bride price and then maintained in the future, including the children that they produce— relatively few men have the surplus resources to afford them. In agricultural societies where resources can be stored and money accumulated, richer men are able to afford the most wives and polygamy increases the variance in male reproductive success. Among the Xavante where monogamy is not socially imposed, the reproductive variance for men is 12.1 compared to 3.6 for women. For the Kung!, male variance is 9.27 compared to the female 6.52. Monogamy, in a perfect world, equalizes the variance of males and females. In a study of the Pitcairn Islands where all the inhabitants are descendants of 28 marooned sailors and Polynesian women, monogamy was legislated and the subsequent patterns of marriage and childbirth documented. Here the variance in reproductive success was indeed equal for men and women (Brown and Hotra 1988). However, before monogamy was imposed, the founding population of males had shrunk from 15 to one, predominantly through murder! In many contemporary societies the greater male variance seems to be the result of serial monogamy, with men being more likely to remarry and produce further children in second and third marriages. Both polygamy and monogamy have been part of the human mating pattern since the beginning of our species. But while males aspire to multiple partners, most fail. Although the roots of male–male aggression lie in their preoccupation with mating, the prizes and the chances of success may have been overestimated.

Premise 2: Male competition causes men’s generalized taste for risk So far we have been concerned with male-on-male aggression. But Wilson and Daly (1985, 1993) have taken the argument further. They propose that as a consequence

Mhoc03.fm Page 70 Friday, December 14, 2001 10:15 AM


    

of mating competition, risk taking in general is higher in males than in females because advertising is a useful strategy for increasing one’s chance of winning. The more fearless a man seems to be, the less likely a rival is to tangle with him. One way a man can do this is to show other men that he positively seeks out situations that other men would avoid. Hence risk taking becomes generalized to non-competitive situations and a crucial ingredient of advertising is that it should be done publicly, the better to intimidate other males. If male fitness derives from success in risky competition, then males are expected to join such competitions willingly. . . . A taste for risk, with or without competition, may also be manifested in many other spheres than the violent conflicts we have thus far considered . . . we furthermore expect an evolved inclination toward the social display of one’s competitive risk-taking skills and again this should be especially a masculine trait. (Wilson and Daly 1985, p. 66)

It is this ‘taste for risk’ that Wilson and Daly (1985) believe lies behind men’s greater involvement in gambling and drug use, and their greater vulnerability to injury and death from external causes, especially car accidents. But these same sex differences would also be predicted by postulating that men’s risky behaviour derives from their relative lack of concern with their own safety compared to women. In addition, this proposal would also account for other sex differences that seem quite unconnected with, and even detrimental to, advertising success in male–male competition. Take health care. Women are more fearful of illness than men (Heikkila et al. 1999). The strongest predictor of preventive health care is gender (Harris and Guten 1979). Women rate the importance of health higher than men, know more about health issues and are more likely to track the status of their health (Umberson 1992; Waldron 1988). Women visit the doctor more often than do men (after controlling for gynaecological and obstetric visits) and more often go for preventive care. Umberson (1992), investigating why the health benefits of marriage are greater for men than for women, found that women are significantly more likely to control the health-related behaviour of their spouse than are men. It stretches the bounds of credulity to argue that men avoid doctors so that they can use an undiagnosed illness as an advertisement of their taste for risk. Compared to women, men also show lower fear of crime. They are more willing to walk unaccompanied in unlit and dangerous areas of cities despite the fact that men are more likely to be victims of crime than women. Given that nobody is there to witness their demonstration of bravado, it seems outlandish to suppose that exposing oneself unobserved to assault and robbery is a form of male advertisement. Then there is suicide. Between the ages of 25 and 44 years, when the quality of life is perhaps at its peak, men commit suicide four times as often as women in Europe and the United States (Department of Health 1994). The greatest disparity in the suicide sex ratio occurs at the same age at which males display other behaviours that

Mhoc03.fm Page 71 Friday, December 14, 2001 10:15 AM

    :   


suggest a relative lack of concern with their own survival, such as homicide and other acts of violence. Do men really kill themselves to demonstrate their bravery, given that they will not be around to capitalize on their new-found status? My aim here is to develop a more detailed understanding of the mechanism that produces the sex difference in risk taking. Wilson and Daly (1993) are agnostic as to the precise psychological mechanism that generates men’s heightened taste for risk but they suggest a number of candidates. Some of these propose an active attraction to and appetite for risk (valuing present excitement more highly than longevity, intensified desire for the fruits of success, finding the adrenaline rush pleasurable and overestimating one’s ability to control the risk). Two others (underestimating objective danger and not caring whether one lives or dies) imply a failure to experience sufficient fear to inhibit risk taking. My argument will be that the data support their last suggestion—men treat their lives more casually than women. But while Daly and Wilson see indifference to personal safety as a tool in the armoury of male–male competition, I will argue that it is equally strongly connected with the fact that male survival is altogether less critical to ensuring their infant’s survival than women’s. Because of this, natural selection has operated on women to raise their inhibitions about risking their own life.

Premise 3: The rewards of female competition are low The male-focused argument tells us that women’s aggression is lower because they do not need to compete with one another for the chance to copulate—men’s polygynous inclinations mean that men are more than happy to oblige any woman in search of sex. In any case, women have little to gain, reproductively speaking, from multiple matings. Once pregnant, a woman gains nothing by wasting energy on seeking extra mates. Females, released from the need to compete for copulations, compete for the resources needed to enable and sustain pregnancy and to raise their offspring—chiefly food. Since, the argument goes, a single piece of food is of little significance, we find that female aggression is less intense than males’. But let us look at the sex difference in the reproductive contest from a typical primate perspective. For a male, we have seen that there is a less-than-perfect relationship between dominance and paternity. Achieving dominance is a very risky enterprise and deaths are not uncommon. If the male succeeds he increases his opportunity to copulate with an oestrus female but she will mate on average with 12 other males as well and experience 138 copulations for every infant born (Wrangham 1993). Half of these may be with males whom he cannot even challenge since they come from outside his group. About 50 per cent of infants born will not achieve adulthood let alone breeding success. Males may have much less to gain by risking their life than we imagined. The corollary is that females have much more to gain than we previously thought. If we look at competition for food from a female’s perspective, is it true that the

Mhoc03.fm Page 72 Friday, December 14, 2001 10:15 AM


    

food pay-off is trivial? For one morsel of fruit, the answer is probably yes. But if dominance buys her several years of priority of access to food, then the question becomes more interesting. Hrdy (1999, p. 83) notes that a female’s reproductive success depends most critically upon ‘how successful she is at keeping alive such infants as she does produce’. The replacement costs of every infant are high—she cannot afford to fail. Her success depends upon her capacity to provision them with food. Now surely it would be in her best interests to strive for dominance if the rewards were a constant and sure supply of nutrients for her offspring? It is ironic that the low fitness variance between women, which according to the male-focused argument would lead to less competition for copulation, leads to the expectation of great competition between them afterwards. Imagine women, quite literally, in a race for reproductive success. The race begins with the birth of a baby and there is little to choose between the contenders as the race starts for they vary less than men do in the number of babies that they produce. But because they are so tightly bunched, even the slightest difference between one competitor and another will have huge fitness consequences. If two women both have two children and one woman loses a single child, her rival now has a 100 per cent advantage over her. Men may be more competitive for copulation because of their greater incidence of both magnificent success and total failure but women should be more competitive for their each child’s survival precisely because of the their greater similarity in rates of reproduction. Females’ high parental investment means fewer children, and fewer children mean greater competition because so much rides on each one of them. When Daly and Wilson (1983, p. 93) wrote ‘The limit on female reproductive potential imposes a limit on the effort that it is worth her while to expend. Why should a seal kill herself for one season’s single pup?’ they got it precisely backwards. A female should not risk death for a copulation as a male might but she should be especially competitive about the survival of any infant once born because she is competing in a game where the survival of just one extra infant can make an enormous difference. Now forget for a moment our view of women as naturally ‘nice’. Imagine only a hypothetical community of females living with their offspring. Though the last few years have been easy and the children are well fed, this year brings with it a famine. The women travel further and further to forage, expending more and more calories, but come back empty-handed. Whenever they find food, they give to their children desperately hoping that they can somehow get them through until the rains come and the plants grow again. One of these women, watching her neighbour giving berries to her children, jumps on her, steals the food and feeds her own children instead. Her children survive but her neighbour’s children do not. In an even worse scenario, she murders her rival’s children to liberate greater resources for her own. Such a gruesome situation has been reported among our closest phylogenetic relatives. At the Gombe Stream Reserve, the female chimpanzee Passion, together with her son and daughter, snatched, killed and ate at least three infants from different mothers in her own community (Goodall 1986). Although initially considered to be a freak

Mhoc03.fm Page 73 Friday, December 14, 2001 10:15 AM

    :   


aberration, killing and cannibalization of unrelated infants has been reported a number of times since that first sighting. Effie, a female gorilla, was also seen eating another female’s infant (Jolly 1985). Goodall and her colleagues conclude that ‘These observations suggest that female infanticide may be a significant, if sporadic, threat, rather than being the pathological behaviour of one female’ (Pusey et al. 1997, p. 830). Disposing of rival females’ infants effectively cuts out the competition. Since most competition occurs between members of the same sex, we would expect that female infants are most vulnerable to infanticide by another female. This is exactly the pattern that is found (see Kano 1992). The fewer of them there are, the more food there is for your own children—to say nothing of the short-term calorie benefits of cannibalism. This kind of ruthless maternal behaviour could have been our evolutionary path too. There is nothing in evolutionary theory that decrees that females should be caring and cooperative to one another. Female nature depends upon the winning strategy in times long past. If lethal aggression increased female reproductive success, it would be retained. And it had a pretty good chance of being selected. Whatever genetic tendency caused a female to compete aggressively would be passed on in her surviving children while her neighbour’s less competitive genes would not. It is very likely that such resource deficiencies occurred often enough to support differential survival of the most competitive and women could have evolved to be highly aggressive. An aggressive strategy should have invaded and dominated a nonaggressive one. But that is not how things have turned out. Something happened that held women back from taking this competitive route. There must have been a strong counter-pressure away from high competition. And that pressure was the evolutionary requirement for the parent that rears the child to avoid dying early. This was elegantly shown by Allman et al. (1998) who compared 10 species of primate that varied in the amount of paternal or maternal effort they contributed to parental care. They hypothesized that whichever sex provides the most parental care should be selected for longer survival in order to guard their reproductive investment to maturity. They plotted the survival curves for each species and found significant sex differences in life expectancy which favoured whichever sex took most parental responsibility. They also ranked the primates in terms of which sex lived longer and by how much. The results showed that the amount of parenting was closely associated with female-to-male survival ratio. Chimpanzees show almost no paternal care and the female-to-male survival ratio is 1.42 to 1. In Titi and Owl monkeys, the father carries the infant from birth and, if the father dies, the mother will refuse to carry the infant, except during brief nursing periods. In these species where the father’s role is critical, the ratio reverses to between 0.83 and 0.87 to 1. But we have seen that male–male competition increases mortality among males. Might it be that Owl and Titi monkeys live longer not because they care for infants but because they do not engage in male–male competition? It is true that Owl and Titi monkeys live in monogamous pairs and so may not

Mhoc03.fm Page 74 Friday, December 14, 2001 10:15 AM


    

need to fight so aggressively for copulation opportunities but gibbons and gorillas also live in single-male groups yet they show no increase in life expectancy. It seems that whoever takes the brunt of infant care lives the longest and this is probably because the survival of their valuable infant depends upon ensuring their own survival. In humans as we have seen it is women who make the greater pre- and post-natal investment. The result is that males are more careless with their lives—they are willing to aggress even where the odds against achieving dominance are high and where dominance is no guarantee of reproductive success. Females, by contrast, seem to avoid outright aggression even where the pay-offs are direct and vital to reproductive success. A female is conservative because the risks to her and to her young and yet-to-be-born infants are too great. A male is risky not because the potential pay-offs are so great but because he has less to lose. In the past we have focused heavily on the prizes for the minority of ultra-successful males and paid less attention to the majority of males who will die without issue. If they die in combat, the chances of success they thereby lose are small (the chance of a male producing no offspring is far higher than that of a female). Nor do they jeopardize the lives of any infant they may have already fathered—the mother will take care of that. Males are in one sense the second sex. Although necessary for reproduction, in many species they contribute little more than sperm. If a proportion die, the remaining males can and will successfully fertilize the females. Males carry their lives more carelessly in their hands because their survival is less critical to the survival of their young and even to the survival of their community. I believe that women’s lower aggression is not simply a default option that results from lower incentives for competition. We have been led off-track by failing to recognize that women face a life history filled with as much potential for competitive violence as men. While the male-focused theory argues that women’s low levels of aggression result from the lower potential number of offspring they produce relative to men, heavy maternal investment leads to the expectation that women ought to be more competitive than men. If this is true, then something held them back from an all-out woman-on-woman war. And whatever that something was, it must have increased their reproductive success.

The female adaptation—fear To explain restraint from aggression, we need to find an emotion that would have the effect of giving a high weighing to the costs of physical combat. That emotion is fear. Fear acts as a signal that we are in above our heads and that our best bet is to back-off. The prospect of a dangerous encounter generates tendencies to both fight and flee but the latter wins out more often in women than in men because the fear response kicks in at a lower level of objective danger. The decision to withdraw is made when fear is the predominant emotion. A systematic bias could be introduced

Mhoc03.fm Page 75 Friday, December 14, 2001 10:15 AM

    :   


between men and women in their willingness to act aggressively if women’s fear rises to threshold faster than men’s. This weighting would mean that aggression was avoided by women unless there was a spectacularly high pay-off capable of counterbalancing the elevated costs (and we shall discuss just such a situation later). In the normal run of things, the same situation that would cause a man to hesitate would make a woman back away. Hundreds of psychological studies have examined sex differences in risk taking. In most of these studies, participants are asked to choose between one hypothetical option and another that differ in terms of the degree of risk involved. Or they are asked to say what the chances of success or failure would have to be before they would pursue a risky course of action (such as having a newly developed surgical procedure). These kinds of studies are predicated on the assumption that people employ cognitive strategies in evaluating risk, consciously weighing up the costs or benefits of various courses of action and multiplying them by their probability. More naturalistic studies use observational methods to examine risk taking in reallife situations, such as making a left turn against oncoming traffic. A meta-analysis of 150 risk experiments shows a significant effect of sex, with females, as predicted, being more risk-averse than men (Byrnes et al. 1999). But the authors make an important statement in their conclusions: It is interesting that the means [average differences between men and women] for certain observed behaviors were considerably larger then the means for certain hypothetical choices. These findings suggest that gender differences may be more likely to emerge when people have to actually carry out a risky behavior than when they have to simply consider the pros and cons of two options. If so, then the processes involved in the translation of cognition to behavior (e.g. fear responses) may explain gender differences in risk taking more adequately than the cognitive processes involved in the reflective evaluations of options. (Byrnes et al. 1999, p. 378)

‘Cold’, cognitively-based sex differences in risk evaluation may exist but the data support the idea that it is fear—an emotional not a cognitive function—that mediates the massive real-world differences in risk taking between men and women. If this proposal is correct, we would expect that women should specifically show greater fear of injury than men. I emphasize this point because there are many things that can cause fear and not all of them are about possible injury. For example, social anxiety is very common and it is a fear of being the centre of attention. People who experience this kind of ‘stage fright’ commonly fantasize about saying or doing something that would make them a laughing stock or invite public ridicule. Given men’s particular sensitivity to status (which we will discuss in the next chapter) it comes as no surprise that this form of fear is more common among men. Fear of one-to-one interaction with strangers would show itself as sex differences in extroversion. This trait is commonly measured by asking people about their enjoyment of social situations and their preference for being with others rather than alone and it shows no sex differences. Nor are women more frightened than men by novel

Mhoc03.fm Page 76 Friday, December 14, 2001 10:15 AM


    

situations or experiences. In fact, primatologists have noted that females are among the first to approach and explore new objects and to invent new uses for available materials. It was a female who discovered that sweet potatoes tasted better after they had been carried into the sea and washed and her novel discovery was taken up more quickly by other females than by males (Nishida 1987). Among humans, Openness to Experience which measures curiosity, imagination and creativity shows no sex differences (Costa and McCrae 1992). We can see just how specific to physical injury women’s fear is by looking at a psychometric instrument called ‘The Sensation Seeking Scale’. Designed to measure people’s desire to engage in exciting but risky activities, it is composed of four sub-scales. Three of these measure attraction to physically dangerous activities. However, one sub-scale—Experience Seeking—measures desire for novel and unusual experiences that do not contain a component of physical risk. Zuckerman (1994) reports that, in 15 out of 17 cross-cultural studies, no sex differences were found on this scale. He concludes that ‘The lack of difference on ES suggests that while men are high on the more active forms of sensation seeking, women are just as open to novel experiences through the senses and lifestyle as men’ (Zuckerman 1994, p. 101). However, as we would predict, men do score higher than women on the three other sensation-seeking scales that measure willingness to take physical risks. These active forms of sensation seeking (on which males exceed females) correlate significantly with a variety of physically dangerous activities such as involvement in crime, dangerous sports, injury proneness and volunteering for drug experiments and hazardous army combat (Zuckerman 1994). We have also seen that there are very significant sex differences in deaths from traffic accidents after controlling for miles driven (Wilson and Daly 1985) and rates of accidental injury on the street and in the home are higher among boys than among girls, even after controlling for exposure times (Christopherson 1989). Such accidents seem to be the result of a boys casual disregard for the danger in their games (Ginsburg and Miller 1982). Phobias are extreme fear reactions that are out of proportion to the actual threat posed by the environment. Humans are far more likely to display phobias in response to some objects in the environment than others and this has been called preparedness (Seligman 1971). Despite the fact that we are at greater risk from cars, guns and electricity, these are rarely the objects of irrational fear. Far more common are fears of snakes, spiders, heights, open and closed spaces and blood. Evolutionary psychologists have argued that phobias reflect an extreme defensive reaction to dangers that were manifest at some time in our evolutionary history. As Marks and Nesse (1997, p. 64) put it: ‘Evolved defences often seem over-responsive because repeated false alarms may cost less than a single failure to respond when the danger is great.’ According to our present argument we would expect to see higher rates of phobias in women than in men—specifically where the feared object posed some real threat to survival. Women are more prone to panic disorder with agoraphobia and to phobias about animals, blood and injury (American Psychiatric Association

Mhoc03.fm Page 77 Friday, December 14, 2001 10:15 AM

    :     1994; Marks 1987). Women seem to have a greater likelihood of overreaction to open spaces (where predation was more likely), to closed spaces (with the attendant danger of being trapped), to potential predators and parasite carriers and to the sight of blood or injury signalling possible death. Evidence of women’s higher level of ‘irrational’ fear is not confined to clinical samples. Survey instruments that assess fear in the general population find that women report more intense fear in response to a variety of stimuli including animals, injury and threatening settings, such as enclosed spaces, darkness, flying and heights. This sex difference persists even when the effects of sex-role (masculine and feminine personality types) and socially desirable responding are removed (Arrindell et al. 1993, 1999; Fredrikson et al. 1996; Gallagher and Klieger 1995). The difference between men and women in their willingness to take risks, including physical aggression, could have resulted from natural selection adjusting two kinds of mechanisms. Risk taking can result from a failure of the braking mechanism (an absence of fear) or from a jammed accelerator (sensation seeking). Daly and Wilson’s choice of the term ‘a taste for risk’ tends to imply the latter but they acknowledge that either one would have the same effect. Though the effect might be identical, the evolutionary logic would not. If the mechanism is an appetitive one, then the focus is on incentives and the implication is that male risk taking, mediated by sensation seeking, resulted from the fitness rewards. By inference, females lower risk taking resulted from a sex difference in the magnitude of what could be gained by aggression and consequent differences in sensation seeking. If the mechanism is an avoidant one, the focus is on costs and the implication is that female avoidance of risk (mediated by fear) resulted from fitness costs. By inference, males higher risk taking resulted from differences in the magnitude of what could be lost by aggression and consequently differences in fear. Fear is a negative motivational state that drives avoidant behaviour while sensation seeking is a positive motivational state that drives appetitive behaviour. Fear makes us avoid things; sensation seeking makes us approach them. But the two need not be seen as polar opposites. The two constructs of fear and sensation seeking are linked to one another by arousal level. As the amount of risk in a situation increases, so do two competing tendencies—one to approach and confront it, the other to withdraw and avoid it. The approach tendency results from the fact that risk in moderate doses can be pleasurable—it increases sympathetic nervous system arousal resulting in what we call an ‘arousal jag’. This effect is what underlies the enjoyment of fairgrounds, horror films and skiing. A small risk such as riding a roller coaster is attractive because the ‘buzz’ of increased arousal energizes approach and is much higher than the experience of fear. But as risk increases, for example by overtaking a car on a blind bend, the level of fear begins to reach the same level as the approach tendency and instead of thrill we experience nervousness and apprehension. In extremely risky situations, such as confronting a burglar, fear overtakes approach and we freeze, panic or otherwise seek to avoid the situation. Natural selection

Mhoc03.fm Page 78 Friday, December 14, 2001 10:15 AM


    

might have worked most strongly on males’ sensation seeking. If so, men take risks more often because their approach tendencies remain above the fear threshold at higher levels of objective danger. But sex differences in risk taking might result from differences in fear. All that is required is for females’ fear parameter to rise faster than males’ so that it overtakes the approach tendency at an objectively less risky point. Or to put it another way, men’s sensation seeking may be a result of the fact that fear rises more slowly in relation to the pleasurable arousal associated with risk. Fear and sensation seeking are emotions. Until recently, the study of emotion was considered a taboo topic for serious psychologists. Emotions evaded discrete categorization, they were internal, subjective, hard to measure, transient and seemingly whimsical. They also had a strong whiff of something female and irrational about them. More productive then to concentrate on visible behaviour or on cognitive faculties that followed rules and could be manipulated in the laboratory. But in recent years (to the delight of evolutionary psychologists who have long argued that emotions serve as vital, precognitive rules of thumb for guiding behaviour) emotions have been subjected to serious study. Much of that work has become from neuropsychologists and pharmacologists who have tried to map the brain areas and the neurotransmitter systems involved in our emotional experience. Here we can find a guide to the relative importance of fear and sensation seeking in aggression. We can also ask the key question—which system shows the strongest evidence of a sex difference?

The biology of fear and sensation seeking A word of warning. We must be cautious in interpreting findings from the biochemistry of aggression. The action of neurotransmitters is still understood only partially. No wonder. Their sheer variety (17 types of serotonin receptor and 5 of dopamine at the last count), the different behavioural effects that they can exert (dopamine controls among other things stereotypic behaviour, the way a female spaces copulation, and the brain’s sensitivity to cocaine), the specificity of their action at different points within the same chemical circuit (dopamine in the nigrostriatal system controls motor activity but in the mesocorticolimbic system controls re-inforcement and motivation), their interaction with one another at the same brain site (serotonin modulates the responses triggered by other neurotransmitters), the two-way causal street between neurotransmitters and behaviour, the complex adjustments that are made both pre- and postsynaptically in the face of sustained transmitter release and the long-term modification of cellular gene expression caused by transmitter activity all make it hazardous to draw conclusions on the current data (see Niehoff 1999). The monoamine family of neurotransmitters has a special role in the brain—to monitor incoming information and prioritize it. Without these chemical pathways we would be overwhelmed by the sheer tidal wave of information that assaults us

Mhoc03.fm Page 79 Friday, December 14, 2001 10:15 AM

    :   


from moment to moment and we would have no way of knowing whether we should finish our meal before or after we repelled an approaching attacker. Because of the vital function that they fulfil, monoamines carry messages from their base in the brainstem and midbrain in two directions. They go forwards to the amygdala and other subcortical structures which react instantaneously to emotion-laden stimuli and on to the frontal cortex where a slower but more reasoned response can be prepared. They also travel backwards to the cerebellum and spinal cord to relay the message to muscles and viscera. The various monoamines jointly infiltrate the whole brain but each travels along its own unique path with its own receptors. The receptor activates second-messenger proteins which can alter the sensitivity of the postsynaptic cell and generate activity among other proteins inside the cell which can trigger gene-switching transcription factors leading ultimately to longterm changes in the cell’s behaviour. The anatomy of aggression is distressingly complex and reaching conclusions about the human system is particularly difficult because even the rat and cat differ somewhat in the structures which subserve aggression (Siegel et al. 1999). But critical to both is the hypothalamus and the same seems to be true of humans. The cat has two distinct aggression systems—one for catching prey and one for attacking conspecifics and intruders. The former predation system is linked to brain centres governing feeding and need not concern us here. It is the latter system—sometimes called the defensive rage system—that is relevant. When activated, cats display the familiar trappings of an angry animal—their fur becomes erect, their ears retract, their back arches, their pupils dilate, their claws are unsheathed and they emit a disturbing (even to human ears) yowl. What triggers these behaviours is electrical activity in the medial and preoptic area of the hypothalamus. Sex differences have been found in this area (Swaab and Hofman 1995). The hypothalamus gets its incoming information from the amygdala and the data travel along a pathway called the bed nucleus of the stria terminalis where sex differences have also been found (Zhou et al. 1995). The messages from the amygdala have a modulatory effect on aggression—they effectively turn the volume up and down. But on what basis is the amygdala offering its advice about whether to proceed with the fight or to withdraw? The amygdala has been strongly implicated in emotion and specifically in the emotional tagging of memory representations of the world. Joseph LeDoux’s (1998) careful studies of fear have clarified the amygdala’s pathways and functions. He used a straightforward, if painstaking, technique to track the neurobiology of fear. First, he trained an animal to anticipate an electric shock at the sound of a tone so that the tone alone provoked all the behavioural signs of fear. Then, starting from the ear, he worked centrally along the neural pathways towards the innermost parts of the brain. At each stage he ablated or destroyed the tissue so that he could establish at what point fear no longer appeared. What he found was surprising. Destroying the auditory cortex did not affect the fear response. What was critical was that the incoming neural impulses, after reaching the sensory thalamus, were passed to the amygdala.

Mhoc03.fm Page 80 Friday, December 14, 2001 10:15 AM


    

When that connection was broken, the fear response disappeared. Somehow the amygdala was receiving the message to ‘be afraid’ even though the experience of noise had never consciously registered in the cortex. The incoming message to the amygdala goes to the lateral nucleus and this initiates activity in the central nucleus which activates the behavioural control system. The central nucleus of the amygdala is critical for every measure of conditioned fear—behavioural freezing, autonomic responses, stress hormone release, pain suppression and reflex potentiation—and each one of these is controlled by a different output from the amygdala. This system is a fast track that allows for an immediate reaction to a fear-provoking event—a response that can occur without conscious recognition that danger has occurred. Milliseconds later the cortex receives the more finely differentiated information (the loudness, pitch, tone) that allows for a slower but more modulated reaction to the danger. That is why we can respond instantly to the shadow of a person caught in the periphery of our vision, even before we have discovered whether it is our spouse home from work early or an intruder. Electrical stimulation of the amygdala in animals causes a fear response. In humans, when the same technique is used to pinpoint the precise focus of epileptic fits, the patient reports a strong sense of foreboding or danger. Patents with amygdala damage cannot be conditioned to fear and are unable to recognize fear on the faces of others. The ‘danger’ information is not only passed on to the prefrontal cortex but messages can also travel in the opposite direction. Imagine the most frightening thing that has ever happened to you. As you do so, the association and sensory cortices work to produce an image fusing visual and auditory information retrieved from memory. At the same time in the prefrontal cortex these images are given emotional ‘tone’ resulting from recognition of the original primary emotions that were experienced at the time these images were placed in memory. At this point we have a representation of both the experience and the emotions that went with it but we cannot ‘re-feel’ the emotional component without the amygdala. When the memory triggers amygdala activity it activates the autonomic nervous system, the motor system, the endocrine and peptide systems and stimulates appropriate neurotransmitter nuclei in the brainstem. We re-experience the emotion. (If you have ever had a truly terrifying experience you will find that it is almost impossible to recall it without physically wincing.) Now the ability to experience these ‘secondary emotions’ plays a big role in decision making. They act as what Damasio (1994) has called ‘somatic markers’, they are our automatic ‘gut’ reaction to a prospective state of affairs. When we make a decision (to fight, to apologize) we rarely have the time to systematically weigh the pros and cons of each course of action. How can a decision be made in the split second that we have to think about it? Damasio (1994, p. 173) explains: But now, imagine that before you apply any kind of cost/benefit analysis to the premises, and before you reason toward the solution of the problem, something quite important happens: When the bad outcome connected with a given response option comes to mind,

Mhoc03.fm Page 81 Friday, December 14, 2001 10:15 AM

    :   


however fleetingly, you experience an unpleasant gut feeling . . . What does the somatic marker achieve? It forces attention on the negative outcome to which a given action may lead, and functions as an automated alarm signal which says: Beware of danger ahead if you choose the option which leads to this outcome.

The emotional valence with which a scenario is tagged affects the likelihood that we will select or reject it next time a choice has to be made. According to the incentive theory, risk taking is positively tagged for males—they seek it out. If this is correct, then we have to look for evidence of the emotionally positive consequences of risk (which we are likely to find in the dopamine system that subserves the experience of pleasure) and for sex differences in these systems. But if I am right—that among females risk taking is negatively tagged—we must look for the emotionally negative aspects of risk (in the noradrenaline and serotonin systems that govern arousal and behavioural inhibition) and for sex differences here. One of the earliest monoamines to come under scrutiny was noradrenaline. It is this chemical messenger of the sympathetic nervous system that rockets up the production of adrenaline in the adrenal gland resulting in the familiar ‘panic’ experience of increased heart rate, sweaty hands and shaking limbs. From its usual baseline firing rate, a sudden danger can increase firing by up to ten times its normal rate. The source of noradrenaline is the locus ceruleus in the brainstem and the fibres travel up through the amygdala, hypothalamus, hippocampus and on to the cortex. When they arrive here we have the conscious recognition of fear and its source in the world beyond. But even before that, the body has been prepared for action—the classic fight or flight response. While the body is primed by the sympathetic nervous system, the brain is woken up too by simultaneous activity of the locus ceruleus. Some of the earliest neurochemical work examined the role of noradrenaline on aggression. As with much biological research, rats were the subjects and the behaviour of interest was shock-induced fighting. Rats placed on electrified grids not only rear up and fight but produce copious amounts of noradrenaline and the greater the stress under which they have been placed in the preceding weeks, the more extreme is both the boxing and the neurochemical reaction. But there is a problem with the paradigm. Boxing, rearing and freezing constitute a very specialized form of aggression. In a real-life rat fight, the attacking male goes for the intruder’s back and in consequence a victim who cannot escape will do his best to protect his back from being bitten. This gives rise to the classic upright boxing stance. In the same way, there is no escape from an electric shock and the rat’s response is similar. Robert Blanchard (1984) has shown that this behaviour corresponds to a fear-provoked form of self-defence that is quite distinct from a full-fledged aggressive attack. It seems likely that the noradrenaline rise that the rats show is not a correlate of aggression per se but simply of fear. Barrett et al. (1990) injected tiny amounts of noradrenaline into the anterior hypothalamus of a cat and measured the amount of current needed at the medial hypothalamus to trigger defensive aggression. Sure

Mhoc03.fm Page 82 Friday, December 14, 2001 10:15 AM

      enough, the higher the dosage of noradrenaline the less current was needed to generate defensive aggression but, once again, it is hard to tease apart the fearprovoking component of the procedure from the specifically aggression-promoting effect. Among humans who have suffered brain damage there are reports that drugs which block the action of noradrenaline decrease aggression. But this cannot in itself clarify whether noradrenaline has direct causal links to aggression or whether the decreased attack is the result of a lower experience of threat. Indeed, others have argued that one particular form of human aggression is in fact associated with a lower than normal fear response. Antisocial personality disorder is a syndrome characterized by an indifference to the well-being of others, an absence of remorse or guilt and a failure to learn from experience which can often result in persistent lawbreaking. These individuals often show signs of disturbance in childhood and may have a history of attention-deficit hyperactivity disorder. Such individuals seem to have too little fear and thus lack the emotional brakes to control their behaviour or to learn from negative experiences. As children, they have lower resting heart rates (Raine 1996), and those who later commit violent offences have the lowest of all (Raine et al. 1990). Their preternatural relaxation is manifest in the low stress responses that they show to anticipated electric shock and their minimal reaction to emotionally charged stimuli that increase the heart rate and skin conductance of normal subjects (Raine 1993). Without fear, they fail to learn from punishment and show a callous disregard for the safety of both themselves or others. Fortunately, psychopathy is not a common condition but it is sexually dimorphic. The diagnostic rate among men is about 4 per cent of the population while among women it is less than 1 per cent. It is rare indeed that a woman’s fear response shows this kind of suppression and in evolutionary terms it is unsurprising—such a careless disregard for her own safety would certainly have compromised the survival of her offspring. The principal function of the noradrenergic system is to raise the psychological and bodily alarm signal. It generates the feelings that allow us to automatically veer the car to avoid a collision, to sit bolt upright in bed when we hear a noise downstairs and to scream with catharsis when we begin the vertical descent on the roller coaster. The feelings alert us to danger but they do not tell us what to do about it. In and of itself, noradrenaline does not guarantee aggression any more than it guarantees running away or freezing. So the alarm reaction may be a necessary component of a regard for our own safety—without it we lack the emotional alerting system that prevents us from jumping into terrifying situations. And the system malfunctions more in males than females, suggesting that in men evolutionary pressures have not been great enough to correct it. Men who behaved this way may have been more cold-bloodedly instrumental than others in their fight for dominance despite the risks of self-destruction that it carried. But another of the monoamines seems to go beyond the ‘Stop what you are doing’ message and adds another one: ‘Don’t act on impulse.’ Serotonin is a chemical

Mhoc03.fm Page 83 Friday, December 14, 2001 10:15 AM

    :   


pathway of behavioural inhibition and has been more closely and directly linked to aggression than noradrenaline. Serotonergic-sensitive neurons make up only about one-millionth of the total population of neurons in the central nervous system, yet they have been in place for over 500 million years and perform vital functions in regulating the response of neurons to other neurotransmitters. In monkeys and humans, the serotonin transporter gene is controlled by a DNA promoter sequence that is unique to anthropoid primates and appeared about 40 million years ago. Differences in this DNA sequence have been linked to individual differences in anxiety, hostility and impulsiveness (Lesch 1998). Arising in the raphe nuclei, projections carrying serotonin (like those for noradrenaline) spread forward to the amygdala, hypothalamus and hippocampus and on to the cortex to areas that are associated with reasoning, feeling, remembering and reacting. One of its chief functions is to tailor responses to the demands of the situation preventing overreaction. Midbrain to forebrain projections using this transmitter appear to have a ‘pacemaker’ function, firing in a steady tonic pattern not unlike the ubiquitous Musak piped into grocery stores and doctors’ offices. The continuous serotonergic background suppresses activity in target cells of the cortex and limbic system, superimposing an interpretative overlay that prevents excessive reactions to sensory information and coordinates an even-handed response just intense enough to meet the demands of the situation. (Niehoff 1999, p. 137)

Aggression is a line of last resort, never a first choice of action, because it is painful and dangerous. One review of over 200 studies of aggression highlights the fact that aggression occurs when inhibitions against it are temporarily suppressed: ‘(Aggression) reflects the behavioural expression of a brain mechanism that is activated when control over the behavior of another organism is lost, and agonistic strategies such as threats, displays and warnings are insufficient to reinstate control. Under such conditions, inhibition over an attack–release mechanism that can be activated from the hypothalamus is overridden and a well-directed attack results’ (Siegel et al. 1999, p. 364). The chemical that holds us back from aggression is serotonin. Low levels of serotonin seems to increase aggression by blocking our ability to inhibit impulsive behaviour. One way to assess impulsivity is by examining passive avoidance learning. For example, a cat with a surgically induced deficit in passive avoidance will run time and again to a feeding bowl where it has been shocked. Active avoidance studies, by contrast, examine the behaviour of animals when they must make a positive response in order to avoid electric shock. Neither of these entail aggression directly but each measures the ability of the animal to either restrain action or accelerate it in response to environmental cues. As Eichelman (1995, p. 61) sums it up: Those manipulations which make an animal more adept at active avoidance or less adept at passive avoidance are frequently the same manipulations which enhance

Mhoc03.fm Page 84 Friday, December 14, 2001 10:15 AM

      aggressive behavior in the laboratory situation. From a neurochemical perspective, those neurochemical manipulations which impair passive avoidance and enhance “active” behaviors (e.g. 5-HT (serotonin) reduction) also appear to enhance aggressive behavior within certain laboratory models.

Low levels of serotonin create an edgy, jumpy, overfast reaction to the environment. The aggression is impulsive, thoughtless and immediate. The ability to stop and think, disrupted by serotonin depletion, causes a disinhibition of aggression. Research on serotonin’s effects on aggression began with mice. Housed alone, mice become increasingly aggressive over time and Valzelli and co-workers (Garattini et al. 1967; Valzelli and Bernasconi 1979) found that as their aggressiveness rose their serotonin levels dropped. Subsequent experimental studies confirmed that artificially lowering serotonin level causes increased aggression while drugs that enhances the uptake of serotonin in the synaptic cleft lower aggression. The effect of 5-HT1 serotonin agonists is aggression specific; it reduces offensive attack while leaving other social behaviours unaffected—even where the dosage is increased to 50 times greater than that needed to cut aggression in half (Miczek et al. 1994a). And, as we might predict, it seems to work by increasing levels of fear and anxiety. Scientists have begun to unlock the genetic secrets of serotonin levels, at least among rodents. By selectively removing critical genes from the chromosome they can create so-called knockout mice. Knocking out the gene (5-HT1B) coding for one of the 17 serotonin receptors produces adults that are far more aggressive than their intact peers (Brunner and Hen 1997; Zhuang 1999). They attack more swiftly and more ferociously without pausing for any preliminary exploration of the situation. In humans too, spinal fluid levels of serotonin are negatively correlated with a history of aggression and violence (Brown et al. 1979, 1982). But further research suggests that serotonin is especially associated with impulsive aggression. This line of work began with studies that linked serotonin to suicide (Coccaro 1995) and recall that suicide shows a very significant sex difference. Asberg et al. (1976) investigated serotonin levels in depressed patients including some who had attempted suicide. They found that the suicidal patients had lower levels of serotonin than the rest and those who had chosen the most violent and determined methods, such as shooting themselves, had the lowest levels of all. What appeared to be happening in these individuals was a failure of inhibition which resulted in violence—in this case turned upon themselves. But suicide and violence are linked—murderers are 700 times more likely to commit suicide than law-abiding citizens. And neurochemical challenge studies concur that self-directed and otherdirected aggression stem from the same basis (Coccaro and Murphy 1990; Markowitz and Coccaro 1995). One study examined imprisoned men serving time for murder or attempted murder. They were classified into two groups according to whether their actions had been premeditated and instrumental (done in the service of executing another crime such as robbery) or impulsive. The lowest serotonin

Mhoc03.fm Page 85 Friday, December 14, 2001 10:15 AM

    :   


levels were found in impulsive killers and were lowest of all among the men in that group who had also tried to kill themselves (Linnoila et al. 1983). But serotonin deficiency has also been examined in people who never get to the point of murder. Impulsively aggressive individuals are primed to see every ambiguous action as a potential threat and every joke as a possible slight. They fly off the handle unpredictability in response to minor provocations that most of us would allow to pass. They are the men who start fights ‘over nothing’ and respond to a spilled drink with violence. It is this kind of overprimed, uninhibited form of violence that characterizes most young male aggression. Unlike psychopaths who feel no remorse for what they have done, these men in the cool light of hindsight often regret their actions (Niehoff 1999). But at the time, they lack the restraint to even contemplate any other course of action. They are not emotionless predators but quite the opposite—men whose emotions directly drive their behaviour resulting in brawls that seem to others (and to themselves the following day) pointless. These men show evidence of reduced serotonin activity in the frontal cortex compared to normal volunteers while the dopamine system continues to work as normal (Siever et al. 1999). But if lower serotonin is the source of male aggression and risk taking then we should expect to find sex differences in this particular neurotransmitter. And we do (Bucht et al. 1981; Heninger et al. 1984; Maes et al. 1989; Odink et al. 1987; Reisert and Pilgrim 1991). In a recent study, using positron emission tomography, 22 healthy, age-matched volunteers agreed to have their brains photographed as a radiotracer was administered that sought out serotonin receptors. Women displayed significantly higher presynaptic serotonin activity than men—especially in the frontal cortex that is responsible for behavioural inhibition (Biver et al. 1996). The argument for a gender difference in inhibition is strengthened by the fact that women are better able to control other social and emotional impulses than are men (see Bjorklund and Kipp 1996). They have an advantage in resisting temptation (girls more often comply with an adult’s request not to touch an attractive toy when they leave the room than do boys) and in delaying gratification (they are better able to wait for a more attractive reward instead of taking a less attractive reward immediately). Girls do better on tests of motor control—as in the childhood ‘Simon says’ game—where the child must inhibit their automatic tendency to comply with the adult’s instruction. They are better able to conceal facial and bodily signals of emotion—for example when asked to pretend that a foul-tasting drink is pleasant or to express delight when they lose a competitive game. They are also better at controlling emotional arousal—one study found that women, when instructed, were better at dampening or exaggerating their physiological reaction to electric shock and another that females were more successful than males at winning arguments because they were better able to control their anger. However, the sexes do not differ in a variety of tasks that assess cognitive inhibition (such as the Stroop test where a colour name, such as ‘red’, is printed in blue ink and the subject is required to respond only

Mhoc03.fm Page 86 Friday, December 14, 2001 10:15 AM


    

in terms of the ink colour while ignoring the semantic information). This suggests that women’s superior inhibitory abilities are precisely tuned to social and emotional domains that have implications for modulating social relations. Intriguing findings in the neuroscience of sexual behaviour suggest an answer to the question of why males have lower levels of serotonin and consequently a weaker ability to inhibit their behaviour. Male sexual behaviour is influenced by both dopamine and serotonin but each of these transmitters acts in opposing and unique ways. Dopamine (as we shall see) is the ‘feel good’ system. No surprise then to find that dopamine enhances sexual motivation in rodents, primates and humans (Bitran and Hull 1987; Gorzalka et al. 1990; Melis and Arigolas 1995; Meston and Gorzalka 1992; Wilson 1994; Zajecka et al. 1991). Dopamine levels rise not just during mating but even when the male rat is returned to a place where he has copulated previously. When a female is introduced, dopamine rises further and peaks during copulation (Pfaus et al. 1990). After copulation, dopamine levels fall and remain low during the refractory post-ejaculatory interval, rising again when the male recovers his interest in sex. Lorrain et al. (1999) inserted a microdialysis probe into the lateral hypothalamus of male rats. Half of them received serotonin while the control rats received an inert solution. The rats who were receiving serotonin showed a steady decline in dopamine concentration and by the time the female was introduced 30 minutes later, dopamine had dropped significantly. When the opportunity for copulation came, control rats had dopamine levels 200 per cent above baseline while the serotonin group were 50 per cent below. This study demonstrates in real time the well-replicated finding that serotonin inhibits and impairs copulation. In untreated males, serotonin rises after copulation and is the chemical messenger responsible for the drop in dopamine that inhibits their ability to be sexually aroused by a female after insemination (Fiorini et al. 1997). (There are good reasons why this refractory period might be useful. In rats, a second immediate copulation would dislodge the sperm plug that promotes sperm transportation to the uterus. A second immediate batch of semen is also likely to contain far fewer sperm than the first.) The inhibitory action of serotonin may explain why some antidepressants (like Prozac) which work on this system have the unfortunate effect of decreasing patients’ sex drive. What all this suggests is that ancestral males may have faced a trade-off. A successful male needs to be ready, able and eager to perform sexually. Serotonin—the ‘don’t rush into things’ chemical—impairs his ability to do this. Males then gave up some behavioural reserve in order to gain a higher sexual drive. (No wonder contraception has been a woman’s responsibility—a sexually excited male lacks the restraint to stop and think about the consequences.) This also implies that other aspects of male behaviour have been affected by the dampening of the serotonergic system—including their ability to inhibit risk taking. Once again it is not that the male has so much to gain from risk taking but that the price of having a robust sex drive is having a poor braking system.

Mhoc03.fm Page 87 Friday, December 14, 2001 10:15 AM

    :   


But if the incentive model of aggression and risk taking were correct, then a prime candidate for this mechanism would be dopamine. In 1954, Olds and Milner found that rats would press a bar to the point of physical exhaustion to receive electrical stimulation of the hypothalamus. Further research demonstrated that it was not this brain structure itself that was critical so much as the pathway of which the hypothalamus formed a part—that pathway was the dopamine system. Dopamine is the chemical that underlies both liking and wanting. If the sex differences in aggression derive from male’s heightened enjoyment of risky situations then the mechanism driving their rewarding properties might be dopamine. So far, so simple. But there is a difference of opinion about just how dopamine does this, which leads to diverging predictions. On the one hand are those who seek a positive association between dopamine and risk-taking behaviours, arguing that the experience of reward depends upon high levels of dopamine. The other camp argues the reverse—that risk taking activities are about stimulating dopamine pathways that are chronically underactive in some people—and hence they would expect a negative association between baseline levels of dopamine and a preference for risky activities such as aggression. Let us take the positive association first. A basic question is whether aggression is itself a pleasurable experience. With rats it is not hard to find out. Electrodes are implanted in areas of the brain governing aggression and the rats are given the opportunity to self-stimulate by bar pressing. If they like the feeling, they will do so and if they don’t, they will avoid the bar. Rats self-stimulated in only six of the 27 sites that produced aggression suggesting that there is little inherent pleasure in fighting (Herndon et al. 1979; Kruk et al. 1984). Nonetheless, the discovery of ‘L-dopa rage’ provoked excitement in the 1960s— very large doses of the dopamine precursor L-dopa seemed to heighten aggressive behaviour in mice. But on closer inspection, the behaviours the mice showed were fragmental and undirected; they bit, lashed out and cowered indiscriminately. What the mice failed to show was a clear and well-organized pattern of attack. In fact, although L-dopa increased defensive reactions to painful stimuli, it suppressed normal fighting behaviour (Geyer and Segal 1974; Karczmar and Scudder 1969; Thoa et al. 1972). Similar findings resulted from the administration of dopamine agonists which enhance dopamine levels. The injected rats showed an increase in hyperdefensive reactions but a suppression of normal coordinated aggressive behaviour involving pursuit, threat and attack (Baggio and Ferrari 1980). All this suggests that the dopaminergic pathway that is being triggered may have more to do with the piecemeal and uncoordinated execution of reactive responses rather than the initiation of spontaneous attack. Furthermore, attempts to control aggression by the use of drugs which suppress the uptake of dopamine are not effective in targeting aggression-specific mechanisms. The evidence suggests that dopamine antagonists interfere generally with the sequencing of motor behaviour. Even more confusingly, drugs which increase rather than

Mhoc03.fm Page 88 Friday, December 14, 2001 10:15 AM


    

decrease dopamine have similar effects. Because dopamine is involved both in the psychological experience of reward and in the motor control of behaviour, it is hard to establish which channel is being affected by dopamine. But the evidence to date suggests that the effect of dopamine on aggression may have more to do with the execution of motor responses than with the intrinsically rewarding properties of aggression. As Miczek et al. (1994b, pp. 101–2) conclude: In spite of their widespread clinical use in the management of aggressive and violent patients with varying diagnoses, preclinical studies have not established a cogent rationale for a specific role of brain dopamine systems and dopamine receptor subtypes in specific types of animal aggression.

If heightened male aggression were mediated by dopamine, two expectations would follow. We might reasonably expect that dopamine would show a relationship to testosterone which has been implicated principally in status seeking and in aggression as a route to that end (Mazur and Booth 1998). Loss of status or defeat leads to reductions in testosterone levels. But dopamine levels at least in the striatum of male rats are quite independent of testosterone levels (Becker 1999). More importantly for our purposes, if dopamine lies at the heart of sex differences in aggression, we would expect to see important gender differences in brain dopamine levels. Here the picture becomes murky and, once again, we have to rely on rodent experiments (Becker 1999). Extracellular dopamine levels are comparable in males and in oestrous females, as we would expect, knowing the relationship between dopamine and sexual behaviour. (The female’s heightened dopamine affects her ‘pacing’ of copulatory behaviour in a way that increases the chance of pregnancy.) When not in oestrus, female levels drop. Does this account for females’ lower levels of aggression? It seems unlikely because although extracellular concentrations of dopamine are lower, females show higher concentrations of dopamine in the brain tissue (J. Becker, personal communication). They seem to pick up more efficiently the smaller quantities of dopamine available at the synapse. Female rats are more active than males even outside oestrus and this motor activity is governed by dopamine transmission, suggesting that they respond more readily to the amount that is available. They also show stronger behavioural effects than males when dopamine is artificially raised and lowered. Cocaine, which increases brain dopamine, triggers more locomotion and behavioural excitation in female than in male rats. In short, dopamine does not look like a strong candidate for explaining sex differences in aggression. But another possibility is that dopamine is connected to aggression indirectly through risk taking. Hamer and Copeland (1998) believe there is a connection. To demonstrate the positive association between dopamine and sensation seeking, they describe a genetic knockout study of mice. Mice without the gene responsible for dopamine became not only uninterested in novelty but lethargic to the point that they would not eat, drink or groom themselves. By contrast, the mice with highly

Mhoc03.fm Page 89 Friday, December 14, 2001 10:15 AM

    :   


efficient dopamine production showed hyperactive exploration, running around the cage in a state of manic activity. Dopamine, it was argued, like cocaine, increases the rewardingness of exploration and novelty. If this is so, then we would expect that mood states associated with dopamine would make people more likely to take risks. Dopamine elevation is associated with positive affect—feeling good. (Its depletion, however, is not associated with sadness but with a flattening of emotion.) Experimental studies have manipulated the mood state of subjects and examined their appraisal of risk taking. The results are surprising. Happy people believe that positive outcomes are more likely than negative outcomes and that winning is more likely than losing (Johnson and Tversky 1983). In the context of a prospective fight, dopamine-high individuals are more likely to think that they would win. But (and this is a big but) happy people anticipate that they would experience more distress if they lost and in consequence avoid taking risks. This is a special case of prospect theory, a phenomenon found by Kahneman and Tversky (1979, p. 279), which they describe as ‘the aggravation that one experiences in losing a sum of money appears to be greater than the pleasure associated with winning the same amount’ or, as Jimmy Connors once described his motivation for tennis: ‘I hate to lose more than I like to win.’ Under positive affect, this effect seems to be enhanced. People are unwilling to relinquish their good moods by risk taking even where they believe the odds are in their favour because the losses are too costly. As Isen et al. (1988, p. 717) describe it: ‘possible gains do not seem more appealing to persons who are happy, but possible losses seem more aversive. Literally then, to a person who is in a positive affective state it may be that taking a chance is just not worth the risk.’ So dopamine-high people are actually more risk averse than others. Other researchers have argued for the reverse relationship—it is low levels of dopamine that make people seek excitement. An experiment undertaken in Israel (Ebstein et al. 1995) tried to unlock the secrets of a gene called D4DR which is responsible for making D4 dopamine receptors. In the central portion of this gene is a segment consisting of 48 base pairs coding for 16 amino acids. This sequence is repeated but the number of repetitions varies dramatically between individuals from 2 to 11. They measured novelty seeking and correlated it with the number of sequence repeats on the gene. People with many repeats had higher novelty-seeking scores. The same result was found in the United States (Benjamin et al. 1996). But here is the trick. The more repeats of the gene, the weaker the ability to bind dopamine. So it seems that individuals with a poor ability to bind dopamine have higher levels of sensation seeking. This result supports the old notion of ‘optimal arousal’—that individuals who are low in cortical arousal seek new and stimulating experiences in order to boost their arousal (or dopamine) to pleasurable levels. But before we conclude that dopamine deficiency drives sensation seeking, at least three subsequent studies have failed to replicate the original findings (see Hamer and Copeland 1998; Vandenbergh et al. 1997).

Mhoc03.fm Page 90 Friday, December 14, 2001 10:15 AM

      One man who has devoted his life to studying sensation seeking is Marvin Zuckerman. As we would expect, under an inhibition model physically risky (but not socially risky) forms of sensation seeking are higher in men than in women and sensation seeking is inversely related to situational fear (though not to generalized free-floating anxiety). The more fearful a person is, the less likely they are to take risks and Zuckerman agrees that risk taking is mediated in part by a weak or underactive serotonin inhibition system. But he also argues for an appetitive role for the dopaminergic system that drives thrill seeking. However, studies which have examined the relationship between dopamine and sensation seeking reveal an extremely mixed pattern of results, with half showing negative and half positive correlations. Zuckerman (1994, pp. 309, 379) hedges his bets. On the one hand he argues that sensation seekers may have sluggish dopamine systems that need external risky stimulation to ‘bring them closer to an optimum level of arousal’. On the other, sensation seekers have ‘a highly active and well-regulated dopaminergic system’. And, just to make sure that all bases are covered, that ‘sensation seekers have a dopamine system that is alternatively over- and under-active, depending on current levels of stimulation’. In short, the association, if it exists, is still remarkably confused. And there is another critical problem in arguing that risk taking is what drives up male aggression. Again I quote Zuckerman (p. 17): ‘aggression is not a high correlate of sensation seeking’. Fighting is not apparently an activity that people engage in to satisfy a need for risk, although mountain climbing, scuba diving and volunteering for hazardous army duty may be. As most violent offenders will tell you, they do not go out looking for a fight and are often perplexed and remorseful about their own violence (Campbell 1986). As they see it, the fights just keep coming to them—when they are insulted or attacked they have no option but to defend themselves. As they overreact to minor provocations that others would ignore, their problem is not sensation seeking but impulse control.

Women, competition and low-risk aggression So it was fear that kept female competition at a less lethal level than that of males. But females still needed to compete. They managed the tension between competition and staying alive by the use of two forms of low-risk tactic—an unwillingness to escalate to direct combat and a preference for indirect forms of aggression. We know that the size of the sex difference in aggression becomes greater as the seriousness of the possible injury increases. Meta-analysis is a statistical technique that allows us to combine studies and to estimate the magnitude of the sex difference based on thousands of subjects. This can provide a more reliable estimation of how big sex differences are than considering studies one at a time. Meta-analysis confirms that the sex difference is greater for physical aggression than for verbal or psychological aggression (Bettencourt and Miller 1996; Eagly and Steffen 1986; Frodi et al. 1977; Hyde 1986; Knight et al. 1996). This is true whether the aggression

Mhoc03.fm Page 91 Friday, December 14, 2001 10:15 AM

    :   


is measured by electric shock giving in laboratory or by a questionnaire asking about personal experience and attitudes. The smallest sex difference that has been found for physical aggression using meta-analysis suggests that 62 per cent of men are more aggressive than the average woman while the largest finds that 82 per cent of men exceed the average woman. For verbal aggression, meta-analyses vary between finding no significant sex difference to an effect size suggesting that 69 per cent of men are more aggressive than the average woman. But meta-analysis can do more. It can tease out why the differences exist. We can ask dispassionate judges to give ratings to each experiment according to, for example, how dangerous aggression might be for the subject and how anxious the subject might feel about aggressing. We can then compare whether the sex differences get bigger or smaller as a result of dividing studies into these different categories of dangerousness. Eagly and Steffen (1986) did just this and found that women judges rated all the experimental situations as more anxiety provoking and more dangerous then men did. This reflects a general tendency for women to experience greater fear than men in relation to anger-provoking situations (Brody et al. 1995). Importantly, the danger that was faced by the subject and the anxiety they would feel had a very marked effect on the size of the sex difference—the greater the danger and anxiety, the more men outstripped women in their level of overt aggression. Ten years later, a second meta-analysis by Bettencourt and Miller (1996) found similar results. It appears that the more women’s fear of aggressive retaliation exceeds that of men, the larger the gender difference in aggression. The fact that differential appraisals of danger by male and female judges were related to aggressive responding is especially interesting given that (a) experimental inductions were objectively similar for female and male participants and (b) with very few exceptions, experimental paradigms provided little if any expectation that targets would have opportunities for retaliation. (Bettencourt and Miller 1996, p. 440)

These last two points are important. Men and women faced objectively the same situation and yet males showed less fear and more aggression. Recall also that all these studies took place in a laboratory—white coats, academic rooms, middle-class researchers and subjects. Compare this to a dark street late at night as a club empties its drunken clientele onto the street. Consider the massively greater likelihood of aggression provoking retaliation in such a setting. The laboratory sex differences would be writ very large indeed. Not surprisingly perhaps, Bettencourt and Miller also found that provocation had a strong effect on the likelihood of aggression for all subjects. They wondered what the effect of provocation would be on different forms of aggression by the two sexes. For men, provocation raised their levels of aggression and it did so equally for verbal and physical aggression. For women, provocation had a much greater effect on their willingness to be verbally aggressive than on their willingness to risk physical aggression. This suggests that women are very much less willing to cross the line between verbal and physical aggression than are men.

Mhoc03.fm Page 92 Friday, December 14, 2001 10:15 AM

      And this restraint cannot be attributed to simply not feeling anger. When men and women are asked to keep a daily record of angry feelings, it turns out that both sexes experience anger about six times a week. Nor do the sexes differ in the intensity of the anger that they feel (Frost and Averill 1982). Psychometric questionnaires that tap anger and hostility by asking subjects to rate their agreement with statements such as ‘Sometimes people bother me just by being around’ find no sex differences (Frodi et al. 1977; Hyde 1986; Tavris 1989). Women experience most anger towards family, friends and workmates followed by anger towards society or depersonalized others (such as arrogant and opinionated people). Issues related to competence and to meeting goals frequently excite anger—when a woman is tired or is expected to achieve goals without adequate help or resources. The next most frequent precipitator is lack of respect or consideration from others followed by other family or co-workers failing to contribute to joint tasks (Denham and Bultemeier 1993). Women seem to be more annoyed than men when they feel another person is talking down to them or behaving in a condescending way (Harris 1993). Both men and women feel most anger when others’ behaviour has a damaging impact on their life and when the behaviour is performed intentionally and could have been prevented (Ben-Zur and Breznitz 1991). Despite this, women are more reluctant to express their anger as outright physical attack than are men. But there are other ways of competing. Indirect aggression is one way and one that does not entail risking one’s life. This is a form of social manipulation where the target is attacked circuitously and the aggressor can therefore remain unidentified. It involves acts such as shunning, stigmatizing and gossiping. It excludes rivals and destroys their reputation without any head-on confrontation and it is the one form of aggression on which females exceed males. Girls are more likely to exclude newcomers than are boys (Feshbach 1969), to destroy others’ property or tell tales on them (Brodzinsky et al. 1979) and to use tactics of ostracizing and manipulating public opinion (Cairns et al. 1989). This form of aggression blossoms as girls polish the social and verbal skills necessary for this tactical and invisible form of victimization. By the age of 11, girls are significantly higher than boys on becoming friendly with someone else as revenge, gossiping and suggesting shunning of another (Bjorkqvist et al. 1992; Crick and Grotpeter 1995). Studies of school bullying also report that girls employ indirect strategies of stigmatization and exclusion that can have devastating effects upon the victim (Ahmad and Smith 1994). Females’ use of indirect aggression continues into adulthood. Bjorkqvist et al. (1994), investigating victimization in the workplace, found that women more than men use indirect forms such as spreading false rumours and not speaking. An ethnographic study in Buenos Aires concluded that women ‘are very competitive, envious and jealous of each other. For example, female competition is expressed through invidious comparisons regarding clothing and appearance . . . we note that several female interviewees expressed similar sentiments about the competitive nature of women, for example “if the office is all men, generally,

Mhoc03.fm Page 93 Friday, December 14, 2001 10:15 AM

    :     there will not be problems. But if they are all women, at any moment problems will arise due to gossip. . . . The women are the ones that start things”’ (Hines and Fry 1994, p. 228). The past emphasis upon physical combat as the gold standard of aggression has until very recently limited our understanding of how women compete. Several years ago, primatologist Sarah Hrdy pointed out that Women are no less competitive than other primates, and the evidence will be forthcoming when we begin to devise methodologies sufficiently ingenious to measure it. Efforts to date have sought to find ‘lines of authority’ and hierarchies comparable to those males form in corporations. No scientist has yet trained a systematic eye on women competing with one another in the spheres that really matter to them. The difficulty is not simply narrowness of vision and the mistaken assumption that female competition will take the same form as competition between males, but also the subtlety of interactions between females. . . . How do you attach a number to calumny? How do you measure the sweetly worded put-down? (Hrdy 1981, pp. 129–30)

Kaj Bjorkqvist and his colleagues were the first to identify this covert form of aggression and it has gripped the popular (as well as the academic) imagination. It provided the perfect counterpoint to women’s growing impatience with the prevailing brand of feminism that portrayed women as devoid of competition. According to ‘distaff’ feminists, women are nurturing, warm, interdependent. Aggression would threaten the very bonds upon which their sense of self was built (Chodorow 1978; Cross and Madson 1997; Gilligan 1982). Now nobody would deny that women all over the world rate themselves (and are rated by others) as being higher in supportiveness and care-taking than men but the problem was that this idealization began to stand in the way of understanding competition between women. It could not be studied because it was effectively obliterated from sight. But social science could not go on for ever ignoring the less benevolent side of women’s nature. In 1987, The Feminist Press published a book entitled Competition: a feminist taboo (Miner and Longino 1987). It was the beginning of a feminist turnaround. Women began to own up to tactics that they used, even with friends, to compete and win. As Kate Fillion describes the female experience: It just doesn’t seem friendly and supportive to contest another woman’s views, decisions or behavior, and open conflict seems to raise the specter of separation. But although we scramble to find common ground, we may wind up competing to be right and competing to be different in indirect, covert ways. . . . We delicately refrain from mentioning anything to the friend with whom we disagree, but run around courting public opinion and appointing third parties as arbiters of an undeclared competition . . . Even when it’s unjust, even when it’s petty, even when we know it simply isn’t true, criticism from other women brings us to our knees. We learn to be nice and sweet so that other women won’t say nasty things about us; we figure out early that the best way to avoid being ostracized is to curry favor and swear fervent assent (Fillion 1997, p. 40 and 54).

Mhoc03.fm Page 94 Friday, December 14, 2001 10:15 AM

      It seems at first glance as if women’s preference for indirect strategies might be a simple result of gender role prescriptions. Since women’s aggression is seen as an aberration not just from decent behaviour (as it is with men) but also from the idealized female stereotype, it is doubly condemned. Because of this, women seek invisible means of expressing competition that is more acceptable. The only systematic work on this idea, however, has produced negative results. In a study of children from five ethnic groups, Osterman et al. (1994) asked children to report on their own and their friends’ use of various forms of aggression. They reasoned that if a form of aggression was socially condemned then the children would be far more likely to attribute it to their peers and less likely to admit to it themselves. They computed an Attributional Discrepancy Index by subtracting self-estimates from peer estimates for various forms of aggression. They found that there was a significant sex difference for physical and verbal but not for indirect aggression. It was boys’, not girls’, physical and verbal aggression that was most condemned. The absence of sex differences for stigmatization of indirect aggression suggests that it is no more socially acceptable for one sex rather than another—yet females do more of it. In fact, indirect aggression has no negative consequences for popularity when used by either boys or girls (Salmivalli et al. 2000). Primate females also use indirect aggression. Females have been observed to engage in sustained, low-level harassment of other females. There is no outright attack but there is much threat and over days or weeks a female is consistently interrupted and displaced as she tries to rest, feed or mate. This harassment takes its toll and the resulting stress suppresses oestrus and can even cause abortion (Chapais 1992; Hrdy 1981; Smuts 1987; Wasser 1983). These indirect tactics diminish the reproductive success of the victim. This means that more material resources are available for the victor and her offspring. And all of this is achieved without risking a direct attack which might result in retaliation.

Maternal aggression—a predictable paradox I have argued that fear is functional for females. It restrains them from exposing themselves to unnecessary danger and from the possibility of sustaining injury or death that would have disastrous implications for their young. But this emphasis upon the costs of aggression to females always has to be balanced against the prospective benefits. It is true that given identical pay-offs, females should be less willing to expose themselves to danger than males. But there is one clear situation where the pay-offs for aggression are much larger for a female than for a male—saving her child’s life. That life has cost the mother dear in terms of time, effort, vigilance and protection. In polygynous species that same life has cost the father a few million easily-replaceable sperm. The cost of replacement for the mother is measured in years while for the father it is measured in minutes. Although animal mothers rarely

Mhoc03.fm Page 95 Friday, December 14, 2001 10:15 AM

    :     go as far as sacrificing themselves for their children, they are more willing to tolerate the risks of aggression for the sake of their offspring’s survival than for any other reason. Maternal aggression has been documented in many species from lemmings to lions and it occurs when another animal approaches or interferes with one of a female’s offspring (Maestripieri 1992). The species in which it has been documented are those where male infanticide is a serious threat to female reproductive success. The mother’s attack is ferocious and immediate. In many primates such as macaques, vervets and baboons, males avoid infants and display fear when one approaches them. Females in these species are extraordinarily sensitive to the presence of males near infants and when an infant shows distress a nearby male may be attacked even if he has displayed no overt aggression towards it (Smuts 1987). Although some researchers have argued that the function of such female aggression may be to defend the nest site or even to test the mate quality of an attacking male, the pattern of maternal aggression is too closely tied to the welfare of the pups to support these views (Maestripieri 1992). If territorial defence of food sites was the reason for aggression, we would expect animals to defend food sites away from the nest which they do not (Maestripieri 1992). If the goal of the aggression was to test the ferocity of the male intruder then it is hard to explain why females are most reluctant to attack large males compared to small ones. It is also theoretically unsatisfactory to propose that a female would abandon her current investment and prefer the uncertain possibility of mating anew with a strange male. The costs of repeating this kind of behaviour during a lifetime would be a severely reduced, if not non-existent, progeny. Maternal aggression begins during pregnancy and continues through lactation, suggesting a strong hormonal component (Rosenblatt et al. 1994). The duration of maternal aggression corresponds with the period when the young are most vulnerable. It disappears when the pups are removed for five hours but is restored five minutes after the pups are returned. If the pups are attacked and killed, maternal aggression switches off immediately. Females discriminate between the danger posed by different intruders, reserving their most ferocious attacks for those males who are most likely to harm the young (sexually naive, newly arrived and recently mated males) rather than paternal males or other females. The severity of maternal attack is directly related to the size of the litter that she is protecting. Does maternal aggression work? Natural selection would only have favoured such a life-threatening act as attacking an intruding male if it enhanced reproductive success in the long run. In the Southern elephant seal, orphans are bitten three times as often as pups with mothers present and the most aggressive mothers reared the largest pups. In Northern elephant seals, mothers whose offspring survived to weaning showed more aggression than those whose pups died. Their pups were less often bitten or shaken and the frequency of these attacks was negatively related to maternal aggression. Maternal aggression has been shown to effectively drive off intruders among polar bears, lions, mice and rats.

Mhoc03.fm Page 96 Friday, December 14, 2001 10:15 AM


    

Primates provide a different slant on maternal aggression because, in some species, threats to the infant may come from other females as well as infanticidal males—an aspect of female reproductive competition that we have already discussed. Pulling, rough handling, kidnapping and direct aggression against young are performed by other females in a number of primate species (Maestripieri 1992). Ninety per cent of infant attacks in macaques come from other females (Troisi et al. 1989) and preventative maternal aggression (interrupting the attack before it can get under way rather than punishing it afterwards) is associated with a reduced probability of infant harassment. In these species, females display a linear dominance hierarchy that is based on matrilines. High-ranking females have little to fear from lower-ranking females and can count on the support of their equally well-placed kin. This leads to a distinct pattern of maternal protection of young that depends on the females’ standing in the group (Maestripieri 1994). Low-ranking mothers protect their infants by restraining them—they prevent the infant from breaking contact with them by holding its tail or leg. High-ranking mothers accord their privileged offspring greater freedom because they have less to fear from launching an attack on any foolhardy low-ranking female that tries to abuse them. When a low-ranking mother is forced to attack a higher-ranking, infant-abusing female she is more likely to be counter-attacked. Yet these low-ranking mothers are far more willing to attack in protection of their infant than for any other reason. The value of saving their infant outweighs the costs of potential injury. In the face of the ultimate danger to their reproductive success, females attack and they attack severely. But what mechanism accounts for this? If fear is as critical as I have argued then fear should be implicated in some way and most scientists think that it is. But they differ quite dramatically in the mechanism that they suspect to be involved. Some believe that it is extreme levels of fear that switch on the especially severe form of aggression used by mothers—let’s call them the high-fear theorists. Others take a diametrically opposed view that it is a temporary reduction in fear that enables the mother to attack with such ferocity—let’s call them the low-fear theorists. At present, we simply do not have enough information to decide between them and it may take several years before we can say anything conclusive about humans. Practically all the experimental work has been done on rodents and even here there are important differences in the way different strains of the same species react (Parmigiani et al. 1999). Administration of the same drug is sometimes reported to increase and sometimes to decrease aggression (Ferreira et al. 2000; Palanza et al. 1996). The effects of a drug can switch from pro-aggressive to anti-aggressive depending on the dosage (Miczek 1974). Nevertheless, we can make a brief visit to the laboratories to find out the nature of the debate about fear and maternal aggression. The high-fear theorists begin by making a qualitative distinction between two forms of aggression. Offensive aggression is used by female rodents principally against other females and these attacks rarely cause severe injury. In rats, offensive attacks are lateral, accompanied by raised fur as a form of threat, and involve sharp nips on

Mhoc03.fm Page 97 Friday, December 14, 2001 10:15 AM

    :   


the opponent’s back. The aim seems to be to warn the intruding female away from the space that the attacker considers her own. Crucially, these attacks involve very low levels of fear on the part of the aggressor (Palanza et al. 1996). But females show a very different response to a male intruder because they are far more likely to have infanticidal intent than females. A dispute with a male is not about space but about her pups’ survival. Here the female uses a defensive attack and it takes a much more serious form. In rats, this form of aggression involves the jump–bite attack— directly facing her opponent, she lunges forward, directing her bites to the most vulnerable regions of his head and ventral surface (Lucion and de Almeida 1996; Palanza et al. 1996). Males evoke far more fear than females (Parmigiani et al. 1988) and the mother’s vicious attack occurs without warning or the use of threats (such as piloerection). It is a desperate last-ditch stand against the loss of her offspring. High-fear theorists believe that these qualitatively different forms of attack are mediated by fear levels. When fear is high, the attack is defensive and severe; when fear is lower the attack is offensive and serves to warn off the opponent (Blanchard et al. 1977; Brain 1981). Any drug that reduces fear ought to shift the pattern of attack down from defensive to offensive. Benzodiazepines are a very effective class of anti-anxiety drug. Highly anxious mice tend to have a poor ability to bind benzodiazepine and have fewer benzodiazepine receptors (Parmigiani et al. 1999). Just as we might expect, the artificial administration of benzodiazepine reduces defensive aggression far more effectively than offensive aggression (Rodgers 1991; Rodgers and Waters 1985). Among mice mothers who had not been selected for being especially aggressive, those who received chlordiazepoxide hydrochloride (CDP) showed both reduced fear and a shift from defensive to offensive attack on male intruders. Mothers who had been screened as being highly aggressive also shifted from a defensive to an offensive pattern although the researchers could not detect any effect of the drug on fear levels—however, they admit that the level of fear in these screened females was already so low that there was little possibility of the drug lowering it any further. As the authors of the study concluded: ‘The aggression against males under CDP treatment became—phenotypically—like that shown against female intruders in the control group, i.e. . . . offensive attack pattern and low fear levels’ (Palanza et al. 1996, p. 181). Other high-fear researchers have also found that maternal aggression is associated with high levels of sympathetic nervous system activation, such as increased and arrhythmic heart rate which is a good proxy for fear (Sgoifa et al. 1995). The highest level of fear, in this model, shifts normally moderate female aggression upwards in a step function to a life-risking attack: ‘Even with a deep strong general behavioural inhibition, the female still performs active behaviors such as lunge towards the intruder; which can be interpreted as reflecting the importance of parental behaviour’ (Lucion and de Almeida 1996, p. 371). If these theorists are right then lactation does not diminish fear but leaves it very much intact. High levels of fear prompt the mother to ‘select’ the course of action that may cost her life (if she fails) but that also

Mhoc03.fm Page 98 Friday, December 14, 2001 10:15 AM

      carries the prize of saving her pups (if she succeeds). In this view, maternal aggression is an adaptation that depends on fear. Mothers launch their lethal assaults not despite the fear but because of it. The low-fear theorists see things just the other way around. They do not distinguish qualitatively distinct types of aggression but see aggression as a straightforward continuum from mild to severe. Consequently, their laboratory measures focus on attack latency and the frequency of a range of aggressive behaviours including biting, kicking and boxing. As they see it, the more fearful the animal is, the less likely it is to launch a severe attack. It is their fearlessness that opens the door to aggression. Immediately after giving birth, rat mothers show a whole host of maternal behaviours: they build nests, they retrieve pups that stray from it and they lick, sniff and nurse their offspring. They also show a significant reduction in fear and anxiety (Ferreira et al. 1989; Fleming and Luebke 1981; Hard and Hansen 1985) combined with increased aggressiveness towards intruders (Erskine et al. 1980; Ostermeyer 1983). Researchers have suggested that these two effects are linked and both are related to an increase in GABA–benzodiazepine neurotransmission that was found in females who had just given birth. This endogenous neurochemical system acts to reduce anxiety and simultaneously to increase the readiness to attack. Especially in nonaggressive females with little fighting experience, the administration of diazepines increase their willingness to attack. This increased aggression is attributed to their reduced anxiety (Miczek and O’Donnell 1980; Mos and Olivier 1989). This chimes with other facts about aggression. Mice from genetic lines that show high levels of aggression also show lower levels of anxiety (Parmigiani et al. 1999) and a high level of anxiety is inversely related to the probability of showing maternal aggression (Maestripieri et al. 1991). Lesions at a brain site that specifically increases maternal aggression to a threatening male (but not a juvenile intruder; Lonstein and Stern 1998) also results in reduced fearfulness (Lonstein et al. 1998). Mothers show more fear of dominant males than subordinates and attack subordinate males more intensely (Palanza et al. 1994). We do not yet understand the neurochemical route to reducing a mother’s anxiety. A number of candidates have been considered. The GABAergic system has already been discussed. Increased serotonin levels can decrease maternal as well as other forms of aggression (Ferreira et al. 2000; Olivier et al. 1995). Normal lactating mice show an increase in the production of nitric oxide in the hypothalamus (Gammie and Nelson 1999). Female mice lacking the gene responsible for the production of nitric oxide show significantly lower levels of maternal aggression although they show no alteration of other maternal behaviours. Since maternal aggression coincides with lactation, a useful place to search for an answer is with an examination of the hormones that are specifically associated with delivery and milk production (Carter and Altemus 1997). Oxytocin, as we have seen, is a neuropeptide made in the hypothalamus that stimulates the contractions that expel the infant from the uterus. It is responsible for

Mhoc03.fm Page 99 Friday, December 14, 2001 10:15 AM

    :     the milk ‘letdown’ reflex and is triggered by the nipple stimulation of suckling. Oxytocin has been called the love and bonding hormone. It has very special effects on mothering (Angier 1999). Administration of oxytocin to virgin female voles induces maternal behaviour—they become solicitous of young, sniffing and retrieving them when they stray. In sheep, a ewe will reject an infant if she does not lick the amniotic fluid from it within minutes of birth. Yet if oxytocin release is artificially stimulated by vaginal stimulation she suddenly renews her interest in the previously abandoned orphan and allows it to suckle. In rhesus monkeys also, this hormone triggers approach, inspection, grooming and kissing of infants. Psychologically, oxytocin promotes a feeling of well-being and tranquillity. It enables the growing sense of love and attachment to the infant. And the more the infant sucks, the more oxytocin is produced. During lactation, the normal hypothalamic stress response is inhibited. When lactating rodents are exposed to stressful experiences (like electric shock) they show a much weaker stress reaction as measured by corticosterone secretion. In humans, stress hormones fall during bouts of breast-feeding and artificial administration of oxytocin by injection to men and women also has the effect of drastically reducing their production of stress hormones. Among women suffering from anxiety and panic attacks, the symptoms decline during pregnancy and remain low during breast-feeding which also suggests that oxytocin is lowering these inappropriate stress responses. Evidence from animals suggests that oxytocin increases exploratory behaviour (by reducing fear), decreases blood pressure and can even block the experience of pain. Oxytocin seems to be a ‘de-stressing’ hormone that activates the parasympathetic nervous system responsible for keeping basic body functions (such as heart rate, breathing, muscle tone and digestion) ticking over in a relaxed state. It is a hormone with a double effect on mothers. It increases their attachment to their infant, promoting the feeling of love that is the profound reward of mothering. It makes her infant valuable to her. It also suppresses the fear that would normally cause her to back-off from threat. It looks like a strong candidate for simultaneously loving an infant and being prepared to fight for its life.

Fear and fitness The differences between men and women in their willingness to move from conflict to violence is tied to the reproductive strategies of the two sexes. The advantages of polygyny for men brought with it direct competition between them for mating opportunities and from this evolved a greater disregard for their own safety. Aggression performed two interlinked functions. It intimidated other men and in so doing it won the victor status. The more successful he was, the greater freedom he had, not only to claim fertile females for himself but also to monopolize other disputed resources such as food. These twin benefits increased reproductive success in direct and indirect ways. Directly, other males were more inclined to withdraw from conflicts over sexual rights to women. Indirectly, his ability to claim the lion’s share

Mhoc03.fm Page 100 Friday, December 14, 2001 10:15 AM

      of food and to use his status to protect mates increased his attractiveness to females. He became a more desirable mate. So for young males, physical aggression brought rewards that made even noble failure preferable to abstaining from the competition. For females, aggression had costs that males did not have to face. With mothers undertaking parental care, a male’s death did not compromise his current reproductive success. For females, injury or death meant the loss of her current offspring. Such a price counselled conservatism in risk taking. Avoid unnecessary conflict and stay away from physical aggression—both threaten the huge investment that a woman had made in rearing her dependent young. I believe that the mechanism that underlies this sex difference is fear. For males, low levels of fear open the door to extravagant displays of bravery and combat that can be used to achieve a reputation in the community. The prize is status and all the benefits that come with it. But wouldn’t status be useful for women too? Might highstatus women not gain better mates, more food and stronger support from allies? If so, why are so few women at the head of multinational corporations and governments? Does the glass ceiling erected by men simply stop them? Are they more concerned with raising children than running banks? Or could it be that status doesn’t provide big enough rewards to outweigh the costs?

Mhoc04.fm Page 101 Friday, December 14, 2001 10:16 AM

 


Who does she think she is? Women and status

Status is a matter of some importance to males—a point that has not escaped most women’s attention. Certainly not that of poet Wendy Cope (1992): One man on his own can be quite good fun But don’t go drinking with two They’ll probably have an argument And take no notice of you. What makes men so tedious Is their need to show off and compete They’ll bore you to death for hours Before they’ll admit defeat. From an evolutionary point of view of course, men engaged in status competition are far from taking ‘no notice of you’. Women are the limiting resource for men and it is reproductive access that drives men’s quest for status. The more dominance a man can gain, the greater his chances of securing the most attractive women. Nobody is suggesting that men consciously know this (although male one-upmanship seems to be energized by a female audience). In fact, as Cope rightly observes, male–male status jockeying often seems to exclude women completely. But 100,000 years of sexual selection have shaped the male mind to be especially sensitive to slights that call into question their social standing and they respond less than philosophically to other men’s attempts to assert their superiority.

Dominance as a male goal As we have seen in the last chapter, the link between dominance and reproductive success grows murkier as we ascend the evolutionary ladder. What of humans? Can we detect any relationship there? Among our own species, status depends upon much more than the ability to physically dominate others. Cultural success may involve the accrual of material goods (the wealthy), the achievement of recognized power over others (in organizations or government) or the possession of a specially

Mhoc04.fm Page 102 Friday, December 14, 2001 10:16 AM

      valued knowledge and competence (doctors, athletes). Cultural success can be recognized by the deference we accord to certain members of society. The routes to cultural success in a meritocratic society depend upon a host of personal factors including determination, natural ability and social intelligence. Looking back through history, there can be no doubt of the extraordinary reproductive success of men whose wealth and power allowed them to keep multiple wives and harems (Betzig 1986; Dickemann 1979). Julius Caesar was happy to find that during his absence a bill had thoughtfully been drawn up legitimizing his union with ‘any woman or women he pleased—for the procreation of children’ (Betzig 1997). His great nephew Augustus, renowned for his humility and common touch, nevertheless was supplied with women by his wife Livia as well as by his friends. ‘His friends used to behave like Toranius, the slave dealer, in arranging pleasures for him—they would strip mothers of families, or grown girls, of their clothes, and inspect them as though they were for sale.’ Ismail the Bloodthirsty, though limited to only four wives, kept 500 concubines each housed in their own accommodation and tended by a female slave and a eunuch (Daly and Wilson 1988). In China, emperors kept royal harems of 1000 women. The system was managed by female supervisors and the women were brought to the emperor on rotation according to a calendar that guaranteed that they were at mid-cycle. One Indian potentate fathered four children in one week and nine more the following week. The key to this kind of unimaginable reproductive success was simply power. Kings with limitless resources could afford as many concubines as they liked, and like them they did. But were the benefits of dominance confined to despots and aristocrats who had absolute power to recreate social arrangements to their own advantage? Does the link between dominance and reproduction survive in societies that have much smaller wealth and power differentials? The most egalitarian societies that have been described are hunter-gatherers and foragers. These people live on the edge of survival in the sense that they cannot accumulate food but subsist on their daily excursions to hunt or gather vegetation. This hand-to-mouth existence evens out differences in rank that accrue from resource surplus and trade. Hunter-gatherers are rarely polygynous for the simple reason that men cannot acquire enough resources to support two wives. The Ache of Paraguay live in this way and their egalitarian lifestyle is exemplified in their sharing of food. Small mundane nutritional packages such as vegetable and insects which are gathered daily are shared only with close kin (Hill and Kaplan 1988). But more exotic, desirable and unpredictable foods such as meat and honey are shared across the whole band. When a monkey has been killed and prepared for consumption, it is cut up and distributed among all the family units. On any given day, 90 per cent of hunt food that is consumed by any given family is killed by a non-family member. This seems to work to reduce day-to-day variation in food intake—to even out the difference between good days and bad days for each hunter. Nonetheless hunters vary in quality. Some do much better than others and a key question is why successful hunters tolerate the poor

Mhoc04.fm Page 103 Friday, December 14, 2001 10:16 AM

        


performance of others by feeding them. Why don’t they give the food selectively to their own families since, on average, they do better than other men? A number of solutions have been proposed. Under the tolerated theft theory, a man might as well share a large piece of food rather than allow it to rot or waste his effort fighting for it. Under the cooperative acquisition theory, individuals share in order to maintain cooperative bonds that they will need for the future. But an intriguing proposal has been made which has implications for both dominance and reproductive success (Hawkes 1990; Hill and Kaplan 1988). First, by sharing with others, a good hunter makes himself indispensable to the group. Ache families often change residence during their lives and the kudos of his hunting may make other band members particularly indulgent to him as a way of ensuring his continued loyalty. This elevation in status has direct implications for the welfare of his wife and children and they may receive better treatment from others. But his generosity and derived status has more carnal payoffs. Although good hunters do not differ from poor hunters in the number of children born to their wives, they are more often reported by women as extramarital lovers, have more illegitimate children and both their illegitimate and legitimate children have higher chance of survival (Hill and Kaplan 1988). This pattern also seems to hold for other groups such as the King San, the Hadza, the Ifaluk and for pre-twentieth century European populations (Hawkes 1993; Hill and Kaplan 1994; Perusse 1993). In a survey of nearly 300 cultures, Gregerson (1982, p. 84) concluded: ‘For women the world over, male attractiveness is bound up with social status, or skills, strength, bravery, prowess and similar qualities.’ But something strange has begun to happen in the last hundred years. Wealthier families do not differ from low-income families in their number of children and in some cases have been found to have fewer children. The relationship has been found by plotting fertility against family income in a variety of Western societies. How could this be? Does this not show that something is seriously awry with evolutionary theory? Daniel Perusse (1993) decided to find out. He collected data on the sexual, reproductive and economic histories of 1,133 French Canadians. He wanted to know if the demographic shift was a result of the imposition of monogamy or the use of contraceptives. First he confirmed that there was no relationship between the social class of the father and the number of children ever born to him (by his current wife, ex-wives or unmarried partner). Then he computed the number of potential conceptions that could have occurred if contraception had not been used—he derived this from a measure of the number of partners the man had had and the number of sex acts that took place with each of them. Here he found a positive correlation—the men with higher social status would potentially have fathered more children if contraception had not been available. But, he reasoned, the relationship that he found was based on all the men in the sample—including married men. These men are constrained by monogamy and so unlikely to seek multiple sexual partners. When he analysed the relationship again, examining only unmarried men he found it was even stronger and did not vary markedly with the age of the man. So it appears that both monogamy and

Mhoc04.fm Page 104 Friday, December 14, 2001 10:16 AM

      contraception reduce the variance in male mating success and that both have the effect of diminishing the relationship between status and reproductive success in men. But just as our evolved liking for sweetness has not (yet) disappeared because sugar has become more easily available, so contemporary social arrangements have not had sufficient time to alter the basic male interest in dominating other males.

Dominance in men’s interpersonal style There are two important dimensions of social behaviour that have appeared and reappeared in a variety of psychological studies. Although they were derived in different ways and given different names, the content of the dimensions is remarkably consistent. On the one hand, we have agency (also called instrumentality, masculinity or dominance). Agentic people strive for individuation within the group. They seek to be independent and autonomous. They believe in meritocracy and the value of competition. They see the rules as a necessary means for regulating competition and are concerned with fairness more than with compassion. On the other hand, we have communion (also called expressiveness, femininity or nurturance). High-communion people strive to be part of a larger whole, to merge themselves into the group. They view relationships in terms of meeting others’ needs rather than a tit-for-tat exchange of favours. They believe in equality rather than equity. They value intimacy and mutuality in their relationships and define themselves in terms of their close friends. We all recognize ourselves in both of these descriptions and with good reason. They are independent of one another. A person’s score on one does not predict their score on the other. We are all a mixture of each. Nonetheless, sex differences have been reliably found on both dimensions (see Campbell 1998). We can trace the twin themes of agency and communion back as far as Confucius (Wiggins 1991). But more recently, psychologists have reinvented them not as philosophical notions but as measurable aspects of personality. The dimensions form the vertical and horizontal coordinates of the circumplex model of personality (Freedman et al. 1951; Kiesler 1983; Wiggins 1979) which used factor analysis to distil large numbers of self-rated traits to two manageable dimensions. Their analysis suggested that interpersonal behaviour can be seen as a circle with north–south indicating the opposing traits of dominance and submission and with east–west identifying the complementary traits of nurturance and hostility. (As we move around the perimeter we find blends of each two contiguous styles so that, for example, between nurturance and dominance we find extroverted gregariousness.) There are significant sex differences with men scoring higher on dominant traits, such as self-confidence and assertiveness, and women scoring higher on nurturing traits, such as kindness and being tender-hearted. Dominance and nurturance are highly correlated with Bem’s (1974) dimensions of Masculinity and Femininity. Bem asked 200 judges to rate the desirability of 200 adjectives in men and women. From the data, she selected the 20 items that were most

Mhoc04.fm Page 105 Friday, December 14, 2001 10:16 AM

        


stereotypically associated with men (such as forceful, individualistic) and with women (such as compassionate, warm) and used them to form the Bem Sex Role Inventory (BSRI) on which subjects are asked to rate how well each adjective characterized them. She found that the two dimensions were statistically independent of each other (as had been found for dominance and nurturance) and that there were significant sex differences in how subjects rated themselves. Men rated themselves higher on traits such as self-reliant, assertive, willing to take risks—a dimension she called Masculinity. Women rated themselves higher on adjectives such as loyal, affectionate and sensitive—which she called Femininity. But subsequent work on the BSRI has revealed problems. Most importantly, she included as items on the questionnaire the adjectives ‘masculine’ and ‘feminine’. These items are very highly negatively correlated. Someone who describes themselves as extremely feminine will almost certainly describe themselves as not at all masculine. This suggests that masculinity and femininity are opposite poles of a single dimension, rather than being independent of one another as we would expect. Furthermore, these two items do not load on the Masculinity and Femininity dimensions. The extent to which you describe yourself as feminine is not highly correlated with whether or not you rate yourself as warm or compassionate. People’s essential sense of their own sexual identity treats masculinity and femininity as mutually exclusive and how feminine or masculine they feel is not reliably predicted by the number of feminine- or masculine-stereotypic traits they endorse. This led to Janet Spence (1985) to critically re-examine her own work that had been undertaken at the same time and in the same way as Bem’s. Her Personality Attributes Questionnaire scales showed an extremely high correlation with Bem’s scales and, like Bem, she also found significant sex differences. However, she argued that neither of them had measured what they thought they had. They had believed that they were capturing the essential nature of masculinity and femininity but in fact they had uncovered two dimensions of social behaviour on which the sexes differed. To prevent further confusion, Spence argued that these dimensions should be called instrumentality and expressiveness and that we should clearly acknowledge that how we score on these two components of personality are distinct from the private ineffable sense of ourselves as masculine or feminine. Instrumentality, dominance or agency is the dimension that is most relevant to understanding sex differences in status (see Cross and Madson 1997). We shall take a closer look at expressiveness and communion in the next chapter.

Growing up competitive Sex differences in dominance behaviour begin early and continue into adulthood. Boys have been found to be more assertive than girls at 13 months of age (Goldberg and Lewis 1969). Between the ages of 2 and 4, boys attack people, fight and destroy things more often than do girls (Campbell et al., in press; Koot and Verhulst 1991). At preschool ages, boys dominate girls (Pellegrini and Perlmutter 1989) and, as we shall see, this is an important reason for the sex segregation that children

Mhoc04.fm Page 106 Friday, December 14, 2001 10:16 AM


    

everywhere spontaneously practice from about the age of 3 years onwards (Maccoby and Jacklin 1987). Boys are insensitive to negative girls’ reactions to their style of behaviour and girls find themselves overwhelmed by the abrasive and dominating tactics of boys (Maccoby 1990). When given a choice, boys prefer to compete while girls prefer to cooperate (Ahlgren 1983; Boehnke et al. 1989; Charlesworth and Dzur 1987; Moely et al. 1979; Strube 1981). This preference is reflected in the play styles that the two sexes adopt. Lever (1978) distinguished between play (a cooperative activity in which there is no winner and no clear end point) and games (competitive interactions governed by rules aimed at achieving an explicit goal). Boys spend 65 per cent of their free time in games compared to 35 per cent among girls. When games are played, boys choose ones that involve direct zero-sum competition while girls prefer turn-taking games such as skipping or hopscotch where the form of competition is more indirect. In fact, girls seem to dislike situations where winning means that another person must lose (Curry and Hock 1981). When children are explicitly told to compete against all the other players, boys are happy to do so but girls still form small friendship cliques (Hughes 1988). These small dyads and cliques are girls’ favourite way of playing but boys prefer large groups in which they can play team sports such as football and baseball. Boys enjoyment of team sports has not altered over the past three decades (Sandberg and MeyerBehlberg 1994; Sutton-Smith et al. 1963). Although there are plenty of arguments during the game, boys resolve these by the application of rules, rarely take the disputes personally, and seem to enjoy the adversarial cut-and-thrust. (Girls, by contrast, will abandon a game if it causes arguments; Lever 1978.) Boys’ love of sport may result from their greater preference for gross motor behaviour and propulsion, which is evident cross-culturally (Eibl-Eibesfeldt 1989; Whiting and Edwards 1988) from infancy onward (Eaton and Enns 1986; Lever 1978). But boys also seem to enjoy the active competitiveness of sport and derive particular gratification from the approval of their team-mates. Sports provide the opportunity for individual excellence in a public forum. There is a marked difference between the sexes in their involvement in roughand-tumble play that involves chasing, capturing, wrestling and restraining. This sex difference is evident by the age of three years (Maccoby 1988) and boys engage in rough-and-tumble three to six times as frequently as girls (DiPietro 1981). Despite its superficial resemblance to fighting, it lacks true hostility. Boys do it more with friends than with non-friends, move easily from a wrestling bout to other forms of play with the same partner and if injury occurs, it is usually accidental. Boys not only seem to enjoy this kind of high-propulsion, contact play, but they also accord it more social significance than girls. Boulton (1996) found that boys, more than girls, can tell who is strongest in rough-and-tumble play, think it is important to win, try to win in order to show that they are tougher and think it is also important to be good at real fighting. This kind of rough play seems to be important in establishing status in the group. Hierarchical dominance relations can be found in boys’ groups from the age of six

Mhoc04.fm Page 107 Friday, December 14, 2001 10:16 AM

        


and once established, most disputes occur between individuals who are adjacent to one another in the pecking order. Boys rate the importance of social dominance higher than girls (Jarvinen and Nicholls 1996). Athletic ability, strength and early maturation are all associated with dominance in boys (Tanner 1978; Weisfeld et al. 1987) and at high-school age, girls also are also attracted to athletic boys. These statusenhancing abilities seem to be established early and a boy’s dominance at age 6 predicts his dominance nine years later (Weisfeld et al. 1987). In a summer camp study Savin-Williams (1987) examined the formation of dominance hierarchies both quantitatively and qualitatively. Dominance once established was remarkably stable over time and over settings. The best predictors of the overall dominance were overt actions such as physical assertion, verbal directives, threat and recognition. Those who ranked highly tended to be athletic, early maturers, physical fit, larger in size and more attractive. The flavour of the dominant position and the camp experience is nicely conveyed by Savin Williams (pp. 77–8): Perhaps because of his status position, Andy usually received what he wanted. On Day 1 he let it be known that he was going to sit beside SW during meals and he maintained that position throughout camp. On Day 10 Delvin had a choice of passing the cake and, hence, decide who would grab the second largest piece, to Andy or to his best friend Otto. He passed it to Andy. When Gar suggested on Day 16 that perhaps the defense and the offense could switch for one day, Andy ended the egalitarian effort by retorting ‘Hell, no one is going to stick me on defense!’ When Andy was in the top bunk, Gar frequently complained that his feet were hanging over the side of the bed in his face. Andy’s usual response was ‘Tough! Move your face.’ After they switched positions at mid-camp, Gar quickly moved his feet when Andy complained they were in his face.

Savin-Williams found that there were five ways in which the boys used speech to assert their dominance; giving orders, calling people names, threatening or boasting, refusing to obey orders, and winning arguments. Maltz and Borker (1982) also found that boys used speech chiefly to assert a position of dominance, to attract and maintain an audience and to assert oneself when another speaker has the floor. They emphasize that dominant boys are not mere verbal bullies. The important thing is not the content of the speech, but what successful boys can achieved by it. Speech is about claiming, holding or disputing status. For boys, maintaining an audience is a real skill because, unlike among girls, the speaker cannot count on a patient and sympathetic audience. Storytelling, joke telling and other narrative performance events are common features of the social interactions of boys . . . The storyteller is frequently faced with mockery, challenges and side comments on his story. A major sociolinguistic skill which a boy must apparently learn in interacting with his peers is to ride out this series of challenges, maintain his audience, and successfully get to the end of his story. (Maltz and Borker 1982, p. 208)

The kind of competitive cut-and-thrust that seems more pleasurable to boys than girls may be reflected in different physiological responses to competitive situations.

Mhoc04.fm Page 108 Friday, December 14, 2001 10:16 AM


    

Ruasta-von Wright et al. (1981) followed 18-year-olds taking their final school examinations. and found that adrenaline and noradrenaline increased to a higher degree in males than females. Among males there was a high correlation between these catecholamine levels and success on the examination. There was also a high correlation with overachievement (defined as a better performance than could have been predicted from IQ scores). Interestingly, the higher the catcholamine level, the higher was the man’s self-image. Among women, correlations between a variety of physiological and psychological measures was zero. The researchers suggest that competitive situations activate adaptive physiological reactions in males, alert them mentally and gear them to better performance. By adulthood, men more than women describe themselves as competitive, independent, dominant and characterize themselves in terms of self-sufficiency and power (Simmons 1987; Wylie 1974). Men rank personal relationships as less central to their sense of identity than do women (Thoits 1992). Men’s self-esteem but not women’s can be predicted two years in advance by knowledge of their sense of independence (Stein et al. 1992). As they enter adulthood, ‘Young men with high self-esteem seemed self-focused and defensively critical, uneasy, and unready for a connection with others’ (Block 1993, p. 28). Status can be enhanced by comparing oneself to less successful peers and men’s self-esteem seems to derive from just such comparisons: ‘The less positively men view their peers, the higher their self esteem’ (Goethals et al. 1990, pp. 170–1). Even in childhood, boys more than girls become friends with children who perform worse than themselves on important tasks (Tesser et al. 1984) and distance themselves from siblings who are better than them at school or in sports (Tesser 1980). Men who believe that their romantic relationships are superior to those of their peers express greater satisfaction with the relationship than men who think that their relationship is more similar to those of other men. Men tend to overestimate both their unique qualities (Goethals et al. 1990) and their abilities in comparison with their objective performance on a task (Lenney 1977). Men’s concern with status differentials has implications for their friendships. Because an element of competition always exists, men are wary about engaging in very self-disclosing relationships with other men. Such information might one day be used against them. As Derlega et al. (1981, p. 445) put it: ‘if men do not disclose personal information, other people cannot understand, predict or control their behavior.’ Although men will disclose facts about themselves they are much less willing than women to share emotions. They are particularly unlikely to share negative emotions that reveal vulnerability such as depression, anxiety and fear (Snell et al. 1989). Men’s conversations with other men tend to focus on shared activities or on depersonalized topics such as politics and sport (Aries and Johnson 1983). Compared to women, they are more likely to define intimacy with other males in terms of shared activities rather than shared emotional experiences (Caldwell and Peplau 1982). However, when they are with women rather than men, men make emotional disclosures

Mhoc04.fm Page 109 Friday, December 14, 2001 10:16 AM

        


more often and they report that relationships with women are more meaningful, intimate, satisfying and pleasant than male–male relationships (Derlega et al. 1981). Men and women differ too in the attitudes they hold to status differentials. Men support group differentiation by holding more positive attitudes than women to conservatism, racism, militarism and violence. They are less favourable than women to social welfare and to minimum wage legislation and more favourable to defence spending, sending troops to foreign countries and the death penalty. These political views derive from what Pratto (1996) calls a stronger social dominance orientation—a preference for hierarchy over equality. Sex differences in attitudes to specific political issues (such as gender and gay/lesbian rights and social welfare spending) are all explicable in terms of the different orientation that the two sexes have towards the basic principle of between-group equality. Men see inequality as a natural result of competition and are unfavourable to policies that are designed to dilute or mitigate that competition. As leaders, men operate in a more autocratic style compared to women by taking charge and discouraging input from subordinates (Eagly and Johnson 1990). Men are also more likely to emerge as leaders in originally leaderless groups, especially when leadership is concerned with direct contribution to the group’s task (Eagly and Karau 1991). This greater task leadership by men was found no matter how leadership was assessed—whether by analysis of verbal behaviour, group participants’ votes, or by productivity. Men inhabit a world that is implicitly competitive and this pervasive sense of competition colours their sense of identity and relationships with others. Too much intimacy with other men can be dangerous—it can reveal vulnerability that might be used against them and it generates a closeness that inhibits their ability to overtly compete with one another. Men crave closeness with other men and achieve it often but they do so without revealing their private feelings—this they reserve for women from whom they need not fear competition. Men’s relationships are more like alliances in which they support one another and share their interests and activities, but always with wariness. One drink too many and even friends can find themselves competing to tell the best story, to offer a better argument for their political views, or to challenge the superiority of the other person. The end of a friendship, which is often devastating to women, does not prompt the same sense of loss because, deep inside, men are always on their own against the world. To lose an ally is not the same as to lose a soulmate.

Why female dominance buys We have seen why dominance matter to males. As Daly and Wilson (1983, p. 279) put it: ‘They wheel and deal, bluster and bluff, compete overtly both for valuable commodities and for mere symbols. Ultimately these male machinations reflect a struggle for access to female reproductive capacity.’

Mhoc04.fm Page 110 Friday, December 14, 2001 10:16 AM

      Note that what males struggle for is mere ‘access’. As we have seen, copulations do not inevitably lead to pregnancy (and if they do, there is no guarantee of paternity), pregnancy does not inevitably lead to live births and live births do not inevitably lead to the production of an adult capable of reproduction in their own right. So the whole edifice of male status competition is built around the mere opportunity of leaving a genetic legacy. Given that maternity is certain, that female reproductive variance is lower than males, that female bodies are highly selective about which zygote will go to term, that maternal investment is high and that only a fraction of infants survive—should we not expect to see even greater concern with status among females? The advantages of dominance for females are not measured in mere ‘opportunities’ to produce young but in the survival of existing young in whom females are already invested. Should competition for status not be more rather than less intense among females? To answer this question, we have to begin by looking at the rewards for dominance in females. Remember that females are not in general competing for copulation. Female primates do sometimes interrupt another female mating with a male but this is likely to be a way of decreasing another female’s likelihood of pregnancy rather than competition for copulations per se. (However, low-ranking vervet females do not approach preferred males when high-ranking females are about, presumably fearing reprisals from above; Keddy 1986.) Even if they did routinely compete for mating opportunities, primate studies have generally concurred that males do not show a strong or systematic preference for high-ranking females (Loy 1971; Packer 1979; Small and Smith 1985) so female rank does not guarantee better quality partners. What females need most is a regular and ample food supply and dominance could provide this. One explanation that has been offered for female’s low levels of direct competition that we discussed in the last chapter is that a single piece of food is not worth risking a life for (Smuts 1987). But what if there were a way to guarantee continuous and permanent priority of access to any disputed food? Such a reward would surely be worth fighting for. Regular food intake is vital in a number of ways. Food intake accelerates growth so that juveniles attain puberty earlier. Low body weight is associated with infertile cycling and in primate studies that have used artificial feeding of free-ranging populations, the additional food increases the number of offspring born. Food is necessary to sustain gestation and lactation. After weaning, food requirements increase even further in species like chimpanzees where the mother continues to provide solid food for her young. Fat stores can also carry a female through periods of famine. Dominance may have other rewards too. Highstatus animals are not subject to harassment (they are usually the ones who inflict it on others) and so are less likely to experience stress-induced reproductive suppression. They are more likely to have allies who will come to their assistance in a dispute. Their infants are less likely to be handled, kidnapped or killed by other adults. And if rank is heritable, then their young will themselves grow up to experience all these rewards.

Mhoc04.fm Page 111 Friday, December 14, 2001 10:16 AM

        


Does dominance increase female reproductive success? Ellis (1995) measured reproductive success with three major indicators: number of offspring born, number of offspring surviving, and the age and point in the breeding season at which first conception occurred. In addition, he also examined the length of between-birth interval, infliction of reproductive suppression, number of copulations, strength of bonds with males, harassment of others during copulation and the length of the females’ life span or reproductive career. Among rodents there is unanimity that higher status is associated with more offspring, percentage surviving to maturity, reproductive suppression (inflicted more frequently by higher status females) and frequency of copulation. Among other mammals the data are more mixed but the general trend is towards increased reproduction by higher status females. Macaques are the primate that has been studied most extensively. Of the 32 studies that examined number of infants born, only 13 found a significant effect of mother’s status. However, about two-thirds of studies found that high-status females had greater infant survivorship and an earlier age at first birth. For the other indicators, about half the studies showed some positive effect of social status. Among baboons, studies generally find an association between status and number of offspring born and about half showed that higher rank predicts infant survivorship, age at first birth and reproductive harassment. But one study that was able to examine lifetime reproductive success found no link with dominance rank (Packer et al. 1995). The advantages of shorter interbirth intervals, improved offspring survival and accelerated maturation of daughters were counterbalanced by reduced fertility and higher rates of miscarriage, perhaps resulting from stress and injury. The authors concluded that the advantages of competition for rank are ‘balanced by significant reproductive costs and may thus be subject to strong stabilizing selection’ (Packer et al. 1995, p. 60). Studies that have looked at other monkeys usually find that higher rank is associated with earlier age at first conception, shorter birth intervals and reproductive suppression and about three-quarters find that rank increases the number of offspring born. Among chimps, higher status females have greater offspring survival, reach sexual maturity earlier, live longer and have a larger annual production of infants surviving to weaning age (Pusey et al. 1997). Although Ellis concludes that the relationship is less firmly established among females than among males, a few caveats have to be born in mind. There are fewer studies of females which make conclusions less certain. And the criteria for reproductive success are much more stringent. Remember that among males frequency of copulation is the major indicator but only a fraction of these copulations will result in a live birth. In females, reproductive success is measured in a far more conservative way by offspring born and surviving. Another important factor, and one to which we shall later return, is that dominance position is often harder to detect in females than in males. Their general level of aggressive interactions with one another is much lower (among chimpanzees adult male aggression is 20 times higher than among females; de Waal 1993) and this is one means by which dominance is

Mhoc04.fm Page 112 Friday, December 14, 2001 10:16 AM

      established. In chimps, although female dominance relations can be detected over several years, they are vaguer and less linear than among males (de Waal 1993). This kind of unreliability in the measurement of dominance and the more stringent criteria for reproductive success diminish the likelihood of finding significant associations for females.

The costs of dominance Although female behaviour was neglected for many years by primatologists, evolutionary theory would predict that dominance ought to be at least as important to females as males and so we would expect to see females jockeying for status in the group. Yet looking at non-human primates we find that only some species show clear dominance hierarchies among females. They occur in female-bonded species such as lemurs, vervets, baboons and macaques (Jolly 1985). In female-bonded animals, females remain in their natal group throughout their life (rather than transferring to a male group in the way that chimpanzees do) and most primates follow this pattern. A maxim of primatology is that social relationships depend principally on females. Males go where the females are and so the way that females disperse themselves gives rise to different forms of male organization designed to maximize their access to matings with these females. Females go where the food is and attempts to explain why most primate species are female-bonded have been driven by examining the feeding patterns of different primate females (van Hoof and van Schaik 1992; van Schaik 1989). If food occurs in well-defined patches that can be monopolized by one or a few animals, females will compete amongst themselves to take control of it, increase their food intake and so increase their reproductive success. In these circumstances, female dominance is important and can be increased by having relatives with which alliances can be formed. These allies also benefit themselves because they share genes with the dominant female and hence her reproductive success is linked to their own. In addition, allies may benefit from assistance from the dominant female when they need it. Dominance relationships tend to be consistent and linear—if female A outranks B then she also outranks C, D and E. Females feeding under such circumstances will have a high emigration threshold—they should be unwilling to leave the natal group. If a female were to emigrate, she would be forced to enter at the bottom of the new group’s hierarchy and she would have no allies to assist her. She is far better off staying at home. If the group becomes too big to sustain itself on the available resources, her best bet is to leave in the company of her relatives and for the whole matriline to set up anew elsewhere. So where food competition is high and must be fought for, females remain with their kin who make the best allies and they show a steep dominance hierarchy. In female-bonded groups, aggression over food is frequent but low-key and has been described as ‘mild bickering’ (Walters and Seyfarth 1987, p. 308). Hrdy (1981,

Mhoc04.fm Page 113 Friday, December 14, 2001 10:16 AM

        


p. 106) notes that ‘females rarely inflict serious damage on one another in their quarrels. Even though female–female quarrels are more frequent than fights between males in most species, encounters between adult males are much more likely to result in one animal being wounded.’ Seyfarth (1976) found that among baboons the ratio of approach–retreat interactions to bouts of overt aggression was 20:1 for females and 1.7:1 for males. Severe aggression is generally agreed to be more frequent among males than females (Smuts 1987). We have seen in the last chapter why females should be unwilling to risk severe assaults. In female-bonded species, everything depends upon which matriline you belong to. Matrilines can be ranked with respect to one another with almost no overlap so that all members of Matriline A dominate all members of Matriline B and so on. There are just three rules that predict a female’s position (Chapais 1992). First, females inherit their mothers’ rank relative to other members of the group. This is important for it means that rank is inherited rather than acquired through direct combat. Secondly, mothers dominate daughters for life. Daughters can rise to their mother’s position in the matriline only upon her death. Third, younger sisters dominate older sisters as adults. This latter rule, which is consistently found to apply, may be the result of mothers favouring the daughter with the highest reproductive value (Chapais 1992) or it may be a strategic maternal decision. To prevent her daughters from forming an alliance against her, she operates a divide-and-rule favouritism so that the younger daughter cannot improve her rank position by forming a coalition with her older sister (Horrocks and Hunte 1983). The rule-governed system of rank inheritance is a million miles away from how males acquire rank. Chimpanzee males stay in their natal group but rather than passing rank peacefully from one generation to the next, it is strenuously and dangerously contested. Challenges depend upon strategic alliance formation and are occasionally lethal and nearly always damaging. Adult males frequently carry severe wounds from past dominance contests. But consider the benefits that could accrue to a female who successfully challenged a dominant female. Not only would she immediately increase her food intake, freedom from harassment and reproductive success, but all of these advantages would be passed down to her daughters. Unlike male chimps, her payoffs would not last merely months or years but through several generations. This makes it all the more remarkable that matrilines are extremely stable over time. Walters (1980) found that a juvenile female’s rank at the time of her birth correctly predicted her adult rank in 97 per cent of cases. In a 400-day study of yellow baboons, Hausfater (1975) found not a single instance of an agonistically induced change of status among females. Chapais (1992) examined female rank challenge experimentally by setting up a number of female groupings to investigate the circumstances of such challenges. He concluded that although competition for rank does occur, it also appears somewhat constrained. In all of the 58 experimental subgroups, actual relative power was in favour of subordinates, i.e. the latter were either larger or more numerous than the single highborn female. Despite

Mhoc04.fm Page 114 Friday, December 14, 2001 10:16 AM

      this, subordinate females outranked the single dominant female in only 56 per cent of the 148 dyads tested. It is noteworthy that they outranked them only in situations where the power asymmetry (relative size or relative alliance power) was pronounced. This was referred to as a ‘minimal risk strategy’ of competition for dominance. (Chapais 1992, p. 44)

So in species where feeding patterns induce competition, we see that females remain with their relatives to increase alliance support from their matriline. Day-to-day aggression over food rarely involves injury because individuals know their position in the group and defer when threatened by a higher-born animal. Unlike male chimps (who also remain in their natal group) they do not show political manoeuvring and complex alliance formation to usurp the dominant animal. Such tactics are dangerous and females are rarely willing to risk their life even in pursuit of longlasting benefits. Even though the payoffs for a revolution from below are enormous, they are very rarely undertaken. Our closest relatives are chimpanzees and they are not female bonded. When they move at adolescence to a new group, they have no kin support with them. Non-female bonded animals feed on foodstuffs that are spread over a large area rather than clumped together. This leads to dispersal of the females over a wide area and animals compete for food not by aggression over a good patch but by dispersing themselves or ‘scrambling’ (rather than fighting) to find and consume the best resources that they can. Because of this there is little advantage to having kin allies and this may be why females are willing to leave their natal group. Matrilines do not exist and individual dominance relationships are unstable and inconsistent (aggression does not occur on a one-way basis between adjacently ranked animals). Female relationships in non-female-bonded species are often described as egalitarian (de Waal 1993). This pattern is clear in wild populations of common chimpanzees. Females spend more than half of their time alone with their infants. Adult females rarely rest in close proximity or engage in mutual grooming. They do not form strong bonds and rarely support one another. However, aggressive conflicts are rare because each animal feeds in her own foraging territory. If food resources become depleted, females emigrate to a new group rather than engage in competition (Mitchell et al. 1991). Dominance relations are ill-defined, weakly-developed, egalitarian and unstable (van Hoof and van Schaik 1992). It is often impossible to assign a rank to the females because ‘dominance behavior in stable groups or stable pairs of females is uncommon and is never observed among some dyads’ (Pusey et al. 1997). Studies of captive chimpanzees show the same pattern. de Waal (1989, p. 53) describes female relationships in this way: ‘[T]he female hierarchy is rather vague. Since status communication is rare among females, it is difficult and almost useless to assign them positions on a vertical scale. The same is true of feral chimpanzee females.’ He notes that he witnessed not a single instance of female status ritual over a six-year period. However, unlike feral chimps, captive females do form close affiliative bonds that are based on personal preferences and shared history. These kinds of friendships

Mhoc04.fm Page 115 Friday, December 14, 2001 10:16 AM

        


rarely develop in the wild because females are forced to spend much of the day alone with their infant, searching for food. Once formed, these female friendships are extremely stable. For example, two females—Mama and Gorilla—had lived together since 1959 at Leipzig Zoo before being transferred together to Arnhem chimpanzee colony and at no time were they seen to turn against one another (de Waal 1989). However, when antagonisms do develop, females are less conciliatory than males. Reconciliation occurs after only 18 per cent of female fights as compared to 47 per cent of male fights. De Waal (1989) attributes males’ greater willingness to reconcile (as well as their more ephemeral alliances and the separation of coalition relationships from social bonds) to the male dominance hierarchy which depends upon political manoeuvring and realignment. Females who do not enter into such powermongering alliances have less reason to forgive and forget. Their relationships seem to be based upon emotion rather than strategy. Bonobos or pygmy chimpanzees share a mutual ancestry with common chimps and the split between them occurred about one-and-a-half million years ago. They are fascinating and highly intelligent animals but what makes them especially interesting is that there is a very different form of social relationships among them and, as we would expect, this seems to have arisen from differences in the foodstuffs that they eat. Common chimpanzees are found north of the Zaire River where the ancestral environment was harsh and changeable. Food was sparse and spaced out which led to mother–infant pairs foraging alone, separated from other females. (This harsh environment produced unexpected benefits—chimps learnt to use tools in their search for termites—as well as driving them towards a partly carnivorous diet.) Bonobos, however, found themselves in a Garden of Eden. They had an abundant supply of plant food in the forest both up in the trees and down on the forest floor. Fruit is still plentiful and one of their favourites, the Treculia fruit, grows to a weight of 30 kilograms—nearly the size of an adult bonobo. They simply have to wait for one of these prizes to fall to the forest floor before they tuck in. So bonobos face a very different daily situation from chimps. The do not have to separate into mother–infant groups to find food. Because there is so little competition, there is no need for either contest or scrambling. They have far more time to devote to their social relationships and these ties are especially strong among the females. One remarkable way that they bond with one another is through sexual encounters. A female will approach another and stare fixedly into her face—this is a sign that she is soliciting sex. One of the two will then lie on her back while her partner mounts her and they rub their genitals together—a pastime dubbed G–G rubbing. (Who lies on top rather than below is arbitrary and unrelated to social seniority.) The clitoral stimulation seems to delight both of them and may be the reason that bonobos, unlike other primates, typically copulate facing one another. Females are continually sexually receptive (though not continually fertile) and this may have evolved to harmonize social relations in the group. There is no dominance ranking among female bonobos. Kano (1992, p. 193) describes their relationship like this:

Mhoc04.fm Page 116 Friday, December 14, 2001 10:16 AM

      Older influential females do not perform dominance displays to show off their high position. (I call them ‘influential’ because we do not use the word ‘high-rank’ with respect to females.) These older females solicit various social contacts from their followers such as grooming, GG rubbing and copulation. They seldom receive threats or attacks from others (including males) and also rarely display aggressive behavior. They are respected out of affection, not because their rank is high.

The absence of a dominance hierarchy and the lack of competition for food means that females live peacefully together. Out of 325 recorded episodes of aggression, only nine occurred between adult females. Primate studies confirm the evolutionary prediction that competition between females is usually about food. The intensity of competition depends on the abundance and spacing of the food. If food can be monopolized, then dominance has advantages. But the achievement of dominance is rarely worth fighting for because the risk of injury or death is too great. So dominance is decided by which matriline a female belongs to and on her age within that matriline. These tight rules prevent females from having to engage in the sometimes lethal competition seen among males. When food is spaced out, females adapt by foraging alone or with their infant. This spacing avoids the need for competition and its accompanying risks even though the price to be paid is a very solitary existence lacking in friendships with other females. When competition is absent, females are strongly bonded with one another even though they share no genetic relationship. Fights are extremely rare and so are the deaths that come from it. Primate females do not show the scheming, plotting and alliance making that males do in pursuit of status. The risks that accompany this kind of behaviour are simply too great. Status may bring advantages and, if it can be won without risk, females will take it. But if status can only be bought by bloodshed, females have good evolutionary reasons to stay out of it.

Women and status Anthropological surveys of traditional societies show that, like apes, humans have generally favoured patrilocal residence which means that females transfer from their home group and lose the advantages of living with genetic relatives (Ember 1978; Foley 1987; Leakey and Lewin 1979; Murdock 1967; Rodseth et al. 1991). Together with this fact, consider van Hoof and van Schaik’s (1992, p. 363) conclusion that ‘we consider the low rates of aggression and agonistic support and the weakly expressed dominance hierarchy as diagnostic of non-FB [female-bonded] groups.’ Taken together, these facts suggest that women should be expected to show less evidence of dominance hierarchies than do men. We have seen that in childhood boys show a preoccupation with dominance in their styles of play, their preference for competition, their rough-and-tumble fighting and the language that governs their social interactions. By contrast, girls in almost every culture are less competitive than boys (Strube 1981). When given a

Mhoc04.fm Page 117 Friday, December 14, 2001 10:16 AM

        


choice, girls prefer cooperation to competition and there is marked rise in this preference at the time of puberty (Ahlgren and Johnson 1979). Girls are concerned with developing shared norms and cohesion within the group (Eder and Sandford 1988) and are more often resolve conflict through discussion (Eder 1990; Youniss and Smollar 1985). Girls use speech to create and maintain relationships of intimacy and equality, to criticize others in acceptable ways and to interpret accurately the speech of others. They show less interruption, more often express agreement and more often acknowledge what another girl has said before beginning to speak (Maltz and Borker 1982). Collaborative interchanges are most common in girls’ groups while domineering exchanges are more common in boys’ groups (Leaper 1991). In groups, girls are more likely to influence others by suggestion rather than by giving direct orders. In brief, the boys structure their directives to emphasize differences between group members. Girls, on the other hand, share decision making during the task activities and formulate directives as proposals which include themselves as parties who are obliged to perform the action at issue. Their directive system during task activities tends to minimize differentiation among group members. (Goodwin 1990, p. 135)

Savin-Williams (1980) reviewed the available information on female groups and reached four main conclusions. First, girls are less likely to form stable and consistent peer groups. Secondly, girl groups are cliquish and composed of dyads or triads. Thirdly, group structure is harder to discern, and fourthly, the female group serves to develop interpersonal skills and sensitivities. Later in a detailed study of summer camp, he concluded that: Boys were more likely to physically assert themselves, argue with others, and, to a lesser extent, verbally and physically threaten and displace cabinmates; girls were more likely to recognize the status of other girls, give unsolicited advice and information, shun and ignore. . . . The dominance structure among the early adolescent girls was not obvious to either group or non-group members . . . Dominance does not appear to have been a highly desirable or discernible trait to them. (Savin-Williams 1987, pp. 124–6)

But this reluctance to engage in direct status competition goes beyond simply abjuring it. Girls seem to actively resent other girls who see themselves as superior. Girls who ‘stick out’ attract a kind of negative halo effect—they are seen as egotistical and likely to betray friendships. AMBER: I know someone who is really pretty, but her attitude is blown way out of proportion . . . is right out of it . . . Yeah, a major attitude problem. STACEY: Snobby. PATTI: Yeah, people who have attitudes . . . Nosy. DIANE: What do you mean, like stuck up? STACEY: Yeah, two-faced. (Brown 1998, p. 136)

Mhoc04.fm Page 118 Friday, December 14, 2001 10:16 AM

      Goodwin (1990) also found that girls criticized and rejected other girls who sounded too sure of themselves or whose behaviour implied that they felt themselves superior to other group members. Girls differ from boys not only in terms of the criteria they employ for making comparisons but also in their attitudes towards the activity of ranking itself. Boys seem to openly encourage statements about relative ranking (although they of course may argue about them). However, a girl who positively assesses herself or explicitly compares herself with others may be seen as showing character and attitudes that the other girls find offensive. Girls constantly monitor each other’s behavior for displays that might be interpreted as showing that a girl is trying to differentiate herself from the others in the group (Goodwin 1990, p. 44).

Among adults too, this reluctance to be seen as superior permeates female friendships. Not only do women avoid bragging about their accomplishments, they actively work to understate them. Most women are well aware that failing to downplay their own success may arouse hostility and personal criticism, and may cost them friends. Furthermore, many who are more successful than their friends—or are successful in different ways—feel guilty about the disparity, as though they are showing them up. Thus, women have good reasons to succeed in silence: not merely to protect other women from feeling bad or feeling jealous, but to protect themselves from other women’s negative reactions. . . . But self-deprecating appeasement doesn’t really forestall competition. Rather, it reverses the usual direction; Nadine competes to lose, not win. Between women this form of downward competition is as habitual as the cycle of complaint and commiseration from which it takes its form. (Fillion 1997, p. 51)

This avoidance of appearing more decisive or more knowledgeable than others may explain why studies of leadership reveal that women are much more likely to become social leaders, responsible for maintaining and supporting good relationships in the group by expressing agreement and showing solidarity, rather than task leaders responsible for getting others to compete the task (Eagly and Karau 1991). This may also explain why women who assume leadership roles prefer to use a democratic style that downplays their own authority in favour of engaging all group members on an equal footing (Eagly and Johnson 1990). When women lead in an autocratic way, they are evaluated less favourably than autocratic men, and women leaders are devalued more by other women rather than by men (Eagly et al. 1992). Using more naturalistic methods, sociolinguists have also examined the way in which men and women handle positions of dominance at work. Tannen (1996) summarizes the findings thus: (Women) will expend effort to assure others that they are not pulling rank, not trying to capitalise on or rub in their one-up position. In contrast, since men’s characteristic rituals have grown out of the assumption that all relationships are inherently hierarchical, it is

Mhoc04.fm Page 119 Friday, December 14, 2001 10:16 AM

        


not surprising that many of them either see less reason to downplay their authority or see more reason to call attention to it—to ward off inevitable challenges. (Tannen 1996, p. 177)

Just as Fillion found for women’s friendships, Tannen reported that female workmates have a conversational ritual through which a woman actively down-plays her superiority. However, Tannen argues that this one-down-womanship gambit is tacitly understood by both partners (at least if they are women) and the listener can be counted on to raise the speaker’s status back up again: ‘Conversational rituals common among women involve each saving face for the other. One speaker is freed to take the one-down position (ritually, of course) because she can trust the other to, ritually again, bring her back up. Neither has to worry too much about casting herself in the best possible light because everyone is working together to save face for everyone else’ (Tannen 1996, p. 146). Tracy and Eisenberg (1990/91) used a role-playing task to examine the styles of criticism used at work by men and women. Men show more concern with tact when criticizing a superior but women are more tactful when they criticize a subordinate. Women dislike female leaders who employ a male (authoritarian) style of leadership, finding them insufficiently supportive and viewing them as having superior airs (Statham 1987). Tannen (1996) found that nurses are willing to accept an authoritarian style from male but not female surgeons. Female surgeons who enlist most cooperation from nurses are those who desist from giving orders and instead emphasize the need for equal teamwork. The most successful women managers are those who report that they avoid behaving like authority figures. Aries (1976) found that women who spoke more than average at one group discussion intentionally diminished their contribution at the next meeting in order to avoid appearing dominating. To make it to the very top of the corporate world requires a particular constellation of traits: initiative, decisiveness, self-assurance, assertiveness, ambition, willingness to take risks and masculinity (Browne 1995). Many of these are of course the very traits that are used to measure dominance or agency. Indeed, one study found that a woman’s score on this dimension was more positively correlated with career achievement than any other background information (Wong et al. 1985). (Her score on communion—including adjectives such as yielding, loyal, sensitive to the needs of others—was negatively correlated with her success.) Top female executives possess the same qualities that top males do. In the corporate world, the very highest echelons are dominated by men and to make it in their ranks women must be willing to compete on their terms. Small wonder then that so many women avoid the fiercely competitive arenas of the private profit-making sector and choose instead to work in public or non-profit-making institutions, despite the lower pay and restricted prospects. And women managers are most often found in staff positions (such as human resources, corporate communications, community and government relations) where the emphasis is upon interacting with others in a supportive way rather than

Mhoc04.fm Page 120 Friday, December 14, 2001 10:16 AM

      in line positions where competition to make the best sales figures or the greatest departmental profit is intense and overt (Browne 1995). At least one feminist lawyer agrees that women simply care less about establishing status differentials in the workplace than do men. ‘Most research suggests that women have placed lower priority than men on objective forms of recognition in employment such as money, status or power’ (Rhode 1991, p. 1774). Although feminists strive to alter the hierarchical and competitive ambience of the workplace, they must achieve some measure of power within it before they can ever hope to alter it (Iannello 1992). Many women see dangers in dominance. As the Australians say, ‘The tall poppy gets cut down’ (Besag 2000) or as the Japanese have it, ‘A nail that sticks up gets hammered down.’ To be dominant is to incur the wrath of other women and to risk their intense and intimate friendship.

Who is the prettiest of them all? But perhaps we have been looking for female competition in the wrong place. For men, rank depends upon those qualities that are attractive to women and that subdue other men (assertiveness, the strategic use of alliances and a willingness to act autonomously). Rather than respect through intimidation, perhaps women seek acclaim from other women for feminine mate-attracting qualities—by being the most beautiful of them all (Benenson 1999). After all, a woman’s social dominance is irrelevant to whether or not men find her attractive (Sadalla et al. 1987). But men everywhere rate physical attractiveness as the most important quality in selecting a partner. Donna Eder (1985) studied middle-school teenagers in the United States and examined the dynamics of popularity among girls. Popular girls were members of high-status cliques that formed during the seventh grade and the chief criteria for entry were attractiveness and being a cheerleader (which also depended in large part on physical appearance). But Eder began to suspect that something was awry when she noticed that the very girls who had been nominated by others as the most popular were also the most disliked. For these girls, ‘popular’ meant visible and well known. Everyone knew the top girls’ names and they received greater attention than other class members. When girls first joined the ranks of the cheerleaders, they experienced a brief rise in social kudos and many other girls wanted to be their friend. But the chosen few began to ignore overtures by non-clique members and rapidly came to be seen as rude and snobbish. The same pattern is reported by Brown (1998) in her study of junior high-school girls in Maine. Lyn (Interviewer) : What bothers you so much about them? I mean, other than . . . Lydia: I hate whiners. Kirstin: They’re conceited. Jane: They treat us like we’re nothing. Kirstin: They’re so conceited, they think . . .

Mhoc04.fm Page 121 Friday, December 14, 2001 10:16 AM

        


Lydia: They’re like the kings of the school. Kirstin: They’re so much awesomer than us. They’re like . . . Lydia: ‘You’re stupid,’ or . . . Kirstin: ‘Get away, you’re a scrub. We’re better.’ They don’t say it, but they imply it so heavily it’s disgusting.

This abandonment of old friends and rejection of potential new ones from the ranks of the lower-status groups was driven by the popular girls’ awareness that association with non-cheerleaders would seriously jeopardize their membership of the top group. In fact they had to monitor their behaviour carefully even within their new clique of friends. If one group member failed to acknowledge or was not especially friendly to another group member, she was frequently accused of being stuck up. The considerable anxiety surrounding the use of this label was evident when one girl would tease another and jokingly accuse her of being stuck up. In most cases, these accusations were taken seriously and denied rather than treated in a light and humorous manner. . . . Girls who were thought to be better than other group members were also accused of being stuck up, even when they had little control over this perception . . . In general, it appears that girls strongly uphold an egalitarian norm and view members who are better or worse than other group members as less desirable friends. Consequently, girls are concerned about being more successful than their friends and about being less successful . . . (Eder 1985, p. 162)

The friendship cliques of girls depend upon physical attractiveness. Cliques form around similarity, in this case of appearance, as has been found in many other studies of teenage friendship (e.g. Brown 1998). Pretty girls tend to associate with other pretty girls. But there is little evidence of girls striving for individual prominence within these groups. Indeed, girls seem to actively work at negating their own superiority to avoid being seen as a snob and being ejected from the group altogether. It is true that the cliques of attractive girls are more visible than others but this visibility is not the same as popularity or respect—in fact quite the reverse. When a girl is promoted into a high-status clique, she loses her old friends. This creates even greater pressure for solidarity within the new clique and an even greater need to debase her own qualities in the interest of equal relationships and continued membership. So even in areas where we might most strongly expect to see status striving, it looks very different from that of males. The aim is to belong but not to excel, to blend in rather than stand out, to be attractive to boys but not to alienate other girls. Brown (1998, p. 138) captures the image held by teenage girls of the perfect young woman: ‘recognizable to anyone who opens the pages of a teen fashion magazine: beautiful, tall, long hair, perfect skin, pretty eyes, nice figure. This girl must also be talented and get good grades; she is humble, modest and liked by everybody, with a personality to match her looks.’

Mhoc04.fm Page 122 Friday, December 14, 2001 10:16 AM

     

The epigenesis of sex differences Sceptical readers will by now be wondering how evolutionary psychologists could so wilfully ignore the power of stereotypes in driving gender differences. Everyone knows how men and women are ‘supposed’ to behave and our natural tendency to conformity pushes us to live up to these images from childhood onwards. Hence, of course, men are more interested in dominance than women because dominance is a characteristically masculine quality. Evolutionary psychologists do not ignore stereotypes but they interpret their relevance in a very different way. The stereotype has had a bad press. The Oxford English Dictionary defines it as ‘an unduly fixed mental impression’ and in everyday usage we treat stereotypes as reflections of oversimplified and prejudiced thinking. Stereotypes are bad, we want to eradicate them. But psychology suggests that we may have a hard time doing so. The ability to categorize the world allows us to assign new instances to known groups and hence to anticipate their nature and inform our response. All of us have a mental representation—a stereotype if you will—of objects as various as houses and dogs. The core representation contains features that characterize a typical house—walls, a roof, doors, windows and an understanding that it is a place where people live. Although any given house might violate our stereotype—an igloo or a tepee—we recognize it as a member of that group and do not discard our fundamental stereotype that accurately describes the kinds of houses that we encounter daily in Europe or America. Stereotypes may be what psychologists call ‘fuzzy sets’ but they are handy for anticipating the unknown. As far as gender is concerned, stereotypes have a lot to answer for. These simplified fictions of male and female, so we are told, drive our socialization of children and our expectations of adults in a way that coerces conformity. Stereotypes are not merely useful interpretative mechanisms but have a causal status in the social world. They drive conformity and so ensure that maleness and femaleness become self-fulfilling prophecies. What is the evidence that stereotypes are ‘unduly fixed’? If we take such a phrase to mean unchanging over time, plenty. Consider the role of women in 1957 and 1995. They have come all the way from the apron-wearing, cookie-baking mom to the power-dressing, full-time board member. Yet our stereotypes of women (and men) have not altered over those forty-odd years (Lueptow et al. 1995). We continue to believe that men are more authoritative, ambitious, domineering and competitive than women. With so little change over time, we can only suppose that structural changes in society have had no impact upon gender stereotypes and that they stumble on with a life of their own, blind to the changes that the personalities of women and men have undergone. Stereotypes have lost contact with reality. But have they? Other studies have investigated not stereotypes (people’s view of the typical or ideal man and woman) but men’s and women’s actual ratings of their own qualities. Here, people are given a list of adjectives (with no reference made to gender) and simply asked to indicate how well each adjective describes them as an

Mhoc04.fm Page 123 Friday, December 14, 2001 10:16 AM

        


individual. Again we find sex differences and no change in their magnitude. This stability is equally evident in standardized psychological tests of masculinity and femininity. It also appears on psychometric test norms for sex-linked traits such as tender-minded and assertive measured in 1955 and 1992 (Feingold 1994). Could it be that while roles, responsibilities and incomes have altered, the basic nature of men and women has not? Evolutionary psychology would be surprised indeed to find that evolved traits had undergone change in a period as short as fifty years. The fact that women work is no reason to suppose that they have altered their personalities. From an evolutionary point of view, the oddity is the fact that there have been historical periods when women did not work. Flintstones caricatures notwithstanding, there was no room in the Pleistocene for women lounging about at home. Without the calories that women provided through gathering tubers, vegetables and honey, families would have starved. What marks our present-day environment as special is not the fact of women working but of women having to leave their children in order to do so (Hrdy 1999). It looks very much as if stereotypes are reasonably accurate appraisals of the differences between men and women. Eagly (1995, p. 154) goes further, noting that ‘the idea that gender stereotypes exaggerate reality has yet to receive convincing empirical support’. If, as has been proposed, stereotypes drive conformity then we would expect that they would be much stronger than actual sex differences. After all, the suggestion is that these stereotypes are idealized and overblown caricatures to which we try to approximate our own behaviour. But Swim (1994) carried out a revealing study that took advantage of the known differences between women and men that had been revealed by meta-analysis of hundreds of studies. She asked her subjects to estimate the size of the difference between men and women over a variety of traits. Given that the participants in the study could not have met every man and woman in the country, they must have relied upon stereotypes to guide their responses. Her results showed that these stereotypes are in fact underestimates of the real differences. Rather than presenting idealized exaggerations of gender, stereotypes seem to slightly underestimate the differences between the sexes. Then there is the developmental question. If sex differences in dominance-related behaviours arise from children’s understanding of gender stereotypes then we should expect to see sex differences only after children know what boys and girls are ‘supposed’ to be like. By stereotypes, we mean not simply the ability to tell the difference between a girl and a boy but inferences about what girls and boys typically like to do. We can approach this question using two different methods. In the first we take a cross-section of children and ask if those who have gender stereotypes show more sex-typed behaviour than those who do not. However, bear in mind that these cross-sectional studies cannot address the cause-and-effect relationship. They show that at a given point in time there is or is not an association between sex-typed behaviour and gender knowledge. But the assumed direction of causality (knowledge causes behaviour) could easily be reversed. It might be the case that children whose

Mhoc04.fm Page 124 Friday, December 14, 2001 10:16 AM

      behaviour is strongly sex-typed are especially sensitive to knowledge about gender differences. We performed a longitudinal study to examine which came first—gender knowledge or gendered behaviour—by measuring children’s understanding of stereotypes and their social behaviour in groups at 24 months and 36 months of age (Campbell et al., submitted). During that year, the children became markedly more successful at correctly labelling boys and girls and they got better at identifying the toys they preferred and the games they liked. (Identifying games was the most difficult and even by the age of 3 most children were unsuccessful). They also showed an increase in sex-typed behaviour with boys engaging in more antisocial and dominating behaviour (they grabbed toys, resisted attempts to grab them and pushed one another around more than girls) while girls showed an increase in spending quiet time in close proximity with their parent. (Both sexes showed an increase in preference for same-sex playmates—a fact to which we will return later.) The real crunch was that the amount of gender-knowledge was unrelated to the amount of sex-typical play at either age and early knowledge did not predict stronger sex-typed play preferences. The message is clear: gender differences can occur without knowledge of stereotypes. Two major reviews of hundreds of studies by developmental psychologists converged on the same conclusions. ‘[T]he literature does not support the proposition that concepts about gender (gender stereotypes) precede preferences or behavioral enactment’, said Huston (1983, p. 409), and eight years later her words were echoed by Martin (1993, pp. 191–2): ‘seldom are individual differences in behavior and thinking explained by differing levels of gender stereotype knowledge.’ What then can be causing sex differences? One possibility is that even by the time they are born girls and boys are already different in their interests and preferences. One particularly poignant case study can illustrate this (Colapinto 2000; Diamond and Sigmundson 1997). In 1966, a pair of identical twin boys aged seven months were scheduled for circumcision to remedy problems with urination that were being caused by a tightening of the foreskin. During the operation one of the boys suffered severe burns to his penis from the electro-cautery needle that was used. What was left of the blackened penis dried up and broke into pieces over the next few days. Seven months later the boys’ parents spoke to a psychologist specializing in gender issues at Johns Hopkins Medical Center. He assured them that there was still time to perform surgery and to raise the child as a girl. He was confident that if this physical and social change of sex could be done before the child was three (and achieved gender identity), there was every reason to believe that she would develop quite normally. Gender, he believed, was simply a matter of how one was socialized. At 22 months, what remained of the penis was ablated and reconstructive surgery created labial folds and a cosmetic vaginal cleft. The child henceforth was named Joan and her parents were instructed not to reveal any of the previous events or any parental doubts that might make her question her status as a girl. Such was the pressure to conform to the doctor’s instructions that her parents in their letters to the

Mhoc04.fm Page 125 Friday, December 14, 2001 10:16 AM

        


hospital emphasized her feminine behaviour. By the time Joan was 7, academic papers and books proclaimed that she had accepted her status as a girl totally. The tabula rasa principle was vindicated. But not for Milton Diamond, an endocrinologist who had argued forcefully in his own work that sex-typed behaviour is originally established by foetal androgens. Since Joan had experienced the normal surge of testosterone during the second and third trimester in utero and for several months after birth, he was astonished that her later behaviour bore no signs of her original gender. Every year, Keith Sigmundson, who had overseen Joan’ psychiatric treatment, noticed an advertisement in the American Psychiatric Society Journal asking whoever was treating the twins to get in touch but he did not respond. At last, in 1991, Milton Diamond contacted him and persuaded him to discuss what had become of Joan. Together they published a full account of the events up until her marriage at the age of 24 based on interviews with Joan, her brother, parents and teachers. Her brother described her like this: ‘I recognized Joan as my sister but she never, ever acted the part. She’d get a skipping rope for a gift and the only thing we’d use it for was to tie people up, whip people with it. Never used it for what it was bought for. She played with my toys: Tinkertoys, dump trucks. Toys like this sewing machine she got just sat. . . . When I say there was nothing feminine about Joan, I mean there was nothing feminine. She walked like a guy. She talked about guy things, didn’t give a crap about cleaning house, getting married, wearing make-up’ (Colapinto 2000, pp. 9–11). A child guidance counsellor noted: ‘Joan’s interests are strongly masculine. She has marvelous plans for building tree houses, go karts with CB radios, model gas airplanes and appears to be more competitive and aggressive than her brother and is much more untidy both at home and in school’ (Colapinto 2000, p. 11). At puberty, Joan had adamantly refused further surgery to construct a vaginal passage. Reluctantly she took female hormones to encourage breast development. She was miserably unhappy, and on occasions, suicidal. In 1980, when Joan was age 14, after her session with a psychiatrist, her father could stand it no longer. He told her the whole story of what had happened. Joan’s reaction was not tears but relief: ‘Suddenly it all made sense why I felt the way I did. I wasn’t some sort of weirdo.’ He immediately changed his name to John, switched from female to male hormone therapy, and had surgery to remove the breasts and construct a rudimentary penis. He later married and acquired two stepchildren. Diamond’s desire to set the record straight grew not just from his own theoretical position but from a desire to spare other intersexuals the psychological and social pain that John had lived through. Clearly John’s brain had been masculinized and no amount of socialization as a girl could alter that fact. His interests and his interpersonal style of relating to others must have been set down very early and many researchers propose that this happens even before birth. When the male fetus develops testes, the production of testosterone begins and testosterone levels are particularly elevated between 8 and 24 weeks in

Mhoc04.fm Page 126 Friday, December 14, 2001 10:16 AM


    

utero and for the first 6 months of independent life. When testosterone is aromatized into oestradiol, it is capable of crossing the blood–brain barrier and affecting the chemistry and the wiring of the brain, shunting it along a male trajectory. (Genetic females are protected from the paradoxical masculinizing effect of their own oestrogens by protective mechanisms that may involve binding oestrogen to alpha-fetoprotein and thus inactivating it or through placental inactivation.) We are still far from understanding the specific sites that are affected although steroidsensitive areas of the brain that also show structural sex differences include the preoptic area of the hypothalamus and associated architecture (that part of the brain responsible for aggressive, sexual and reproductive behaviour; Zhou et al. 1995), the hippocampus, the amygdala and the corpus callosum. John’s experience of a mismatch between brain and body arose as a result of an unforeseeable accident. Other children have had similar experiences but from anomalous hormonal overdoses not from failed surgical procedures. Congenital adrenal hyperplasia (CAH) is one such condition. The developing fetus produces large quantities of androgens—a masculinizing hormone—from the adrenal glands. (It is caused by a deficiency of the enzyme 21-hydroxylase which is used by the adrenal cortex to produce steroids such as cortisol. When the homeostatic system detects insufficient circulating cortisol, it responds by increasing production of its chemical precursors. Because the precursors cannot be converted to cortisol, they are taken up by the androgen pathway which does not require 21-hydroxlase for its synthesis—resulting in raised levels of the male hormone.) One child with CAH is born for every 5,000 to 15,000 live births. This excess of male hormone is often undetected in infant boys but in girls it can be recognized by masculinization of the genitals at birth. Typically the condition is ‘corrected’ by surgery to feminize the external appearance of the genitals. But surgery cannot ‘correct’ the organizing effects that androgens have had on the brain. In consequence, in childhood these girls show more ‘tomboyish’ behaviour than their peers (Berenbaum and Snyder 1995; Collaer and Hines 1995; Dittman et al. 1990; Kuhnle et al. 1997). They enjoy masculine activities such as athletics and rough-and-tumble games, they choose male-typed toys and prefer boys as friends. They show decreased interest in feminine clothing, cosmetics, doll-play and infant care. These differences are sustained into adolescence with CAH girls showing higher levels of aggression (Berenbaum and Resnick 1997), greater interest in male activities (such as football, model building, working with engines), lower levels of female activities (such as reading fashion magazines, cheerleading, keeping a diary) and more interest in typical masculine rather than feminine careers (Berenbaum 1999). Synthetic progestins have been used to sustain at-risk pregnancies. These manmade compounds come in two forms; some are progestin-based and consequently anti-androgenic or feminizing in their effects while others are androgen-based and androgenic or masculinzing in their effects. As we would expect, genetic females who have been exposed to these androgen-based progestins before birth show

Mhoc04.fm Page 127 Friday, December 14, 2001 10:16 AM

        


more athletic interests, a preference for masculine playmates and toys, and are more aggressive (Money and Erhardt 1972; Reinisch 1981). Reviewing dozens of studies on the relation between prenatal hormones and sex-typed behaviour, Collaer and Hines (1995, p. 97) conclude that ‘Evidence is most consistent for a developmental influence of androgens on sex-typical play. There is also some evidence supporting a role for androgens in the development of tendencies toward aggression and for androgens, or estrogens derived from them, in the development of sexual orientation.’ Notice here the triumvirate of playstyle, aggression and object of sexual attraction. As we shall see, how children play affects who they play with and this in turn affects which sex becomes the object of their desire. But what psychological mechanism lies behind the early differences in the characteristic pattern of rough-and-tumble play, high-energy expenditure and competitive interactions of boys? Exactly what is it that early exposure to androgens is affecting? We are not certain of the answer but one promising candidate is inhibition. Boys may be less able to control their impulses than are girls and, if this were indeed the root of sex differences, it would gel with the idea that boys have a less effective braking mechanism, as I argued in the last chapter. Developmental psychologists have been fascinated by the response of infants to an apparently simple task originated by the famous Swiss writer Jean Piaget. Hide an object from an infant by covering it with a cloth or container. The 8 month-old infant will lean forwards and try to uncover it. Now move the object—in full view of the child—and hide it in a new location B a few centimetres away. The child continues to search in the old location A even though she has watched the object being moved. This is the A-not-B error. The longer the time delay between the object being moved and the child being allowed to search for it, the greater the probability that the error will be made. Even at one year, a delay of 10 seconds can cause the child to search in the original location. Piaget believed that the child had developed a sensorimotor habit that it could not shake off—the baby had developed a coordinated pattern of behaviour that ‘stuck’ even when the situation had altered. But this habit hypothesis seemed unlikely since infants made the error after simply watching someone else find the object at location A (Butterworth 1975). Others believed that it was a memory problem—the delay before being allowed to search effectively eliminated the new memory trace. But infants made the same error when memory was not even required—when the uncovered and clearly visible object was moved to location B (Bremner and Knowles 1984). One particularly intriguing solution is that the problem is not one of memory but of inhibition—with age the child achieves the ability to withhold their prepotent response and adjust their behaviour to the new situation. When children make the A-not-B error, it is not because they fail to remember where the object is but because they cannot inhibit the pattern of motor behaviour that worked before. In rhesus monkeys, Diamond (1985, 1988) has shown that lesions of the frontal lobes which govern inhibition cause monkeys to make a similar error but the same lesion has no effect on memory. The effect of delay is more marked on

Mhoc04.fm Page 128 Friday, December 14, 2001 10:16 AM

      boys than on girls and it may be that they are maturationally slower in acquiring inhibitory control of their behaviour. Perhaps this is why boys achieve toilet-training later than girls (Martin et al. 1984). It may also explain why girls are better able to tailor their behaviour to the situation—girls but not boys decrease their activity level when they play indoors in a confined space. The high activity level of boys may in part be explained by their slower maturation but not all of it. Eaton and Yu (1989) investigated children aged between 5 and 8 years of age and found that the activity level of boys was higher. But they also noticed that the girls were more physically mature (as assessed by the proportion of estimated adult height that the child had attained) and that greater maturity was associated with lower activity level. When, instead of comparing similarly aged boys and girls, they compared similarly mature boys and girls the sex difference dropped. Nevertheless, it was not completely eliminated. Boys may be slower to ‘grow up’ in the sense of gaining control of their energy but the possibility remains that as a sex they never achieve the behavioural and emotional control characteristic of girls. While uncontrolled tantrums of the ‘terrible twos’ begin to decline in girls at the age of 36 months, they remain high in boys. Sex differences in inhibition are still apparent in the third and fourth years of life (Goodenough 1931). When preschool children have their activities interfered with by peers, boys are more likely than girls to yell, scream, stamp their feet or lash out (Eisenberg et al. 1994) as if they are unable to restrain their emotional reaction. One study (Kochanska et al. 1996) tested impulse control more directly in a variety of ways including keeping a sweet on the tongue for a period of time without chewing or swallowing it, performing a motor activity deliberately slowly, speaking softly and waiting for a command to ‘Go’ before starting an activity. The boys scores were about one-third lower than those of the girls. As Maccoby (1998, p. 103) puts it: ‘It would seem that high-energy behaviour in boys is more likely to have an unrestrained, hyperactive quality than is the case for girls’, and boys difficulty with emotional and behavioural inhibition is most apparent when it reaches clinical levels. Attention deficit hyperactivity disorder (ADHD) affects between 3 and 7 per cent of children. It is associated with a pattern of difficulties that seem characteristically masculine: poor school performance, aggression, conduct problems, substance abuse, driving accidents, speeding violations and difficulties in adult social relationships, marriage and employment. Its main diagnostic features are impulsive and hyperactive behaviour patterns. Empirical evidence is now converging upon the proposal that a series of higher-order ‘executive’ problems with working memory, self-regulation and planning result from an underlying difficulty with behavioural inhibition (Barkley 1997). Boys have trouble, as I discussed in the previous chapter, putting the brakes on, and ADHD, like aggression, is associated with reduced activity in the serotonergic system (van Goozen et al. 1999; Stevenson and Williams 2000). ADHD children talk more than other children, have a higher activity level, find it difficult to control their behaviour in accordance with instructions and to resist temptation. They prefer a smaller reward immediately to a larger reward in the future. Behavioural inhibition

Mhoc04.fm Page 129 Friday, December 14, 2001 10:16 AM

        


can be broken down into three major areas, all of which present difficulties to children with ADHD. They find it hard to inhibit a ‘prepotent’ response—that response which is most automatic as a result of past habit, rewards or punishments. For example, they are very susceptible to error under the Stroop paradigm. Here subjects are initially asked to name the ink colour in which a series of pseudo-words are printed (ZYP, QLEKF) and typically find the task simple. It becomes much more difficult, however, when they are presented with real words that are the name of colours that do not match the ink colour (RED for example is printed in blue ink). Here the ‘automatic’ tendency to read the word has to be inhibited and replaced by the new self-instruction to name the colour. Hyperactive individuals have particular difficulty with this. But they have difficulty also in simple child-like games where they have to stop what they are doing and wait to be instructed to start again. They also have trouble stopping an ongoing response once it has been triggered and so have little opportunity to pause, reflect and modify their actions. For example, in the Wisconsin Card Sort subjects have to categorize cards bearing colours and symbols according to an unknown criterion that they must work out for themselves assisted only by feedback from the experimenter. The unknown ‘rule’ may be a simple one—to sort by colour. If they pile cards together based on shape, they will be told that they are ‘wrong’ and most subjects stop and think out a new strategy. Hyperactive individuals, however, tend to make perseverative errors—they go on employing a rule that they already know cannot be correct. They seem unable to pause and plan. Even if they succeed in delaying their immediate response, they are easily distracted by external events during the delay period and their attention is deflected away from whatever goal-directed plan they have tried to develop. ADHD is three times more common in boys as in girls. This is the pattern we would expect if there were small but systematic sex differences in behavioural inhibition with consequences for activity level—sex differences would be most apparent at the very extreme end of the distribution. Boys’ characteristic rough-and-ready play style has assumed a central place in the thinking of Eleanor Maccoby who has struggled to understand the mechanisms that underlies sex segregation—why it is that boys and girls the world over prefer to associate with their own sex. Despite adult attempts to encourage cross-sex interaction, children when given a free choice pick their own sex as their preferred partners (Serbin et al. 1977). This is as true in Africa and India as it is in the United States and Europe (Dunn and Morgan 1987; Omark et al. 1973; Whiting and Edwards 1988). With these preferred partners, children seem to have a more enjoyable time—they engage in genuinely cooperative play rather than merely playing in proximity to one another or standing watching the other child (Serbin et al. 1994). When three-yearold children are randomly assigned to play with a peer, twice as much social behaviour is directed to a same-sex than to a cross-sex partner (Jacklin and Maccoby 1978; Lloyd and Smith 1986). The same-sex preference is quite visible by the age of three and increases in magnitude until by the age of 6, children are spending as much as

Mhoc04.fm Page 130 Friday, December 14, 2001 10:16 AM

      eleven times longer with same-sex than with opposite-sex peers (Maccoby and Jacklin 1987). Maccoby believes that the forces driving this are complex but occur in a kind of chain reaction that is initially triggered by differences in behavioural inhibition. These differences manifest themselves in the different play styles of boys and girls. Boys’ behaviour is more active, unrestrained and rough while girls greater behavioural control means that their interactions are more modulated, harmonious and less physical. Although children might be drawn together by their shared preference for toys, studies suggest that sex-congruent toy preference is not associated with higher than normal levels of same-sex interaction. But interaction style is associated with a tendency to prefer to interact with members of one’s own sex (Bukowski et al. 1993; Serbin et al. 1994). Boys who show the most rough-and-tumble play and who are regarded by teachers as the most disruptive and active are the same boys who show the greatest preference for playing with other boys. Girls who are regarded as being the most socially sensitive and controlled are the ones who most often associate with other girls. This initial preference for others who enjoy the same ways of playing is then enhanced by a positive feedback loop—boys display even higher levels of activity and increased amounts of threats, challenges and competition when they are playing with other boys (Fabes et al. 1997). Both sexes are responsive to the reactions of their own sex and relatively unaffected by the reactions of the opposite sex to their behaviour (Fagot 1985). This has the effect of increasing the different nature of interactions in girls’ and boys’ groups. Boys—especially as a group—are boisterous, unrestrained and rowdy. Girls avoid them, sensing and withdrawing from boys’ dangerous inability to control their wild activity level. They are less able than girls to engage in the kind of sustained behavioural inhibition that is needed to take turns or engage in a protracted role-playing game. Sex segregation is seen first in girls as they avoid their overexcitable male companions but after a year or so it is boys who take the lead in sex segregation, refusing to take part in girls’ games. Once formed, boys’ and girls’ groups then become increasingly distinctive. A new dynamic begins—in-group allegiance and out-group hostility (Tajfel 1982). This phenomenon has been created in the laboratory on the basis of allocating people to groups on the flimsiest of reasons (their preference for one of two paintings). Once formed, groups seem to reinforce their identity by contrastive comparison with the other group. The ‘other’ group embodies a number of negative traits that do not apply to the in-group. Furthermore, while the in-group is seen as richly differentiated in its membership (‘We are all different’), the out-group is homogenized as well as denigrated (‘They are all the same’). These processes lie behind the formation of social identity and, if they can be generated arbitrarily in the laboratory, imagine how much stronger is the impact on children’s sex-segregated groups that spend every school day together. The sense of the distinctiveness of boys’ and girls’ groups is apparent in the way children describe one another. Between the ages of 5 and 8, girls think that boys ‘scratch, fight and wear trousers’, ‘boys are naughty’, ‘they’re horrid’,

Mhoc04.fm Page 131 Friday, December 14, 2001 10:16 AM

        


while boys believe that girls are ‘dumb’, ‘rubbish’, ‘horrible’, ‘a little bit soppy’ and ‘teachers pets’ (Abrams 1989, pp. 65–6). Maccoby’s strong point is that gender segregation cannot be understood as an individual difference variable but is a grouplevel effect. Girls as a group avoid boys as a group. To explain this, a group-level social process such as the forging of social identity is required. So strong are the between-group dynamics that by middle childhood, there are rules that govern association with the enemy (Sroufe et al. 1993). It is permissible only under clearly defined circumstances. First, the contact is accidental (if you bump into a member of the opposite sex). Secondly, the contact is incidental (you have to line up next to members of the opposite sex in order to get a drink at the water fountain—conversation is not allowed). Thirdly, the contact has a necessary purpose—you may ask a member of the opposite sex to pass something to you provided that no interest in them is expressed. Fourthly, contact is acceptable provided that neither sex is alone—two boys may speak to two girls as long as same-sex contact is more evident than cross-sex contact. Fifthly, interaction is acceptable if it is accompanied by disavowal—a child may verbally insult or physically intimidate a member of the opposite sex. What lies behind these strictures on cross-sex interactions—interactions that by adolescence will not only be sought out but become a central part of most teenagers’ lives? An intriguing suggestion has come from Daryl Bem (1996). He took on the task of trying to explain sexual orientation and his answer involved a five-step chain of causal links. Rather than arguing for a direct genetic basis to sexual preference (Hamer and Copeland 1994), he argued that the principal genetic effect is upon childhood temperament, especially aggression, rough-and-tumble play and activity level. Temperament affects the kinds of activities and games that children like and draws children together with others who share those preferences. In the majority of cases, these are same-sex peers. Once in these groups, children develop a strong feeling of being different from the opposite sex. While Maccoby sees this step as a social psychological one involving in-group–out-group differentiation processes, Bem conceives of it as a less social and more personal experience. He notes, for example, that among adult homosexuals 71 per cent of gay men and 70 per cent of lesbians retrospectively recalled feeling different from their same-sex peers. (Among heterosexuals 51 per cent of men and 38 per cent of women also said that they had felt different. But unlike the homosexual sample who cited gender-linked preferences such as liking sport (lesbians) or not liking sport (gays) as the factor that made them feel different, heterosexuals cited reasons that were unrelated to gender-typing such as differences in intelligence, extraversion or physical attractiveness.) Now comes the step that gives its name to his theory ‘Exotic becomes erotic’. He argues that we experience non-specific arousal in the presence of peers that we feel are different from us. For heterosexuals, this is the opposite sex and for homosexuals, their own sex. Bem admits to uncertainty about the mechanism behind this eroticization. He proposes three candidates. The non-specific arousal hypothesis takes as its starting

Mhoc04.fm Page 132 Friday, December 14, 2001 10:16 AM

      point the finding that autonomic arousal, however artificially generated (for example by riding a bicycle or watching an anxiety-provoking movie), tends to augment attraction towards physically desirable others. This mechanism is based upon a classic psychological study that argued that emotion is a combination of autonomic arousal (increased respiration, heart rate and sweating) and cues in the immediate environment that informs the individual how to label the felt state. The very same arousal may be felt as fear, hate or attraction. However, as Bem notes in studies of attraction, arousal also has the effect of intensifying rejection of individuals rated as unattractive—in short arousal intensifies both attraction and repulsion. More problematically, there is no indication of what causes the arousal in the first place and why it is reliably associated with one sex rather than another. His second suggestion is even more speculative. He notes that many psychological processes appear to operate on a homeostatic mechanism so that one psychological state (e.g. pain) is overridden and reversed by the initiation of an opponent process (e.g. endorphin release). According to Bem, an effeminate boy taunted by his peers experiences negative affect which is subsequently opposed by a hedonic counter-state of attraction. Finally, he discusses sexual imprinting but his examples in the main deal with avian studies of species-appropriate (rather than sex-contingent) sexual choice. He mentions but does not give extended consideration to a phenomenon originally proposed by Westermarck (1891), who noted that individuals who had spent their childhood together tended to avoid sexual intimacy in adolescence and adulthood. In Taiwan it was the practice for families to ‘buy’ a girl child who would live in the family and provide domestic labour until she reached adolescence when she would be married off to the son of the family. These marriages had a high rate of failure, non-consummation and childlessness. Children raised in communal nurseries on kibbutzim also do not later marry (Sheper 1971). For evolutionary psychologists such a finding strongly suggests the operation of an evolved module geared to incest avoidance. The peers with whom we would normally be raised would be siblings and hence such a mechanism would ensure that we did not feel attraction to them at puberty. The fact that sex avoidance also occurs in sim pua marriages and on kibbutzim suggests that the module is triggered by extensive early contact rather than by some form of kin-recognition system. In a sense the module is ‘fooled’ by the input which is supposed to modulate incest avoidance. This same mechanism, however, can be exploited as an adaptation for eroticizing the opposite sex if it operates under the general rule ‘avoid sex with childhood familiars’. We would find attractive those individuals with whom we had not spent extensive time as children and, in the heterosexual case, this would be the opposite sex. If Bem is right, then pre-adolescent cross-sex interaction might be seen as indicating either homosexuality or precocious sexual interest. On the first point the literature is clear—sex role rigidity is far stronger for males than for females and boys suffer the consequences of difference more extremely (Archer 1992). Boys who

Mhoc04.fm Page 133 Friday, December 14, 2001 10:16 AM

        


prefer to play with girls or enjoy female-type games are subject to considerably more ridicule than girls who show masculine preferences. Tomboys are acceptable but effeminate boys are not. This intolerance is most marked among male peers. CARL: There is no way I’d be mates with a poof. ND (Inteviewer): What about you? CARL: If he was queer? I’d slap him, I would. I would not have him coming near me. ADIE: It’s right. I would do the same, not hit him, but tell our mates and we’d probably all get him. Let him know. (Duncan 1999, p. 108)

Boys reject feminine boys quite brutally—ridiculing, shunning and referring to them as ‘girls’ (Thorne 1986). Boys who display such characteristics are positioned far down the boys’ status hierarchy (Best 1983). Fathers, as well as male peers, are alarmed by evidence of inappropriate gender role behaviour by boys, concerned that they may grow up to be homosexual. Adult homosexuality is seen as a likely outcome of effeminate behaviour in boyhood, but not of boyish behaviour by girls which is seen as a temporary phase (Martin 1990). Fathers are more likely than mothers to refer their sons to psychologists and psychiatrists for such behaviour (Archer 1984, 1992). A follow-up of such boys who were referred to a gender clinic programme in California suggested that these paternally forecasted outcomes were often accurate; most boys referred at the age of 12 or younger developed homosexual or bisexual preferences within the next ten years (Green 1987; Zucker 1990). In addition, retrospective data corroborate the fact that early preference for female companions and play styles is associated with later homosexual preference among males (Bailey and Zucker 1995; Bell et al. 1981). A preference for playing with girls is likely to provoke especially strong suspicions of homosexuality as boys grow older. This paradox—a boy who spends too much time with the opposite sex must be gay—which was quite evident to teenage boys, caught one researcher off-guard (Duncan 1999, p. 117). CARLTON: If a boy was really popular with all the girls and he used to really hang around and talk to them all the time, he’d be called a poof or something. ND (Interviewer): Cos he’s going out with loads of girls? BEN: It depends . . . AARON: Not in my case it doesn’t. Well, it depends if he’s hanging around with loads of girls all the time, then I suppose he would be. ND (Interviewer): What, because he was just friends with them? AARON: Yeah. CALTON: Yeah. ND (Interviewer): Right, I see. What you are saying is that if a lad wasn’t trying to go out, like boyfriend and girlfriend, but was friendly and nice to girls, and being able to talk to them about things they liked, have a laugh with them, and sat with them in class, that he might be called a poof by the rest of the lads? CARLTON: Yeah. AARON: Yeah.

Mhoc04.fm Page 134 Friday, December 14, 2001 10:16 AM

      Playing with opposite-sex peers when not subject to ‘gay’ teasing is likely to provoke ‘they’re in love’ teasing. In many cultures, versions of the kiss-chase game are played in which members of one sex must tag a member of the other and then kiss them. The game seems to generate a frisson of excitement and taboo despite the fact that boys particularly are inclined to treat a kiss from a girl as a form of contamination. But less overtly, any cross-sex encounter—even the most innocent—is liable to be interpreted as romantic interest. A boy who entered a girls’ tent at summer camp to retrieve his radio was greeted with taunts from his friends wanting to know if he had been kissing them (Sroufe et al. 1993). A boy who harasses a particular girl too much (chasing, pushing or calling her names) is equally likely to be accused of being attracted to her—a form of behaviour that has been eloquently named ‘pushing and poking courtship’ (Schofield 1981). As Maccoby (1998, p. 69) explains: ‘Children themselves seem aware that a declaration of “hating’’ a given child of the opposite sex is by no means a denial of interest in that other child, and indeed can be very close to “liking’’.’ Cross-sex encounters seem to both attract and repel children, combining the anticipatory pleasures of future romantic intimacy with the recognition that at present they are too young to fully enter this sexual minefield. As well as maintaining social separation in preparation for later eroticization, sexsegregated peer groups set the scene for peer scrutiny of sex-appropriate behaviour. Years of research have made it clear that peers are more influential than parents in children’s social development (Harris 1995). And with good reason—peers are drawn from that generational cohort in which a child’s allies, enemies and mates are likely to be found. The childhood peer group is salient to children for the same reason that the adult peer group is salient to adults. Social success is about relationships with our own cohort. Just as adults do not slavishly seek acceptance by their child’s peer group, no more should we expect children to care about their popularity with adults. Children may acquire their basic feminine and masculine activities and preferences from natural and sexual selection but the specific ways in which they are expressed is likely to be moulded by their peers. Sex segregation both enhances the differences between boys’ and girls’ groups and minimizes variability between individuals in each group. Boys in contemporary Western culture prefer fourwheel-drive cars to chariots and guns to lances. Girls prefer Barbie to rag dolls and toy microwaves to cauldrons. The ways in which masculinity and femininity are expressed change across time and culture and much of the ‘fashion setting’ is done by peers. The trigger for this process of sex segregation (and all follows from it) seems to be the emerging boisterous and exuberant behaviour of boys focused around the establishment of a dominance hierarchy—one that girls neither seek to be part of nor replicate within their own sex. Yet that very quality which girls abhor—superiority at the cost of reciprocity, conspicuousness at the cost of intimacy—is something that will later prove peculiarly attractive to them in the opposite sex.

Mhoc04.fm Page 135 Friday, December 14, 2001 10:16 AM

        


Why winning is dangerous Dominance brings rewards—others defer to you, you have the pick of the best resources, your life is freed from the daily stress of appeasing others and hustling for a share of what is left over. The rewards are just as great for females as for males— arguably greater because resources fuel the survival of offspring in which they have already invested while for males it merely buys a ticket in the copulatory lottery of possible fatherhood. We would all opt for dominance if it cost nothing. But dominance is a zero-sum game. To win you have to compete and competition means conflict. Among primate males these conflicts are dangerous and can be deadly. Alpha males do not give up without a fight and today’s allies are tomorrow’s pretenders. We have seen why females should be especially reluctant to risk their life and those of their young to attain it. The fear of direct physical assault runs higher in females than males. In female-bonded primates, some females have found a way to have it all. Rank is inherited rather than fought for and the benefits are a happy accident of birth not a result of bravery. But what of those who were born to low-ranking mothers? Why is there such extreme reluctance to fight what seems to be a despotic and unfair system of privilege? The answer is fear. The costs of a failed revolution are too great and only under the most extreme conditions—when they massively outnumber the opposition—will low-status females risk the little they have for a larger share. In common chimpanzees, females have no kin allies. They do not fight for food but disperse over large territories in search of anything they can find to sustain themselves and their young. The cost in terms of sociability is high. They forage alone and their only close relationship is with their offspring. If it is a daughter, then she too will leave them one day to seek her fortune in another group. Daughters return now and then as if to acknowledge their relationship with their mother who has sacrificed so much for them. This solitary life may be hard but it is better than risking life and limb to fight with other females over each piece of food. Bonobos also leave their natal group but nature has been kinder to them. They feast on the fruits of the forest and need not compete for the resources they need to survive. They can spend time together and forge close female friendships. Unlike males, these friendships are not shot through with concerns about dominance. Although some individuals are more influential than others, such influence comes through age and personality rather than fighting ability. The differences in relationships between chimps and bonobos are a testament to primate adaptability. Bonobos more than chimps resemble the common ancestor from which chimpanzees and humans descended (Kano 1992) but given the extraordinary plasticity of primate adaptation it would be wrong to conclude that humans resemble bonobos more than chimps for reasons of genetic similarity. Bonobos are specialized for survival in their own particular ecological niche and were not subject to the strong environmental pressures that gave birth to tool use and other adaptations

Mhoc04.fm Page 136 Friday, December 14, 2001 10:16 AM


    

seen in common chimps and humans. If we resemble bonobos more than chimps it is a result of convergence in the way we have adapted to our environment rather than because of a closer phylogenetic relationship. The common thread for females from all three species is their reluctance, compared to males, to engage in direct competitive aggression. Girls’ development is marked by this avoidance of physical combat. But more than this, it is marked by a relative indifference to public status and an active aversion to flaunting dominance over others. Newspaper journalists are not an infallible guide to reality but I was struck by a female-authored article that recently appeared castigating the British Government’s ‘Minister for Women’s Issues’ in the following terms: I don’t believe women actually want a minister for women. The women we truly admire are not high-achieving doers, but muddlers-through; we have little time for unsympathetic prescriptive types. The naïve assumption that we respond better to another female is flawed, since women reserve their most ardent dislike for members of their own sex, and their greatest outrage for female variants of Harry Enfield’s Mr. You-Don’t-Wanna-Do-It-Like-That. (The Sunday Times, 18 February 2001)

The emphasis upon egalitarianism is so strong that girls and women manage their self-presentation to minimize the possibility of being seen as stuck-up or snobbish. To incur the resentment of others is to risk conflict, and conflict might lead to confrontation. Although in current society outright fighting is rare, this aversion to conflict is an inherited adaptation that was forged in an environment much harsher and more ruthless than the one we currently inhabit. If chimpanzees and gorillas are a guide, it was a world where infanticide and attack by other females was an ever-present threat. But conflict carries with it another danger too—banishment and exclusion. A woman who thinks she is superior to the rest risks becoming the victim of the specialist female strategy of indirect aggression. She is ignored, shunned and rejected. Now this condemnation to solitary existence would hardly constitute a major fear for chimpanzees but it does for humans. In the next chapter we shall try to answer a real evolutionary puzzle—what made women so dependent on their friendships with one another?

Mhoc05.fm Page 137 Friday, December 14, 2001 10:16 AM

 


Like a sister: Women and friendship

Women’s friendships are deeply enigmatic for evolutionary theory. We come from a prehistory in which women left their relatives to live with strangers in an unknown community. Yet despite this lack of blood relationship between them, the women became friends. Indeed they became more than friends, they became intimates. They forged not just exchange relationships with one another (‘I’ll help you if you help me’) as men did, but something stronger than this—they forged an emotional interdependence that went beyond mutual back-scratching. Their relationships were intense, trusting and communal. How did this happen? In the last chapter we saw that there are two dimensions on which men and women differ in their relationships: agency and communion. Agency is about dominance over others and this dimension characterizes men’s relationships more than women’s. Communion, which we will discuss now, is about bonding and interdependence with others and this interpersonal orientation characterizes women more than men. Women in nations around the world are described as warmer and more caring than men (Williams and Best 1990). Women also rate themselves higher on adjectives that define communion such as ‘kind’ and ‘sensitive’ and rate themselves higher than men on a range of items connected with interdependence with others (Simmons 1987; Wylie 1974). Feingold (1994) examined data from the thousands of responses that were obtained on personality questionnaires in order to construct national norms. He found that women exceeded men on trust, tender-mindedness and gregariousness and these sex differences were invariant over age, educational level and nation. Women’s friendships have a different flavour to those of men both quantitatively and qualitatively. From childhood on, girls prefer to have one or two intimate friends while boys play in larger groups where the emphasis is upon the group rather than the pair (Belle 1989). This is not to say that boys never engage in two-person interactions. In one study, it was found that boys and girls spent about equal amounts of time in dyadic interactions but for girls these exchanges occurred with a smaller range of people and were more extended in time (Benenson et al. 1997). Boys moved from one partner to the next, spending less time with each of them. The difference lies in the intensity and exclusivity of boys’ and girls’ friendships.

Mhoc05.fm Page 138 Friday, December 14, 2001 10:16 AM

      In adolescence, girls’ friendships are very intimate and some behaviours, such as holding hands or hugging, are reserved only for best friends. Because of adolescents’ increasing awareness of sexuality, best-friend relationships have to be defended against rumours of homosexuality about which girls are extremely sensitive (Brown 1998). One girl habitually signed notes to her best friend ‘Love you dearly but not queerly’ (Eder and Sandford 1986). In one particular sense, female same-sex friendships appear to be more intimate than men’s. As one teenage girl put it: ‘girls can talk to each other, boys just keep things into themselves, they don’t tell no one nothing’ (Brown 1998, p. 50). Intimacy is often measured by self-disclosure—the extent to which we confide private details and experiences to another person. Self-disclosure can be broken down further, however, into descriptive self-disclosure (in which we merely share facts about our lives) and evaluative self-disclosure (in which we reveal the emotional impact that these experiences have had upon us). Men’s friendships seem to include a comparable degree of factual disclosure but less evaluative disclosure than women’s (Cross and Madson 1997). Men are especially reluctant to disclose negative emotional experiences that include depression, sadness, anxiety, anger and fear. This sex difference does not seem to be linked to different motivations—men express just as much interest and desire to have close relationships as women do. But studies of their actual patterns of self-disclosure suggest that they reserve it for women rather than sharing it with men (Dindia and Allen 1992). Conversations with women are rated by men as more meaningful, more intimate, involve more self-disclosure and are viewed as more pleasant and satisfying than men’s interactions with other men (see Cross and Madson 1997). In same-sex relationships, intimacy between men revolves around experiences that are relevant to their shared interests. They discuss public-domain issues that they have in common such as sport, business and government, while women are more likely to talk about personal topics such as feelings, relationships and problems (Bischoping 1993; Caldwell and Peplau 1982). Men believe that these common interests and activities are the most important factor in their friendships while women report that similarity in attitudes and values is more central. Men tend to respond to conversations about intimate topics by using non-verbal signs of withdrawal while women tend to show non-verbal approach tendencies (Schwartz and Shaver 1987). In adolescence, girls rate interpersonal relationships and sensitivity as more important than boys, while boys rate social dominance and toughness as more important (Rosenberg 1989). In adulthood, women rank relationship-related aspects of their identity as more important than men (Thoits 1992). Women also describe themselves more than men in terms of their connection to other people (Lyons 1983). When men and women were asked to collect together photographs that described themselves, women selected pictures of themselves taken with valued others more than men did (Clancy and Dollinger 1993).

Mhoc05.fm Page 139 Friday, December 14, 2001 10:16 AM

  :   


The central role that relationships occupy in women’s lives has its downside. Failure in this domain can have negative consequences for mental health. Self-esteem for women, but not for men, is correlated with an interest in close personal relationships, and how highly a woman values her relationships predicts her self-esteem two years later (Stein et al. 1992). Stress in relationships has a more damaging impact upon self-esteem in women than in men (Moran and Eckenrode 1991). When told that they lacked sensitivity and interpersonal skills, women with high self-esteem engaged in compensatory bolstering of their threatened sense of self-worth while men with high self-esteem bolstered themselves when their sense of independence was threatened (Josephs et al. 1992). Problems in relationships are especially hard for women. Women report that not being forgiven by a friend that they had hurt would have a greater effect on their self-esteem than men did (Hodgins et al. 1996). Girls and women are more prone to depression and anxiety than are males (Brody and Hall 1993). Under stress, women tend to internalize their problems while men tend to externalize in the form of antisocial behaviour and hostility. Among third- to twelfth-grade students, girls more often endorse internalizing symptoms such as sadness, loneliness, fatigue, worry about things going wrong, not liking themselves and wanting more friends (Leadbetter et al. 1995). But depression can be more finely differentiated. One form is called interpersonal depression and it appears more often in people who are preoccupied with personal relationships and fear abandonment by others. This of course bears a close correspondence with communion and, as we would expect, there are significant sex differences in his form of depression. (Self-critical depression, on the other hand, is seen more commonly in people who are preoccupied with issues of self-worth and are particularly concerned about failing to meet personal goals or standards or appearing incompetent. Men show greater depressive vulnerability in this area than do women.) People with high interpersonal depressive vulnerability also have anxious or insecure attachments to others, place an intense value on emotional closeness and are preoccupied by fears of being abandoned or neglected in relationships. This preoccupation with relationships can lead to depression in two ways. On the one hand, depression can result from empathy or engagement with other people who are experiencing negative life events. Girls and women experience more distress than men when something bad happens to a friend, family member or neighbour (Kessler and McLeod 1984). On the other hand, women fear losing a valued relationship. Girls who lose a parent are over nine times more likely to report serious depression than other girls (Reinherz et al. 1989). Girls display stronger associations between behavioural problems and stressful events in their relationship with families and friends than boys do (Cohen et al. 1985) and girls report more stress than boys in connection with close relationships (Gore et al. 1993). In part this stress may result from being the target of indirect aggression. Tactics include ostracizing, excluding her from a friendship or even making friends with somebody else as punishment. As Goodwin (1990, pp. 47–8) describes the girls in her study of childhood friendships:

Mhoc05.fm Page 140 Friday, December 14, 2001 10:16 AM

      Within the girls’ group, there were continuous processes of coalition formation as the girls vied with each other over who would be friends with whom, and who would be excluded from such friendship arrangements . . . The girls talked extensively about other girls behind their backs, and the he-said-she-said disputes that emerged from this were more elaborate and more extended than any of the disputes that occurred among the boys. Moreover the girls imposed far more powerful sanctions on parties they judged to be offenders than the boys ever did. For example, they ostracized Annette from the play group for a month and a half, a situation that led her mother to consider moving from the street.

This kind of treatment is rated by girls as more hurtful than direct physical aggression (Allison 1999) and studies of bullying show that these tactics are not only used more frequently but can lead to severe depression and even to self-harm or suicide in the rejected girl. In adolescence, girls place a premium on the trust and loyalty of their friends at a time when personal reputations can be undermined by the betrayal of an embarrassing secret or the publicizing of an ill-judged sexual adventure. As girls struggle for self-respect, their friends can be their most supportive ally or their worst enemy (Brown 1998). Female friendship can both protect and undermine girls in their social relationships. All girls stressed the importance of trust in their friendships. In a world in which a girl’s reputation is constantly under threat both from girls and boys, trust is the most crucial attribute of friendship. A small group of close friends becomes essential in a world where other girls and boys will openly criticize you, talk behind your back, gossip about your every move and spread rumours. The primary demand in such a close group is trust and loyalty, but that trust is precarious. The gravest crime is to betray a confidence. The gravest risk is that a confidence will be betrayed. (Lees 1993, p. 98)

Best friends hold a special status for women. These relationships can endure for many years and partners are privy to the most personal experiences and innermost emotions of one other. As Fillion (1997) describes one typical female friendship: They entertained and sustained each other, sharing even the most mundane details of their lives—especially their love lives—and remembering to ask how each other’s big meetings had gone. Both were successful, but disaster was the staple of their conversations: men who said they would phone but never did, mothers who turned every phone call into an inquisition about their marriage prospects, male colleagues who got the promotion that would have been theirs if they too had been given the opportunity to bond with the boss on the golf course. The anecdote of failure became something of an art form between them; the narrator usually depicted herself as a hapless naif, bumbling through crises with gallows humor. But there was always a feel-good ending: the more they confessed, the closer they felt. . . . One night by way of concluding a baroque tale about an unreliable ex-boyfriend, Anna sighed and said, ‘I think I’ve given up on men. Too bad you’re a woman, we’d make a perfect couple.’ Of all the possible subtexts to her remark, Julie thought the most important one was this: men come and go, but women

Mhoc05.fm Page 141 Friday, December 14, 2001 10:16 AM

  :   


friends last forever. Monogamy was not the label she attached to her expectations, but it fit. The best friend is, after all, singular by definition. (Fillion 1997, p. 5)

The puzzle of female friendships Women often describe their closest friend as being like a sister to them. But of course they are not. They are one-time strangers who, over a period of time, became so close, trusted and dependable that they began to feel like members of the same family. But in the animal world the general rule is that we are closest to our blood kin. In a fight, they will back you up, when in need they will help you, when you are hungry they will feed you. It is in the interest of any animal to remain with its own kin if it possibly can because they make the most dependable allies and, in a social species, allies are the key to survival. Kin care because they share genes. By saving you, they increase the probability of saving a proportion of their own genes. If you are a sister or daughter they save 50 per cent of their own genes. If you are a niece or a granddaughter they save 25 per cent. They care because genetically the two of you are in part the same person. But women did not remain with kin forever. The problem with sexual reproduction is that it requires the mixing of unrelated genes. Inbreeding raises the probability of genetic mutation and to keep the gene pool fresh some members of the group must leave home. In most primate species, females stay at home with their kin and the males, after reaching juvenile status, have to leave and find a new group of females within which to seek their mates. Because females are a limiting resource for males and because females disperse according to where the resources are, males are obliged to fit in with the females’ plans. But among a few species, males won and females dispersed. This pattern occurs in hamadrayas baboons, red colobus monkeys, some spider monkeys, chimpanzees— and humans. In a world survey of human societies, Murdoch (1967) found that two-thirds were male philopatric and less than a fifth were female philopatric. In hunter-gatherers, whose lifestyle typified human communities for hundreds of thousand years and who are among the most egalitarian of all societies, 62 per cent involve female transfer (Ember 1974). Female transfer becomes increasingly common (71 per cent) as societies move from hunter-gatherer lifestyles (in which women make a major contribution to food production) to more formal economies involving agriculture, division of labour and trade (Hawkes et al. 1998). The reason for females being the dispersing sex in our own and other species is not fully understood. Since remaining with kin confers a significant advantage, it is unlikely that females chose to leave. As Rodseth et al. (1991, p. 230) put it, ‘philopatry, consanguinity and cooperation tend to coincide’, and it would be a foolish animal that gave up these advantages without a struggle. And even in chimpanzees, where females voluntarily transfer, they seem to do so with understandable reluctance. One day the young female, when she is in oestrus and near to a neighbouring group of

Mhoc05.fm Page 142 Friday, December 14, 2001 10:16 AM


    

males, will leave and join her new group. But she is likely to be on the receiving end of considerable aggression from her new female companions (who resent the additional competition for resources presented by the new arrival) and she may return temporarily to her natal group as if seeking comfort or reassurance (Pusey and Packer 1987). Why did females simply refuse to emigrate? As yet we have no definite answer. Male philopatry could simply be the result of males banding together more cohesively and refusing to leave. Wrangham and Peterson (1996) offer one explanation of how males bonded so tightly. As chimp parties travel, there is a predictable pattern to their daily movement. Although the party may set off together, the mothers that are carrying infants with them walk more slowly and tire more quickly than the males. On a 20-minute expedition they arrive at food patches five minutes behind the males. Summed over weeks and months, males spend more time in each other’s company than females do. And this extra time is enough for males to forge tighter bonds than females. Unencumbered, childless females can keep up with the main party and, as Wrangham and Peterson would predict, they are more social than the mothers. According to this ‘cost of grouping’ theory, male solidarity is simply the outcome of the amount of time they spend together. But how did males succeed in establishing such a high degree of group cooperation (and thus establish the leverage to drive females out) when they were natural competitors for mating opportunities? How did group-level commitment develop in the face of within-group male competition? Surely, competition for females stood in the way of males achieving the kind of common interest they would need in order to form coalitions with one another? Even chimpanzee brothers will dispute access to an oestrous female despite the fact that they share 50 per cent of their genes (although the disputes are less fierce than among unrelated males). But it is from observations of chimp political manoeuvring that we can begin to see how the problem of male mutual hostility was solved. Males began to form alliances with one another to challenge the dominant male. These alliances were opportunistic and shifting at first. Having overthrown their mutual enemy, the two allies would not hesitate to fight one another but the loser would then enter another alliance with a less dominant male and the pattern would begin again. Often male allies would use their power to coerce females into copulations, both within their own group and beyond it, and females became increasingly vulnerable to the threat of attack not just by one but by two angry males (Goodall 1986). Most importantly, chimpanzee males use this solidarity to cooperate with one another to defend and enlarge their territories (Pusey and Packer 1987). This territory is important to them in attracting new females to the group and expanding their feeding areas. Because of the value of territory to males, male bonding became strong and, coupled with their greater size and strength, females may simply have been forced out by the males’ refusal to leave. This hypothesis is supported by comparing chimps with orang-utans. Both have the same diet and the same size range but orang-utan males do not cooperate

Mhoc05.fm Page 143 Friday, December 14, 2001 10:16 AM

  :   


in territorial defence and consequently have less reason to remain together. In this species, it is the males who emigrate (Pusey and Packer 1987). Once such a pattern was established it would be hard to change. Female exogamy probably extends far back to the common ancestor of chimps and humans (Wrangham 1986). If it has existed so long it is likely to have been subject to very heavy selection pressures so that even if female philopatry became advantageous it would be hard to counter such a strong evolved tendency. The shift from one system to another would be extremely costly in the short term with effects on inbreeding, group size and group defence (Foley and Lee 1989) and these costs might well be so great as to massively decrease species numbers to the point of possible extinction. During human evolution, men also became increasingly successful at sustaining alliances with one another in the face of the mutually competitive tension between them. Men’s capacity to set aside inter-individual conflict over reproductive opportunities may have been enhanced by their increased capacity to engage in strategic planning that freed them from the immediate short-term imperative of competition. This same long-term planning ability may also have been the driving force that moved humans away from despotic rule by a single, privileged male and towards democracy and power sharing. Dominant males began to recognize that the best way to ensure faithful and loyal followers was to refrain from monopolizing female reproductive opportunities and to tolerate monogamy among their subordinates. To keep other men pacified and faithful, it is wise to ensure that they could meet their basic needs and the most basic of all is reproductive opportunities (Betzig 1986). In this way, male–male competition for mates was reduced within the group and men were able to act in concert to launch territorial takeovers against neighbouring groups which had the twin benefits of extending their feeding area and of acquiring females from the vanquished group who often remained in their familiar territory under new male protection. Men’s ability to act in concert opened the door to a variety of intergroup relations. Between hostile groups, we see the kidnap, abduction and rape of women from other rival groups (Chagnon 1983). But the ‘friendly’ exchange of women in marriage could also foster alliances between hostile groups while blood-related males stayed together (Berndt 1964; Hayano 1973). Women were used as commodities—as gifts to cement ties between groups or as trading tokens between friendly bands (Levi-Strauss 1969). But this kind of expropriation and bartering in female lives could only appear after men were able to form strong intragroup alliances among themselves and so to unite against other groups. So, under female exogamy, women at adolescence found themselves cast out and alone in a new community, deprived of the strong kin ties that had protected and supported them. This is what makes the pattern of strong female relationships that we have seen so paradoxical. Other women, to whom the migrating female has no blood ties, are nevertheless turned into a kind of pseudo-kin. On the one hand, it is tempting to think that this is an inevitable example of individuals compensating for

Mhoc05.fm Page 144 Friday, December 14, 2001 10:16 AM

      their loss of close relatives. But there is nothing inevitable about it. Female gorillas and hamadrayas baboons also leave their natal group but they develop only the most rudimentary relationships with other females. Instead they concentrate their social efforts almost exclusively upon the breeding male in the group—and wisely so. Hamadryas baboons live in groups composed of a single male and several females. These breeding units associate together in bands which include some unattached males who are eager to seek copulations. The breeding males constantly herd their females way from such interlopers and, if a female strays too far away, the male targets her with threatening stares. Should she fail to instantly move back towards him, he launches an immediate biting attack on her neck to discipline her. The female is less concerned with her co-wives than she is with avoiding attacks by the male. The common chimpanzee is also patrilocal but again shows no special interest in the other unrelated females with whom she finds herself. She spends almost 75 per cent of her time alone with her offspring and when in the group is more attentive to the males than the other females. But consider the bonobo (or pygmy chimpanzee). A newly arrived adolescent female is careful to avoid unpleasant interactions with her new female companions. She avoids the disputes of others, does not initiate any aggression of her own, refuses to counter-attack if picked on and rarely vocalizes. She does, however, show a lively interest in the older established females. She initiates friendships by ‘peering’—staring directly at the face of the other female. This behaviour is a prelude to sexual activity and signals her willingness to engage in genital rubbing. Afterwards, she takes her food away and feeds quietly on the outskirts of the group. Later she may approach older females to groom them. She is actively working at ‘making friends’. Having left her natal group at the age of 8 years, she is firmly established with her new friends by the age of 13 (Kano 1992). She will spend 60 per cent of her time associating with other females and two-thirds of her time grooming, making body contact and playing with females (Parish 1996). Despite their lack of kinship ties, bonobo females are among the most amicable, supportive and mutually protective of all the primates.

Varieties of friendship The kinds of friendships that men and women prefer correspond to an important distinction that psychologists have noted between different forms of social relationship. And these two forms in turn map on to a distinction that evolutionary theorists have made between different kinds of altruism. By examining these in more detail, I want to suggest that women’s friendships are an attempt to re-create a particularly close bond of communion that mirrors the kind of relationship most often found among blood relatives. One type of relationship is called communal sharing (Fiske 1992), interdependence (Cross and Madson 1997) or mutual aid (Kropotkin 1972). Here, people’s individual identities are merged into a greater whole. The driving motive is ‘From each according

Mhoc05.fm Page 145 Friday, December 14, 2001 10:16 AM

  :   


to his ability, to each according to his needs’. People are given what they need without regard to their ability to repay it at a later date. Goods are held in common and used as they are needed. At a party, for example, the guests drink out of a communal punchbowl without anyone keeping track of how much each individual has consumed. When it is applied to work relations, people work collectively without assessing individual inputs and the product of the joint labour is a collective resource. It typifies the spirit in which people will dig out survivors of a collapsed building—they work until the job is done without apportioning specific amounts of work to one person or another. In group decision making, communal sharing is expressed in a search for consensus, unity and conformity and contributions are about finding a joint solution that will transcend the separate attitudes of the participants rather than claiming particular expertise or special authority. This requires a degree of humility and a desire to please other participants. The overriding motivation is not to be different from others, not to draw attention to one’s self as a unique or special individual. We see it most evidently in very close attachments such as between lovers or between a mother and child where this de-individuation is enhanced further by the physical closeness and even interpenetration of bodies. The greatest fear and distress is caused by the threat or loss of this closeness. But communal sharing also has its less attractive aspects. People can lose a sense of individual identity and be unable to separate themselves from others, leading to a personally destructive dependency. Within the group, a misfortune that occurs to one member may be seen as contaminating and contagious, leading to ostracism. As Fiske (1992, p. 699) describes it: ‘Americans who have suffered a major misfortune often feel isolated and different; they themselves withdraw from outside contacts, and others treat them as tainted or contagious. People feel vaguely uncomfortable around them without quite understanding why. When someone commits suicide, their kin and friends often feel cut off in this way.’ In communal relationships, friends feel mutual responsibility for each other’s welfare and provide benefits either in response to need or to demonstrate concern for the other person. The aim is not to selfishly get something out of the relationship or even to ensure that both parties receive an equal benefit but rather to meet the needs of the other person. Margaret Clarke studied communal relationships experimentally by leading subjects to believe that the stranger whom they had just met was likely to become a close friend (Clarke and Reis 1988). She found that when such an expectation was in place, people paid little attention to how much each of them had contributed to a joint task and, when the partner tried to immediately recompense them for favours they had given, they felt insulted and liked the partner less. They did not feel exploited when they gave help but received none in return. They were more likely to focus on the other’s needs, to offer help and to increase the amount of help given in line with the other person’s distress. Another kind of relationship has been called equality matching (Fiske 1992), reciprocity (Polanyi 1957) or exchange relationships (Clarke and Reis 1988). The focus

Mhoc05.fm Page 146 Friday, December 14, 2001 10:16 AM


    

is upon the rewards and costs in a relationship. Social psychologists who first took this econometric view of relationships argued that people seek to maximize their rewards and minimize their costs—the so-called minimax principle. However, this strategy is highly exploitative and although we may behave in this way with strangers it is unlikely to result in a sustained relationship. With acquaintances or colleagues, with whom we must maintain an ongoing relationship but with whom we are not emotionally involved, we tend to follow a rule of equity, according to which two people seek to maintain a similar ratio of rewards to costs. According to this view, if you put twice as much into a relationship as I do, we are still in balance as long as you get twice as much out of it. However, in long-term relationships where we experience a strong positive bond to the other person, the relationship becomes more finely tuned to one of equality (Deutsch 1985; Mikula 1980). The rewards and costs are equal for both parties. If I agree to act as chauffeur one night that we go out together, then it is only fair that you forgo alcohol the next time we go out. My cost has to be balanced at a later date by your cost and the immediate reward that you receive of being able to have a glass of wine has to be repaid to me sooner or later. The way it is repaid need not be exactly in kind—you might agree to babysit my child for a couple of hours instead. The point is that exchange relationships depend upon a close monitoring of who is benefiting most at any given time. When the costs become too great for one partner, the relationship will begin to show the strain and eventually may dissolve. Anthropologists have noted that it can also regulate some working relationships—neighbours may come to assist in threshing your grain on the understanding that you will do the same for them at a later date. The underlying notion of fairness or balance also guides co-worker relationships so that people feel aggrieved if they have a greater workload than their equally paid colleagues. When one party receives a favour they feel obligated to return it. This back-and-forth symmetry can be seen in the way that we send Christmas cards to one another or the way that we feel obliged to return a dinner invitation. Exchange relations are so universal that they carry with them both a moral force (as a principle that is binding so that a violation of it requires some apology or explanation) and a norm (a regularity of behaviour that people expect of one another). Gouldner (1973) argued in a classic paper that the norm of reciprocity is universal and since then no societies have been found in which it is absent. The aim of social exchange is to do as well the other person and to avoid coming out worst. The ideal of social exchange is exemplified in the process of blind justice where the aim is fairness regardless of the particular parties involved (Rawls 1971). But exchange relationships are not all fairness and harmony. The idea of mutual back-scratching has its flip side in an eye-for-an-eye. Much retaliatory violence and feuding stem from this same emphasis upon fairness of outcomes to all parties. And although exchange relationships may be egalitarian, they set the scene for very unequal relations in the group. If one person can dispense unique rewards that cannot be

Mhoc05.fm Page 147 Friday, December 14, 2001 10:16 AM

  :   


returned by others, a debt is set up that is paid back in deference, loyalty and submission. Clarke has studied exchange relationships experimentally by introducing strangers and leading them to believe that they could not expect to form a strong or enduring relationship with the other person (see Clarke and Weis 1988). She expected that this manipulation would result in subjects anticipating an exchange relationship. Sure enough, these subjects tended to forge a relationship governed by principles of equity. They liked it when they received immediate payback for favours given and when the other person requested immediate return of favours that they themselves had accepted. On joint tasks, they carefully monitored each individual’s input and felt exploited when help was not reciprocated. In these relationships, people like to know where they are and want to avoid both feelings of obligation (so they repay debts immediately) and feelings of exploitation (when another person does not immediately return a favour).

The evolutionary roots of communion and exchange relationships Armed with evolutionary theory we can go beyond a mere description of communal and exchange relations and examine the logic of their natural selection. When evolutionary theorists look at social relationships they begin by examining behaviour rather than emotion. They look at who helps whom and try to explain why. In effect, beneficent relationships are seen as special case of altruism—where an individual incurs costs in order to provide benefits to someone else. (For evolutionary thinkers this is distinct from cooperation where both parties benefit immediately from the transaction.) Altruism was a real problem for Darwin because it was unclear how such a trait could evolve. Let us suppose that one person in a community behaves altruistically; she places a greater importance on the lives of others than on her own. She spends her time helping others at the expense of herself while they benefit but experience no costs at all. Her genes will die out and theirs will survive. Yet we see acts of altruism all around us, from small favours to dramatic events. William Hamilton (1964), armed with knowledge of genetic transmission that Darwin did not have, solved his altruism problem not only with great elegance but in a way that changed our whole conception of the unit of analysis in evolution. He pointed out that the unit of genetic transmission is the gene and the same gene can reside in many bodies. It is especially likely to reside in the bodies of close relatives. He built his theory on Wright’s (1922) coefficient of genetic relatedness, r, which is the probability that a particular gene will be identical to another because they share a direct ancestry. This value is 0.5 for parents and their biological children, 0.25 for uncles or aunts with nieces and nephews and 0.125 for cousins (and 1.0 for monozygotic twins!). Hamilton argued that natural selection would favour altruism to kin where the cost to the altruist was less than the fitness benefit to the receiver multiplied by the genetic

Mhoc05.fm Page 148 Friday, December 14, 2001 10:16 AM

      relatedness of the pair (r). In short, altruism is more likely between those who are closely genetically related than between strangers. We are more likely to donate a kidney to our child than to someone we have never seen. This revolution in thinking about the unit of transmission is of course what led Richard Dawkins to title his seminal book The selfish gene (not The selfish individual). But with animals we see altruism among apparently unrelated individuals and even between different species. Robert Trivers (1971) was particularly fascinated by the behaviour of fish at underwater cleaner stations. Some large fish allow small cleaner fish to enter their mouth and feed on the ectoparasites within it. This in itself is a simple case of cooperation. Both fish benefit immediately—the small fish gets a meal and the host gets a free teeth-cleaning. What seems to require explanation is why the host fish does not end the encounter by simply swallowing the cleaner. Essentially, the host forgoes this benefit in order to save the small fish’s life. Indeed the host goes to some lengths to carefully signal to the smaller fish that he is about to close his mouth and his guest should leave. He even chases off other fish who represent a danger to the cleaner. But why? Trivers argued that it was because the two fish set up a long-standing relationship with one another. The cleaner is worth more dead than alive. If the big fish ate the cleaner, he might be unable to find another or at least one that was equally good. He would certainly encounter some expense and danger in seeking a new cleaner at a different location. Apparent altruism emerges but in fact it is in the big fish’s own best interests to behave in this way. Trivers speculated that humans evolved reciprocal altruism, as did other animals, because it was in their own best interests to do so. Altruism is a situation in which one individual dispenses a benefit to another that is greater than the cost of the act to the donor. If you are starving and I am replete and I give you a piece of food, the benefit to you is very great and the cost to me is small. But at some future time you will repay me on an equally grand scale. This kind of mutual back-scratching is especially likely to flourish in communities where the population does not disperse too widely—after all there is no pay-off for me if I help you, only to find that you have moved to another group and I will never be repaid. It is especially valuable in species with long lifetimes, since there is ample time for such altruistic acts to be returned. A good long-term memory helps (to keep track of debits and credits) as well as a particular ability to discriminate and recognize faces—an ability that humans possess and that seems to be underpinned by a specific psychological module that differs from the one we use to recognize objects (Sacks 1985). But wherever there is an adaptation, there is a counter-adaptation to challenge it. If most people are altruists then a cheater may flourish. After all, she can benefit from being helped by others while not having to risk her life or expend energy for anybody else. Cheaters do exist and psychopaths are classic examples of individuals who serially exploit others (Mealey 1995). The fact that they exist in small numbers suggests that the rest of us have specially evolved mechanisms for detecting their presence and avoiding them (Cosmides and Tooby 1992). Trivers suggests that

Mhoc05.fm Page 149 Friday, December 14, 2001 10:16 AM

  :   


cheat detection has also been responsible for a range of universal emotions. Anger and aggression are the responses to injustice, unfairness and lack of reciprocity (we can hear them in children’s furious cry ‘It’s not fair’) and they function to regulate cheating by the threat or actual delivery of harm. Guilt (an emotion so conspicuously absent in psychopaths that it is one of the diagnostic criteria for the condition) serves to motivate the cheater to repay the altruism that he has received. On the other hand, gratitude is an emotion that alerts us to the current credit that we have received from another person and sympathy may be calibrated to the degree of need that another person is in. For Trivers, friendship and liking are emotions that regulate long-standing patterns of reciprocal altruism between people. We are more altruistic to those that we like. But there are occasions when cheating is the most successful strategy. One method that has been used to discover when is the Prisoner’s Dilemma game. Here the players are asked to imagine that they are criminal partners who have both been arrested. Although the prosecutor does not have enough evidence to charge them with the serious crime that they have committed, he does have enough evidence to put them both away for a year on a lesser charge. In order to charge them with the more serious offence, the prosecutor must persuade one of the men to confess and to ‘finger’ the other. The men are in separate cells and each is faced with the same dilemma: ‘If you confess and your partner does not, then you are free and I will use your testimony to put your partner away for ten years. But if you confess and your partner does too, then you will both get three years. If neither of you confesses, we will put you both away for one year on the lesser charge. Exactly the same offer is now being made to your partner.’ (Although altruism is concerned with extending benefits to others, this dilemma effectively treats refusal to betray the partner as equivalent to such an act. It addresses the same underlying question of whether it is better to trust the other person—do them a favour, refuse to betray them—or not.) What is the best course of action? If your partner betrays you and you stay silent, then you will get ten years but if you also confess you will get three. So, the best solution if he betrays you? Betray him. Now suppose that he stays silent. If you do the same thing then you will get one year. But if you betray him, you will go free. The best solution if he does not betray you? Betray him. Things look good for the cheater strategy. Axelrod (1984) invited computer programmers to write the most successful program that they could for the dilemma. The programs would be run against one another to see which one won over 200 rounds with every other program. The winning program was called Tit-for-Tat and it worked embarrassingly simply. It began by non-betrayal and then did whatever its opponent had done on the previous encounter. (Since that time even better programs have emerged. The problem with Tit-for-Tat was that it got into a long series of betrayals when it encountered a betrayer. Forgiving Tit-for-Tat has a slight amendment. It forgives its partner for betrayals at a variable but fixed rate. Another version of much the same thing is

Mhoc05.fm Page 150 Friday, December 14, 2001 10:16 AM

      Tit-for-Two-Tats which, as its name suggests, is more slowly provoked to retaliation; see Dawkins 1989; Ridley 1997.) Over the long haul, mutual altruism pays, even though on a one-off encounter, it does not. There is something else about the Prisoner’s Dilemma that makes Tit-for-Tat so successful. It is a feature that is extremely common to human life and one that Trivers identified in his formulation of reciprocal altruism. It is a non-zero sum game—one player’s gain is not equal to the other player’s loss. The value of your life being saved from starvation (100 points) is not equal to the cost of the food that I was too full to eat (1 point). And if you reciprocate at a later date on an identical exchange then both of us have a net gain of 99 points. Non-zero sum also means that both parties can exploit the prosecutor by cooperating. It is the third party that suffers while the two cooperators benefit. Richard Dawkins (1989) gives a wonderful example from English soccer where three teams, playing on the last day of the season, were each striving to avoid relegation to a lower division. One of them would have to go down. Sunderland were playing a fourth team that was safe from the threat of relegation while the two equally desperate teams of Bristol and Coventry met one another. If Sunderland lost, then Bristol and Coventry needed only to draw to stay in the top division. Two minutes before the end of the game, the electronic board displayed the news that Sunderland had lost. Immediately both teams gave up any attempt to win and settled back into a final few minutes of distinctly lethargic play. Alliances can happen even between competitive individuals if by cooperating then can successfully defeat a third party. Chimps who work together to overthrow an alpha male are acting in their own best non-zero sum interests. Unfortunately, having overthrown him, there can be only one alpha male which translates their former cooperation into a more ruthless zero-sum contest. Now there is tendency to see these two forms of altruism—Hamilton’s kin selection and Trivers’ reciprocal altruism—as quite distinct. But a moment’s reflection reveals that they are not. Although we will go far out of our way to help a sibling in need, there is always a limit to that beneficence. Even among kin, altruism is not infinite. On the other hand, there are some relationships with non-kin that seem to stretch the bounds of reciprocation very far indeed. We continue to care for a terminally-ill friend even when there is no possibility of later reciprocation. Step-parents may not indulge their acquired offspring with the same degree of care that they do their own, but nonetheless they exhibit staggeringly high levels of care, given the absence of genetic relatedness. Among married partners, the degree of experienced emotional closeness approaches that felt for genetic relatives. The two forms of altruism shade into one another. It is as if our ‘kin responder mechanism’ can be fooled by particularly close, trusting or long-standing relationships so that we act altruistically without expectation of repayment. This should come as no surprise given that reciprocal altruism evolved out of kin selection. Kin selection was a necessary precursor to reciprocal altruism for a very simple reason. Reciprocal altruism cannot get started as a strategy in a population of

Mhoc05.fm Page 151 Friday, December 14, 2001 10:16 AM

  :   


individuals who exploit others. Imagine a hypothetical community of people in which 95 per cent refuse to act for the good of anyone else and indeed take advantage of anyone who does. The few altruists rarely encounter one another so their opening helpful move on any Tit-for-Tat encounter is met by betrayal. Eventually, as Darwin saw, the trait of altruism is pushed into extinction. In order to survive, altruism must reach a critical mass where altruists meet each other often enough to increase their pay-offs and their chances of survival. Where might such a critical mass of altruists be found? The answer is among individuals who are genetically related to one another. If you have genetic tendency to altruism there is a good chance that your sister has it too. Since reciprocal altruism is advantageous, we should expect that individuals in such families would be reproductively more successful than others and in this way the tendency for altruism might spread. In the end it might be present in the whole community. Once that happens, altruism has a second advantage. Betrayers or cheaters do not benefit from being in a majority of the population because they keep meeting others like themselves. Altruists, however, flourish where they are in the majority. As Robert Wright (1994, p. 200) puts it: ‘Simple conditional cooperation is more infectious than unmitigated meanness.’

Men and women as relational strategists Recall the distinction that we made between communal and exchange relationships. It is easy to see the correspondence between kin altruism and communal relationships on the one hand and reciprocal altruism and exchange relationships on the other. These can be seen as extreme ends of a continuum along which relationships can be ranged in terms of their closeness and the form of altruism on which they depend. Although some evolutionary psychologists seek to maintain a clean demarcation between categories of relationships, they also note that communal sharing is not restricted to kin and that linguistic usage is often employed to blur kin and non-kin relationships as in ‘Brother, can you spare a dime?’ (Daly et al. 1997; Haslam 1997). Individuals, rather than being discretely demarcated into kin and non-kin, can be gradated in terms of intimacy ranging from true and close genetic relatedness (a sibling) through extremely close friends (recall the fictive ‘aunts’ that were in fact your mother’s best friend) to well-liked acquaintances and strangers. These relationships map on to a continuum of altruism that ranges from kin selection in its most extreme form (as in a mother towards her child where no element of reciprocation is anticipated) to reciprocal altruism at its most extreme (where payment of a plumber is immediate, business-like and without any expectation of further interaction). There is some aversion among evolutionary psychologists to dimensions of this sort. Haslam (1997, p. 302) claims that relationships are governed by ‘discrete cognitive structures and do not simply represent poles on continuous relational dimensions’. But he goes on to note that ‘many relationships are governed by some

Mhoc05.fm Page 152 Friday, December 14, 2001 10:16 AM

      combination of two or more [relationship categories]’. Likewise, Daly et al. (1997, pp. 271–2) claim that ‘cognitive structures do not merely locate relationships in a space defined by a couple of dimensions such as “intimacy’’. They distinguish relational categories.’ But they also go on to say that ‘A long-standing mateship becomes increasingly like a genetic relationship . . . ’ (Daly et al. 1997, p. 278). This determined preference for categorical distinctions (in the light of an irritatingly continuous perceptual world) derives from a commitment to locate and describe evolved psychological modules. These separate modules (for cheat detection, for sexual jealousy, for taste preferences and so on) are thought to be activated by specific external stimuli (potential cheaters, signs of infidelity, sugar). Some evolutionary psychologists believe that adapted behaviour must be responsive to qualitatively discrete inputs and that the notion of dimensions destroys this categorical input. But does it? Take colour perception. It is true that everywhere in the world people discriminate the boundaries of colours at the same point in the hue wavelength spectrum (suggesting discreteness) but that does not mean that they are perceptually unable to tell the difference between a green that is nearly blue and a green that is nearly yellow. When we say ‘She is like a sister to me’, we are not distorting the objective input in the sense of really believing that we share genes with that person but we are redefining the position of that friend towards the kin end of our continuum. While evolutionary psychologists are happy to endorse the fact that we discriminate between fine degrees of kinship (as Hamilton’s theory requires), they are less happy to acknowledge that we discriminate equally fine degrees of friendship. It would seem on the face of it to be grossly maladaptive to suggest that we treat all non-kin in the same way. Indeed Trivers has noted how emotions such as gratitude and guilt serve to modulate, discriminate and refine our behaviour between members of the ‘friend’ category. Over time, friendships can gradually move from a Tit-for-Tat exchange relationship into a communal ‘pseudo-kin’ closeness. And it is along this continuum that men and women seem to differ in their friendships. Women’s close friendships are more communal and men’s more exchange-oriented. What is especially paradoxical is that women formed kin-like relationships with non-kin while men who shared genes with others in their group often looked more like reciprocal altruists. Let us take men first. Status striving, as we have seen, is greatly facilitated by alliances. These are often of a shifting and short-term nature as de Waal’s (1982) chimps have shown. A male chimp’s enemy’s enemy is his most useful ally for they both have something to gain by deposing an alpha male. But if the ally now assumes dominance, his erstwhile partner will seek a new ally to overthrow him. Furthermore, if our status-seeking chimp finally achieves his heart’s desire and becomes dominant, he will probably find his two prior allies lined up shoulder-toshoulder against him. All this Machievellian strategizing means that there is little room for trust—defined as a continued willingness to show altruism (such as coming to the defence of a threatened ally) when there has been no immediate payback

Mhoc05.fm Page 153 Friday, December 14, 2001 10:16 AM

  :   


of previous good turns. Males then should be keenly aware of reciprocation and less willing to endure long periods of self-sacrifice without seeing a return on it. There is no advantage to a male of risking his life for a partner over and over again when that partner not only fails to reciprocate but achieves dominance at the altruist’s expense. Given the shifting and opportunistic nature of these alliances, they should also be careful to conceal vulnerabilities. An ally who knows your weak spots is one thing—an enemy who knows them is quite another. But among male philopatric species such as chimps (and humans) surely this tendency for mutual distrust ought to be modulated by their kin relatedness? If a senior brother achieves dominance and sires sixteen offspring then his younger sibling increases his own inclusive fitness eightfold. Why should not the younger brother bask in the reflected glory and the accompanying, vicarious genetic success? The answer seems to be that, while chimps may share less valuable resources such as food with relatives, the value of reproductive dominance is simply too great to be compromised by kinship. And reciprocal altruism for all its debit/credit emphasis upon coming out even has the paradoxical quality of opening the door to unequal relationships. The more resources an individual can accrue, the greater is his ability to help others. Recall that under reciprocal altruism the costs to the giver are lower than the benefits accrued by the receiver. In the long run, the assumption is that the costs and benefits average out because although one individual may be rich in food today (and can be generous at low cost to himself), another may be oversupplied another day and so able to return the favour. But in a primate dominance hierarchy or in a human society where rewards are chronically unevenly distributed, those at the top have more ‘giving power’ than those below. Debtors who cannot repay in kind must repay in some other currency such as respect and deference. Perhaps this is why men think twice about incurring the kind of debt they cannot repay—a man who offers to find a job in his company for his friend’s son uses his status to perform a favour of such magnitude that the beneficiary may never be able to repay it fully. We call unwillingness to accept such a large debt ‘pride’. (Women often see it as selfishness because it penalizes the needy child in order to protect the father’s self-esteem.) The same premise explains the curious ‘round buying’ ritual seen in British pubs. Men will carefully monitor who bought the last round of beer and whose turn it is now, despite the fact that the monetary sums involved are relatively small. What the men are doing is ensuring equality of status among friends. People who ‘flash their money’ by buying round after round of drinks are not popular—quite the reverse. Their behaviour is read as an attempt to indebt others and to lower their status. Men who cannot ‘stand their rounds’ often stay away from the pub altogether rather than become indebted beneficiaries of a largesse that is read not as generosity but as a status display. Women are also reciprocal altruists with colleagues and acquaintances. They are aware of the debts and credits they have. When they say ‘It’s my turn for the next dinner party’ they demonstrate their sense of justice. But there is something special

Mhoc05.fm Page 154 Friday, December 14, 2001 10:16 AM

      about the one or two women that are singled out as best friends. Here the relationship is characterized by a much longer leash of trust—a willingness to extend credit for a long period of time based upon assessment of another person’s need. This trust is meted out with care. It may take years of intimacy to achieve and it is rarely extended beyond one or two close friends—a mistake can be costly. But such a friendship can afford a sense of solidarity and closeness that men may experience only with a spouse. If women are trust specialists, what particular features should we expect to see in their relationships with others? First, we would expect considerable selectivity about the beneficiaries of trust. It is a strategy that can backfire if the recipient simply takes advantage and deserts. Among primates, kin altruism is glued together by the mutuality of genetic relatedness and blood is much thicker than water. Relatives can be trusted because their genetic interests overlap to a large extent with your own. So a friend who will be like a sister must be selected with care. We would expect that women should be careful about who crosses the line between acquaintance and best friend. Recall that evolutionary theorists have suggested that altruism beyond the kin group must have depended upon selecting beneficiaries who were likely to repay their debts. One such mechanism is to choose friends who are similar to yourself. If they resemble you in other respects then there is a better than average chance they also resemble you in their tendency to reciprocate help. And this idea is corroborated by one of the most robust findings in social psychology—the tendency of friends to resemble one another. (Though this may seem obvious, many years were spent working on the hypothesis that friends were selected to be complementary rather than similar. If you are an introvert you might compensate by becoming friendly with an extrovert that could widen your social circle. This would not have been the prediction of evolutionary psychologists and they were right.) Now if reciprocal altruism depends upon similarity, how much more true this should be if we are speaking of ‘pseudo-kin’ altruism where trust lies in place of immediate payback. And as we would expect, female friendships depend more than males’ upon similarity of personality and attitude (Cross and Madson 1997). Boys like other boys who share the same interests so that they can focus their play on things that are satisfying for both of them. Girls are less concerned with common interests than with common character. Girls also spend more time talking to one another than do boys. These conversations are about their lives, feelings and aspirations involving increasingly high levels of self-disclosure (Dindia and Allen 1992). They are a testing ground for establishing similarity and trust. Teenage girls spend hours in one another’s bedrooms experimenting with their appearance, listening to music and chatting about themselves and others (McRobbie and Garber 1976). Once a best-friend relationship has been established, it is jealously guarded, as Eder and Sandford (1986, p. 293) describe:

Mhoc05.fm Page 155 Friday, December 14, 2001 10:16 AM

  :   


Tami at one point interrupted her story to ask Heidi if she had been combing Peggy’s hair the other day. Heidi denied it but Tami persisted saying ‘Yes, you had.’ Heidi denied it again and just then Peggy happened to walk by. Heidi asked her if she had been combing her hair and Peggy said, ‘no.’ Then Tami said to Heidi, ‘Well, whose hair were you combing?’ Heidi said ‘I wasn’t combing anybody’s hair!’

There is a high degree of exclusivity in these friendships and friends are acutely concerned with maintaining them. To spend too much time with a third girl is to threaten the bond between friends and much of the gossip that takes place concerns who is best friend with whom and whether that relationship is in jeopardy. Because girls invest so much of their private lives and personal feelings in the revelations that they make to one another, the loss of a best friend threatens not only loneliness but also the betrayal of the confidences that have been made. The importance of trust runs through the literature on girls’ and women’s friendships and the greatest betrayal is the abuse of trust. It is trust that distinguishes a best friend from other friends: I probably trust her the most of anybody and she trusts me. I mean, of course people have told me things that they don’t really want anyone else to know, and I’ve told someone else and everybody does that. But something Jane tells me I couldn’t tell anyone else. Something that I told her, I hope she couldn’t tell anyone else . . . Like Nicky [another friend] and me are very close but I couldn’t trust her at all. We’re basically not very alike. She doesn’t really know how much something means to me when I tell her something, and, I mean, she’s very bitchy behind my back. (Lees 1993, p. 76)

Because of the intensity of the relationship and what is at stake in terms of betrayal, when female friendships break up they often do so with acrimony and the split is long-lasting, sometimes irreparable (Davies 1984; Duncan 1999). (Interestingly, captive chimpanzee females show this same unwillingness to forgive; de Waal 1993.) A second prediction is that women should be more socially sensitive than are men. In an exchange relationship, a failure to successfully analyse character is less disastrous than in a communal relationship with its attendant measure of trust. The evidence is fairly clear that women more than men are relationship specialists. Women’s interpersonal sensitivity has been examined by looking at the ability to read non-verbal information from other’s behaviour such as posture, vocal inflection and facial expression. Many studies have examined sex differences in accuracy and the results clearly favour women (Hall 1984; Hall et al. 2000). This difference in the accuracy of facial expression processing is already apparent in infancy (McClure 2000). As well as decoding emotion, women seem to be more effective senders of signals that foster intimacy and closeness. They demonstrate greater involvement in the conversation and a greater interest in tracking the state of the other person’s mind than do men by smiling more, gazing more, displaying greater facial expressiveness, reflecting their own emotional state more clearly in their facial

Mhoc05.fm Page 156 Friday, December 14, 2001 10:16 AM


    

expression, using more hand gestures and approaching others more closely. In short, women seem keen to establish a mutuality in their conversations by engaging and monitoring their partner’s state of mind. Nowhere is the difference in sensitivity to another person’s experience more evident than in the sex ratio of two related disorders—autism and Asperger syndrome. Autism is an impairment of social development with deficits in communication and imagination, the presence of narrow, restricted interests and repetitive activity. The overwhelming majority of those affected are male—official figures put the ratio at about four to one (Rutter 1978). Three-quarters of those affected also have mental handicaps and if we take ‘pure’ cases of autism where intelligence is unaffected (often called Asperger syndrome), the ratio rises to nine males to one female (Wing 1981). One specific ability that is affected in autism is ‘mind reading’. Despite the terminology, there is nothing mysterious about this ability. It is simply the capacity to accurately infer the desires and beliefs that lie behind another person’s behaviour. In normal children, this ability manifests itself quite spontaneously by about the age of four. A typical diagnostic task is to show the child two puppets. One of them, Sally, places a marble in a basket and leaves the room. While she is away her partner changes the location of the marble from the basket to a box. The child is asked simply where Sally will look for her marble when she returns. The answer to this deceptively simple question lies in understanding that the absent puppet cannot have known what occurred when she was not in the room despite the fact that the child herself knows. It requires an ability to recognize the differing perspectives of the puppet and herself. Autistics compared with both normal children and a control group of learning disabled children show a specific impairment in this task. Yet they excel in some tasks at which normals fail such as identifying hidden drawings of objects embedded in larger pictures. Like Asperger’s sufferers, they have a selective difficulty with people while retaining an excellent memory for objects and a fascination (bordering on obsession) with non-social stimuli such as trains, cars or numbers. They are object-oriented with little understanding of the motivations of other people, which leads to deficits and inappropriateness in their social behaviour. Recently it has been suggested that such an impairment can be viewed as a pathological degree of ‘maleness’ (Baron-Cohen and Hammer 1997). One serendipitous but fascinating fact is that the same gene which controls the development of masculine reproductive organs also controls the length of fingers. Men typically have a larger index than ring finger while in females this ratio is reversed, and this ratio seems to predict a number of male characteristics including sperm count and testosterone level. As we might expect, autistic males show a more ‘masculine’ index to ring finger ratio than non-affected males (Manning et al. 1999). Mind reading, as a cognitive ability, leads naturally to empathy as an affective response to the information that is uncovered. (Autistic children, as we would expect, show very little response to the distress of another child.) Empathy can be defined as an affective response to the emotions of others (Jordan et al. 1991) and

Mhoc05.fm Page 157 Friday, December 14, 2001 10:16 AM

  :   


we would expect women to show higher levels if indeed they engage in more communal relationships where the other person’s need must be assessed and responded to. Sex differences in empathy have been found in infants in response to the sound of another infant crying (Hoffman 1977). In preschool children, girls respond to the distress of another child with greater emotional arousal reflected in their facial expression, statements of concern and gestures (Zahn-Waxler et al. 1992). In adulthood also, there are large differences favouring women where self-report measures are used or where facial expression is monitored in response to interactions designed to elicit sadness, sympathy and distress (Eisenberg and Lennon 1983; Eisenberg et al. 1989). However, when physiological measures such as heart rate, sweating and blood pressure changes are used, the sex difference disappears. Other researchers (Baumeister and Sommer 1997; Cross and Madson 1997) have noted that the experimental designs used to examine empathy frequently use a written story format or videos of people unfamiliar to the subject and that this may contribute to the absence of sex differences in physiological measures. We would expect that women should show selective responsiveness to those with whom they have established a communal relationship and that sex differences should be much less apparent when men and women are compared in their response to strangers. Furthermore, the tendency to empathize may have different motivations in the two sexes (Ickes et al. 1986). Men tend to adopt a partner’s perspective infrequently and usually in order to gain instrumental control of the interaction while women do so frequently and are motivated by the intrinsic satisfaction of sharing the other person’s mental state. A related question is whether men and women show differences in the circumstances in which they help others. Eagly and Crowley (1986) reviewed many studies of helping and concluded, contrary to our expectation, that men exceed women in their tendency to help. But remember that these studies were all experimental in nature—in most cases subjects were exposed to an unknown person in distress and their behaviour was evaluated. This is a very special form of helping. It is often construed as an emergency, it is extremely public and is a one-off encounter with a stranger. Men are more inclined to step forward because such circumstances call for leadership and decision-making skills in a forum where there is some scope for heroism. If we turn instead to the private and long-term sphere such as caring for family members, the sick, the old or the very young, women clearly predominate (Hochschild 1989; Lerner and Mikula 1994). Furthermore, Gilligan (1982) has argued that women, more than men, subscribe to a moral ‘ethic of care’. By this she means that women’s evaluation of moral dilemmas is more guided by the personalities and needs of the participants and less by stringent issues of fairness and the blind application of law. Gilligan’s idea came from her dissatisfaction with previous work on morality which she believed had erroneously given the impression that women were morally inferior to men by using male-biased tests and a male system of scoring. For example, on one standard test of

Mhoc05.fm Page 158 Friday, December 14, 2001 10:16 AM

      moral reasoning (‘Should a man steal a drug that can save his wife’s life if he cannot afford to pay the extravagant price asked by the manufacturer?’), men typically score at Stage Four. This stage emphasizes maintaining the social order by upholding the laws to which citizens have agreed. It takes the view of the system as a whole (‘What if everyone did that?’). Males tend to subscribe to the view that justice inheres in ignoring the identities and relationships of the actors and focusing instead upon the formal qualities of the act which are then evaluated in terms of the application of legal rules (although these rules may themselves be challenged if they are deemed to be unfair in their motivation or outcome). At the very heart of this approach is the formalization of an exchange relationship of morality. Citizens agree to subscribe to a common law so that, in the event of a dispute, a dispassionate law can ensure that cheaters are punished and equity is restored. Women typically score at Stage Three, which emphasizes the individual in relation to other individuals. It reflects a belief in the Golden Rule and highlights both the needs of others and the importance of mutual loyalty. This viewpoint harmonizes with what we know about women’s communal relationships. But according to the original scoring system, women had not yet managed to escape from a purely relational viewpoint—they were unable to stand apart from the actors involved and consider the dilemma from a dispassionate societal point of view of fair exchange. Gilligan’s work has attracted criticism. It has been argued that the sex differences she highlights are much less pervasive than she suggests and a number of metaanalytic reviews conclude that there are few differences in the moral stage at which the two sexes are capable of reasoning (Colby and Damon 1983; Thoma 1986; L. J. Walker 1984). Using real-life moral dilemmas, Gilligan and Attanucci (1988) found that while both sexes were able to use the justice and the care viewpoints, women tended to use the latter more then men. So, she argued, it is not a cognitive limitation that ties women to a communal analysis—it is a preference to think in terms of the needs and the mutual obligations of care that bind people together. A meta-analysis of 113 studies confirmed that there is a tendency for women more than men to use a care orientation when thinking about moral problems (Jaffee and Hyde 2000). Another criticism is that Gilligan places women on a pedestal—that her work implies that women’s behaviour must always be self-sacrificing (Coney and Mackey 1997). But this is a misunderstanding of what Gilligan is arguing. She is addressing the way in which women choose to think about moral dilemmas not about what they actually do. The relationship between moral reasoning and moral behaviour is far from strong (Kutnick 1986; Siegel 1982) but from the present perspective we are more interested in the thought processes of women rather than whether they live more selfless lives than men. Any evolutionary psychologist would be surprised if consistent self-sacrifice was a winning strategy for either men or women (MacDonald 1988). The debate about Gilligan’s work continues but studies tend to support her view that women place a high value on their close relationships with others and that

Mhoc05.fm Page 159 Friday, December 14, 2001 10:16 AM

  :   


an individual need-based analysis is more congenial to their way of thinking than the exchange-driven social contract view of men.

What made women communal? And so we come to the critical question—why did women form these exclusive intense relationships with other women to whom they were not related? The final answer is uncertain as female relationships have only recently come under scrutiny. But for evolutionary psychology, an important place to start is by considering what kin relationships offered females before they were forced to leave their natal group. Whatever benefits they obtained, it seems reasonable to guess that they were and are trying to recreate them in their new group. Kevin MacDonald (1995) suggests that the key to understanding women’s communal relationships is their maternal investment. Because females invest more heavily in offspring than do males, they are expected to have a greater elaboration of the affective mechanisms related to parental investment. He notes that women typically score higher than men on the personality dimension of nurturance/love and he proposes that this dimension of personality underlies the formation and maintenance of intimate, long-term relationships. But what this proposal amounts to is the suggestion that because mothers are good at loving their infants, they are also good at loving other women. From an evolutionary perspective, this is problematic because we would expect that individuals tailor their behaviour quite sensitively to the individual with whom they are interacting. If females displayed unconditional, need-responsive love to everyone around them, they would have little time and energy left to devote to their own lives and reproductive success. And as we have seen, females are quite capable of engaging in competition with other females when they represent a threat to their infants, or where there is competition for food. Indeed, MacDonald himself notes that ‘people can have radically different relationships with others depending on the context because different biological systems are involved’. If he is correct about this compartmentalization of relationships then we need a much clearer explanation of the precise nature of the interaction between the nurturance and hostility among both males and females. At present it is not clear how the tendency to specialize in communal relationships which he attributes to women is selectively turned on and off by the nature of the parties involved and why it remains generally higher in women than it does in men. We need an explanation of how maintaining a few intimate ties with other females actively benefits women rather than treating these relationships as a ‘spillover’ from their love for their own children. David Geary (1998) offers a different proposal and one which is based on a very different reading of the sex differences data. He concludes that women are more characterized by reciprocal exchange than are men and that this results from women entering new groups as strangers without kin. He argues that the maintenance of

Mhoc05.fm Page 160 Friday, December 14, 2001 10:16 AM


    

friendships depends more heavily upon strict reciprocity than does the maintenance of kin relationships. Under this proposal, women are more careful than men to ensure that they maintain equality in their relationships by monitoring their debits and credits and this increases their social solidarity. Because men reside with kin, they do not have to adhere to such strict principles of reciprocity. The logical deduction is therefore that it is males that are the communal sex. If that is so, then how can we explain the fact that among male chimps levels of aggression are twenty times higher than among female (de Waal 1993), that the quest for dominance among male primates involves the careful trading of favours in the formation of coalitions or that human males strive to avoid high levels of intimacy in their relationships with one another? Although evolutionary theory suggests that blood is thicker than water, reproductive opportunities achieved through dominance are considerably stronger than blood ties, resulting in fierce competition even between related males. In fact Geary himself acknowledges that communal relationships characterize females more than males: ‘girls are more consistently communal—manifesting greater empathy, more concern for the well-being of other girls, more nurturing, intimacy, social–emotional support and so forth—than the relationships that develop among boys’ (Geary 1998, p. 242). This communal orientation has been described by researchers from a variety of disciplines, including anthropology, psychology and gender studies (e.g. Bakan 1966; Clarke and Reis 1988; Fiske 1992; Spence 1985), and all agree that the hallmark of this orientation is that it is responsive to need rather than to the principle of equal exchange of services. Although Geary and I differ on the interpersonal profile of women’s friendships, he agrees that females form especially close relationships with one another and he proposes a reason. He suggests that the driving force is the well-being of their children. He argues that children thrive best in a stable social atmosphere and that women cement their friendships in order to provide such a tranquil environment (see also de Waal 1993). But if this was the goal, it could be accomplished as easily by simply attenuating levels of female–female aggression. His argument hangs on an absence of discord not upon the positive establishment of communality. Chimpanzee females minimize intrasexual conflict by foraging at different sites alone with their infants and they, like humans, are male philopatric. Why did women not adopt the same asocial solution? Peace and harmony requires only the absence of conflict but we chose something more than that—women carefully foster warm relationships with one another. Geary (1998, pp. 250–1) has a second suggestion. A female immigrating into a community is a stranger. She has no allies and she may well have potential enemies. Other females may resent the arrival of a strange female who represents yet another competitor for resources in their territory. It may be in her best interests to establish the warmest relationship that she can with her new comrades in order to avert social isolation (at best) or aggression (at worst). If she can persuade these strangers that she represents no threat to them or even better that she is like a sister to them, she

Mhoc05.fm Page 161 Friday, December 14, 2001 10:16 AM

  :   


may be able to rely on their acceptance and be freed from the anxiety of potential harassment. In this view, it is females that represent the threat to her, and her behaviour is designed to ingratiate herself to them. But again we are faced with the question of why a female should want to bond so tightly with others rather than pursuing a tranquil if lonely life foraging with her infants. One answer that holds true not just for women but for most physically vulnerable animals is that there is safety in numbers. Even birds flock closely together, driven by the desire to work their way to the centre of the group where the risk of predation is low. And the possibility of predator attack has been offered by others as an explanation of female grouping, at least in female philopatric species (Dunbar 1988; van Schaik 1989). The predator hypothesis depends in part on establishing just how vulnerable early hominids were. Chimpanzee females forage alone and return to the group to sleep—the time at which they are most vulnerable—but on the savannah the risk of predation may have been much higher than in the forests inhabited by chimpanzees. But since males are stronger and more aggressive than females, surely it would have been even more advantageous to seek male friendships rather than female ones? Among gorillas, females emigrate from their natal group and join a new one dominated by a single breeding silverback. The females have only the most rudimentary relationships with one another. Instead they seek the protection of the silverback, especially when they have young (Stewart and Harcourt 1987). But here the silverback male knows that he is the father of the female’s offspring and may be more likely to offer protection than among promiscuous species such as the chimpanzee where paternity is far more uncertain. Primatologists have ideas about female relationships based upon another fundamental need—food. Richard Wrangham (1986; Wrangham and Peterson 1996) suggests that for some species the food on which they depend, chiefly fruit, grows in massed, clumped areas that can be defended by the group against rival groups. In such a situation, females need strong allies and the most dependable are kin. For these female-bonded species, females remain in their natal group and support one another against incursions by neighbouring bands of females. But there are problems with this—some species of fruit-eating primates (for example common chimpanzees and spider monkeys) do not form female-bonded groups and many mammals that do show female philopatry are solitary and do not depend upon cooperative resource guarding (Pusey and Packer 1987). In addition, despite the term female-bonded, relations within the group are not harmonious. There is daily low-level antagonism between matrilines and this suggests that we have to consider within-group as well as between-group competition. Van Schaik (1989; see also van Hoof and van Schaik 1992) also proposes that female primate grouping depends upon food—but the driving force is finding a way for every member within a group to get enough food to enhance her reproductive success. Where food grows in fairly large patches, females adopt a conflict approach to getting their share. They compete against one another to control access to the

Mhoc05.fm Page 162 Friday, December 14, 2001 10:16 AM


    

food patch and females remain with their relatives to increase their number of allies in these competitions. They cooperate with kin but there is frequent low-level bickering between matrilines which occupy different positions in a well-defined and enduring dominance hierarchy. The characteristics of this kind of group (which corresponds to Wrangham’s female-bonded groups) are frequent displacements over food (subordinate animals are excluded from access), a very clear dominance hierarchy that depends upon the size of the matriline (because a larger set of relatives provides more allies), a stable ranking of both matrilines and the individuals within them and, if a territory becomes depleted of food, a trend for entire matrilines to emigrate together thus maintaining their kin support. This pattern is common in Old World monkeys. But where food cannot be defended (because the patches are small and widely spaced or because the patch is too big for a single matriline to control access), scramble competition takes the place of conflict. Here animals range over the available territory eating wherever they can. There is no advantage in competing and hence no need for kin to remain together. In these non-female-bonded species, aggression is rare. If a dominance hierarchy can be established it is vague, unstable and not dependent on fighting power. Females spend little time in one another’s company and when food is depleted, they are likely to emigrate alone to a new area since allies confer no particular advantage. This is a pattern that is often seen in common chimpanzees, langurs, howlers, colobus and gorillas. In this model, the driving force of female relationships is food availability and distribution. Kin bonding appears as a solution to controlling particular kinds of defensible foods. But we are still left with the puzzle with which we began—how did women who left their natal group at birth (‘non-female-bonded’ in the kin sense) manage to establish such close relationships with one another (‘female-bonded’ in the social sense)? One way to approach the puzzle is by looking at the bonobo. Like chimpanzees, females leave their natal group but there the resemblance ends: Female relationships are the most disparate between the two species. Among female common chimpanzees, mutual social interactions, such as grooming, were found to be rare since the beginning of research. In contrast to the asocial female common chimpanzee, female pygmy chimpanzees are extremely sociable and always feed, travel, and rest in clusters; social interaction between them is also lively. (Kano 1992, p. 205)

What makes female bonobos so friendly? Originally it was suggested that they were displaying mere tolerance rather than bonding (Wrangham 1986) and that this was driven by the need to attract males to travel with them. Because two mates are more attractive than one, females tolerated each other primarily as a means of attracting males. However, Amy Parish (1996), after hundreds of hours of close observation, found that on 60 per cent of the occasions where an adult female was close to another adult group member, it was to another female. Females followed

Mhoc05.fm Page 163 Friday, December 14, 2001 10:16 AM

  :   


other females more than seven times as frequently as they followed males. When females did approach males, the interaction was terminated by the female on threequarters of the occasions which hardly suggests a keen interest in the opposite sex. When Parish examined affiliative behaviour, such as body contact, play and grooming, she found that two-thirds of this activity by females was directed at other females rather than males. This suggests not mere tolerance but preference and it also suggests that female affiliation is not just an instrument for gaining access to males. Even more remarkably, Parish found that males did not dominate females. In fact, quite the reverse. When aggression occurred between the sexes it went in one direction only—from females to males. And their massed attacks could be severe with fingers, toes and testicles bitten off. But how did such solidarity come about between unrelated females? Parish, like van Schaik, thinks the starting point may have been food availability. Separated from chimpanzees by the Zaire river, bonobos live in a setting where food is abundant and where its distribution can support large numbers of animals feeding together. Unlike chimpanzees, they are not forced to separate into mother–child units to find food and so females are not isolated from one another during the greater part of the day. So the first step towards pseudo-kin female bonding was spending time together. Step two was the fact that groups of females acting together were able to defend food resources from males by driving them way. The third step was that this cooperative activity further enhanced the bonds between females and proved adaptive because all females could increase their reproductive success. It increased their food intake and gave them crucial leverage over the reproductive costs of male aggression. Neither infanticide nor forced copulation has ever been reported among bonobos. By sticking together females can decrease the dangerousness of male attacks, increase their food intake and enhance their reproductive success. All this depends upon cooperation and upon trust. Females can only dominate males when they act in concert and each one has to be sure that she has female back-up before she risks an attack. Only numbers can prevail over the greater size and strength of a bonobo male. A female who goes forward only to find herself deserted by her friends is staring death or at least severe injury in the face. Bonobo females have to be certain that they will not be deserted and that certainty comes from trust. It is cemented by mutual grooming and by the genital rubbing that maintains the affectional ties among the female group. Bonobos do not have dominant females who coopt subordinates to follow them—their aggression towards males and their friendly sexuality towards one another are based upon equality. The most striking thing that Parish’s data show is that kinship is not a necessary condition for bonding between females. Where females have the chance to spend time together and where there are benefits to mutuality, unrelated females can form powerful sisterhoods. And her observations help to make sense of observations of bonobos’ cousins, the common chimpanzees. In captivity, they are provisioned with food and hence

Mhoc05.fm Page 164 Friday, December 14, 2001 10:16 AM


    

females no longer have to separate and forage alone. The more time they spend together, the stronger the bonds between females become (de Waal 1993). But for common chimps, this pattern of female bonding is a facultative adaptation to new feeding conditions and hence it is not so evolved, elaborated or extreme as among bonobos. Females never achieve sufficient solidarity to dominate males. But they do show friendships that are directed at one or two other females and that are very stable over time. They involve commitment not just to a female ‘friend’ but also to her offspring. These friendships are cemented by grooming and association and seem to be built upon personal preferences and shared history—unlike the more instrumental and shifting alliances of males. While female ‘best friends’ show strong and enduring mutual commitment, females also display systematic exclusion of others that they do not choose as ‘friends’. It all sounds so familiar and so human. One or two close friends but not quite the larger cohesive solidarity of the bonobo.

Men, women and female friends Why were women unable to form this powerful group-level alliance with one another? Barbara Smuts (1995) suggests that two primary forces were involved: male control of resources and wealth differentials between men. Like chimp and bonobo females, women emigrated from their natal group. But unlike all other primate species, women do not provide for themselves but rely upon males. In hunter-gatherer societies, this dependence is limited to meat as the women perform most of the local foraging for roots and berries. But meat is rare and highly valued and among the Kung San, women yearn for the taste of protein-rich meat and the sizzling fat that it contains (Shostak 1981). However, with the advent of agriculture and animal husbandry, not only diet but the very structure of societies began to change. Women no longer roamed freely during the day but were confined to the local fields or to the home where they tended cattle. It became easier for men to monitor and control their whereabouts. At the same time, as men made a greater contribution to women’s food supply, they became increasingly concerned with monitoring women’s sexuality to ensure their own paternal certainty. And as women came to depend more and more on the resources provided by males, it became increasingly risky to resist male control (Irons 1983). Under agriculture, wealth differentials increased between men because surplus resources could be stored and traded—unlike the day-to-day, hand-to-mouth existence of hunter-gatherer peoples. So differences in status and power among men became more marked. In these circumstances the value of cooperation between women was reduced because each woman could do better by competing for the best male than by creating alliances with other women. The bonds between human females could not become superordinate because of the divide-and-rule strategy of male wealth and resource control. But nonetheless contemporary women resist this isolation from one another. Although young women may compete for mates, after marriage women often make

Mhoc05.fm Page 165 Friday, December 14, 2001 10:16 AM

  :   


and retain strong and intimate friendships with one another. What benefits do they obtain from friends and how do they replace the benefits offered by kin? I believe the answer is safety—not from predators or from rival groups or from other females—but from male aggression. We have seen that among bonobos male aggression against females is extremely rare. Among humans, this is not the case. In extreme female exogamy, women were cut off from their kin and from the natural support and protection of their family. In agricultural and industrial society, the family became increasingly atomized constituting a unit of its own, subject to the economic control of the husband. Even where women made substantial contributions to economic production (as in agricultural societies), they lost the right to control the fruits of their own labour. Under these patriarchal arrangements by which women were divided from one another and dependent on a man, men assumed the right to control family affairs in the privacy of the home. And many men used this privacy to beat and abuse their wives. In all biparental species, males exhibit marked concern with ensuring paternity and avoiding cuckoldry. In some animals, this entails extreme mate guarding during oestrus and in others, tactics of sperm competition have evolved. In human societies the requirement of virginity at marriage, chastity belts, purdah, infibulation and clitoridectomy are all practices designed to control the sexual life of women. The independent development of adultery laws in the Mediterranean, the Far East, the Andes, Mexico, Northern Europe and Africa all converged in defining adultery exclusively in terms of the marital status of the woman involved (Daly et al. 1982). In criminal law, the ‘justifiable’ rage of a sexually betrayed husband provides mitigation for his violence and the Anglo-Saxon law of torts specifies adultery, loss of consortium, enticement, criminal conversation, alienation of affection, seduction and abduction as wrongs for which a husband may seek compensation (Wilson and Daly 1992). Studies of wife abuse reveal that it is very closely associated with the husband’s desire to control his wife’s sexuality. Men who abuse their wives often believe that she is having an affair, or that she plans to leave them, or that she seeks a more independent lifestyle that will diminish the man’s ability to monitor her whereabouts. The exercise of this kind of ‘patriarchal terrorism’ depends upon isolating the wife from sources of social support and this process often begins in courtship before the violence has begun. Battered women typically interpret this early behaviour as deeply romantic—the fiancé wants to know where she is at all times, phones her often, meets her from work and seems besotted with her. Later, she realizes that this behaviour is not love but a deep desire to isolate and control her (Browne 1987; Pagelow 1984; L. E. Walker 1979, 1984). Such men not only cut the woman off from friends but are equally eager to isolate her from her family—both can both offer an escape route from his control. We were two miles from the village. He allowed me half an hour to go up to the village and half an hour to walk back and ten minutes to get what I needed from the shops. That was what I was allowed, an hour and ten minutes. If I was nae back inside that hour and

Mhoc05.fm Page 166 Friday, December 14, 2001 10:16 AM


    

ten minutes I got met at the door saying where the f— hell have you been. . . . I used to marvel at women that could go down shopping, you know, and their friends say, ‘Oh, come in for a cup of coffee.’ And they spend an half an hour chatting and having a cup of coffee and then come home. But I couldnae do that. (Dobash and Dobash 1979, p. 129)

What protection can a woman find? State intervention, orchestrated by men, has waxed and waned over the years depending upon the prevailing political attitude to the family and, by extension, to men’s right to control women by force. The greater the rights and privileges of the traditional family, the lower the interest in the criminalization of the family. When family violence is seen as a crime that threatens not only its victims but also the social order, support for criminalization of family violence increases. Thus the conditions that impede the criminalization of family violence are increased respect for family privacy, diminished enthusiasm for the state’s responsibility to intervene in the family, and vigorous defense of the family ideal, a range of beliefs about the sanctity of the family that serves to shield the home from public view and state intervention. (Pleck 1989, pp. 20–1)

Family and friends offer a more dependable source of support than the state. Despite the fact that female exogamy has meant that women enter new communities effectively as strangers, in some non-western societies women inhabit a sex-segregated world where eating, working, resting and leisure time is spent primarily with samesex. Even though women may be subordinate to men, on a daily basis they have limited interaction with them. ‘Women inhabit a women’s sphere that buffers them from male coercion’ (Burgess and Draper 1989, p. 82). At the same time, women often maintain ties with their own kin through visits and only become completely isolated from them under conditions of extreme patrilocality, patriarchy and warfare (Rodseth et al. 1991). Family and friends offer a degree of protection from male abuse by virtue of their nearness and the flow of information and gossip between them. Women maintain kin relationships after exogamy except where they are prevented for doing so (Rodseth et al. 1991). Women make the most active efforts to maintain contact with kin (Hogan and Eggebeen 1995; Schneider and Cottrell 1975) and know more about their own genealogy than do men (Salmon and Daly 1996). In accounts of contemporary marriage, women rely on the advice, help and support of their natal family, especially of their mothers (Komarovsky and Philips 1962). Indeed, the long post-reproductive lifespan enjoyed by women has been attributed to the importance of grandmothering in enhancing daughter’s reproductive success (Hawkes et al. 1998). Mothers are more apt to offer assistance to daughters rather than sons because a daughter’s line of descent is uncompromised by uncertainty of parenthood. Daughters, more than sons, keep in contact with their family of origin when they leave home and they are more likely to stay in touch with maternal rather than paternal kin (Salmon 1999). Friends offer a proximal substitute for the kin that matter so deeply to women. In a study of 300 Californian women (Essock-Vitale and McGuire 1985), respondents

Mhoc05.fm Page 167 Friday, December 14, 2001 10:16 AM

  :   


were asked about who they received the most help from during the preceding year. Friends were nominated by 35 per cent of the sample and were a greater source of immediate help than blood relatives (19 per cent). Where female strength is less than males, a woman’s best hope lies in allies. But these allies must be more than fleeting or useful acquaintances. Since her safety may depend upon them, she must create bonds of trust that are as strong as those she feels with her own family. In a Milwaukee study, 43 per cent of battered women obtained help from family and 52 per cent from friends (Bowker 1984). Frieze et al. (1980) also found that 55 per cent had sought help from relatives and 52 per cent from friends. Social support can be an important determinant of coping and readjustment (Janoff-Bulman 1985). They can provide sympathy, confrontation with the aggressor, referrals to agencies and shelter. But perhaps the main help from friends is preventative. Friends join a woman to a wider community and close friends are those with whom a woman can share her most private experiences. They can render semi-public, events that would otherwise be private. In this way they can act as an effective deterrent to abuse. Now it may seem as if I have overstated the dangers of male partners. In contemporary monogamous societies, surely men are protectors and defenders of women? Male protection is often identified by evolutionists (Buss and Schmitt 1993; Symons 1979) as a critical component of females’ mate selection and no doubt this is true in many species. Chimpanzee females, when they are in oestrus, are frequent targets of aggression by males and are vulnerable to attack when they are foraging (either alone or with their infant) by bands of rival males. Infanticide by immigrant males is also a real threat in many primates. For females, the price of male protection from these equally-made dangers is sex. Even originally non-sexual ‘friendships’ among male and female primates (which include protection of the female and her offspring) seem to be undertaken with a view on the male’s part to future copulation (Small 1993). Among humans, paying for male protection means guaranteeing exclusive sexual access to a man over many months or years. But there is a flip side to a husband’s protection. Concealed ovulation and high paternal investment increase his motivation to ensure female fidelity at all costs. In our own species, the irony is that a woman is more likely to be attacked by her partner than by a male stranger. The effect is to turn her partner from being a woman’s protector to the greatest threat to her life. In the United States in 1994, 71 per cent of female victims of violent crime were attacked by an intimate, friend or acquaintance and 46 per cent of these acts occurred in the home. Sixty-eight per cent of women who were raped were attacked by a spouse, ex-spouse, partner or male friend. Women homicide victims were murdered by a male in nine out of ten cases. Thirty-one per cent of women were murdered by a spouse, ex-spouse or boyfriend and a further 24 per cent by a friend or acquaintance (Craven 1996, 1997). The social arrangements under which we live make a long-term monogamous partnership the most attractive option for a woman who wants to raise children. Without a second provider, women must work in order to pay for full-time childcare or are

Mhoc05.fm Page 168 Friday, December 14, 2001 10:16 AM


    

forced into poverty via state dependence. And a husband can offer not only material support but emotional support also. In Essock-Vitale and McGuire’s study, 40 per cent of women identified their husband as the person who had given them most help during the preceding year. But male proprietary feelings for women run deep and, especially among men who doubt their ability to meet the needs of their wives, this insecurity can erupt into violence, turning them from protectors to assailants. It is the links that women can forge not with men but with other women that are likely to offer her protection.

Lord help the sister . . . In this chapter, I have argued that female bonding requires an explanation in its own terms. It is tempting to see it as simply the result of female’s lack of interest in dominance—if individuals are not engaged in status struggles, surely friendship is bound to ensue? We know that this explanation will not suffice. Dominance and nurturance are not two opposite ends of the same dimension. To be low on one dimension does not mean that one must be high on the other. Chimpanzee and gorilla females show little interest in dominance but at the same time show little interest in female friendships either. Women did not forge communal bonds just because they lacked dominance motivation. They did so because these bonds offered a measure of safety in a world where patriarchy and male control of the family made life dangerous for a female alone. In the end, women found a path between the male-worshipping isolation of the female gorilla and the aggressive sisterhood of the bonobo. Women can protect and support one another despite the divisive competition for mates that their loss of self-sufficiency has engendered. And this competition for men, either ignored or denied by feminists, is another outcome of imposed social arrangements that women have found a way to live with, as we shall see in the next chapter.

Mhoc06.fm Page 169 Friday, December 14, 2001 10:17 AM

 


But she that filches from me my good name: Women and competition

In South-East Asia, thousands upon thousands of male fireflies sit in a group of mangrove trees in the forest and flash in synchronous rhythm through the night. The effect is so dazzling that it is used as a navigational aid by ships and aircraft. The fireflies care not a jot about the safety of passing humans but care deeply about how attractive they are to firefly females. They gather together to flaunt their brilliant lights and so to entice females to visit and to select the very best of them for mating. Females are more likely to come and survey a group of males rather than a lone ‘flasher’ and they know where to come because the display ground is passed from generation to generation. Among birds, this type of communal male display is called a lek. The lek site is situated apart from the usual nesting or feeding territory and is common in species where males and females meet only for the purpose of copulation. Four hundred or so male North American sage grouse occupy a hectare or more of land during the mating season. Each bird monopolizes his own area that can vary from 10 to 100 square metres. But only the males at the very centre of the lek are attractive to females. Because of this, about 10 per cent of the males achieve 75 per cent of the copulations. The males are naturally very competitive about who will gain the best real estate and the young males do better and better each year. When a prospective female buyer arrives, the male inflates his chest sac, tilts his body forwards and holds his head high. He erects the plumes on the side of his head and expands the yellow combs above his eyes. Then he performs a noisy, two-second ritual. After making a brief cooing sound, he extends his wings forwards and then backwards over his chest, raises and then abruptly drops his chest sac to reveal two bare yellow patches of skin on his chest. The second time he does it, the sac is compressed and air is released into the pockets of yellow skin which inflate forwards like balloons and then collapse. The whole operation produces two sharp snapping sounds which can be heard several hundred metres away. His territory makes him initially attractive to the female but among the centrally located prime males, it is the strength and grace of the chest display that clinches the deal in favour of one male rather than another.

Mhoc06.fm Page 170 Friday, December 14, 2001 10:17 AM


    

After several days of window shopping, the female crouches down before her chosen mate and he obliges with the only gift she will ever receive from him—his genes. It seems on first sight a rather arbitrary way to select a mate. Yet it serves the female’s purpose well. She is not choosing a husband but a set of genes. The male will play no part in defending, feeding or caring for the offspring. She is not interested in finding a good parent. Male parental care is rare—it occurs in only about 5 per cent of species and among primates male help is restricted to carrying infants now and again. Even this may have little to do with genuine attachment or concern. In macaques and baboons males carry infants to deflect aggression from other animals (a strategy called agonistic buffering), to enlist support if they are attacked or to increase their future sexual access to the infants mother (Clutton-Brock 1991). Whatever a female grouse wants it is not long-term commitment, so what makes one male better than another? Darwin firmly believed in the importance of female choice. Many of his followers were less confident, preferring instead to highlight the importance of competition between males for female favours—elaborate displays were to intimidate opponents not to seduce females. More than a hundred years after Darwin, Donald Symons (1979, p. 203) wrote that ‘Human physical sex differences are most parsimoniously explained as the outcome of intrasexual selection (the result of male–male competition) and perhaps natural and artificial selection, not intersexual selection or female choice’, despite the fact that female choice has been demonstrated unequivocally in a wide variety of different species (Andersson 1994). Nonetheless Darwin was vague about why some characteristics were preferred over others. Colourful plumage or flamboyant displays were thought to ‘charm the female’ as a result of some innate sensory bias towards the particular trait. So for Darwin, female preference was a priori and preceded the appearance of the trait in males. There is evidence that he may have been right. Ryan (1991), studying tungara frogs, found that although other related frog species did not display the trademark tungara ‘chuck’ noise at the end of their mating call, the females of these species when played a sound recording of them evidenced a distinct preference for the sound by hopping eagerly towards it. So it may be the case that there are inchoate female preferences and successful males simply ‘discover’ and exploit them. In this view, female preferences are essentially frivolous—they are not geared towards any particular advantage. This approach, championed by Fisher (1930), has come to be known as the sexy son hypothesis. Once a preference, however arbitrary, begins to emerge it takes on a dynamic of its own called runaway sexual selection. If most women prefer men who are tall, then any woman that bucks the trend and goes for short men will be less likely to produce tall sons herself. She will also be less likely to produce daughters who share the common trend of tall-man mate preference. She loses on two counts—her sons get fewer mates and her daughters, like herself, choose undesirable partners. The dynamic of contagious female choice can be witnessed at a lek. Female grouse are more likely to mate with a male who has just

Mhoc06.fm Page 171 Friday, December 14, 2001 10:17 AM

  


finished mating with another female. In some fish, females prefer the male who they have just seen courting another female rather than swimming about solo. In humans also, women’s ratings of men’s attractiveness are influenced by those of other women, especially when the target men are rated negatively by others (Graziano et al. 1993). It is hard not to see teenage girls’ contagious and spiralling hero-worship of a rock star in much the same terms. It is a kind of market-forces model—the more people buy a given commodity, the more desirable it becomes to others. Nothing succeeds like success and fashion creates its own dynamic. The only cost to females in this scenario is the likelihood of becoming so picky that they waste good time that they could be spending in reproduction. The lek offers a perfect solution to the problem of comparative shopping. All the goods are on display at the same store and the female does not have to waste her time trudging from shop to shop. Among peacocks the female preference for huge decorated tails led to the evolution of ever-more enlarged tail displays until at some point natural selection (greater vulnerability to predators as a result of slower running speed caused by carrying the weight of a large if unfurled tail) countered the sexual selection pressure. At some point tall men cannot get taller without incurring a greater risk of weakened bones and more fractures. There is line beyond which the advantages that accrue from mating success are offset by the disadvantages of compromising survival. But there is a theoretical problem for the sexy son hypothesis called the ‘lek paradox’ and it is simple: if females all want the same thing, they will all mate with males who possess the trait. After a few generations, all their sons will have it and all their daughters will want it and there will be no variability left to drive selection. At that point female choice becomes unimportant. But genes are subject to random mutation and mutations are in the main bad news. The more complex the trait that is coded for, the less likely is a genetic mutation to improve it. Most mutations are harmful. They have the effect, therefore, of introducing variation into an otherwise ‘perfect’ genotype and over generations they maintain variability (Pomiankowski et al. 1991). But there may be a reason for choosing elaborate ornaments that is more utilitarian. And here we see a hint of a second approach to female choice. Perhaps it is not arbitrary after all—perhaps females choose males for a good reason. This view is called the good genes approach. This camp argues that the traits that females prefer are true indications of the genetic quality of the male—which after all will determine 50 per cent of her offspring’s own qualities. Certainly sons will inherit their father’s sexiness but offspring of both sexes will also inherit a host of other traits that will fit them for survival to a greater or lesser extent. In this view, male ornaments are designed to advertise the fitness of the bearer. And the very trade-off between sexual and natural selection that had troubled Fisher’s theory became a central part of the good genes theory in the form of the handicap principle. Zahavi (1975) reasoned that peacocks’ tails represent a distinct disadvantage in terms of survival. Peacocks would be better off without them. For sexy son theorists, they emerged because

Mhoc06.fm Page 172 Friday, December 14, 2001 10:17 AM


    

females whimsically liked them. But for Zahavi they emerged because what females like is good genetic quality and what his huge and brilliant blue tail said to them was that here was a male who had been subjected to even greater predatory threats than his less colourful peers but who had nevertheless survived. He had proved his genetic superiority by staying alive despite the massive threat posed by possession of that tail. What females use as an index of genetic quality may be more than the presence of gaudy plumage. It may also be its quality. Parasites represent a constant threat to health and survival. Fortunately animals are equipped with resistance to many parasites. But parasites replicate at a rate which far exceeds that of their hosts and in each generation find new ways around their host’s defences. The Red Queen effect is a response—by reproducing sexually, animals produce offspring that vary in their resistance to different strains of parasite. If females select males for disease resistance and if the biggest threat of disease comes from different parasites in different generations, then females will be changing their tastes over time. The question is: how do females manage to read healthiness simply from inspecting males? Hamilton and Zuk (1982) found that the most brightly coloured species of birds were the ones that carried the greatest load of blood parasites. It seems, therefore, that the quality of ornaments is a good way of establishing the health of a male. They proposed that male health could be read off the state of his secondary sexual characteristics. Hamilton and Zuk tested this idea experimentally by infecting young jungle fowl with a parasitic worm. These males grew more slowly and, when they matured, sexually-selected traits were more affected by their poor health than were other bodily characteristics. The parasites made their comb smaller and less brightly coloured but did not affect other neutral traits like ankle length to which the females were indifferent. Compared with non-diseased males, the parasite group was selected by females only half as often as mates. If spots are painted on a male grouse’s air-sac to mimic the effect of lice infestation, the male’s reproductive success drops (Boyce 1990). Other studies find that parasite infection is associated with deterioration of sexually-selected characteristics, poorer survival, lower female preference and inferior infant survival rates (Moller 1994; von Schantz et al. 1996; Wedekind 1992). Another line of work which favours the good genes approach has come from studies of symmetry. Symmetrical bodies are harder to grow than asymmetrical ones—any slight error in genetic translation or an environmental perturbation such as stress or malnourishment causes development to become uneven. Females searching for high-quality males would do well to pay attention to their symmetry. In the barn swallow, males have long symmetrical tail feathers. Although these feathers are a handicap to the male (and his mate) in terms of food foraging, they are a good predictor of a male’s speed of mating, likelihood of producing a second brood in a single season, the number of extra-pair copulations he obtains, the quality of his mate’s parental investment and his reproductive success (Moller 1994). Tail feathers are an index of parasite load and physical health. The female chooses her

Mhoc06.fm Page 173 Friday, December 14, 2001 10:17 AM

  


mate by establishing his health which she reads principally from his tail feathers. And this relationship between symmetry and reproductive success has been found in 78 per cent of the 41 species so far studied, including humans (Moller and Thornhill 1998). In male ornaments, symmetry is a sign of good genes and so is size. If the good genes argument is correct then symmetry should be greater in larger sexually-selected characteristics. The sexy son position, however, argues that these characteristics tell the female nothing about the quality of the bearer and so there is no expectation that symmetry (health) should be associated with size (sexiness). Moller (1992) artificially shortened and lengthened the tail feathers of males as well as tinkering with their symmetry. Those with long tail feathers mated sooner and produced more offspring but whether their tail feathers were short or long, symmetrical birds were more reproductively successful than asymmetrical birds. In another ingenious study, Moller (1990) showed that long-tailed fathers produce not only longer-tailed but healthier offspring. He switched some of the nestlings from long- and short-tailed fathers so that half were raised by adoptive rather than biological parents. He infected half of the nests with a parasite. The best predictor of the nestlings’ resistance was the length of their biological father’s tail. So it seems that male ornaments are an honest signal of genetic quality and females are acutely attuned to these differences between males. Females are able to read genetic quality not only from appearance but from the extended phenotype—physical evidence of a male’s quality manifest in the artefacts that he produces. Bowerbird males build elaborate bowers of twigs and ferns decorated with brightly coloured flowers or objects to attract females. A special favourite of the females is a rare bright feather that comes from a bird of paradise. Once a male has found or taken one, he must guard it jealously for his rivals will attempt to steal it. What the female seems to be doing is selecting males who are brave enough to capture and hold the feather against rivals—a useful sign of his genetic quality (Diamond 1991). Among primates, behavioural cues and social relationships seem particularly important in female choice. Primates live in social groups and even though they may not be monogamous, males are likely to be around her for some time. Female choices are geared to ensuring not only genetic quality but a safe environment for herself and her offspring. One general rule of thumb is that females prefer mature males and although females may rather disinterestedly tolerate juveniles’ attempts to mate, they restrict such matings to the beginning or the end of sexual cycles when they are least likely to become pregnant (Small 1993). A second rough rule is that they tend to prefer more dominant males. The advantages of choosing a dominant male lie in part in gene quality. A male who is capable of subduing other males and resisting challenges to his power is one who must be strong, healthy and vigilant. He will pass these qualities on to his sons—but only if they live long enough to reach maturity. Dominant males are in a better position to protect the female and her infants from attack by other group members. Among chimpanzees, chivalry is not

Mhoc06.fm Page 174 Friday, December 14, 2001 10:17 AM


    

the norm—males attack females twice as often as they attack other males (Goodall 1986). Sometimes this occurs in disputes over food, sometimes in sexual encounters (or attempts at them) and sometimes male attacks have a distressingly ‘kicking the dog’ quality whereby the male simply takes out his irritation on a passing female. In olive baboons also, males attack and can severely injure females. Her offspring are vulnerable to being kidnapped by both males and females. A female’s choice of mate may be driven by her need to ensure a defender who will protect the pair (Dunbar 1984; Smuts 1987). Dominant males can offer such protection but not all females choose them as mates. Sometimes they are forced to accept them whether they like it or not because of the sheer brute force of male pressure. The power and ruthlessness of dominant males can be seen in the disturbing phenomenon of infanticide. When a new male dominant enters the group, he may kill existing offspring in order to replace them with his own. At the very moment when we would expect the female to repel his mating attempt most vigorously, she acquiesces (Hrdy 1979). The logic is simple—a male strong enough to take over the group and kill her offspring is a male with whom she would do well to cement a relationship. Whether she does so out of fear, respect or a simple desire to ensure the survival of her next child is not known. But primate females do not always choose the most dominant male. In her search for a safe haven for herself and her offspring, she may develop a special friendship with one or two males. Baboon females demonstrate clear preferences for some males by following them, sitting with them and grooming them (Smuts 1985). Often these friendships may be with a dominant male but not always. In 90 per cent of occasions when a female or her infant were defended from attack, the assistance came from a male friend. The male friend regards these relationships as special—he shares food with her and spends time playing with her infant even if he is not the father, although he rarely goes near the offspring of non-friend females. But there is a downside; male friends can be jealous and attacks on the female are not uncommon if she strays too far away from him. What the male gains from his friendship seems to be a sexual relationship with the female. They are far from monogamous but males copulate with friends twice as often as they do with non-friends and are more likely to ejaculate during these copulations. And females also have a preference for strangers. In one group of chimpanzees, 50 per cent of offsprings were fathered by a non-group male (Wrangham 1997). This preference may be driven by a simple desire for novelty borne out of inbreeding avoidance but it can also be strategic. In this case, the females mated with members of a neighbouring group who were launching attacks on them. Perhaps it was in the females’ best interests to anticipate the takeover with its concomitant infanticide and make a pre-emptive sexual strike. Females can sometimes use their sexuality to influence whether or not outside males are allowed to enter the group. Gelada baboons live in harems of about ten females and one male. If a new male has ambitions to take over the group, he can work by stealth (following the group as a subordinate and

Mhoc06.fm Page 175 Friday, December 14, 2001 10:17 AM

  


developing grooming relationships with the females and infants) or he can launch a full-scale junta. In either case, his success depends critically on the amount of female support and allegiance that he can muster (Dunbar 1984).

Short-term relationships: what women want The lek is a rather improved form of sperm bank. Instead of simply being given an anonymous curriculum vita by which to choose the father, the sage grouse female is able to see him in action. If women could choose fathers for their children in this way, they might be allowed to watch through a one-way mirror as a group of males interacted with one another and then select the one that they liked best. This is a somewhat artificial scenario—women want far more than sperm from a prospective father in most cases and we shall discuss why in a later section. But let’s look at female infidelity—here a woman chooses a man not in terms of his willingness to make a long-term commitment but simply on the basis of his physical and social qualities. In these cases, what do women want and is there any evidence that their choice favours men who are genetically and physically superior? Not surprisingly perhaps, women like attractive men. Attractiveness is an important quality in both sexes but one which for years has been considered to be too subjective (different people have different preferences) and too culturally variable (different countries and ethnic groups vary in their preferences) for serious scientific study. In fact, correlations between different judges’ ratings of people’s attractiveness show near-perfect agreement (Langlois et al. 2000). Even when people judge the attractiveness of people from ethnic groups and cultures different from their own, the agreement remains just as high. We view attractive people more positively than unattractive people (we think they are better at their jobs and at relating to others) and in the main we are right. Attractive people are indeed more successful at work, more popular, have better social skills, enjoy better physical and mental health, and obtain more dating and sexual partners. But what makes a man attractive? Attractive men are taller than the average. In a study of personal ads, 80 per cent of women who mentioned height asked for a man who was six feet or taller (Lynn and Shurgot 1984). They are an inverted V-shape in which the shoulders are wider than their hips (Barber 1995; Lavrakas 1975; Singh 1995a) and the waist is slightly smaller than the hips—a 0.9 waist-to-hip ratio. Male models and mannequins are usually about six foot two with a 42-inch chest and 30-inch waist (Etcoff 1999). Women like men who appear tall and powerful—they prefer those features that are maximally sexually dimorphic and signal a man’s strength relative to other men. As for faces, women seem to seek symmetrical features. Galton was the first psychologist to blend photographic images to form composites through which he attempted to develop prototypical types—in his case he wanted to examine the stereotypical criminal face by fusing photos of robbers, burglars and killers (see

Mhoc06.fm Page 176 Friday, December 14, 2001 10:17 AM


    

Etcoff 1999). The surprise was that the average face turned out to be better looking than any of the specific faces that had been used to compose it. In the last decade psychologists have followed his example and produced digital images of composite faces (Grammar and Thornhill 1994; Langlois and Roggman 1990). And they follow a simple rule, the more faces that are fused, the more attractive the composite becomes and the lower the likelihood that any one of the faces used will be more attractive than the group face. The reason seems to be that averaging has the effect of increasing symmetry and it is symmetry rather than averageness that is attractive. Recall that among animals symmetry was argued to reflect both good genes and strong disease resistance. Low (1988, 1990) developed the ingenious hypothesis that if men lived in an area with many parasites, only the genetically superior males would be able to develop symmetrically and this would cause large differences between men in their attractiveness to women. Gangestad and Buss (1993) confirmed that women in high-parasite cultures rated the importance of physical appearance in a mate as more important than did women in more parasite-free societies. This should mean that some men would be unacceptable as husbands. Instead women should be prepared to accept a polygamous marriage to a healthy male in preference to a monogamous marriage to a less healthy specimen. Examining 185 cultures, Low found that the number of pathogens in the environment was strongly related to the degree of polygamy. Fluctuating asymmetry, as it is called, can be assessed by careful measurement of bilateral body and facial features such as eyes, ears and wrist size. Symmetrical males report fewer minor illnesses and less anxiety and depression than asymmetrical males. They also have lower metabolic rates, are more muscular, taller, heavier, more vigorous, higher in social dominance and have a slightly higher IQ than their peers (Furlow et al. 1997; Gangestad and Thornhill 1997; Manning 1995; Manning et al. 1997). Not only are they healthier but, as we would predict, they are preferred as mates. Symmetrical boys have sex three or four years earlier than their peers and as adults have twice as many sexual partners (Thornhill and Gangestad 1994). They bring their wives to orgasm more often and although symmetry is associated with bigger, taller and more muscled bodies, it was symmetry over and above these factors that predicted orgasm frequency. It was a stronger predictor than the partners ratings of their love, their commitment to the relationship, their frequency of lovemaking or the amount of sexual experience they had (Thornhill et al. 1995). So symmetry matters but so too do the sexually dimorphic quality of the facial features. A good-looking man is not just symmetrical, he is distinctively male in his symmetry. There are characteristic differences between men’s and women’s faces and the features that are most attractive in the male face are those which are distinctively different to those of women. Just as averaging average faces makes them more attractive, averaging a subset of only handsome faces makes them especially attractive (Perrett et al. 1994). And what makes a man’s face handsome are strongly etched cheekbones and chin, together with large eyes and a generous smile (Barber 1995;

Mhoc06.fm Page 177 Friday, December 14, 2001 10:17 AM

  


Cunningham et al. 1990). Mazur and his colleagues (Mazur et al. 1984; Mueller and Mazur 1996) showed women photographs of young men who had attended a military school in 1950. They were asked simply to rate how dominant the faces were. The women showed a very high degree of consensus. Dominant faces were oval or rectangular (not round), had strong brow ridges, deep-set eyes and prominent chins. The women also found these men to be the most handsome. The presence of strong brow ridges, jaws and cheekbones is the hallmark of masculinity and maturity. These features develop during puberty and are dependent on levels of circulating androgen (Tanner 1990). Levels of testosterone can be depleted by stress hormones resulting from a poor environment or from parasite infestation and have the effect of ‘demasculinizing’ these facial features (Thornhill and Gangestad 1993). Handsome men, as we might predict from the good genes theory, are indeed healthier, less vulnerable to disease and had fewer developmental problems in their childhood and adolescence (see Geary 1998). But dominance may be about more than just physical strength, it may also signal future financial prospects. Mazur and his colleagues found that his dominant cadets, in later years, achieved higher military rank and went further in their professional careers. As we have seen, dominance to some extent can be read from the face but these cues are indirect and women in general are less preoccupied with looks than are men (Buss and Schmitt 1993; Townsend and Levy 1990). A better gauge of dominance is by viewing a male in social interaction. Sadalla et al. (1987) experimentally manipulated the social dominance of men and women (through body posture and position as well as active rather than passive behaviour) and found that while dominant women were not rated as especially attractive by men, women’s judgements of attractiveness were strongly affected by social dominance. In another study, they found that characteristics associated with dominance were as important to a woman’s choice of a date or a lover as they were to her choice of a long-term partner (Kenrick et al. 1990). When women are exposed to pictures of males, varying in both dominance and physical attractiveness, and then asked to rate their satisfaction with their current partner they show a strong ‘contrast effect’ of male dominance. They become less content with their present partners although their ratings are unaffected by male attractiveness (Kenrick et al. 1994). In the world outside the laboratory, dominant-looking men were found to have had sex earlier than their peers (Mazur et al. 1994). But dominant faces have their costs too. Although hyperfeminine faces are more attractive to men, hypermasculine faces are less attractive to women (see Etcoff 1999). Exaggerating sex-typical male features does not create the highly attractive ‘supernormal stimulus’ that it does in women’s faces. Men with facial features that are rated as high in strength and dominance are also rated as less warm, cooperative, honest, friendly and good (Perrett et al. 1998). Most significantly, they were rated as likely to be less good parents. It seems that too much dominance makes a man attractive as a short-term mate but raises the spectre that his very good looks may make him an

Mhoc06.fm Page 178 Friday, December 14, 2001 10:17 AM


    

unreliable long-term partner. Dominant males have more sexual partners than their peers (Perusse 1993). Women rate good looks as more important in a short-term sexual relationship (where genes are all they are likely to receive) than in a long-term marriage prospect (Buss and Schmitt 1993). If a man possesses extreme good looks women may be more inclined to engage in short-term affairs with him and this popularity in turn detracts from his willingness to commit to a long-term partner and their progeny. Perhaps that is why some women prefer a husband’s face to be softened by some child-like features such as large eyes and wide smiles (Cunningham et al. 1990).

Biparental care and two-way sexual selection: what men and women want We have discussed the dynamics of the tension between males and females that most often leaves the female ‘holding the baby’ and allows the male to desert and seek new mates. But in a few species, males do not desert but remain with the female and her young providing food and protection. What would make a male take this responsible option when faced with the attractive alternative of independence, new mates and greater sexual opportunities? Maynard-Smith (1977) set up a model to understand this problem, based upon birds because about 80 per cent of the 9000 bird species investigated show biparental care. According to this model, the male should only stay if the number of eggs that survive under biparental care exceed the number of eggs that survive with maternal care only multiplied by the mating success of the deserting male. The temptations of irresponsibly philandering his way to reproductive success are balanced against the prospect of increasing his offsprings’ survival by staying with a single female and their brood. And the weighting of these options depends both on his ability to attract new mates and on a female’s ability to raise the offspring successfully by herself. Among many bird species, the nestlings require brooding which detracts from the time that the female can spend in searching for food so a male who can share the tasks offers a distinct advantage, especially where predators are present, where food demands are high or where food is scarce. Among polygynous birds, the male offers little help to his secondary female and these females have lower reproductive success (Clutton-Brock 1991). Where fertilization occurs internally, female care is usually greater than male care because of the problem of cuckoldry. Consequently, in internally fertilizing species, biparental care is strongly associated with monogamy. Males are prepared to offer assistance when they have a reasonable guarantee that the child that they are aiding is their own. So women may have traded their natural desire for sexual variety for the advantages of monogamy and biparental care. Humans are an unusually slow-developing species and one in which the calorie demands of supporting our huge brains are high. Male assistance in provisioning enhances the probability of child survival especially in adverse environmental conditions, where predation and

Mhoc06.fm Page 179 Friday, December 14, 2001 10:17 AM

  


parasites are a threat and where there is intense competition with conspecifics (Clutton-Brock 1991; Hill and Hurtado 1996; Lancaster 1989). Under biparental care, we have a situation in which what females want from males should be far more sophisticated than good looks that reflect good genes. What they should be evaluating is the resources that a male can offer and the degree to which he is likely to commit these resources exclusively to her. There is strong evidence that evolutionary theory is right. In cultures as widely spaced as the Kipsigis of Kenya (Borgerhoff Mulder 1990) and the Yanomamo of Venezuela (Chagnon 1988), the richest men are preferred as husbands. David Buss (1989) began a now-blossoming line of work on mate preference by surveying 10,000 men and women from 37 cultures. He asked each of them to rate and rank the importance of various traits in the person that they would like to marry. Women rated the importance of good financial prospects higher than did males in all cultures and the importance of ambition and industriousness higher in 34. College students from the United States, Japan and Russia showed the same pattern; women placed a higher value on success, earnings and status than did men (Hatfield and Sprecher 1995). Feingold (1992) performed a meta-analysis of 26 questionnaire studies of mate preference conducted in 82 independent samples. The results strongly confirmed the idea that women place a much higher value on qualities such as socio-economic status and ambition. His analysis of men’s and women’s lonely-hearts advertisements showed the same pattern—women more than men stipulate high income or status in their prospective partners. In a representative sample of Americans, irrespective of age or race, women wanted husbands who were better educated and who earned more money than they did (Sprecher et al. 1994). This preference for resource-rich men pays off. Compared to women who select young or less educated men as husbands, women who marry older and better educated men have more children, are less likely to get divorced and report greater marital satisfaction (Bereczkei and Csanaky 1996). But money alone is not enough. Women also seek men who show signs of warmth and character that are associated with making a long-term commitment to her and her children. Buss found that women valued a man who had the combination of being kind, understanding and intelligent higher than a man who was financially successful but who had none of these traits. Feingold’s meta-analysis revealed that women place a higher value then men do on the character of their mate. Dunbar (1995) analysed personal ads placed by men and women. He found that 45 per cent of women (compared to 22 per cent of men) stipulated traits that signalled commitment and parenting qualities such as loving, warm, good sense of humour, family-minded, gentle and dependable. Men who are prepared to invest heavily in their children are likely to try to impress women with their resources but they are also more likely to offer her help, give her things she likes, be sympathetic to her troubles and emphasize their fidelity (Cashdan 1993). If women are seeking men who demonstrate commitment, then we would expect men to engage in more deception than women about

Mhoc06.fm Page 180 Friday, December 14, 2001 10:17 AM


    

these qualities in an attempt to sell themselves to the opposite sex. Men do indeed feign not only the possession of dominance and resources but also they exaggerate their sincerity, trust and kindness. They admit to saying things to ‘butter up’ the opposite sex and pretending to be more humble, polite, trusting and considerate than they really are. They also feign disinterest in immediate sex and pretend to be interested in starting a relationship when they are not (Tooke and Camire 1991). So women on the lookout for marriage pay attention to status, resources, ambition and commitment. But they may not always get it. If a man expects only a sexual onenight stand, he will drop his higher marital requirements for intelligence, friendliness, kindness, excitingness, health, easygoing temperament, creativity, emotional stability, sense of humour, educational attainment and social status (Kenrick et al. 1990). But under monogamy it pays a man as well as a woman to be choosy in their choice of long-term partner. A universal feature of men’s preferences is age. After a period of adolescent sterility, women become most fertile at the age of 25 from whence their fertility declines until it reaches zero by the age of about 45 to 50. Thus when adult men are asked about age preference they consistently choose someone who is younger than themselves and marriage patterns indicate that the typical age gap is about three years (Buss and Schmitt 1993; Kenrick and Keefe 1992). But this three-year gap is not a constant; as men age, they prefer younger and younger women. By the age of 60, they prefer women who are on average 15 years younger than themselves. Teenage boys, by contrast, rate a woman five years older than them to be the perfect partner (Kenrick et al. 1996). The data suggest that what men are trying to do is to capture the most fertile women that they can, regardless of their own age, and for men who marry at the normal time that means a woman about three years younger than themselves. Men place a greater premium upon physical attractiveness than women (Buss and Schmitt 1993; Feingold 1990) and, for men, physical attractiveness is closely bound up with age. Facial features that reflect youth include shiny hair, unwrinkled skin, large eyes, small nose and full lips. (Etcoff 1999). Like women, men show a preference for facial symmetry. But a very strong factor in men’s preferences is not just faces but bodies. Women with symmetrical breasts are rated as more attractive (Singh 1995b), and however whimsical such a preference may seem, asymmetric breasts are associated with lower fecundity (Manning et al. 1997; Moller et al. 1995). Men’s favourite body shape is a small waist in relation to the size of the hips—a 0.70 waist-to-hip ratio (Singh 1993). Although men’s preference for fatness or thinness varies over cultures, this particular ratio remains constant. The significance of this preference remains unclear although there are a number of candidates. At puberty, a girl’s body is transformed from its androgynous shape to a curvier silhouette by the appearance of a waist caused by the enlargement of the breasts and hips. After 30, however, her waist begins to thicken. When a young woman puts on weight, it accumulates on the breasts and buttocks

Mhoc06.fm Page 181 Friday, December 14, 2001 10:17 AM

    while in men, young girls and post-menopausal women it is deposited on the stomach and limbs. This gynoid distribution of fat is necessary for the hormonal changes associated with female fertility (anorexic girls not only find themselves with a low waist-to-hip ratio but they also cease to menstruate). According to Singh, men select small-waisted women as a means of ensuring fertility. Or perhaps narrow waists serve to dupe the male into believing that a woman has greater reproductive potential than she really has (Low 1979). The male might assume that large breasts signal a greater milk supply and that large hips diminish the likelihood of birth complications. (In fact he would be wrong since breast size is uncorrelated with milk supply and it is pelvic size, not hip size, that ensures an easy birth. Notwithstanding, males’ preference is likely to be as vulnerable to runaway selection pressure as is females’.) But a third answer points in a different direction and one which is also of special concern to men. Ridley (1993) believes that a thickening waist is the first sign of pregnancy and that males may avoid women who seem to be carrying another man’s child. In all of these accounts, waists are a clue—honest or dishonest, direct or indirect—to current reproductive capability. Despite men’s greater emphasis upon youth, looks and fidelity, it is important to bear in mind that men and women are perhaps more similar than different in what they are seeking in a long-term partner. In comparing the minimum acceptable standards in a marital partner between men and women, there were no sex differences for a number of qualities, including intelligence, aggressiveness, sexiness, exciting personality, creativity, friendliness, sense of humour, easy going, healthy, religiosity, desire for children, kindness and being understanding (Kenrick et al. 1990). Both sexes in all of the 37 cultures surveyed by Buss (1989) ranked kind, understanding and intelligent higher than earning power and physical attractiveness. Both men and women seek mates who are easy to live with. Despite the conflicting reproductive interests of men and women, they are nonetheless locked together under monogamy and both sexes try to maximize the day-to-day cooperation that such an arrangement requires. Under monogamy, then, both sexes are selective about their partners but each sex wants the same set of sex-specific attributes. Inevitably, some are bound to be disappointed. But what you receive is inevitably connected to what you can offer. In the mating market-place, buyers and sellers find their own level. Women with youth and good looks on their side can select the males with the highest resources, just as rich men have their pick of attractive female partners. The less blessed must eventually settle for the best that they can get. This is the phenomenon of assortative mating.

Flexible mating strategies Recall the familiar mantra of evolutionary theory; the reproductive success of men depends upon the number of partners that they can inseminate and the reproductive

Mhoc06.fm Page 182 Friday, December 14, 2001 10:17 AM

      success of women depends upon their ability to secure the resources that they need to ensure the survival of their children. Under monogamy and biparental care, this predicts that women should generally favour long-term relationships and men should favour short-term arrangements. There are many data that support this general viewpoint. Men’s increased sex drive, their preference for a larger number of sexual partners, their willingness to drop their standards in order to procure short-term mates all testify to the men’s polygynous urges. Unlike our primate sisters, women do not generally fend for themselves and fare better when they receive assistance in child-rearing. This assistance is usually and chiefly economic, with men being far less involved in day-to-day care of children than are women. In the environment of evolutionary adaptation, women depended upon men to provide high-quality meat which augmented the calories that they themselves were able to provide by foraging. In more extreme northern environments, men’s fishing may have been the chief means of sustenance and one which precluded participation by lactating women or mothers of young children (Miller 1994). In our own society, the inflexible labour arrangements imposed by agriculture, industry and service economies make it difficult for a woman to simultaneously rear children and engage in full-time labour. So, it is argued, women seek resources by attracting a male who is able and willing to provide long-term investment. Data showing that women are more choosy in their selection of partners, prefer to wait longer before sex and value resource and parenting qualities in men all support this proposition. But there are some problems. One is logical—if women engage only in long-term relationships, where are promiscuous men finding their mates? Another is a theoretical problem thrown up by the evidence of our everyday experience—why do women ever acquiesce to short-term relationships which do not carry the benefits of resource provision and why do so many men marry and remain faithful? In recent years, evolutionary theory has begun to grapple with the problem of intrasexual variability. Women differ among themselves (and at different points in their own lifetime) in the sexual strategies that they favour. The same is true for males. The solutions that have been offered depend on the concept of conditional strategies. A conditional strategy has a number of properties (Gross 1996). First, there are several behavioural tactics available from which one is chosen. Secondly, the choice that is made is responsive to specific cues in the environment. Thirdly, all individuals are genetically monomorphic—they are all genetically able to enact the full range of strategies. Fourthly, when these tactics evolved, they were not equal in their average adaptive value. Fifthly, as these tactics evolved the selected tactic tended to have a higher fitness value for the individual, given the current environmental conditions, than other tactics. So human beings are equipped with multiple strategies from which they choose during their lives. Some may favour a single tactic throughout and genetic factors may influence this behavioural preference. But strategies may also reflect environmental cues or life-history variables which are either constant (in the case of consistency) or shifting (in the case of flexibility). A number

Mhoc06.fm Page 183 Friday, December 14, 2001 10:17 AM

  


of proposals have been made to explain these variable mating strategies. Although they are not incompatible they have different emphases—some depend on tracking the behaviour of other individuals and others upon tracking environmental and ecological cues. Dawkins (1989) has proposed a model of conditional mating strategies that rests on the choices that other women and men in the population are making. And it is here that we meet the idea of frequency-dependent selection. Frequency dependence means simply this: the adaptive value of a strategy depends upon the proportion of the population also using it. The value of one strategy compared to another decreases as its relative frequency increases. A rare strategy might be successful but if the majority of the population adopt it, it becomes increasingly less successful. An oftquoted example is mimicry in moths. If a handful of moths are able to mimic a species which its predators find unpalatable, they will successfully evade predation and their numbers will increase because it is not worth the predators’ time and trouble to distinguish mimics from the truly unpalatable moths. Eventually, however, almost all moths will be mimics and the predator will have to eat something in order to survive. The mimicry strategy becomes increasingly less useful. At some point a stable equilibrium will be reached. Dawkins proposes that short- and long-term strategies are frequency dependent and therefore reflect the strategies that predominate among both males and females. He weighs up mathematically the pay-off of a successfully raised offspring against the twin costs of wasting time in prolonged courtship and raising the child. In a monogamous population, men and women do equally well—they halve the cost of child-rearing by sharing it between them and both pay the price of courtship and reap the same reward. If an unrestricted female comes on the scene and all the men are monogamous, she does well at first. She avoids the price of courtship but still halves the cost of child-rearing and gains the reward. This strategy is more successful than monogamy and it will start to be adopted by other women. But as unrestricted females become more common, the male’s optimal strategy changes from monogamy to providing nothing more than genes. Such a strategy earns the man the reward of a successfully raised child but he pays none of the costs of either waiting for sex or child-rearing. But now the unrestricted female pays the entire cost of child-rearing alone (although not the costs of delayed mating) leaving her in serious deficit. A restricted female now starts to do better. Admittedly she will at first encounter only bounders but if this happens she at least does not lose—she will refuse to mate and so will not suffer the costly experiences of her unrestricted rival. As the number of restricted females rises, uncommitting males will find it increasingly hard to find partners and their pay-off will be zero. Finally they will have to pay the price of parental care and the society will veer back to monogamy. The problem for this model is in accounting for the swing back from promiscuity to monogamy. When the majority of both men and women are promiscuous, Dawkins argues, the restricted female does better than her less choosy sister. His

Mhoc06.fm Page 184 Friday, December 14, 2001 10:17 AM

      argument rests on the premise that there is a cost associated with a woman raising a child alone, whereas the restricted female pays no such cost. True, but the restricted female also produces no offspring—at least if there are almost no monogamous males in the population. This can hardly be considered a good outcome from the point of view of inclusive fitness. In reality, conditional strategies do not allow one option to become completely dominant (if they did the less popular strategy would be selected out of the behavioural repertoire eventually) so there are always likely to be some monogamous males about. But nonetheless she will spend a great deal of time meeting promiscuous males and rejecting them and a woman’s reproductive life is a relatively short one. Each male she rejects costs her valuable time. Dawkins glosses over this difficulty and moves on to the assumption that eventually monogamous females will predominate, finessing the question of how this could occur. Once monogamous females are in the majority, he argues, males will waste too much time looking for an unrestricted female and so will be forced into monogamy. But note that while he admits the time-wasting cost for males, in the previous step of the argument he ignores the same cost for monogamous females searching for a faithful male—despite the fact that waiting costs are greater for females because menopause looms ahead, truncating their reproductive future. In a subsequent account of flexible strategies—the strategic pluralism argument— a woman’s choice between good genes and good parenting occupies a central position (Gangestad and Simpson 2000). Good genes are reflected in good looks which in turn depend upon fluctuating asymmetry which is heritable. As we have seen, good looks in men are rated as more important when a woman is considering a short-term relationship than a long one (Buss and Schmitt 1993) and looks are also important when women contemplate infidelity. In other species too, more symmetrical males are preferred as extra-pair mates. Although barn swallows offer biparental care and appear monogamous, approximately 35 per cent of nestlings are sired by another male. These males are more symmetrical than the female’s regular mate and presumably more physically attractive since they do not compensate her in any other way than through their genes. More symmetrical avian males provide less in the way of material benefits to their mates—attractive males perform a smaller proportion of feedings than do unattractive males. In humans also, symmetrical males have a greater number of extramarital partners and they are more often chosen as extra-pair partners by women who are engaged in an ongoing relationship (Gangestad and Thornhill 1997). Ovulating women prefer the smell of a T-shirt worn by a symmetrical male over that of an asymmetric male (Gangestad and Thornhill 1998; Thornhill and Gangestad 1999). Women experience not only more orgasms but more high-retention orgasms (those that occur immediately prior to ejaculation and have a greater chance of resulting in pregnancy) with symmetrical males (Moller et al. 1999; Thornhill et al. 1995). And these highly attractive symmetrical males, like their avian peers, invest less in their relationships than other men. In a study of dating couples, Gangestad and Thornhill (1999) found that attractive men were less honest with their

Mhoc06.fm Page 185 Friday, December 14, 2001 10:17 AM

  


partners, spent less time with them and sexualized other women more. So here we have it—the empirical distinction between Cads and Dads. Cads offer their genes as the prize while Dads offer long-term resources, security and commitment. Women then have a choice to make. At any given point in time, they can opt for one or the other (or, if they are deceitful, both). Women who opt for short-term relationships can garner the best genes at the cost of paternal care; women who opt for long-term relationships may have to sacrifice some genetic quality in order to secure resources and commitment. Women were given a questionnaire to measure their willingness to engage in short-term sexual relationships (Gangestad et al. 1999). ‘Unrestricted’ women find symmetrical men more attractive particularly as short-term mates. They also care more about a man’s appearance and, when forced to choose between dating an attractive man who is not very loyal versus a loyal man who is averagely attractive, unrestricted women choose the former while restricted women choose the latter (Simpson and Gangestad 1992). But sexual strategies are variable—today’s unrestricted woman may be tomorrow’s long-term strategist. Much depends upon the circumstances in which she finds herself. Where the environment is harsh, unpredictable and dangerous, a woman may do best to find a long-term mate. But where the chief danger to children arises from parasites and pathogens, she would do better to seek evidence of genetic quality in the father. Over 29 countries, men and women in regions containing more pathogens rated the value of physical attractiveness higher. And the greater the pathogen prevalence, the less importance was placed upon mate dependability, character, disposition, emotional stability and desire for home and children (Gangestad and Buss 1993). Where environmental uncertainty is the chief threat to women’s reproductive success, they should look for evidence of male resources and commitment. And here we meet an ongoing and heated debate. What about women’s own access to resources? If a woman is able to ride out environmental demands through her own independently accrued resources, it is argued that her need for a provisioning male will be diminished (Gowaty 1992). Instead she might look for good genes (short-term strategy) rather than a good provider (long-term strategy). A number of studies show, however, that women in high-paying jobs tend to value male resources as much or even more than women in low-paying jobs (Buss 1989; Townsend 1989; Wiederman and Allgeier 1992). This has been interpreted as evidence that women’s desire for resource-rich males was established so far back in evolutionary time that it is maintained today. In short, this position supposes that women either do not use conditional strategies or, at least, do not use with them with regard to resources. On the other hand, in a cross-cultural study, Eagly and Wood (1999) report that as women become more ‘empowered’ (as measured by nationwide statistics on women’s earnings, representation in government and their participation in professional occupations), they place less importance on the value of a prospective partner’s earnings. The jury, as they say, is still out on

Mhoc06.fm Page 186 Friday, December 14, 2001 10:17 AM


    

the question of whether women’s interest in male resources is conditional or constant. Unlike Dawkins’s model, the driving force behind conditional strategies is not what others are doing but what the environment signals to women. It is women who drive the system and their preference is a function of ecological threats (parasites or a harsh environment). Men follow women’s preferences while capitalizing on their own assets. When parasite infestation is a problem, attractive men are selected and do well. When resources are scarce, less attractive but faithful men do much better. So the advantages of both strategies are maintained over time as the threats to infant survival alter. By focusing upon the different qualities that men are able to offer, Gangestad and Simpson’s model does a better job than Dawkins’s of explaining why monogamy should be attractive to the majority of males (they lack the drop-deadgorgeous quality that would enable them to pursue a short-term strategy). And in emphasizing women’s choices as dependent on the environment rather than on male supply, they simplify Dawkins’s recursive model into one where ecological forces take the foreground. But the model makes the assumption that pathogens and environmental difficulties rise and fall independent of one another. Even speculating on ancestral evolutionary environments, such a supposition seems doubtful. In the harsh subzero climate of the Arctic, pathogen prevalence is likely to have been low but at the same time food resources would have been scarce. In warm African climates, food would have been more abundant but so also would have been parasites. And in today’s environment, poverty is associated equally with resource scarcity and poor health. The rise in single-parent families that we have seen in the West in postwar years, according to their model, would be a function of increased pathogen load—at the very time when immunization is common, prenatal care improved and health care more widely available. And the model makes two related and critical omissions—it neglects the fact that raising a child alone doubles the cost the mother must pay in child-rearing and the fact that infant survivorship decreases without paternal assistance (Blurton-Jones et al. 1997; Hill and Hurtado 1996; McLoyd 1989). Another account of conditional mating strategy depends upon the number of men and women in the population because this has an effect upon which sex achieves their preferred reproductive modus operandi (Guttentag and Secord 1983; Pedersen 1991). Imagine a community of men and women of marriageable age. Now suppose that an imbalance begins so that there are too many women relative to men. A skewed operational sex ratio includes not just the loss of men but the specific loss of marriageable men. This can occur when the birth rate rises. Because women want to marry men who are older than they are, they will find fewer eligible men around. But men have no such problem since they are seeking their mates from the expanding pool of younger women. It can also result from a high proportion of males who are drug addicts, alcoholics, unemployed and long-term sick. Although available, they are unlikely to be chosen as marriage partners since they represent a

Mhoc06.fm Page 187 Friday, December 14, 2001 10:17 AM

  


drain on the women’s assets rather than an economic advantage. Too many women means that men are in a good market position to enforce their preferred reproductive strategy and that strategy is likely to be one that minimizes paternal investment and increases maternal investment (Guttentag and Secord 1983). Just such a shift in the sex ratio occurred in the United States between 1965 and the 1970s and, as predicted, divorce rates, illegitimate births and single-parent families increased, more permissive sexual attitudes predominated and the number of women in the workplace rose. When there is an oversupply of men, there is reversal in these trends with an increased commitment of men to marriage and lower divorce rates. In this scenario, women’s preference for a long-term partner is taken as given (it halves the cost of child-rearing) as is men’s preference for a cad strategy (which maximizes reproductive success at zero cost in child care). Women are forced into an unrestricted strategy by the sheer paucity of males. All of these arguments suggest that women track the environment to ascertain parasite load, ecological difficulties, the proportion of cads and dads or the number of marriageable males. By highlighting the flexibility of women’s choices, these theories pay less attention to the possibility that environmental factors can also produce consistency in a woman’s choice of mating strategies. Draper and Harpending (1982) have offered an explanation of how such a preference might be set up early in their development and how this preference might be influenced by indirect cues to the proportion of provisioning males. They argued that the absence of a consistent provisioning father in the home signals to the girl that paternal investment is rare and that she is unlikely to secure a stable pair-bond. In these circumstances, her best strategy might be to seek a series of short-term mates and to undertake child-rearing alone: ‘a young women who waits for the right man to help rear her children may lose valuable reproductive opportunities at a time when her health and physical capability are at their peak’ (Belsky et al. 1991, p. 653). This idea has been given a more clearly psychological treatment by Belsky et al. (1991). There are steps in the developmental sequence leading to a short-term mating preference. The girl’s home is marked by father-absence, marital discord or inadequate resources which create a climate of stress. This stress manifests itself in harsh, rejecting or inconsistent parenting practices and results in the child forming an insecure attachment to the parent and developing an internal representation of relationships characterized by mistrust. In girls, this is manifested in anxiety and depression (internalizing symptoms) while in boys it is more frequently externalized as aggression and non-compliance. In both sexes, these psychological states induce early physical sexual maturity. Subsequently, sexual relationships tend to mistrustful, opportunistic and short term. This model has much data to support it. Resource shortage and stress have both been linked to poor parenting practices (Belsky 1984; McLoyd 1990; Patterson 1986) as has marital discord (Jouriles et al. 1988). Maltreated children are at high risk of developing insecure attachments not only to their parents but also to peers

Mhoc06.fm Page 188 Friday, December 14, 2001 10:17 AM


    

(Howes and Eldredge 1985; Lyons-Ruth et al. 1987). Family conflict and father absence (Moffitt et al. 1990; Surbey 1990) are both associated with early menarche in girls because neuroendocrine functions that precipitate menarche are responsive to early childhood events. Boys and girls who mature early initiate intercourse at an earlier age than their peers (Smith et al. 1985). Furthermore, girls from never-married or divorced mothers are also more likely to be sexually active at an earlier age (Jessor et al. 1983; Newcomer and Udry 1984). Compared to girls from intact families they have more sexual partners, although girls who lose a father from death do not (Demo and Acock 1988; Hetherington 1972). Dating, sexual activity and marital stability are all associated with exposure to divorce in childhood, and children of divorced parents are themselves at greater risk for divorce (Keith and Finlay 1988). Nonetheless we cannot rule out a simpler genetic effect—mothers who engage in short-term relationships give their genes to daughters who share a similar preference. Comparing these models, we see that in most of them selection of a short-term sexual strategy by a woman is a less preferred option resulting from some curtailment of her choice. Women select short-term sexual relationships when there are two few men or when her upbringing has signalled that men are unreliable investors in their progeny. For Dawkins, short-term sexual relationships are an unalloyed disadvantage to females, for while they may save on courting delays, they more than pay for it by the massive burden of raising the child unassisted. Only in the Gangestad and Simpson model do women spontaneously select such a route—in response to pathogen prevalence—but they present no data to show whether this strategy results in greater child survival nor whether it compensates for the massive cost of raising a child alone and the diminished survival rate of children that is associated with the absence of a supporting male partner. First and foremost women seek resources. For this, they often forgo the best genes in favour of the best survival rates for their children. Short-term relationships for women often amount to serial monogamy in response to a population of males who cannot or will not provide sustained economic and emotional commitment. And if she can maintain her attractiveness to a series of males in the face of her increasing age, decreasing looks and the handicap (from a prospective partner’s viewpoint) of already-born children, she can also gain the advantage of genetic diversity and perhaps better genetic quality in her children. But the most secure and stable route is to attract a male who will commit. Unfortunately that idea has occurred to other women also and she is in a competitive market-place.

Looking good Although systematic studies of men’s mate preferences may be relatively recent, it seems that women knew all along what men liked. A youthful body and face are exactly what women compete with each other to achieve. In the United States 88 per cent of American women aged over 18 wear make-up and twice as much is

Mhoc06.fm Page 189 Friday, December 14, 2001 10:17 AM

  


spent on personal grooming products as on books. Women’s fascination with enhancing their appearance is far from new as Nancy Etcoff describes: Women conceal, bleach and blush. They have applied poisonous lead and mercury to their skin mixed in with egg whites, lemon juice, milk and vinegar. They have attached leeches to themselves and swallowed arsenic wafers. To mimic skin translucency the Greeks and Romans and, later, Queen Elizabeth I painted the veins on their breasts and their forehead blue. For two thousand years European face makeup was made from white lead, which was combined with chalk or used in a paste with vinegar and egg whites and applied thickly to completely mask the skin’s surface and colour . . . Over this pale canvas, women apply exclamation points of red to their lips and their cheeks. Red, the colour of blood, of blushes and flushes, of nipples, lips and genitals awash with sexual excitement, is visible from afar and emotionally arousing. (Etcoff 1999, p. 101)

Make-up is in short designed to mimic youth, correct asymmetries and signal sexuality. And it is used almost exclusively by women and not by men. If ever there was a prima facie case for the idea of two-way sexual selection, here it is. The cost, time and energy women devote to enhancing precisely those facial features that attract men makes it abundantly clear that women are far from passive in the mating game. Although cosmetics can help to conceal the ravages of ageing, they are fighting a powerful enemy. In youth, skin cells are replaced every two weeks but, as we age, the cycle slows down. The cells on the surface of the skin remain there longer and begin to look dull and tired. Natural oil production decelerates and collagen and elastin, which give the skin elasticity, break down, especially after menopause. Pockets of fat beneath the skin’s surface become thinner. The overall effect is a face that is wrinkled, sagging and bony. Some turn to a more radical approach—surgery. Of all cosmetic surgical procedures in the United States, 89 per cent are performed on women—85 per cent of eyelid surgery and 91 per cent of face lifts (Etcoff 1999). Surgery is also used to make women’s bodies conform to male preferences. Every year in the United States approximately 125 000 breast implant operations are performed. As well as the large-breasted, men also prefer slim women. Comparing women who married upward in the class structure with women who stayed demographically where they started, researchers have found a significant difference in weight with hypergamously married women being thinner. Given a choice, men prefer slim women but they do not favour the skinny (Singh 1993). But this fact has been lost on the 5 per cent of young women who suffer from bulimia and anorexia. The certainty that, despite their protruding ribs and sunken cheeks, they are overweight and their obsession with food has become an end in itself—a way of life. No longer connected to attracting mates—what use are they anyway in a hospital bed?—their plight is the reductio ad absurdum of the cultural ideal of thinness. Eating disorders are nine times more common among women than among men. Men and women in a competitive market-place do whatever they can to increase their attractiveness to the opposite sex. Tooke and Camire (1991) found that

Mhoc06.fm Page 190 Friday, December 14, 2001 10:17 AM

      women, far more than men, are likely to deceive the opposite sex by alterations to their appearance. They judged tactics such as wearing make-up, painting their nails, getting fake tans and wearing tight clothing to be the most effective in securing male attention. In a similar study, Walters and Crawford (1994) asked their undergraduate subjects about tactics that they used in order to compete with members of their own sex. No explicit mention was made of competition for mates but nonetheless women most often nominated, performed and rated as effective the tactic of attracting attention to their appearance. (Men more often competed with one another in terms of taking risks in sports, demonstrating resources and engaging in drinking contests!) Attractiveness seems to be the currency of female competition even when no mention is made of what the competition is about. Using diaries to examine the everyday experience of competition, in which subjects were asked to record their thoughts and actions, Cashdan (1998) also found that while men competed with other men in the arena of sports, women competed with one another in the currency of looking attractive. Whenever they can, women take an indirect approach to competition—through the manipulation of their faces and bodies to attract and retain men. David Buss (1988) asked men and women what acts they used to keep the attention of the opposite sex. Predictably he found that men made efforts to be considerate and to impress their partner with their money while women showed that they cared about their partner and made an effort to improve their appearance. He also examined the tactics that were least used and rated as least effective. Men were more likely than women to pick fights with rival suitors, while for women these directly aggressive acts (‘I hit the girl who made a pass at him’, ‘I picked a fight with the girl who was interested in him’) were the least frequently used and were judged as the least likely to be effective. But the subjects of this study were relatively privileged undergraduate students. Playing nice is nice—if you can afford it. We turn now to more extreme tactics that women may be forced to use when times are hard and good men are scarce.

When push comes to shove Women come to blows far less often than men do because the risks are usually too great. Instead women prefer to restrict their competition to less lethal contests for sexier faces, clothes and figures. But the confluence of some important factors can create a recipe for female–female assault. Let us examine the ingredients. First, age. Young women fight more than older ones. And if my thesis—that much female aggression is about mate competition—is correct, it is a tribute to the strength of their inhibitory tendencies that women are not much more violent than they are. Here’s why. Everywhere, young men between the ages of about 20 and 24 fight much more often than more mature men (Gottfredson and Hirschi 1990; Kruttschnitt 1994). Daly and Wilson (1988) have offered a persuasive reason why

Mhoc06.fm Page 191 Friday, December 14, 2001 10:17 AM

  


this should be. At this age men are seeking mates and, especially for those who lack material resources to signal status, physical dominance is an important attribute. It causes other men to back-off and their fearful respect provides evidence to women of the man’s dominance. Adolescent bellicosity testifies to the importance of mate choice at this critical age when young men enter the sexual arena. Traditional theories of mate selection followed the axiom that males compete and females choose (Andersson 1994; Bateman 1949; Trivers 1972) and this one-way form of selection may be adequate to describe the behaviour of species where males invest little more than sperm. But if we look at human data we find something that such a position would not predict. Plotting age against assault rate we find a remarkably similar age–assault curve for both sexes. In fact the correlation between them is 0.89 for US data and 0.98 for British data (Campbell et al., 2001). The female rate is only a fraction of that of males—women commit about 15 per cent of criminally violent acts—but nonetheless there is a clear correspondence between the age at which violence is most likely in the two sexes. The obvious explanation for this astonishingly high correlation is that aggressive forms of sexual selection are a two-way street, although males engage in such risky strategies far more frequently than women. Yet if we take the idea of two-way selection seriously, what is surprising is the astonishingly low rate shown by women. Men are fighting, though not consciously, for the status that can buy them reproductive opportunities. Yet recall that men become increasingly attractive to women as they gain more resources and this is likely to be a function of increasing age. Consider too that women are far less choosy about their partner’s physical appearance (which declines with age) than men are. Also bear in mind that a man’s ability to father children extends throughout his adult life span because, unlike a woman, his career is not cut short by menopause. (In 2000 the novelist Saul Bellow fathered a child at the age of 82.) Taken together, age should have less impact on a man’s mating opportunities than on a woman’s and it should take some of the urgency out of his sexual competition. But for women’s chances, age is of the essence. Between the ages of 15 and 25 a woman reaches her maximum attractiveness to men. Never again will she have such a good chance to acquire the best of the available men. Men’s preference for youth places serious constraints on the mate opportunities that are open to her after her ‘best’ years—a younger, more attractive cohort of women is already snapping at her heels. And from a purely biological viewpoint, a woman of 20 is looking at a remaining 25 years of reproductive possibility compared to about 50 years in a comparably aged man. Age matters—but more so to women than to men. Hence it is all the more surprising that female competition, however extreme it may be in the area of achieving physical perfection, rarely reaches the boiling point of violent confrontation. But if a woman is ever to attack another woman it is more likely to happen in her teens than at any other age. Girls mature earlier than boys and menarche occurs relatively late in the sequence of developmental changes—approximately 18 months after the growth spurt has

Mhoc06.fm Page 192 Friday, December 14, 2001 10:17 AM

      reached its peak. Menstrual periods occur without ovulation for about 12 to 18 months. Thus from the time that a girl begins to show signs of sexual development, she has as long as three years during which time she is unlikely to conceive. This provides her with a safe period of sexual experimentation as she trawls the market for a long-term mate. It also means that we would expect girls to show not only an earlier interest in establishing relations with the opposite sex (they do) but also a slightly earlier involvement in competitive aggression. As we would expect, girls’ involvement in assault peaks about two years earlier than that of boys whether we use official criminal statistics (Kruttschnitt 1994; Piper 1983) or self-reported involvement in aggression (Elliott et al. 1983). The next ingredient is the degree of competition that a woman faces. In societies where the number of women exceeds that of men, competition becomes more intense. Of particular importance is the local sex ratio since we know that a woman is likely to marry a man from her own geographical area who matches her in social class, ethnicity, education and intelligence (Hill et al. 1976; Tharp 1963). It is within this pool of available men that the woman must compete. In poverty-level, inner-city areas, male mortality is high and decreases the available pool of males. To make matters worse, those men that are available may not constitute acceptable partners. Incarceration, drug addiction, alcoholism, mental illness, homelessness and unemployment effectively remove a proportion of males from the realistic marriage market. Though available, such men are likely to constitute a drain on the woman’s own meagre resources rather than providing any additional ones. In these areas, there is marked variation in men’s ability to support a wife and children. With the exodus of upwardly mobile and well-educated men from these areas, those that remain are often forced to take irregular, menial or low-paid work. The few wealthy men often make their money from illegal enterprises and such men are highly valued by women. Carl Taylor’s (1993, p. 130) interviews with young black women made clear that the source of a man’s income was less important than its size and his willingness to share it. Dope guys is straight if they think you ain’t dissing them. They got coin and they will spend on you, and that’s better than getting messed over for nothing. At least dope guys will buy you dinner a some place besides Mickey Dee’s . . . I date whoever is treating your girl the right way. Me, if a guy got some paper well, it’s okay with me. I like fellas that’s rolling, least they making it.

The marked variation between high rollers and losers (compared to the smaller variation among middle-class males) also increases competition between women. The difference between marrying a doctor or an accountant is less extreme (and much less worth risking violence for) than the difference between a successful drug dealer and a drug addict. As the disparities in male wealth increase, the tension between women increases also. Added to this is a further dynamic that feeds off the paucity of good men. The fewer of them there are, the stronger is their ability to pursue their preferred mating

Mhoc06.fm Page 193 Friday, December 14, 2001 10:17 AM

  


strategy. Men’s greater interest in uncommitted sex and short-term relationships can flourish since women have little leverage against it. ‘If you won’t give it to me, someone else will’ is a coercion that is all too true and that drives women to accept temporary sexual relationships that, from their point of view, are far from ideal. In underclass communities, men are able to get their way and the women fight, not only for a man who will be faithful and committed but even for a temporary liaison with a man who can, at least in the short-term, provide material resources. These women understand the equation of sex for money, the rarity of long-term relationships and the dynamic that drives male choice. I know that’s why some girls ration out the sex, that’s all you got and the dogs be sniffing and you got to let ’em smell it (laughing); they want them boots and you got to keep ’em as long as you can. When men get what they want then you never get what you want if you didn’t play it right . . . I tell her take all his paper, all of it, ‘cause it’s just a matter of time and he’s gonna do some rotten dog shit on her . . . Got to get it when you can. You never know when it’s gonna stop and you better get much as you can while you can . . . When fellas get tired of your pussy, it’s good-bye girl, naw, it’s get the fuck out of my life bitch! Next bitch! (Taylor 1993, p. 131)

A woman’s attractiveness is another part of the violence equation and it is a two-edged sword. While pretty girls do not need to fight to secure the attention of men, they are also likely to be the target of other girls’ jealousy. A British girl describes the dangerous effects of her friend’s attractiveness to men: They were jealous of Beverley because she could get quite a few boyfriends and half of them (the other girls) were too fat or too skinny. They were following us and we got up and started walking off and they was calling out names and I turned round and said ‘Just because you are too fucking fat, Smith’. They come over from the pub and I was walking ahead of Bev and next minute I heard one of them slap Bev round the face. There was a whole gang of them onto her. (Marsh and Paton 1984, vol. 1, p. 36)

Precocious sexual maturity can also make a girl that target of other girls’ resentment. Girls who reach menarche early are popular with boys yet are rejected by other girls (Faust 1963). A number of sociological and psychological explanations have been offered including asynchronicity as a source of awkwardness among adolescents who prize conformity (Jones and Mussen 1958), a girl’s early curvaceousness being too threatening in a sexually repressive society, and the association of early menarche with shortness and chubbiness which are devalued as body types (Tobin-Richards et al. 1983). Perhaps there is a more direct explanation for both boys’ attraction and girls’ hostility— early-maturing girls have a distinct advantage in terms of mate selection. They garner a disproportionate amount of male interest and by entering the mate market earlier, have a wider choice of prospective men. Not only that but, by starting early, they can look forward to a longer reproductive career than their developmentally more leisurely peers.

Mhoc06.fm Page 194 Friday, December 14, 2001 10:17 AM

      Aggression like any other strategy comes more easily to some than to others simply because they are familiar at it. And aggression in girls as in boys shows considerable stability over time (Cairns and Cairns 1984; Cairns et al. 1989; Pitkanen-Pulkkinen 1981; Pulkkinen 1987). A tendency to aggression has a strong component of heritability (Rushton et al. 1986) but its expression depends upon circumstances. The backgrounds of violent offenders, both male and female, show a consistent thread (see Moffitt 1993). They come from families which are turbulent and unstable. The parents as well as the children use aggression as a coping strategy. Abuse and slaps displace reasoning and debate when conflict erupts. The early years of the child set the parameters of what they can expect from the environment around them and what they learn is that the pre-emptive strike wins out over withdrawal or tolerance. The world is a competitive place in which access to money, food and clothes depends upon a willingness to shout louder or hit harder. Children from such homes begin to show higher levels of aggression even before they are school-aged. Expecting others to be hostile, such children generate conflicts in response to the slightest perceived threat. What is to be feared—more than injury in an attack—is becoming a victim. These girls achieve reputations as school bullies and are both deferred to and feared by classmates as well as coopted as allies by those who wish to settle a score with others. The more they fight, the more adept they become at controlling their fear and vanquishing opponents (Campbell 1984). By their teens, they are seasoned veterans with a reputation to defend and an armoury of combative techniques. For these girls, fighting has become if not routine then at least a useful strategy. The ingredients of aggression do not travel independently but together. The same impoverished neighbourhoods that have more than their fair share of male unemployment and single mothers also have daughters who each menarche early, begin their sexual careers sooner and have greater experience of fighting for what they need. Each ingredient depends upon the others. The absence of supporting males drives women’s short-term sexual strategies that results in female poverty and a hand-to-mouth existence. These in turn set the selection of strategies by children in their early years who learn to fight for what they want. Daughters in father-absent families become sexually mature at an earlier age and pursue their own search for short-term mates with the aggression that has been honed in an environment where forbearance and patience produce few rewards. We know precious little about the nature of female fights. We can gain some rather impersonal perspective from official statistics. Robert O’Brien (1988) generated predictions about the expected distribution of male-on-male and female-on-female crime based on the proportion of each sex in the population and known facts about the sex of most violent offenders. He found that women commit simple and aggravated assault against other women more often than expected by demographic predictions. This is especially true for simple assault where no weapons are involved and where the victim is a non-relative. British victimization data of 12–15-year-olds

Mhoc06.fm Page 195 Friday, December 14, 2001 10:17 AM

    also suggest that ‘victims of violent assault usually cited perpetrators of the same age and sex as themselves’ (Home Office Research and Planning Unit 1995, p. 8) and this pattern is confirmed by large-scale surveys of bullying in secondary schools (Ahmad and Smith 1994). Female victims reported that their attacker was another woman in 20 per cent of all reported incidents of contact crime (Home Office Research and Planning Unit 1993; Home Office Statistical Bulletin 1996). They are attacked by other women in 22 per cent of stranger assaults, 39 per cent of acquaintance assaults, 33 per cent of home-based assaults not involving partners, 35 per cent of street assaults, 22 per cent of work-related violence and 56 per cent of assaults in pubs and clubs. George (1999) obtained data from a random sample of 1455 British male and female respondents concerning incidents of attacks by women during the preceding five years. Seven per cent of women reported an assault by another woman and a further 4 per cent reported a threat of assault. Female assaults were most commonly directed at and committed by 15- to 24-year-olds, occurred predominantly between friends and acquaintances and the most frequent forms of attack were pushing, shoving, grabbing, tripping, slapping, kicking and punching (M. George, personal communication). I conducted a questionnaire study of 251 girls aged 16 from cities in Britain that attempted to gather information not just from victims but from the attackers themselves (Campbell 1986). All of the girls had witnessed a fight, 89 per cent had been involved in at least one and 43 per cent had fought during the preceding year. The majority of fights (73 per cent) were against another girl (see also Campbell et al. 1998), and occurred most often in the street (47 per cent) or at school (29 per cent). Only 10 per cent of the fights involved weapons and the most common tactics were punching (80 per cent), kicking (79 per cent) and slapping (57 per cent). One-quarter of the girls reported no injuries from the fight while others had bruises (41 per cent), scratches (22 per cent) and cuts (11 per cent). Taken together, these studies suggest that female–female attacks often occur between young and similarly-aged acquaintances in drinking establishments or in the streets and involve non-weapon, hand-to-hand tactics. But what are they fighting about? I posed this question to the girls who answered my questionnaire. The most common category which accounted for 46 per cent of fights was an attack on the girl’s personal integrity which included instances where there had been allegations about the girl’s promiscuity, false accusations or pejorative gossiping behind her back. The next most common category was loyalty where the girl fought to defend the name of a friend or relative who had been the butt of an integrity attack. The third most common category was jealousy about a romantic partner (12 per cent). (As part of this study I also gave the same questionnaires to girls aged 16 to 20 in a juvenile prison and to adult imprisoned women. In these two older samples, the category of jealousy accounted for 25 and 42 per cent, respectively, of the responses.) The importance of sexual reputation was underlined when I asked girls about the last remark that was made before the first blow was struck. Equal proportions of responses (31 per cent) cited accusations of promiscuity (‘slut’,

Mhoc06.fm Page 196 Friday, December 14, 2001 10:17 AM


    

‘whore’) and dispute-related responses to such epithets (‘Come on then and get some, slag’, ‘Takes one to know one’). This suggests that female fights are indeed about mating issues—often an attempt to retrieve a sexual reputation that has been verbally impugned by another girl’s gossip and sometimes directly about jealousy over a boy. Marsh and Paton (1986, p. 62) found a similar pattern in their ethnographic study of British teenagers, concluding that ‘Highest on the list, however, was “rumour spreading’’ or sexual insult. It was agreed that it was wholly justifiable to use physical violence in the defence of one’s sexual reputation: “If anyone slags you off—calls you a tart or something you’ve got to be able to do something about it.’’’ Their interview transcripts clearly support this interpretation (Marsh and Paton 1984): If a girl’s going out with a boy and he’s two-timing her, they tend to take it out on the other girl. Well, it takes two doesn’t it? (Interviewer) Are the fights generally over boyfriends? Mainly The thing you said earlier about girls being accused of being sluts, they fight for their reputation . . . (Interviewer) What do you reckon is the reason girls fight? Jealousy Yeah (Interviewer) What over? Boys usually Like a lot of girls say they fight to protect their reputation. Yeah . . . (Interviewer) Why do girls fight? Over boys—yeah, there ain’t really any other main reason is there?

A similar theme emerged in interviews that I conducted with girls from several British cities (Campbell 1982): (Interviewer) What do girls fight about? Boys Ripping up one another’s clothes and calling each other names. Jealousy (Interviewer) So do girls and boys fight about the same things or not? Boils down to the same thing really ’cos girls fight over boys, boys fight over girls. Fight for their pride, things like that and boys fight about the same thing.

And the same conclusions about girls’ fights come from teachers (Davies 1984) and from male peers (Duncan 1999, p. 94): ND (interviewer): What do you think girls fight about? CIARAN: Calling names. BILLY: Well if one girl is nicer looking than another girl, then that girl will call the nice looking one slut or slag or something. CIARAN: Whore.

Mhoc06.fm Page 197 Friday, December 14, 2001 10:17 AM

  


ARTHUR: Or if one girl knows that a girl fancies a boy and she has gone out of her way to ask this boy out and the girl fancies him as well, well that might cause friction. ND (Interviewer): So jealousy of boyfriends, stuff like that? ARTHUR: Yeah.

Clearly mate selection plays a major role in female aggression. But we can be a bit more specific and identify three motives that should be especially likely to ignite violence: the management of a woman’s sexual reputation, the degree of competition for desirable partners and the expression of jealousy in response to another woman’s attempt to ‘steal’ a man whose resources are already committed.

Sexual reputation A woman’s chances of securing a desirable long-term mate depend in large part upon a man’s evaluation of her likely fidelity, which will affect his future certainty of paternity and paternal investment. Terms such as ‘slag’, ‘tart’ or ‘whore’ are powerful triggers to aggressive incidents (Campbell 1982; Duncan 1999; Lees 1993; Marsh and Patton 1986). Despite the achievements of the women’s movement and the manifest inequity of the sexual double standard, women still respond angrily to accusations of ‘loose’ sexual standards in a way that men do not. ND (Interviewer): What do girls get called if they go out with lots of lads? SPIRO, GARAHAM, FAZAL, NASUR: A slag! (Laughter) FAZAL: A slut. They hate it an’ all, the girls do. ND (Intervewer): Is that (being called a slag) a bad thing? SPIRO, GARAHAM, FAZAL, NASUR: Yeah. NASUR: You don’t know where they’ve been, do you? (Duncan 1999, p. 53)

Wilson (1978), in a study of working-class teenage girls, noted the ‘repressive triangle’ of love, sex and marriage. Girls took love to be a prerequisite for sex and this in turn was a precursor to marriage. Sex without at least lip service to marriage placed the girl in danger of developing a sexual reputation. It was the girls themselves who were most vocal in enforcing this code. ‘The girls regulated their contact with other girls who were known as “lays’’ in order to preserve their own reputations. In fact they openly ridiculed the girls referring to them as “whores’’’ (Wilson 1978, p. 70). And lest we think that things have changed dramatically, fifteen years later teenagers were still alert to the distinction between nice girls and tarts. ‘If you don’t want to get married and want to live a free life . . . everyone will call you a tart, like you’ve got to go out with a bloke for a really long time and then marry him’ (Lees 1993, p. 105). And boys continue to make the same distinction between prospective wives (‘Attractive, clever, nice personality, trustworthy, older, not someone that everyone knows. Not someone who’s been out with people I know’; Lees 1993, p. 139) and slags (‘You wouldn’t go out with her, you would just knock her off. She is just easy, she is just easier to get off with . . . So you would use her . . . Yeah, skank her’; Duncan

Mhoc06.fm Page 198 Friday, December 14, 2001 10:17 AM

      1999, p. 54). And girls continue to avoid friendships with sexually available girls for fear of reputation-by-association. Indeed girls themselves actively collude in enforcing the double standard through gossip, rumour spreading and ostracism. Information given in an intimate friendship between girls is always dangerous because it can be used later as a particularly potent form of attack. The most risky confidences centre around sexual behaviour and feelings. One reason why so few girls even talk to their closest friends about sexual desire or actual sexual behaviour is through fear that their friends might betray them and gossip—spread the rumour that they are a slag. There is no parallel for boys to the risk of betrayal which can destroy a girl’s whole social standing. (Lees 1993, p. 80)

Lees goes on to analyse a fairly typical response by one of her female informants to being called a slag (in this case by a male): ‘I’ve been walking down the street and someone’s said to me, across the road, being rude, and says “You slag,’’ and I think, “How do you know? You’ve got no evidence.’’ That makes me angry, ’cos, like, you see someone and you’re meant to know whether they sleep around or not.’ Here Kate is angry that the boy calls her a slag without having any evidence. Even when the boy does not know her, a girl reacts by denying the accusation rather than by objecting to the use of the category. For them what is important is to prove that you are not a slag: what they unquestioningly accept is the legitimacy of the category of slag. In other words, the category has uncontested status. (Lees 1993, p. 267)

As one girl remarked to me ‘A girl that’s been called a slag is the same as a boy that’s been called a chicken’ (Campbell 1982, p. 142) and indeed from the viewpoint of threat to their future mating opportunities both epithets are damning. But while a boy can demonstrate he is not a chicken by fighting anyone who so impugns him, a girl is unable to demonstrate in any public way that the accusation is false. This places her in a difficult, if not impossible, situation. A girl can attempt to prove her innocence by publically confronting the boy she is alleged to have had sex with. But in these situations it is his word against hers, he has every reason to lie and in general it is his word that prevails. As one teenage girl said: ‘If they are regarded as a slag, no one is going to look at them, no one is going to be nice to them are they? They are going to say she’s a slag. Everyone is going to believe it’ (Duncan 1999, p. 52). If, however, she fails to respond to the remark, she may be taken to have admitted the truth of the accusation. She could try to ignore it, but here she runs the risk of demonstrating a freedom-from-convention that may add further fuel to her image as a non-conformist. The potency of calling a girl a slag rests on the anxiety girls experience about their sexual reputation. Ignoring insults can be an effective way of counteracting the abuse. If girls can recognise the double standard, then they will not be so shaken by it. Yet being independent in thought or deed carries risks. Being free and independent can signal promiscuity. The freedom of women can be equated with prostitution. What is lacking is a language through which the legitimacy of slag as a way of censoring girls can be contested. (Lees 1993, pp. 275–6)

Mhoc06.fm Page 199 Friday, December 14, 2001 10:17 AM

  


The best she can do is to forcefully repel anyone who labels her as a tart and so minimize the likelihood of such a reputation attack being repeated. A major factor in driving girls’ fights is the spreading of rumours which triggers a violent reaction from the target of the gossip. She started spreading rumours about me saying that I used to sneak out in the middle of the night in my night-dress and meet ten boys or something really stupid . . . Well, we were arguing with each other about the rumour mainly and she was saying she didn’t say it . . . and then she started calling me names like that and then she started to walk off across the road and she said ‘I’ll get you sometime, you fucking bitch’. And that made me mad because if she was going to get me, she was going to get me there and then. I mean there was no point in getting me later and so I kicked her in the back and she fell flat on her face and I said ‘If you are going to get me, get me now’ and we started fighting. (Marsh and Paton 1984, vol. 1, p. 8) I was walking round to your house and you know one of the girls we used to walk to school with said ‘Oh my God. You look a right slut. Look at all that make-up you’ve got on’. Now I never batted an eyelid. I just told her to shut her mouth—there was three of them. I said ‘Oh, shut your mouth’. Yet that was on my mind all night and we got up to Scamps (night-club) didn’t we? And she was there and I just went over to her and hit her round the face. ‘Now say it’ I said and she poked me in the eye. (Marsh and Paton 1984, vol. 1, p. 33)

Feminists have struggled to understand the potency of accusations of sexual accessibility. In this day and age, with the inequity of the double standard acknowledged, ‘slut’ should have lost its power long ago. But even today, young women cannot afford to shrug it off and attempts to turn it on men simply provokes laughter (Duncan 1999, p. 53). ND (Interviewer): So are the lads called ‘studs’ by the girls in the school? FAZAL: Yeah. ND (Interviewer): And is that a bad thing to be called? FAZAL, GRAHAM: No. (Laughter)

Sexual conquests enhance rather than detract from a young man’s reputation— his past desirability to other women increases his future success rate. When we learn that an unattractive man has bedded a hundred women, we suppose that he must have hidden qualities that have escaped us. But when we learn the same about a women, we guess she is desperate. The most desirable women can pick and choose— they don’t need to resort to the ‘cheap trick’ of sex. A woman who is discriminating in her choice of partners signals that she believes she is worth waiting for. And it is men’s choosiness about who they will accept as a long-term partner that drives the dynamic. Paternal investment is costly and it precludes the opportunities a man would otherwise have for multiple short-term, cost-free alliances. Most importantly, if he selects an unfaithful partner he runs the risk of a lifetime’s investment in another man’s child. Men, under monogamy, are as choosy as women and

Mhoc06.fm Page 200 Friday, December 14, 2001 10:17 AM


    

high up their list of desirable qualities comes fidelity—which he can only estimate by the ease with which other men have gained sexual access to her in the past. So men, as the saying goes, may want a wife who is ‘a whore in the bedroom’—as long as she is their whore and nobody else’s.

Competition Competition among women for men has been documented cross-culturally. Glazer (1992) argues that female dependence upon men for resources drives competition among women to extreme levels. ‘The more subordinate women are to men and the more dependent they are on patriarchal social structures, the more injury they inflict upon one another’ (Glazer 1992, p. 164). To illustrate her point, she compared levels of female–female aggression at two time periods in Zambia. In pre-colonial times, communities were matrilocal—women inherited farming rights and had major responsibility for food production and distribution. Their closest cross-sex relationships were with their brothers who had every genetic interest in encouraging their sister to produce children (for whom they were jointly responsible) without the attendant male proprietal jealousy of a marriage. Women’s sexuality was encouraged and divorce was easy and frequent. During this period, when women were not economically dependent upon men, rates of intrafemale aggression were low (Schuster 1983). Similar findings have been reported in other matrilineal societies including the Mundurucu of Brazil and the Hopi North American Indians (Benedict 1934; Murphy and Murphy 1974). With colonial rule, female economic dependence on men increased while at the same time men were unreliable and uncommitted fathers. Female–female aggression rose as women competed for men with urban jobs and steady incomes. The intensity of competition was inversely related to the financial independence of the woman: ‘Physical violence between women is common in confrontations among the poor, where the fight is for economic survival’ (Schuster 1985, p. 190). The competition between females appeared also at a class level. Elite women could not ‘marry up’ and so were forced to search for mates among men of their own class. Some subelite women, however, strove for hypergamy and presented a direct threat to them. Other less ambitious sub-elite women would accept the best and most ambitious of the lower class but this brought them into conflict with lower-class women who witnessed the cream of their potential mates being lured away from them. In a similar argument, Burbank (1987) surveyed 137 societies in the Human Relations Area File, which collates cross-cultural anthropological observations, to locate episodes of aggression initiated by women. She concluded that female aggression is largely directed against other adult females (over 90 per cent of attacks) and often appears to be a means of competing for men or subsistence products. Out of a total of 297 female–female fights in which reasons were recorded, 121 were about men and 67 about subsistence concerns, including food, crops, essential domestic goods, money, tools and implements.

Mhoc06.fm Page 201 Friday, December 14, 2001 10:17 AM

  


Schuster and Burbank both argue that it is chiefly when women cannot provide for themselves that they are forced into competition for men who have access to the necessary resources. In support of this, Campbell et al. (1998) found that female within-sex assault in the United States is closely related to female resource shortage as indexed by rates of female unemployment and AFDC (Aid to Families with Dependent Children) receipt. Among the American underclass, well-resourced males are in short supply. Taylor’s interviews with young women in Detroit show that they make a clear differentiation between drug users and drug dealers; the former are rejected as the ‘living dead’ while the latter are in a position to furnish abundant, albeit temporary, resources. As Taylor describes their viewpoint: ‘They can see the power of the gang, the celebrity status. This is real, it can happen to people just like them. These women remember seeing “that girl’’ at school, in the shopping malls, driving a new Mercedes, BMW, Corvette, sporting Gucci, Louis Vuitton, Fendi and smelling of expensive perfume, going to Auburn Palace to see the Pistons, meeting John Salley at parties’ (Taylor 1993, p. 198). The desirability of access to such material resources means that, even among college girls, drug dealers are considered desirable partners worth fighting for. As one girl explained: ‘It’s hard to get a good man and girls grab any fella that treat you special . . . It’s just tight out here, the campus is fucked up ’cause we ain’t got nothing but girls, girls, girls and the guys got their pick. We just start fighting each other over the same guys’ (Taylor 1993, p. 129). Male preference for novelty is not lost on these girls. Girls who are involved in mixed-sex groups or gangs are acutely sensitive to incursions by other females who—by virtue of their novelty—have an in-built advantage over the ‘home girls’. Fights can occur when outside females seek to enter the group or to establish relations with males who are already spoken for. As a British girl described it: If you get some girls what come in the Oranges (pub), they have been in there once and they come in. They really start slagging themselves around. They ain’t been in there for about two years kind of thing and they just start hanging around the boys and showing off. I cannot stand that. And that attracts the blokes more than anything (Marsh and Paton 1984, vol. 1, 51) Well, most of the fights are in the Oranges and Lemons (pub) between the girls. Get a bird in the bog (toilets) and beat her up. You’ve done it. Rosie has done it. We are really possessive. If there is a bloke, we like think ‘Oh, he’s eyeing her up’ so I follow her into the toilet. (Marsh and Paton 1984, vol. 1, p. 156) There was a fair and there was all these girls up there and there is this girl called Della— she is supposed to be a right old scrubber and she went up to her, in front of her and that, and then she pushed her so Jackie had a fight with her. (Marsh and Paton 1984, vol. 2, p. 231)

An almost identical scenario was described by a member of a New York City mixedsex gang called the Sex Girls:

Mhoc06.fm Page 202 Friday, December 14, 2001 10:17 AM

      I had a fight with this girl, her name was Sugar, right? So she was new on the block . . . She just came on the block, right, and you know how the guys are. I used to be going out with Chico? And the guys, they see a new girl? This guy—our guys when they see a new girl, they be checking her out you know. We don’t like that. So she came to the block but she didn’t come to make friends with us, she just came to make friends with the guys you know? . . . So then one day, she was on the corner and Big L. was in the corner with her and she was over there you know real flashy and everything, like nothing happened. And I had that in my mind and then I was in the corner and Booby was in the corner and Booby’s talking and we was laughing. And I see her laughing and I tell Booby, ‘I bet you I’ll go over there and slap her in the face’. I tell Booby and Booby tell me ‘Go ahead, Weeza.’ I went over there and I almost going to hit her and Big L get in my way. He goes ‘Hey, Weeza, you hit her? I’m going to hit you,’ and I say ‘Go ahead, Big L, hit me. If you’re going to hit me, hit me now.’ And he didn’t do nothing and then he took her to his house like always . . . Then she came with a sweat shirt, her hair was all grease like she was ready to fight me. So I don’t know what happened. Big L said ‘OK, Weeza, she’s ready for you.’ And I tell her, ‘Oh yeah? She’s ready for me?’ And I took my shoes off and I rolled my pants up and I said ‘Oh yeah? Come on. You ready for me?’ I just hit her once. The fight only last three minutes. That’s it. I just hit her and I bit her over here. . . . When I spit, I spit only blood. That’s it. They got to take her to the hospital. She used to go with them all. That was finished. (Campbell 1984, pp. 147–8)

The Turban Queens of Brooklyn told a similar story: They (the boys) go to where they think we ain’t gonna find out. No matter how far they do it, there’s always somebody so scared of us that they’ll come running to tell us. We’ll still find out. We’ll always find out. They’ll swear on their mother, their father, their sister, their brother, ‘I didn’t do it, I wouldn’t do it with that bitch. I wouldn’t make good with that bitch.’ Then they try to soup. But I already know the deal with them. ‘Alright, yeah, yeah, yeah’. And that’s when I go. I go up to the girl. And they won’t even bother hitting us ’cos they know they’re gonna get worse. I would just go up, ‘Hey, I hear you made it with my old man.’ This and that. And blat, that’s it. The whole thing is over ’cos they don’t even raise their hands. They put their head down and they cut out fast.

Aggressive female competition of a less serious nature has been documented in a variety of cultures. In Buenos Aires the principal reason for female fights are over a man. While both men and women in Argentina are competitive, a recurring theme throughout the ethnographic interviews is that women in Buenos Aires are very competitive, envious and jealous of each other. For example, female competition is expressed through invidious comparisons regarding clothing and appearance. Women are very concerned with remaining thin and attractive. They place great emphasis upon dressing fashionably, often in clothes that accentuate their figures. Both female and male informants concurred that competition among women is generally over men, either a particular man, or else for the attention of men in general. Fighting among women results according to one man from ‘competition for a man for sexual reasons’. (Hines and Fry 1994, p. 228)

Mhoc06.fm Page 203 Friday, December 14, 2001 10:17 AM

  


Among the Zapotec of Mexico, women attack other women more often than they attack men and these episodes result predominantly from disputes over men, triggered by malicious rumours about rivals (Fry 1992). Bellonese women from the Solomon Islands live in a society of extreme male domination and have specific fighting techniques that are only used against other women. The attacker grabbed the other person’s hair and hanked it back and forth, trying to pull her antagonist to the ground. Once the opponent was grounded the victor raked her long fingernails down the loser’s face, neck or chest. Such a treatment was felt as a serious humiliation and could only be used by a woman toward another woman. (Kuschel 1992, p. 180)

Jealousy But attracting a man is not the end of the story. Women also strive to maintain ‘ownership’ of males against incursions by other females. This has been noted most forcibly by anthropologists in the resistance expressed by women to prospective co-wives and to the conflicts that are attendant on polygyny (Collier 1974). Wherever polygyny is practised, conflict between wives accompanies it. Burbank (1994) points out that the co-wife relationship is so implicitly filled with venom that it is reflected in the very vocabulary of aggression. In East Africa the Gusii word ‘engareka’ means ‘hatred between co-wives’, in Luo co-wives call each other ‘nyieka’ which means ‘partner in jealousy’ and in Surinam the word for fight is literally translated as ‘act like a co-wife’. They compete for food, for money, for their own children’s inheritance and ultimately for sex as a route to all of these, as Holmberg (1969, p. 46) explains among the Siriono of Bolivia: Since food and sex go hand in hand in Siriono society—and there is a scarcity of the former— the wives with whom the husband most frequently has sex relations are also the wives who generally get the most to eat. Consequently co-wives frequently vie with each other for the sexual favours of their husband. This sometimes leads to bitter fights and quarrels.

Burbank’s finding that 90 per cent of recorded instances of female aggression worldwide were against other women is considerably more surprising than the fact that co-wives, sexual rivals and ‘the other woman’ were the most frequent victims. Among co-wives, the three principal reasons for fights were jealousy, the husband’s distribution of favours and the acquisition of an additional wife. As Burbank notes: Where co-wives are competitors, a man’s acquisition of an additional wife presents one of the most immediate threats to the women of the household and thus creates a reason and a target for female aggression. To find this reason accounting for a proportionately greater amount of the reasons for physical aggression between co-wives is therefore not surprising. (Burbank 1987, p. 94)

Murdock (1949) noted the potential for conflict between co-wives and found that where the co-wives are sisters they may share the same household but that where the

Mhoc06.fm Page 204 Friday, December 14, 2001 10:17 AM


    

women are unrelated they frequently do not. Burbank (1987) found that even in societies with sororal polygyny at least 20 per cent of co-wife aggression was between sisters. So even a close genetic relationship may be insufficient to contain the explosive jealousy between women who share the same man. Tension between women resulting from their mutual economic and social dependence on a man can be seen intergenerationally between women who are not even sexual rivals. Collier (1974) and Lamphere (1974) have shown how patrilocal extended families, where the power and authority reside with men, can give rise to intense competition between women for preferential access to and influence over the man. The archetypal example is the tension that exists, in all societies, between mother-in-law and daughter-in-law. Glazer (1992) describes traditional Chinese family structure as one that was almost guaranteed to produce conflict between women. From about 1000 BC women lived a cloistered existence in patrilocal, patrilineal and patriarchal households with little opportunity to travel, work or socialize beyond the home. A superfluous daughter was often sold into slavery or prostitution and, in her place, the family adopted a new girl whose labour they could exploit as a child and who they could later marry off to their son. Adopted and natural daughters were natural enemies. The mother of the man-of-the-house had the right to control the labour of her daughter-in-law and because daughters-in-law could be sold or pawned, they too were natural enemies. Sisters-in-law were also at loggerheads since a wife was deprived of her husband’s earnings if his younger sister needed to be married off. Wealthy husbands might also take a concubine, adding further competitors to the struggle to control the man’s resources. Within these households, beatings by the mother of daughters-in-law, daughters and adopted daughters were common. Mutual dependence on the economic support by a single man drove division through the whole female household. Victoria Burbank (1994) has provided a detailed ethnography of Aboriginal women and their involvement and attitude to fighting. She uncovered 147 episodes of female–female violence and was able to ascertain reasons for 50 of them. Again, men occupied a prominent place. When it comes to ‘men’ as a reason for aggression however women’s attacks on each other are much more frequent. Thirty cases of aggression between women were attributed to jealousy or conflict between unmarried women over men, a husband’s real or suspected adultery or a husband’s attempt to bring another wife into the household— all situations in which women are pitted against one another in competition for men’s favours, if not their resources . . . what we would call ‘sexual jealousy’ and the knowledge or suspicion that a husband is having an adulterous affair are the most common. It should also be noted that twenty one of the thirty cases of aggression between women for these reasons involve fights. . . . ‘Jealousy’ of both married and unmarried women is clearly the cause associated with the most amount of injury. (Burbank 1994, pp. 104–5)

In Western society too women strive to hold on to their men. Indeed, among adolescent girls, a sense of ownership may not require the bonds of marriage,

Mhoc06.fm Page 205 Friday, December 14, 2001 10:17 AM

  


cohabitation, engagement or even an ongoing courtship (Marsh and Paton 1986). The most serious disputes arise in the course of steady dating relationships which girls feel they have a right to protect against blatant takeover attempts. This mate of mine, her boyfriend was sat talking to another girl. And my mate she was really drunk and she’s the type of person—she’s really possessive over her boyfriend. I mean I can put up with it to a certain extent but anybody will lash out. And she followed this girl into the toilets and I was sat in there and she threatened her with a bottle and this girl started pulling her hair and Karen was slapping her and it just amazed me. The bouncer come in and tried to split it up. (Marsh and Paton 1984, vol. 1, p. 28) Say she was leading him on, I would tell her to shut her mouth or something like that. She would either lay off or if she carries on, she gets the worst doesn’t she? It just goes into an argument and then it could end up in a fight . . . Well, I was going out with this boy called Steven. This was on holiday, because I met him on holiday, and this girl I was going about with—her name was Mary—she was flirting round with him and that and I didn’t like it and one night I saw her just about to kiss him so I went up and hit her. (Marsh and Paton 1984, vol. 4, pp. 2–3)

On occasions, the full brunt of anger may be borne not by the rival herself but by anyone who carries unwelcome news of a boyfriend’s betrayal. Duncan (1999) recounts the story of Roza, a secondary-school girl of legendary fighting ability, who was informed by her friend Helen of her boyfriend’s involvement with an older woman. Roza interpreted this confidence as evidence that Helen was ‘stirring things up’ in a jealous attempt to interfere with the true course of young love. Helen reacted angrily to this interpretation of her actions and proceeded to spread the news not only of Roza’s cuckoldry but of an abortion that Roza had had at the end of her previous affair. This was a clear attempt to sully Roza’s reputation and some confrontation became inevitable. Helen and her friend Tammy challenged Roza to a fight. Roza declined but appointed her friend Mary to fight in her place. Mary, lacking the social skills to either extricate herself or defuse the situation, lunged at the pair leaving them battered and bleeding on the ground. Helen and Tammy were excluded from the school for a week. Though everyone expected further violence, the stand-off was broken by the permanent exclusion of Roza’s boyfriend from the school (arising from a fight with local youths) and shortly thereafter the departure of Roza who found herself pregnant. Thus can general mayhem arise from a whispered piece of gossip. And a girl can feel a sense of proprietal rights over a boy both before he is aware that they have a relationship and long after he believes it has ended. At one extreme of this ownership ‘spillover’, girls believe that merely being attracted to a boy should deter any other girls from coming on to him. As Duncan (1999, p. 97) found in his study of secondary-school girls: ‘Prospective paramours were sometimes told that they should not go out with a certain boy because another girl had a claim on him. Even if the boy did not agree, the clique would maintain that he was doing one of their friends an injustice and any girl who broke the embargo would suffer

Mhoc06.fm Page 206 Friday, December 14, 2001 10:17 AM


    

accordingly.’ The point is not lost on boys who have witnessed the phenomenon at first hand. Supposing that two girls fancy him, then there is a fight between the two girls to get in first . . . If he went to a disco and she went with him and the next disco this other girl went with him, the two girls would fight about it. That’s what happens. ‘He’s mine.’ ‘No, he isn’t, you bitch.’ (Marsh and Paton 1984, vol. 1, p. 104)

At the opposite end of the relationship spectrum, girls may continue to fight for a romance that has officially ended. Me and my boyfriend finished for a while. I went to London. Then I came back and if I see him with any bird I just went mad and then seeing him with her as well . . . I walked up to her and I says ‘You fucking leave him alone or else you’ve had it.’ I really went mad at her. (Marsh and Paton 1984, vol. 1, p. 134) This girl (Suzie) has been going out with his boy for a year and she packed him up and her friend (Linda) goes out with him so Suzie keeps calling the girl a slut all the time but the girl won’t say nothing. Well. Its only because Linda can’t have her (beat her up) though, isn’t it? Because she knows that she would probably get beaten up. (Marsh and Paton 1984, vol. 2, p. 215)

Duncan (1999) recounts just such an incident involving a popular secondary-school boy called Pinxi who had, several months before, gone out with Anna, a fringe member of a ‘hard’ group of girls. Jackie, a younger girl, assumed that since their relationship had not only ended but Pinxi had dated a number of other girls in the intervening period, it was safe for her to be seen with him. Warnings were issued for her to stay away from him. Foolishly she ignored them only to find herself surrounded and assaulted by Anna and five friends outside the school disco. So serious was the assault that she had to be placed in the custody of staff in the Special Needs centre until tempers cooled. It is perplexing that young women direct their attack towards their own sex rather than to the errant boyfriend or husband who would appear to be just as guilty as the ‘other woman’. Burbank suggests that women may be fearful of the retaliation of which a man would be capable yet this does not seem to deter young women from ‘slap-the-cad’ assaults on their partners (Archer 2000). Alternatively, she proposes that a man’s love cannot be physically compelled while a female rival can at least be dispatched by violence. Among gang girls, the reason for their choice of target was more clear-cut. Men, they reason, cannot be expected to turn down the offer of sex from a woman. It is in men’s nature to take sex wherever they can find it and indeed failure to take advantage of an unsolicited offer would cast serious doubt on his manhood (Campbell 1984). Men in this regard, despite their macho posturing, are viewed as weak. It is women who are the controlling players in mate retention because of men’s priapic lack of discrimination when it comes to one-night stands. Women know all too well what men want. And when other women take advantage of men’s weakness, they know too how to protect their own interests.

Mhoc06.fm Page 207 Friday, December 14, 2001 10:17 AM

  


Our picture of mate selection has come a long way since it was first proposed that males were eager and females coy. Where females undertake almost all of the parental care and live an isolated life, what they should seek chiefly is quality genes and, given their high investment on which the male’s genes hitch a free ride, they should be highly selective in their choice of mate. But when men and women not only cooperate in rearing offspring but face an extended future together because of the long period of infant dependency, choices become more complex. Men also become choosy because commitment to a single woman is a choice of reproductive quality over quantity. And where committed and well-resourced men are in short supply, this can engender competition between women. Rivalry can escalate from cosmetic improvement, to verbal attacks on sexual reputation to outright violence. Even when circumstances militate against acquiring a permanently faithful and committed ‘good man’, it is a dream that never goes away. Girls in street gangs, when asked to describe their perfect future, come back to the traditional desire for a man who will provide resources, security and unconditional love (Campbell 1984). Short-term partners are a default option not an ambition. Even girls who will freely admit to theft, robbery and assault, bridle at the suggestion of cheapness or prostitution. Shakespeare as usual put it best: Good name in man and women, dear my lord, Is the immediate jewel of their souls: Who steals my purse steals trash; ’tis something, nothing; ’Twas mine, ’tis his, and has been slave to thousands; But he that filches from me my good name Robs me of that which not enriches him, And makes me poor indeed. (Othello, Act III, Scene III)

Mhoc07.fm Page 208 Friday, December 14, 2001 10:17 AM

 


In the underworld: Women and crime

The female criminal rarely fails to capture media interest. For the past twenty years, newspapers have reacted with repetitive and predictable shock to women’s lawbreaking. ‘Sugar N’ Spice But Not At all Nice’ (The Sunday Times, 27 November 1994) and ‘Gentle Sex Indulges in Thrill-Seeking Violence’ (The Observer, 24 September 1995) trumpeted two British newspapers, each more appalled than the other about the unladylike involvement of women in crime. Despite the massive changes in women’s opportunities and roles in society, the media still invite us to be outraged that the hand that rocks the cradle could also steal, defraud and fight—as if womanhood was an automatic prophylactic against lawbreaking. Because women so seldom offend in relative to men, female criminals have been treated as a kind of freak show and nowhere more so than by one of the founding fathers of criminology Cesare Lombroso. Much influenced by Social Darwinism, he held that criminals of both sexes represented atavistic throwbacks to a less evolved human state—they were, in short, ‘born criminals’. He searched for proof in the faces and bodies of female criminals, measuring their skull and jawbones, height, weight, arm span, hand size, facial proportions, moles, prehensile feet, hairiness and voice pitch. Unfortunately he found rather few ‘evolutionary anomalies’ among the women he studied. Undeterred by the uncooperative data, he proceeded to explain why so few women look like born criminals. Because women as a sex have not evolved so far as men, their evolutionary ‘regression’ is consequently less extreme and less evident. But, he argued, women who are born criminals are notable for their resemblance to men. Such a woman is a ‘monster’ who ‘belongs more to the male than to the female sex’ (Lombroso and Ferrero 1895, pp. 152–3), combining the worst aspects of womanhood (cunning, spite and deceitfulness) with the criminal inclinations and callousness of men. The majority of women lawbreakers were ‘occasional’ criminals who were fully female but, by virtue of their sex, were weak, morally inferior and easily led into crime by men. And so Lombroso and Ferrero introduced an idea that would endure well into the next century—criminal women were either pseudo-males or inadequate personalities, putty in the hands of the evil men. The bad and the sad. W. I. Thomas (1967) was the next to take up the explanation of female crime and he aligned himself, as we can see by the title of his work The unadjusted girl,

Mhoc07.fm Page 209 Friday, December 14, 2001 10:17 AM

  :     steadfastly with the ‘sad’ camp. Humans are driven by four wishes or ambitions: for new experience, security, response and recognition. The way in which these wishes are fulfilled are dependent on social regulation. Delinquent girls lack such regulation and seek to fulfil their wishes in ways that society recognizes as deviant. Girls are especially driven by a wish for response—maternal and sexual love. When such a drive is not channelled appropriately into marriage, it leads to prostitution and immorality. Female criminality thus became sexualized and the female criminal was to be pitied rather than blamed. She was a pathetic figure searching for love through sex: ‘I have seen . . . the life history of a woman who has had sexual relations with a number of men. . . . While she was sexually passionate her concern was mainly to satisfy the sexual hunger of others, as she satisfied their food hunger’ (Thomas 1967, p. 21). Traditionally, he argued, society had acted to control the expression and organization of the four wishes, but the ‘modern revolt’ had led to an anomic situation where social control had diminished. This was especially true among women who had been the most restrained and hence reacted with the greatest ‘unadjustment’ when the restraints were slackened and moral decisions were no longer dictated by the community. In Thomas’s writing, we see the beginning of two more images that became increasingly common in later thinking: female criminality as the result of eroded social control of women and the female criminal as victim not agent. Otto Pollak (1961) took an emphatically less sympathetic view of female crime. According to him, women are far more criminal than we might suspect. They appear to commit less crime only because the kinds of offences that they choose are rarely reported to or detected by the police and because both the police and courts are more lenient to women than to men. Women are devious, cunning and highly manipulative—engaging in hidden crimes against their own families and secretly instigating men into criminal activities. Women’s natural deviousness derives from menstruation and the nature of women’s sexuality. Women learn to hide their monthly periods and to tolerate their discomfort in silence. Unlike men, they can engage in sex in the absence of sexual arousal or orgasm without risk of detection. Their skill at concealment and manipulation can be exploited in criminal conduct. Men display misguided chivalry towards them because they view them as passive and harmless—but this is only because men need to deceive themselves about women’s strength and potential for revolt. Pollak’s depiction of women as ruthless killers, manipulatively evading detection by exploiting their virtuous feminine façade, continues decades later in popular books with titillating, if inaccurate, titles such as Deadlier than the male (Kirsta 1994). In 1975, Freda Adler’s book Sisters in crime provoked renewed attention to female crime and much controversy. Her thesis was straightforward—the rise in female crime in the United States was a consequence of women’s liberation. The greater the opportunities afforded to women and the weaker the traditional controls upon their behaviour, the more like men they had become in their rates of crime. (This theme

Mhoc07.fm Page 210 Friday, December 14, 2001 10:17 AM

      has remained popular with the media despite a resounding lack of empirical support for this thesis, see Muncer et al. 2001. The British media never miss the opportunity to describe female offenders as ‘ladettes’, allegedly aping the ‘laddish’ behaviour of male hooligans.) The core of Adler’s argument was twofold; women were seeking greater criminal opportunities just as they were seeking greater legitimate economic opportunities and they were adopting male traits of which criminality was but one manifestation (‘The trend in the last two decades has been toward female adoption of male attitudes, traits, vocations, prerogatives etc., as a means of raising their status also’ Adler 1975, p. 106). These arguments generated heated debate (Box and Hale 1984; Chesney-Lind 1986, 1989; Smart 1979; Steffensmeier and Steffensmeier 1978; Weis 1976). For many feminist criminologists, Adler’s position amounted to a condemnation of equal rights for women and their outraged response was driven more by ideological than by empirical considerations. Others gathered data showing that, contrary to her thesis, female offenders held remarkably conservative ideas about women’s role. Still others demonstrated that criminal women were in fact drawn from the very sector of society that had been least affected by changes in women’s occupational aspirations and opportunities. The desire to discredit the ‘liberation’ argument (and hence preserve the gains made in women’s rights) as well as to encourage a less condemnatory attitude to female offenders spawned a whole new set of proposals about female crime. Some argued that Adler had got the story precisely backwards. Any rise in female crime (a rise that turned out to be restricted to petty property offences rather than violence) was the result not of women’s increased status and greater opportunities but rather of their increasing proportionate share of poverty (Chesney-Lind 1989; Daly 1989; Miller 1986; Richie 1995; Steffensmeier 1980). Women’s crime was born of the desperation of poverty, not the ambition of equality. She invited sympathy not condemnation. A more extreme form of this thesis was that female criminals were appropriately seen as victims, not only of abstract economic inequality but directly at the hands of men. They were subject to physical and sexual abuse in their homes and communities, exploited by drug dealers and pimps on urban streets and unfairly treated by the criminal justice system where police and courts took a more punitive stance towards women by virtue of their ‘double deviance’ against their sex role as well as the legal code (Chesney-Lind 1989). Some went beyond the portrayal of women as victims to view women criminals as heroines, who had not only experienced the abuses and injustices of male oppression but had survived or even directly challenged male control of women (Chesney-Lind 1999; Gilfus 1992; MacDonald 1991). And so today we have a host of contradictory images of the criminal woman. She is both villain and hero; she is both weak and strong; she is to be pitied and to be condemned; she escapes proper punishment and she is punished too extremely; she is economically liberated and financially impoverished; she is confrontational yet devious; she denies her femininity and yet exploits it; she is a victim of societal

Mhoc07.fm Page 211 Friday, December 14, 2001 10:17 AM

  :     repression and she is a symbol of female resistance to control. In short, the study of female crime is in a mess. This muddle has arisen, I believe, for three reasons. First, there has been a confusion about exactly what questions academics are trying to answer. Some writers are asking a question about sex differences: Why do females commit less crime than males? Others are asking about within-sex differences: In what way are criminal women different from non-criminal women? Still others are asking if theories of male crime can be usefully applied to explaining female crime: Do we need a separate and distinct theory to explain women’s offending? Others focus upon the alleged rise in female crime, particularly in Western societies: What factors explain historical trends in women’s involvement in crime? The explanation of female crime that I will offer is capable of addressing each of these questions through the same basic set of evolutionary premises. A second problem is that the kinds of answers that have been given are driven very strongly by feminist ideology and practice. Feminist academic criminologists tread a carefully measured line in their view of the female criminal. Anxious not to see equal opportunities snatched away and to further the safety of women, they present her to us as a victim, pointing to the kinds of male abuse and control that drive women into lives of crime and to the continuing oppression that they suffer as they move into the equally patriarchal criminal underworld. At the same time, sensitive to the danger of ‘overvictimizing’ the female sex, they move towards viewing women criminals as heroic survivors whose acts of theft or prostitution are symbols of their resourcefulness, resistance and strength. From an evolutionary perspective, however, women, like men, must secure the resources they require for survival. There are a variety of ways to achieve this and some are against the law. But because evolutionary theory emphasizes the parallel natural selection processes operating on both men and women, there is no reason to be surprised by women’s engagement in crime any more than we would be surprised to find that women as well as men need food to survive. Survival behaviour is gender-neutral, although undoubtedly human culture overlays this neutrality with a variety of expectations about gender-appropriate behaviour. What evolutionary approaches suggest is that our values about criminal behaviour have to be separated from our explanation of that behaviour. Altruism was selected because of its individual survival value not because it was a ‘nice’ trait. Similarly, crime has remained as a behavioural option because it meets survival needs under desperate circumstances. Natural selection does not jettison a useful strategy because it is ‘unfeminine’. Closely aligned to ideologically selective portrayals of women criminals is the third problem—new feminist epistemology. Rejecting as a myth the notion of value-free objective science, feminist methodologists insist that political agendas suffuse the very questions that scientists ask (as well as the answers that they give), that there can be no ‘objective’ description of human behaviour, that attempts to separate researcher and researched are oppressive and authoritarian, and that variables cannot be studied in

Mhoc07.fm Page 212 Friday, December 14, 2001 10:17 AM

      isolation without the risk of reductionism. The net result is that social science can only describe (rather than explain) and, since there is no objective and uniform reality independent of lived experience, these descriptions can only be provisional and subjective accounts offered by the researcher who cannot be expected to free herself from her own ideological standpoint. At its most extreme, this opens the door to selective and subjective descriptions of small numbers of women who may or may not be representative of criminal women in general. Now I have no problem with ethnographic research and nor should any rightthinking social scientist. We cannot investigate what we cannot describe and the first step in any research programme is to establish the nature of that which we want to explain. (It was for just this reason that, some years ago, I conducted an ethnographic study of female gang members whose identities and lives were quite opaque to us; Campbell 1984.) Nor do I have any quarrel with taking seriously criminal women’s own reports of their lives and experiences (though I am worried when these accounts are heavily overlaid with the researcher’s own ideological interpretations). A full account of crime must address the way it is experienced by the actors themselves. But at this point in time, our view of female crime is so thoroughly muddied that we would do well to go back to basics, establish a few facts that we know to be true and work from there.

Female crime: some facts Fact 1 Everywhere in the world men commit more crime than women (Ellis 1988; Gottfredson and Hirschi 1990; International Criminal Police Organisation 1994; Simon 1975; Simon and Baxter 1989). This is one of the few facts on which criminologists agree and this disparity is true across all age groups and at all times for which reliable data are available. Nor is this a function of the fact that men are more often detected or prosecuted. Self-report studies confirm higher rates of male involvement and victimization studies, which tap ‘invisible’ crime that goes unreported to the police, concur. In general, the size of this sex difference varies with the dangerousness of the crime. Crimes can be rank ordered in terms of their relative risk of physical injury or death. A violent confrontation is the most immediately risky of criminal acts since it involves direct physical engagement with a hostile victim. Property offences carry less immediate and evident risks but the expropriation of others’ resources is always subject to individual retribution and collective punishment or ostracism. Data confirm that the magnitude of the sex difference is greatest for violent crime and the least for petty property crimes (Steffensmeier 1993). According to 1997 US data (United States Department of Justice 1998), the female percentage of all arrests is low for weapons offences (8.2 per cent), robbery (9.8 per cent), burglary (11.9 per cent), homicide (10.3 per cent), simple (21.3 per cent) and aggravated assault (18.8 per cent). Women’s rates of involvement are higher, though still below

Mhoc07.fm Page 213 Friday, December 14, 2001 10:17 AM

  :     the level of men, for larceny-theft (34.6 per cent), forgery (38.6 per cent), embezzlement (47.1 per cent) and fraud (46.1 per cent). The same pattern is evident internationally. International crime statistics (International Criminal Police Organisation 1994) indicate that the female percentage of all arrests (averaged over a number of nations) is lower for robbery (5.27 per cent, N = 53), burglary (5.52 per cent, N = 50), assault (8.72 per cent, N = 63) and murder (9.46 per cent, N = 63) and higher, though still well below the male rate, for larceny (12.91 per cent, N = 53) and fraud (14.14 per cent, N = 60). Fact 2 In places and times where male crime rises, female crime rises too. The correspondence between the rates is staggeringly high. Taking data from 48 US states (Federal Bureau of Investigation 1996), we computed correlations between rates of male and female crime for property and violent offences separately (Campbell et al. 2001). For violent crime, the male–female correlation is r = 0.95 and for property offences r = 0.99. For data from 43 police force areas in England and Wales (Home Office 1980), the correlation for violent crime is r = 0.98 and for property offences r = 0. 99. Similarly high correlations appear over nations (International Criminal Police Organisation 1994): murder (r = 0.98, N = 67), assault (r = 0.99, N = 66), minor theft (r = 0.96, N = 54) and fraud (r = 0.95, N = 65). Others have noted the remarkable stability of these crime patterns over time. Historical data dating back to the thirteenth century suggest that male and female crime rates rise and fall together (Beattie 1975; Hanawalt 1979). Boritch and Hagan (1990, p. 586), analysing crime data from Toronto between 1859 and 1955, also reported extremely high correlations between males and females noting ‘an overriding uniformity of male and female patterns of arrest’. Steffensmeier and Allan (1996, p. 465) reported US arrest rates between 1960 and 1990, concluding that ‘Statistically when the female rates for a given group are regressed on the male rates for the same group across time or across crime categories, the results for most comparisons do not differ significantly from a prediction of no difference.’ Fact 3 Female property and violent crime rise and fall together, just as male crime does. Once again, the figures are strongly confirmatory of this pattern. We examined data from 48 states in the United States (Federal Bureau of Investigation 1996) and the correlation between violent and property offending for women was computed, giving a value of r = 0.91. Data from 43 reporting districts in England and Wales (Home Office 1980) show that among females the correlation between violence and property offences is r = 0.96. International crime statistics from 53 nations (International Criminal Police Organisation 1994) were used to calculate the correlation for females between theft and assault rates with a resulting value of r = 0.92. There is then a remarkably high covariance between property and violent offending in women.

Mhoc07.fm Page 214 Friday, December 14, 2001 10:17 AM

      These facts give us some clues as to what an adequate theory of female crime must look like. First, the effects we have described appear across the world. This suggests that explanations of both the sex difference in crime and the fact that property and violent crime are fellow travellers reflect fundamental principles of behaviour that are not dependent upon a particular culture or historical period. They are as true today as five hundred years ago and they occur whether the country in question is agricultural or industrialized, rural or urban, east or west. The stability of this pattern does not guarantee that an evolutionary account is required but it raises it as a very strong contender in any global explanation. Secondly, the difference between male and female rates of crime suggests that there is something fundamentally different between the sexes in their willingness to resort to criminal behaviour. Although the absolute magnitude of the sex difference may vary across nations, there are no cases where women outnumber men as criminals in a given culture. Thirdly, men and women clearly engage in crime as a function of similar prevailing ecological conditions. Whatever factors drive up male crime have exactly the same effect on female crime. Fourth, theft and violence respond in the same way to these ecological conditions. Whatever function violence serves, it is likely to be similar to the functions served by theft. We are led to a theory which posits similar causes of crime for the two sexes but which recognizes that there is a threshold difference between them. Women as a sex are everywhere less likely to engage in crime but the factors that propel them into it are the same factors that drive men. And these criminogenic factors exert a similar force on theft and violence with the threshold for violence being higher than for theft.

Explaining the sex difference The most widely accepted view of sex differences in crime emphasizes the greater social bonding and integration of girls within the family and the school. Social control theory began with the work of Travis Hirschi (1969) who profoundly changed criminological thinking about delinquents when he insisted that the correct question to ask was not ‘Why do they do it?’ but rather ‘Why do we not all do it?’ For him the gratifications of crime were immediate, obvious and required no special learning—revenge without court delay, consumption without labour, sex without dating. Crimes of all varieties are marked by the immediate gratification of desire— they are manifestations of a mind that will not wait. What prevents the rest of us from taking this route, he argued, is our bonds to conventional society and our understanding that lawbreaking will bring with it the rupture of those bonds to family, friends, employment and community. The root cause of delinquency lies in the fact that some individuals fail to establish strong enough social bonds to control callous selfishness. This deficiency in bonding begins with parents who fail to enmesh the child in a warm relationship that commands their desire to live up to parental expectations. They neither monitor their children’s whereabouts nor care about

Mhoc07.fm Page 215 Friday, December 14, 2001 10:17 AM

  :     their activities and companions. Later, Gottfredson and Hirschi (1990) extended this argument to highlight the importance of internalized social control in the form of self-control. It is self-control that is most strongly implicated in criminality beyond childhood when parental monitoring ceases. They argue that the greater family bonding and control experienced by girls results directly in lower levels of delinquency during adolescence and, indirectly, in a higher level of self-control in adulthood that accounts for their lower engagement in crime. This belief in the pro-social effect of surveillance is visible also in Adler’s ‘liberation’ hypothesis of female crime. She stressed that as women’s economic opportunities paralleled those of men, girls would be allowed increasing freedom to prepare them for their equal place beyond the home. The slackening of surveillance on girls was thought to lead to increased delinquent participation. Power-control theory (Hagan et al. 1987) is a reformulation of the social control position which similarly argues that girls have been subject to greater social control than are boys. Power relations between husband and wife derive from their economic roles and are linked to the kinds of socialization practices used with sons and daughters. In traditional patriarchal families in which the father is the main provider, it is the mother who exercises major control over the children. Mothers control their daughters more than their sons because girls must be kept closer to home to prepare them for their future roles as homemakers and consequently they experience tighter surveillance. Sons are especially likely to be free from control in families where the father has greater authority (within and beyond the home) than the mother. In addition, boys in traditional families are encouraged to develop social lives beyond the home and to take risks in preparation for their future lives in the market-place. In less traditional homes, where the mother works in a position of some authority herself, the power of husband and wife is more balanced. The egalitarian relationships between husband and wife attenuates this pattern of differential child-rearing and there is less distinction made between the way in which sons and daughters are treated. This leads to the expectation that sex differences in delinquency should be less marked in non-traditional families. Attempts to test the theory have not provided unanimous support (Hill and Atkinson 1988; Jensen and Thompson 1990; Morash and Chesney-Lind 1991; Singer and Levine 1988). This is not to say that social control is unimportant—data have confirmed the importance of parental monitoring on delinquency in both sexes (Hagan 1989; Hirschi 1969; LaGrange and Silverman 1999). But there is no strong support for the idea that sex differences in crime result from unbalanced power relations between parents. These theories are unanimous in at least one regard—that girls and women are more tightly controlled and constrained in their behaviour than are boys and men. This sex difference is ascribed, in some form or another, to patriarchy and paternalism. Evolutionists not only concur that females have been the object of special control (exercised directly by males or by related females who are themselves dependent upon male resources) but offer an explanation of why this has occurred (Hrdy 1997;

Mhoc07.fm Page 216 Friday, December 14, 2001 10:17 AM


    

Smuts 1995). Male control of female sexuality has been an abiding feature of human history stemming from concealed ovulation, uncertainty of paternity, males’ monopoly of resources and the rise of patriarchy. The value of women has been traditionally established by their reproductive worth. Females have been guarded more closely than males (Wilson and Daly 1992), though they have often attempted to subvert such control (Hrdy 1997). But surveillance is not socialization. Is there evidence that girls are actively encouraged to behave in a more law-abiding way than are boys? Sex differences in socialization, which have been examined in hundreds of studies, were the object of a meta-analytic review by Lytton and Romney (1991). They examined 20 areas of socialization (including encouragement of dependency and achievement, and discouragement of aggression) and found no significant differences (with the exception of sex-typed chores or toys) between the degree of socialization directed at sons and daughters. The specific trait that is postulated, by at least two theories, to be subject to differential socialization is risk taking. Hagan’s power-control theory argues that boys are socialized into risk taking in preparation for their future roles in the workplace while girls are encouraged to avoid risk and hence reproduce the gender relations of their parents’ generation. Gottfredson and Hirschi’s general theory of crime conceptualizes low self-control individuals as ‘impulsive, insensitive, physical (as opposed to mental), risk-taking, short-sighted and nonverbal’ (Gottfredson and Hirschi 1990, p. 90). Do these qualities really coalesce to form a single dimension that we might call ‘self-control’? This disparate assortment of traits has been the subject of conceptual and empirical examination and no final consensus has yet been achieved. Suffice it to say that impulsivity and risk taking have emerged as the most important of these proposed predictors of delinquency and they are also the ones on which sex differences are greatest (Arneklev et al. 1993; Longshore et al. 1996; Piquero and Rosay 1998). In a study which decomposed self-control into seven subscales in order to examine links with different forms of offending, the authors concluded that ‘The single dimension on which the two gender groups differ most, reported as the most influential predictor, was identified as “risk seeking” for delinquency and for all of the offense-specific subtypes’ (LaGrange and Silverman 1999, p. 58). However, while mainstream sociological accounts emphasize that such differences are the result of differential parental treatment and therefore subject to cultural variability, an evolutionary viewpoint suggests otherwise. Similar sex differences are also apparent in most species that show imbalanced parental investment (Daly and Wilson 1988). In humans, sex differences in risk taking are universal (Brown 1991). Developmentally, they appear by the second year of life, well before girls and boys are capable of reliably discriminating the sexes or knowing which behaviours are more characteristic of one than the other (see Hoyenga and Hoyenga 1993; Ruble and Martin 1998). We have already seen, in Chapter 3, that sex differences in fear are predictable in a species, such as our own, with biased parental investment.

Mhoc07.fm Page 217 Friday, December 14, 2001 10:17 AM

  :     A women’s lower threshold for fear derives from the greater importance of her own survival if she is to realize the full potential of the restricted number of offspring that she can produce, relative to a man. So sociologists’ emphasis on risk taking is entirely compatible with my own argument about fear because risk taking can be seen as an inverse measure of fear. But if sociologists are right about the predominance of nurture over nature then they would have to find a remarkable uniformity across cultures in the degree to which boys and girls are differentially socialized to explain the remarkable uniformity of the difference in their offending rates. And so far the data have not been persuasive.

Theft and assault: a common denominator Female crime can be seen as a desperate competition for access to scarce resources. Although female–female competition is underemphasized in previous accounts of crime, we have seen that access to resources is the chief limiting factor in female reproductive success. Although both sexes clearly require resources sufficient to sustain themselves, among females these needs are augmented by the extra calories required to sustain ovulation and support pregnancy, lactation and the additional needs of a growing, dependent child. Females’ strategies for acquiring resources lie at the heart of any evolutionary analysis and especially so for humans, a slowmaturing species where children remain dependent for 15 to 18 years. A woman has two avenues to obtaining necessary resources: through attachment to a provisioning male or through her own independent efforts. The long period of infant dependency means that women typically benefit from assistance with provisioning during their reproductive years (Lancaster 1991). The child’s father is usually expected to assist in this provisioning. In many cultures, this stricture, together with the implications of concealed ovulation for paternal uncertainty, results in the claustration of women and in restricted access to economic independence and selfsufficiency. The level of social control imposed upon women is extremely high but the corresponding benefit is that their male partners are required to provision them. These benefits are so great that mothers and kinswomen in some cultures will collude in surgical procedures and social arrangements designed to ensure virginity and prevent infidelity (Dickeman 1997). However, in some cultures females provision themselves and their infants with only sporadic assistance from males. In human hunter-gatherer societies, women provide the bulk of calories consumed by the family because they can obtain adequate supplies of food by foraging in the vicinity of the village with nursing infants. Although men supply meat on an irregular basis, this provisioning may have less to do with family support than with men’s demonstration of status-enhancing hunting prowess (Hawkes 1990). In contemporary Western society, unlike pre-industrialized societies, single women cannot easily combine child care and productive labour. In some Western societies, state aid is provided for single mothers. The availability of such aid may depend upon the

Mhoc07.fm Page 218 Friday, December 14, 2001 10:17 AM

      number of children the women has already produced and the financial stipend is small—consistent with the government’s aim of encouraging mothers to be selfsufficient as early as possible. Women receiving such incomes effectively live in poverty with food supplies bought on a daily basis and with reliance on loans from extended family and friends (Stack 1974). Resource shortage drives innovative and risky ways of securing resources. One such avenue is crime (Cohen and Machalek 1988). Serious, victimful crime is highest in the lowest social classes (Braithwaite 1981; Elliott and Ageton 1980; Elliott and Huizinga 1983; Ellis 1988) and indicators of acute poverty (unemployment and welfare dependency) are most closely associated with greater involvement in crime (Brownfield 1986; Farnworth et al. 1994). Studies suggest that the relationship between crime and social class is similar for males and females (Canter 1982; Carlen 1988; Elliott and Huizinga 1983; Triplett and Jarjoura 1997). The typical female offender is ‘poor, undereducated, disproportionately a member of a minority group, and dependent on her limited resources for her own support and often the support of her children’ (Wilson and Herrnstein 1985, p. 124; see also Chapman 1980). Property and violent crime are twin routes to solving problems of resource shortage. While property crime is a woman’s direct attempt to secure her own resources, female–female violence often results from competition for resource-rich males. What makes this proposal novel is that criminological theories to date have forced an explanatory separation between the two forms of crime. Criminologists are aware that the two are highly correlated over time and space and equally aware that most offenders show a cafeteria approach to crime, engaging in both forms during the course of their offending career (Chaiken and Chaiken 1982; Osgood et al. 1988; Petersilia et al. 1977), yet they have been unable to offer a unitary account of their etiology. Typically property crime is linked to poverty or relative deprivation while violent crime has been explained by a variety of sociological and psychological constructs including subcultural norms, status frustration, high or low self-esteem, psychoanalytic defence mechanisms, genetic predisposition and brain malfunction. Sociological explanations, which are predicated upon social structure as the key variable, typically ignore individual and sex differences. Psychological explanations avoid the question of why violence is concentrated in some sectors of society more than others. The argument that I will offer suggests that poverty drives both forms of crime but that individual differences are also be expected. Fear varies between individual women and some will be more disposed to select risky strategies than others. Ecological factors such as the availability of resource-rich males and her ability to attract them also play a part. But the common denominator is competition driven by resource scarcity.

Women and property crime The past thirty years have seen a marked increase in women’s involvement in petty property and drug-related crimes relative to the increase among men

Mhoc07.fm Page 219 Friday, December 14, 2001 10:17 AM

  :     (Chesney-Lind and Shelden 1992; Steffensmeier and Allan 1996). Women’s rates began to approach those of men by 1997 with the percentage of female involvement at 46 per cent for fraud, 47 per cent for embezzlement, 39 per cent for forgery and 35 per cent for larceny-theft (United States Department of Justice 1998). Lest we fall into the trap of imagining a female executive busily transferring funds and stocks into her account on a computer (Daly 1989), these offences encompass passing bad cheques, defrauding an innkeeper, other thefts of service, credit-card fraud, welfare fraud, stealing from an employer and shoplifting. Criminal women are more likely to be lying to welfare agencies or stealing tonight’s dinner than transferring funds to offshore accounts. They are small-scale, low-take offences committed by women who are undereducated, unemployed, receiving welfare benefit and supporting dependent family members (Chesney-Lind 1986; Daly 1989; Steffensmeier 1980). Women’s profit from each offence is relatively small, leading to sporadic but highfrequency offending in times of need (English 1993). Studies of women’s fraud and embezzlement find that women are frequently motivated by emotional responsibility to the family and by economic need (Daly 1989; Zietz 1981). Gilfus (1992, p. 86), interviewing women in prison, found that ‘the women in the study considered their illegal activities to be a form of work which is undertaken primarily from economic necessity to support partners, children and addictions.’ This is borne out by British work with imprisoned women which similarly highlights the crucial role of poverty and their view of theft as a form of incomegenerating work: I’d get my children ready for school in the morning, go to work, collect the children after school, take them home, cook them supper, entertain them, bathe them, put them to bed—and, Bob’s your uncle, it was my turn to go and have some fun. . . . I always went to work sober though. By work I mean shoplifting: I thought of it as my work time, and I treated it accordingly. (Carlen 1988, p. 78)

In her study of Milwaukee street hustlers (women involved in prostitution, fraud, forgery, embezzlement and larceny), Miller also notes: ‘Looked at from the point of view of women themselves, however, hustling is work that someone who has not shared their precarious lives has defined as in violation of the law’ (Miller 1986, p. 10). Between 1960 and 1990, the rise in female crime occurred almost exclusively in the area of minor property crimes (Steffensmeier 1993; Steffensmeier and Allan 1996). Women’s criminal involvement in property crime (24 per cent of all female arrests in 1997) became concentrated in this area relative to that of men (12 per cent of all male arrests). This 30-year period corresponds to the period where illegitimacy and divorce rates rose, creating a population of mothers and children living in poverty (Chesney-Lind 1997; Daly 1989). Steffensmeier and Allan (1996, p. 470) argue that petty property crimes ‘are less likely to result from workforce gains than from economic pressures on women that have been aggravated by heightened rates of divorce, illegitimacy and female-headed households.’ Boritch and Hagan (1990,

Mhoc07.fm Page 220 Friday, December 14, 2001 10:17 AM

      p. 569) concur: ‘increases in female property crime rates are more a result of economic marginalisation of women than an expansion of their criminal opportunity.’ Both cross-sectional and longitudinal studies of female arrests for property offences confirm that the increase is a function of adverse economic circumstances affecting females, including occupational segregation, a rise in female-headed households and rises in the rates of illegitimacy and female unemployment (Steffensmeier and Streifel 1992). Despite women’s entry into the workforce, they remain in low-paid, irregular and part-time work. As a result of divorce and male abandonment, they are less likely to be supported by male partners and take more responsibility for child care than they did 30 years ago (Dabelko and Sheak 1992; Kitson and Morgan 1990). The offending style of women is brought into sharper focus when we compare them to men. Unlike the low-profile, mundane involvement of women, men use their criminal profits to furnish a conspicuous and lavish lifestyle. For men, the aim is not simply to pay the rent but to broadcast their over-the-top lifestyle in the local community and thereby to gain indebtedness, prestige and respect. Boone (1998), using Zahavi’s (1975) notion of strategic handicap, argues that apparently ‘wasteful’ deployment of economic resources functions to elevate status among males. As one convicted criminal put it: ‘They think I’m crazy but they look up to me because I dress good, I keep cash, I’ve got women, and I don’t work’ (Willwerth 1974, p. 100). To neighbourhood men, accustomed to long bouts of unemployment and reliance on their partner’s welfare checks, such ostentatious style (reflected in partying, gambling, drinking, drug use and the purchase of expensive brand-name clothing) makes a deep impression. Successful criminals have not simply achieved the middleclass dream (home ownership, steady income) but have far exceeded it: Straight people don’t understand. I mean, they think dudes is after the things straight people got. It ain’t that at all. People in the life ain’t looking for no home and grass in the front yard and shit like that. We the show people. The glamour people. Come on the set with the finest car, the finest woman, the finest vines. Hear people talking about you. Hear the bar get quiet when you walk in the door. Throw down a yard and tell everyone to drink up. (Wideman 1985, p. 131)

Robbery is a distinctively male crime in which resource acquisition and violence are merged in the pursuit of status enhancement (see Campbell 1993; Lejeune 1977). As one commentator expressed the attractions of robbery: ‘Unless it is given sense as a way of elaborating, perhaps celebrating, distinctively male forms of action and ways of being, stickup has almost no appeal at all’ (Katz 1988, p. 247). The thrill associated with the moment of confrontation and dominance is an important attraction. Indeed robbers often begin their careers with less hazardous forms of action such as mugging and jackrolling. Unlike most other crimes, robbery requires a declaration of criminal intent and this opens up the possibility that the victim may not comply. When the offender demands that the victim ‘hand it over’ or ‘give it up’

Mhoc07.fm Page 221 Friday, December 14, 2001 10:17 AM

  :     he is referring not only to money but to control. Robbery is about exploitation and humiliation and robbers are acutely sensitive to this aspect of the crime. To be exploited is to invite contempt not sympathy and to rob is to demonstrate superiority. One robber explained: ‘The amount of money don’t make no difference: you just can’t let people get out on you. It don’t matter who you are or what you are, they’ll try and you gotta stop them. At all times you gotta stop them’ (Allen 1978, p. 181). Although injury in robberies is broadly related to degree of victim resistance (Cook 1987; Zimring and Zuehl 1986), violence may be employed gratuitously before or after the crime in the interests of impressing peers or deterring the possibility of a double cross from co-offenders (Katz 1988). The financial motive that spurs men to robbery is not borne out of desperation. Only 18 per cent of robbers say that they need the money for themselves or their families (Walsh 1986). A core concern for robbers is the lavish and conspicuous disbursement of money (Conklin 1972). As one street woman described it: ‘With money comes power, with money comes respect. You are a big dope man, you know what I’m saying, one of the best. And you had to kill a few people to get there, you had to rob a few people to get there. That even gives you more respect out there’ (Taylor 1993, p. 87). The spending of surplus cash both impresses and indebts others. It also sets in motion a cycle of ‘earning and burning’ in which robbery proceeds are immediately spent on high living, triggering imminent poverty and loss of status which requires another robbery to re-establish the high-roller, status-enhancing lifestyle. It ain’t the money or the cars or the women. It’s about all that but that ain’t what it’s deep down about. Cats blow a thousand a night when they on top. The money ain’t nothing. You just use the money to make your play. To show people you are the best. (Wideman 1985, p. 131)

Such posturing and unabashed ostentation does not provoke unalloyed admiration from their female counterparts. For women, the aim is to conceal rather than advertise illegal income: we ain’t wearing no street signs in diamonds like the fellas, guys like to show off too much, guys do shit that’s unnecessary. Girls like to save their monay; now you got ho’s that fuck up their money but more girls save and do smart shit rather than buy six, seven cars and go to Vegas and just blow big monay trying to impress their boys . . . Now these here boys is out of control. They may as well have a sign saying I am rolling and living large. See, that’s what girls won’t do, shit like this here . . . these niggahs is showing the world. The hook gonna be all over them, look at that Vette . . . Look at their Vette, it’s got spoilers, the hubcaps cost more than some poor peoples pay for their whole damn car. (Taylor 1993, pp. 66, 97)

For men, the status engendered by a high-rolling life exceeds the risks of retribution and incarceration. It comes as no surprise given what we know of evolutionary theory that, for women, it does not. Relative to men’s property crime, women’s is more responsive to personal or family need (rather than desire for a hedonic lifestyle), is viewed as a form of work

Mhoc07.fm Page 222 Friday, December 14, 2001 10:17 AM

      (rather than adventure), involves more frequent offences with smaller returns and is characterized by concealment (rather than the advertising) of their criminal activities. Women need resources as an end in themselves. Men use resources as a means to status and respect.

Women and drugs Drug sales offences are distinct from property crimes but they share the same characteristic of being significantly less risky than violent offences. Drug selling (from the point of view of the offender, if not of society at large) can be seen as a victimless crime—the buyer voluntarily parts with his or her money in an economic, although illegal, transaction. The crime involves no force or threat of force and as such is unlikely to lead to counter-violence or direct injury. But drugs are also distinctive because the motivation can derive from physiological as well as financial need. Many low-level dealers are also users and acquiring money is no longer volitional but driven by the dealer’s drug habit. Initially, the pleasurable effects of drugs derive from their direct action on the dopaminergic reward system of the brain but once addiction has set in, the user needs the drugs as much as she wants them. Drugs are essential to simply feel normal—addicts still experience a temporary high after smoking crack but this quickly evaporates into a return to semi-normal functioning prior to the onset of withdrawal symptoms. The motivation for criminal involvement is the necessity of procuring more drugs rather than the more obvious survival needs (food, shelter, clothing) that motivate other forms of crime. Human relationships become casualties of drug use and especially so with crack cocaine, as one addict explained: My kids hate me . . . They should, I ain’t shit. I tried to sell my daughter, who is fine, she is just beautiful, and I was trying to sell her for some rocks three years ago. My children was taken from me, and they was right, the judge said I was evil; well I don’t know ’bout all that, I know I am a dope fiend . . . AIDS, fiending, beatings, fuckings, shootings, whatever . . . I am already dead, even sometimes it’s hard to remember when things were so bad. The dope just takes over, it’s wild, kinda strange. (Taylor 1993, pp. 74–5)

During the 1980s, women became increasingly involved in the US drug economy and this involvement paralleled the rising popularity and availability of crack cocaine. In New York City in 1988, 56 per cent of the 2280 inmates identified as crack addicted were women (Maher and Curtis 1992). Women represent 17 per cent of those arrested nationally for drug abuse offences (United States Department of Justice 1998). Do we see in women’s involvement in crack dealing any evidence that they have progressed to the higher echelons of the crime or do these women, like their property-offending counterparts, show a profile of low-level, small-scale involvement driven by poverty? The organization of drug selling is structured hierarchically. Once the drugs have been smuggled into the country, the drug ‘owners’ employ a

Mhoc07.fm Page 223 Friday, December 14, 2001 10:17 AM

  :     number of managers to act as conduits to street-level sellers. Sellers in turn are assisted by lower-level employees in the roles of runners (who keep the sellers supplied on a continuous basis and lower the quantity of drug that he is holding at any one time), lookers (who check for police or other threats to the drug transaction), steerers (who relay customers to the sellers), holders (who are charged with guarding either the money or the drugs) and enforcers (who intervene if the situation becomes disordered or violent). Adler (1993) found that women are rarely involved in the highest level drug dealing, such as smuggling, which carries with it the highest profits. Instead they are small traders (transactions of less than ten dollars), but many women are doing approximately twenty such deals every day. They report hectic daily lives complicated by children, pregnancy and unemployment (English 1993). In Maher and Daly’s (1996) study of Brooklyn crack women, they discovered not a single women who owned her own drug business. Again, women were concentrated in the lowest levels of work. When women were allowed to sell the drug, it was as a result of a strategic decision taken by male owners and managers in response to a shortage of male workers or because the area was under close police surveillance and women were less likely to be searched by male officers. Even then, women had to apply for such temporary openings. If they were accepted they were entrusted with only small quantities of drugs which limited their potential sales and they were given daylight shifts, which were far less lucrative and more risky than night shifts. As one woman explained: ‘I am currently working White Top (crack). They have a five bundle limit. It might take me an hour or two to sell that, or sometimes as quick as half an hour. I got to ask if I can work. They say yes or no’ (Maher and Daly 1996, p. 475). The majority of women work as steerers, recommending a particular brand of drug to neighbourhood newcomers in exchange for a few dollars. To augment their income, the women also sell syringes. Some are procured from family members, diabetic acquaintances or needle exchange schemes, but often the women collect used syringes from the neighbourhood, clean them with bleach and water and resell them. Other women make small amounts of money or obtain free drugs by sharing their crack stems and pipes with buyers who are reluctant to be caught with drug paraphernalia. Others sell condoms, which they receive free at health centres to drug customers, for as little as 25 cents each. More ambitious women who try to set up their homes commercially as ‘galleries’ for drug consumption are usurped by male dealers. Initially agreeing to pay them rent or a share of the profits, the men default on the agreement and take over the apartment. The women find themselves forced out as a result of the unsavoury behaviour of the customers, direct violence or police action (the police seem especially punitive towards female-owned galleries as a result of their perception that women cannot control the male customers). Many women make small sums of money ‘copping’ drugs for neighbourhood ‘outsiders’. This makes them highly vulnerable to arrest but provides a useful, if unpaid, service to street sellers. Sometimes the woman makes an outright charge to the customer for this service, at other times the woman inflates the cost of the drugs

Mhoc07.fm Page 224 Friday, December 14, 2001 10:17 AM

      and pockets the difference. The latter is a more successful strategy since men often refuse to pay up when the women return with the drugs. Sometimes women adulterate the drugs after purchase and keep the excess for their own use. Or they return to the client minus the drugs and the money claiming that they have been ripped off. But the chief means of financing their drug habit for most women is prostitution. Given the ‘institutionalized sexism of the underworld’ (Steffensmeier and Terry 1986), this is one of the few money-generating niches open to them. As we shall see in the next section it is also one that generates strong competition between women.

Women and violence It takes a sizeable income to maintain a drug habit. Prostitution offers one avenue of money-making and one that is particularly attractive because it requires no special training, no specific equipment, no initial capital and no planning. Prostitution can be undertaken as and when the need arises and the rates that are set can vary depending upon the woman’s desperation and the customer’s ability to pay. During the 1980s and early 1990s, crack use in New York City peaked. Women became more and more desperate for money at the same time as male crack users became less and less able to pay for sex. The sheer number of women willing to prostitute themselves led to a buyer’s market in which prices for sexual services were driven lower and lower. Women are reluctant to admit to ‘crack dates’ (the direct exchange of sex for drugs), but many acknowledge the reality of their desperation. Okay. If I was stressin’ real bad and a guy tell me that he’s got two or three caps I might do it. I’m sorry. I’m no better than anyone else—shit, I don’t put myself on a pedestal, if I was stressin’ bad enough I’d do it. But then after I do it—I think—you motherfucker man you give me two bottles and I gotta suck your dick. (Maher and Curtis 1992, p. 242)

Others are willing to drop their price to rock bottom to undercut the opposition and get business. Normally you try to ask fifteen for a blow job. These girls do it for five—less. That’s, that’s—you know—self degradation. You can’t even blame the men because of course if you can get a blow job for five instead of ten or fifteen, you’re going to take it . . . We got a girl we call ‘Two Dollar Mindy’. She go out . . . Mindy will stay with a date forty-five minutes and come back and wanna get down on (share) crack . . . Two Dollar Mindy don’t care about nobody, nuttin—long as she got two dollars. (Maher and Curtis 1992, p. 240)

The steep competition for customers together with price deflation intensifies competition between women. They are effectively vying for the financial resources of men. Among these women sustained friendship is almost unknown—‘there’s no friends, only associates’ (Maher 1997, p. 38). The women view one another with hostility and mistrust. This explicit economic competition between the women for ‘dates’ effectively eliminates female friendship. Indeed, women will even rob one

Mhoc07.fm Page 225 Friday, December 14, 2001 10:17 AM

  :     another when dates become especially hard to get. ‘Increasingly, some of the interactions between the women are characterized by mutual hostility and competition— to the point where when they are not ‘vicced’ first by dates or male users, some will even ‘vic’ each other for increasingly hard-to-get cash or drugs’ (Maher and Curtis 1992, p. 235). In addition to robbery, women described direct assaults on rivals who undercut the already low rates. Miller’s (1986) description of street women also captures the sense of intense competition between prostitutes but the object of their rivalry is not a customer but ‘their man’. In contrast to Maher’s Brooklyn study, the Milwaukee women very rarely operated without some form of relationship with a man. Solitary hustling only occurred after a woman had been badly victimized by a previous partner or when he was imprisoned. None of the women who had tried solo hustling repeated the experience. A ‘man’ is not exactly a pimp. A pimp is a flamboyant character who manages a stable of prostitutes as an entrepreneur and administrator. He works exclusively with prostitutes, often in the more desirable economic locations that are used almost solely for the purpose of prostitution. A ‘man’ is a woman’s lover. In partnership, they exploit whatever criminal opportunities present themselves in their local, usually severely depressed, neighbourhood but a regular feature of their lifestyle is prostitution. Unlike a pimp, a man is a line manager who is on hand to supervise and (ideally) protect the woman. The women themselves clearly distinguish between the two and rarely refer to their man as a pimp except when seeking to convey a negative impression of him: ‘He was nothing but a pimp’ was a regular complaint when such relationships terminated acrimoniously. A woman’s relationship with her man is an important first step in her life on the streets and women are extremely sensitive to the threat posed to that relationship by other women who may seek to take him away. Miller notes that the women share a strong desire for a good man who will provide for them, assist in raising their children and, above all, be faithful. Other women pose a constant threat to this dream. And I could just sit with men and just talk. I couldn’t say that about a lot of female friends. We would be partyin’ and the next thing you know this woman would try to steal my “man”. That’s enough right there for me to really hurt somebody. I caught my friend one time in bed with my “man”. I was mad enough to want to kill them both, but, instead, I jumped on her. Women are just so petty. (Miller 1986, p. 161)

The competition for the attention of a good man leads to considerable mistrust and hostility among the women, as Miller explains: Moreover, if the street women interviewed distrust individual men, they distrust women as a group even more. Again and again, women told me that they had few, if any, female friends and that women, generally, were “snitch-bitches.” Women are seen as jealous gossipy troublemakers who are both more emotional and less rational than men. There is, of course, a great deal of competition among them for good ‘men’ and they often undercut each other’s attempts to sustain relationships with ‘men’ . . . There is then little sense of

Mhoc07.fm Page 226 Friday, December 14, 2001 10:17 AM


    

camaraderie among female street hustlers, little sense of political identification with women generally. (Miller 1986, pp. 160–1)

If the woman is unable to lay exclusive claim to her man, she may reluctantly share him rather than losing him completely. She may agree to him taking a second wife—known as a ‘wife-in-law’. She then becomes ‘bottom woman’ and the assistance she receives with both household chores and income-generating prostitution means that she can work less hard. The possibility of one or more wives-in-law means that the ideal situation of a ‘monogamous’ hustling partnership is rarely long-lasting. This is a cause of considerable distress to many women who depend heavily upon the exclusivity of this attachment for romantic and emotional support. Many have come from families where they have experienced rejection and abuse that makes them particularly vulnerable to the affection that the man initially proffers. The advent of co-wives leads to intense rivalry and makes heavy demands upon the man. He especially needs to meet the sexual and socio-emotional needs of his women without creating in-fighting and jealousy among them. . . . If he is unsuccessful in achieving this delicate balance or is arrested, the pseudo-family dissolves. It may also break up due to the active intervention of one of the women. It is not uncommon for a jealous woman to put one of her wives-in-law in a situation that leads to her arrest. (Miller 1986, p. 41)

The tense and competitive relationships between marginalized women are visible also in the street gang. About 10 per cent of gang members in the United States are young women. Although almost always affiliated with a male gang, female gangs operate as semi-autonomous units with their own rules, disciplinary procedures, initiation rites and roles. Gang members, like street women in general, are remarkably traditional in their desires—they dream of having a nice apartment, beautifully dressed children and a man who will love and care for them. The level of violence among female gang members is much lower than among males and while males will exploit their violence for profit (through robbery, intimidation and acting as enforcers for local drug operations), the girls confine themselves chiefly to fist or knife fights and their targets tend to be other girls or female gangs. During the time that I was doing ethnographic work with these girls I was struck forcibly by the extent to which their aggression was motivated by jealousy or proprietary ownership of boys (Campbell 1984). Expecting to find independent and sexually unrestricted young women, I found a set of traditional values whereby enemies were disparaged as whores, manifest promiscuity within the gang was subject to sanction and attachment to a provisioning male was very highly valued. Within-gang jealousy between the girls was both intense and frequent. I told you I choked a girl because I had a feeling that something was going on. That’s the one I said I can’t stand. And I wanted to choke her to death. I went and I broke a pizza shop window because I thought the owner had been my friend and I thought he cared for that chick’s life. Because I knew she ran in there and she probably said ‘Connie wants to kill me.’ . . . I ran to the corner and I didn’t see like nothing, so I looked and I saw a pizza

Mhoc07.fm Page 227 Friday, December 14, 2001 10:17 AM

  :     shop and waited outside. She came out ‘Oh, Connie, please, please. Please.’ You know, ‘What did I do? What did I do?’ To me, Gino is going like this, showing her how to go on a bike and I had never even done that. So to me, I thought he was showing her special attention. (Campbell 1984, p. 73)

But girls outside the gang can generate even more anger. Girls feel a strong sense of ownership about the boys in the gang to whom they are affiliated and sexual relationships are usually confined to them. The boys, however, are always looking out for new women and newcomers to the community represent a special threat to the gang’s ownership of ‘their’ boys. Sometimes, the whole gang would converge on a single unfortunate girl who had shown too much interest in one of their male counterparts. Like when we hang out, all of us got a guy that hangs out with them. That’s her man and that’s her boy. Like all of us already had one of the guys from the Turban Saints. All of us had one of them. We used to fight a lot when they used to cut out and go see other girls, you know. That’s when we really used to go off. (Campbell 1984, pp. 257–8)

One of the gangs that I spent time with was the Five Percenters, a group who had adopted the teachings of the Nation of Islam. According to their laws, men are ‘gods’ or ‘suns’ and women represent the earth. According to the ‘Six Rules of the Black Woman’, women must not prostitute themselves, may not use birth control, may not have sex with anyone but their ‘god’, must obey their god in all things, must raise her children according to the teachings of her god and must maintain scrupulous personal hygiene. The gods, however, may have as many earths as they wish providing that they can care for them and their children. Sun-Africa was a 16year-old girl who had become the third wife of a Five Percenter called Shamar. Tensions between the co-wives were frequent but especially so between Sun-Africa and his first wife Lady who refused to fully subscribe to Islam and was described by Sun-Africa disdainfully as ‘flashy’. Shamar was killed in the course of burglary and his death precipitated an emotional and violent confrontation between the two wives when Sun-Africa emerged from his mother’s house the morning after his death. I was just talking to myself. ‘Please, he just going to knock on the door, he going to come home, he going to come home.’ And he just never come home and everything, you know? I spent the night there, you know? And I waited and waited and waited. And then the next morning when I woke up, that’s when we had the fight. Me and the other girl—Lady, you know. Soon as I got up, cleaned myself up, and went outside, there was a whole hallway full of people, talking about him being cremated. And they were arguing and everything, right? And so I had on his crown [a knitted hat worn by the gods to cover their head out of respect for Allah], right? I had put it on my head and everything ’cos it make me feel a little better. I went outside. And then Lady asked me for it, she asked me for it in the wrong way ’cos she said ‘You got Shamar’s thing?’ And then she just started getting all smart and stuff, and then all her girlfriends started talking ’bout ‘You ain’t giving it up? Then you going to have to get your ass kicked.’ And all kind of shit. And, you know, I was saying, ‘Yeah. You all try it ’cos you know I don’t play that. That’s right.’ And I was ready

Mhoc07.fm Page 228 Friday, December 14, 2001 10:17 AM

      for all their little shit. And being that they started it off, I finished it for their ass. She was up in my face, right? Like she was putting her breasts against me, you know? And I was standing up and she was saying ‘You ain’t giving it to me?’ And I said ‘No!’ And I said ‘No!’ real smart, you know, right? So she get up in my face and I just punched her. I had to hit her first, you know. I just had to hit her and like we was fighting and everything and she was pulling my hair, right? And like, she pulled a whole patch of my hair out, and I felt my hair rip up from my skin and I just broke. I took her head and started banging it against the concrete. (Campbell 1984, pp. 199–200)

As with American street hustlers and women from all over the world, sharing a husband leads to trouble. Women are in competition with one another and where men are a necessary resource for economic survival, they are the objects of a strong sense of ownership and a violent reluctance to share. Marriage is a zero-sum game for co-wives; the more one takes, the less is left for the others.

The development of crime The factors that drive males and females into crime show far more similarities than differences which is perhaps not surprising in light of the very high correlation across place and time in their offending patterns. Given the covariation in their rates, it would be strange indeed if the very same behaviour were predicted by one set of factors in men and quite another in women. Boys and girls who become delinquent show a marked similarity in their background (Elliott et al. 1985; Hindelang 1973; Liu and Kaplan 1999; Loeber and Stouthamer-Loeber 1986; Messner and Krohn 1990; Smith and Paternoster 1987; Thornberry et al. 1991). The list of variables that predict later criminal behaviour is a familiar one—crime is an area where social science has confirmed common sense. Those most at risk for crime come from impoverished and disorganized neighbourhoods composed of large, singleparent or parentally-conflicted families who exercise little effective control over the children’s behaviour and are inconsistent, lax or abusive in their parenting practices (Farrington 1998; Werner and Smith 1992). Family processes have been found to be one of the strongest predictors of later delinquency (Farrington 1995). The difficulty for researchers has been that these risk factors tend to travel together and consequently it is difficult to find the unique impact of a single variable. (Usually this must be done statistically rather than by finding anomalous families who possess a single risk factor.) The key question is why these variables correlate and the answer is resource shortage. Teenage births occur disproportionately among the poor where the availability of resource-rich men is shortest and where deferment of child-bearing in the hope of finding a supporting male is a risky strategy. Young women in such neighbourhoods frequently place a high value on traditional female roles and motherhood is intrinsically rewarding as well as providing a stepping-stone to the adult world. A high proportion of men are unemployed or sporadically employed and cannot offer

Mhoc07.fm Page 229 Friday, December 14, 2001 10:17 AM

  :     the sustained economic resources necessary to support a child. Yet children of teenage mothers are more likely than those of older, married women to employ poor child-rearing methods, creating a cycle of inadequate parenting that increases the risk of delinquency (Morash and Rucker 1989). Where a child begins life with two parents, she may lose one through separation or divorce. Episodes of marital conflict frequently revolve around issues of housekeeping, money, children, social activities and sex (Straus et al. 1980). Living on the edge of poverty triggers disputes about the allocation of resources, workload and rights to recreation that become all the more critical where children are involved. Poverty and overcrowding not only provide the fodder for arguments and fights but exert a diffuse daily stress on those who must live with it. This stress in turn leads to emotional explosions which can become normative within a family or even a community so that raised voices, blows and tears become mundane and the volatility assumes an upward spiral (Baum and Paulus 1987; Bernard 1990). Evidence strongly suggests that it is parental conflict rather than divorce that places children at greater risk for delinquency (Block et al. 1986). Homes that are broken by a parent’s death (rather than their abandonment) are much less likely to produce delinquent children (Rutter and Giller 1983; Wells and Rankin 1991). Rates of delinquency are similarly high in broken homes without an affectionate mother and in homes characterized by parental conflict. They are much lower in unconflicted two-parent homes and in broken homes where the mother is affectionate to the child (McCord 1982). Even in intact homes, parents’ preoccupation with the stresses of crowding and poverty reduce the amount of time and positive attention that they can give their children (Conger et al. 1994). So fundamental is this effect that it has even been observed in animal studies. When food is scarce and rhesus monkey mothers are required to spend increased time and energy foraging, they become less solicitous of their offspring’s needs (Rosenblum and Paully 1984). In humans also, poverty and its associated stressors is associated with child maltreatment, the use of powerassertive disciplinary practices, and a decline in the expression of affection and maternal responsiveness to infants’ needs (Elder et al. 1985; McLoyd 1990; Sampson and Laub 1993). These practices—poor supervision, harsh and erratic punishment, parental neglect and lack of affection—are all associated with a higher risk of delinquency (see Farrington 1998). Physical abuse of children is also significantly associated with later criminality including both violent and non-violent offending (Malinosky-Rummell and Hansen 1993) and, like the other variables we have discussed, this is true for both boys and girls (Widom 1989). To make matters worse, parents living in such conditions are more likely than others to have to cope with a ‘difficult’ baby, leading to parental stress and a further deterioration in the child’s chances of a successful future. Maternal drug use and poor prenatal nutrition and care can disrupt neural development in the foetus (Moffitt 1993). Postnatally, it can be further disrupted by inadequate stimulation,

Mhoc07.fm Page 230 Friday, December 14, 2001 10:17 AM

      absence of affection and by neglect and abuse. These children show a variety of characteristics including overactivity, clumsiness, inattention, irritability, impulsivity, schedule irregularity, slowness in reaching developmental milestones such as walking and talking, poor verbal skills and slow learning. These low-birthweight infants arrive before their parents are ready for them, their crying is more disturbing and irritating, they are less satisfying to feed and hold and they demand more care than healthy babies (Hertzig 1983; Tinsley and Parke 1983). The tragedy, as Moffitt (1993, p. 681) notes, is that ‘Unfortunately, children with cognitive and temperamental disadvantages are not generally born into supportive environments, nor do they even get a fair chance of being randomly assigned to good or bad environments . . . Vulnerable infants are disproportionately found in environments that will not be ameliorative because many of the sources of neural maldevelopment co-occur with family disadvantage or deviance.’ As they develop, the strain that such children place upon even the best parents should not be underestimated. In an ingenious study, Anderson et al. (1986) watched normal and conduct-disordered boys interact with mothers of boys from their own and the opposite group. There was no difference between the two sets of mothers in their reaction to conduct-disordered boys—both were more negative in their response than they were when interacting with ‘normal’ boys. Little wonder then that over time, mothers of ‘problem’ children become less and less involved in directing and enhancing their children’s behaviour (Maccoby and Jacklin 1983). These difficult children have a markedly higher rate of delinquency than do nonimpaired children (Moffitt 1990). This package of poverty, disrupted families, disturbed children and poor parenting practices derives from resource shortage. The route from poverty to child criminality is probably twofold. On the one hand, the link may be an indirect one through the parents’ reaction to deprivation. Poverty, overcrowding and conflict create a situation that makes effective and affectionate parenting difficult. Absorbed by adult concerns (finance, domestic arguments, stress), parents have little time to devote to stimulating or directing the course of their child’s development. Children, dependent and vulnerable, become the first casualties of family poverty. They grow up unconstrained by parental expectations or affectional ties, free to engage in the short-term gratification of acquisitive or aggressive impulses that they have not been taught to control. Bearing in mind that their neighbourhood peers are, to a greater or lesser extent, living in similar circumstances and acting in the same way, this behaviour is likely to be further reinforced by the effects of peer conformity. But resource shortage may also have a more direct link to children’s criminality. Children are astute readers of their local ecology. On a daily basis, they witness parental fights over money, the way in which their parents have no time or attention left over for them, the fact that better-off children have material goods that their own parents cannot buy, the way in which their parents spend money as soon as it arrives and do not plan for the future. From these observations children can decipher

Mhoc07.fm Page 231 Friday, December 14, 2001 10:17 AM

  :     resource shortage directly. The message that it sends is clear—the future is uncertain, money is scarce and unpredictable, take it whenever there is an opportunity and spend it when you have it. ‘Because the child perceives rewards as undependable, he or she learns to grasp immediately opportunities for short-term gratification rather than to defer them for future rewards’ (Vila 1994, p. 333). But these responses are modulated by sex differences in fear. This is a high-risk strategy and boys more than girls adopt it.

Women and crime: a unified account I noted earlier that the study of female crime had become confused because there was a lack of unanimity about what question was being posed and by the fact that a theory designed to answer one question does not necessarily do an inadequate job of answering the others. Does the formulation that I have offered go any way towards solving some of these problems? Most immediately, the present proposal speaks to the gender ratio problem— why do females commit less crime than males? The answer I have given dwells not upon socialization differences between the sexes but upon differences in fear and risk taking that are products of evolutionary pressures and manifest themselves cross-culturally. There are differences in the surveillance and the control of daughters and sons but the point I want to emphasize is that these differences should not be confused with the idea that parents inculcate different values about lawbreaking in the two sexes. Studies show that parents are not more punitive towards female than male aggression (indeed some studies suggest that parents may respond with greater disapproval to boys aggression). Girls and women are more closely controlled than males both within the family and beyond it. The reproductive value of girls is closely guarded even in today’s more egalitarian society and the fear of an unmarried pregnancy is very real in parents who have upwardly-mobile aspirations for their daughter. The state also colludes in enforcing a double standard that treats precocious female sexuality as more deviant and dangerous than that of boys (Chesney-Lind 1989). Girls are more often referred to court for status offences such as incorrigibility, running away from home, being in need of supervision and being beyond parental control. Such offences account for 25 per cent of juvenile female arrests compared to only 8 per cent of boys. The same double standard of treatment for women is as evident in the underworld as in the straight one. Male criminals assert that women make undependable and weak partners in crime (Steffensmeier and Terry 1986). When men involve women in crime, it is to exploit particularly female characteristics. As sexual objects they can keep a pimp, as apparently respectable home-owners they can supply a stable base for drug-dealing arrangements or they can be used sporadically to fill gaps in street sales when regular male sellers are imprisoned. That said, there is little evidence from ethnographic studies that women feel a deep sense of frustration or thwarted ambition about failing to reach

Mhoc07.fm Page 232 Friday, December 14, 2001 10:17 AM

      the highest echelon of the criminal subculture. Most criminal women hold traditional views about women’s role and many would give a higher priority to a stable relationship with a good man than to climbing to the top of an international drug cartel. A second stream of researchers have asked about within-sex differences—the ways in which criminal women are different from non-criminal women—while a third asks if women need a separate and distinct theory to explain their offending. We can consider these two issues together. Men and women share a marked similarity in their rates of crime across space, age and time. They also share the same set of risk markers for crime. Taken together, these facts suggest that women do not need a separate explanation of crime. They do, however, need some mechanism that explains why, given the same backgrounds, they are less likely than men to commit crime. This is particularly important since according to official statistics more women than men live in poverty and female-headed households massively outnumber father-headed households. This threshold difference, I believe, is a function of differences in fear and risk taking. Resource shortage drives crime and this is true for both males and females. Resource shortage is not only a part of their immediate motivation but plays a role in their early development where ecological factors set the parameters for a readiness to exploit short-term opportunities for crime. The chief differences between criminal and non-criminal women centre, as they do in males, upon variables related to resource shortage. But even in conditions of poverty we would expect variation in women’s willingness to engage in crime. These differences, I believe, stem from individual differences in fear. Women who are low in fear are more predisposed to the risk taking that goes with crime (LaGrange and Silverman 1999). Other researchers focus upon the rise in female crime and try to identify factors that explain this change. There is now considerable unanimity that the rise in female crime has been most marked in petty property crime and that this rise is best explained by the increased number of women living in poverty. Far from endorsing the view of the ‘new female criminal’, this view suggests that such women live troubled and chaotic lives characterized not so much by liberation as by entrapment in poverty and single motherhood. Far from yearning to head-up international corporations, such women have more modest and more traditional desires—for stability, a regular income and a supportive partner. Some of those who cannot achieve these desires turn to crime as a means of survival. Feminist criminologists’ tendency to turn them into heroines or victims will do little to materially improve the quality of these impoverished women’s lives.

Mhoc08.fm Page 233 Friday, December 14, 2001 10:18 AM

 


A coincidence of interests: Women and marriage

Appearances can be deceptive. For example, we think of depression as something that goes wrong with our view of the world. Depressed people do not see exciting opportunities lying ahead, they focus only on the terrible things that can go wrong. In fact, depressed people see the future in more realistic terms than the rest of us (Alloy and Abramson 1979). Nevertheless, none of us wants to be depressed because our rose-tinted everyday view is what makes life not just bearable but enjoyable. So it is with marriage. We see it as a partnership, a pooling of our resources, a striving for mutual happiness. Two become one. From an evolutionary and genetic perspective this is far from the truth. (But let’s not abandon our optimistic if erroneous belief—it is what keeps us striving each day for that just-over-the-next-hill fusing of souls.) The reproductive goals and strategies of men and women are different. Marriage is the triumph of compromise over individual satisfaction. Let us view this conflict between men and women at a gene’s eye level, starting with the pair of sex-determining chromosomes. Females receive an X chromosome from both their mother and father. Males receive an X chromosome from their mother and a Y from their father. This form of sex determination is not the only possible option. But once our species switched to a genetic method of sex determination (instead of relying on, for example, the temperature surrounding the egg as some reptiles do) it opened the door for sexually antagonist genes. These are genes that are advantageous for one sex but disadvantageous for the other. When X and Y chromosomes evolved, genes coding for features that were good for males but bad for females could position themselves on the Y chromosome and female-benefiting genes on the X chromosome. A genetic battle between the sexes could now begin. Y chromosomes only ever inhabit male bodies and can therefore carry traits that are good for males—no matter how bad they are for females. X chromosomes spend two-thirds of their time in female bodies and only one-third in male bodies. This means that they are selected twice as often for genes that are good for females rather than males. Suppose that the Y chromosome carried genes for killing sisters, it would never find itself in a female body and consequently could never act destructively to its own sex. The opposite could also occur. A brother-killing gene could be carried

Mhoc08.fm Page 234 Friday, December 14, 2001 10:18 AM

      on the father’s X chromosome. Since the father contributes a Y chromosome to sons, the X would never be passed to them and hence could not act self-destructively. But if the X received by a girl was somehow labelled as to its parental origin (more on this later) then it could arrange to be expressed only when it came from the father. (Remember it would not be expressed by the father because he received his X from his mother.) Another interesting consequence of the arrangement of sex chromosomes is that because X chromosomes outnumber Ys, they are in a much stronger position to fight back against any female-deleterious Y-carried gene. The Y chromosome is a sitting target for attack by its partner X chromosome. This seems to be the reason that the Y chromosome is so small (carrying two or three dozen genes compared to several thousand on other chromosomes). The less information it carries the better it is able to avoid attack. It serves principally to carry the SRY gene that determines the manufacture of testes and the inhibition of female reproductive organs. The embryonic testes release testosterone which is carried in the bloodstream and, once converted to oestrogen, is able to act on the developing brain. A great deal of maletypical behaviour, including sex and dominance, seems to be the indirect result of testosterone action rather than being genetically specified by the Y chromosome. SRY is one of the fastest evolving genes. Comparing humans to other primates, there is ten times more variation on this gene than on other genes. Vacquier et al. (1995) have tried to understand why this should be by studying the abalone, a marine snail in which the female lays her eggs in water to be subsequently fertilized by a male. The male’s sperm carries a protein called lysin which is released from the front of the sperm when it encounters an egg. Lysin’s job is to bore a hole though the matrix that surrounds the egg so that the sperm can enter, fuse and fertilize. Lysin is an asset to males. But it is a problem for females—it creates a channel that permits too many sperm to enter and it is also an open door to pathogens. These are twin threats to zygote survival The very rapid evolution of lysin seems to be a function of an escalating arms race between what males want and what females want. Lysin causes faster sperm entry and the female counters by changing the egg matrix which slows the sperm down. The male then evolves a new faster form of sperm entry that the female resists by changing the glycoprotein matrix to protect her egg from multiple fertilization. The arms race drives evolution at a faster and faster pace as the male and female attempt to out-manoeuvre one another. These conflicts move even faster in animals with internal fertilization, resulting in a very rapid evolution of the male and female reproductive tracts and genitalia (Eberhard 1996). Such evolutionary conflicts are called Red Queen arms races after the famous passage in which Alice finds out that merely staying in one place takes all the running one can do. A brilliant study by William Rice shows just how dangerous it can be to pause for breath. To examine this co-evolutionary process, he chose to work with houseflies since their generation length is short enough to allow us to watch the

Mhoc08.fm Page 235 Friday, December 14, 2001 10:18 AM

   :     effects of selection over several generations. The male’s seminal fluid is mildly toxic to females but, as females co-evolve with males, they develop tricks to counter the effects. But what would happen if females were removed from the Red Queen contest and males could pursue their own interests unhindered? Rice (1992, 1996) separated out the female flies and prevented them from evolving while the males were given new batches of females to mate with. After three dozen or so generations, the males had the uncontested upper hand. When the ‘time-out-of-evolution’ females were returned to them, the females began to die from the effects of sex. The males’ unopposed sperm had become more and more toxic. To make matters worse, these males had evolved the ability to make females re-mate with them (against the females’ own interests) and even to deny subsequent matings to safer male competitors who might have displaced the lethal sperm. Females need to hold their own in the battle of the sexes just to stay alive. Males do their best to deter a female from mating again after they have gifted her with their semen. In domestic cats the male’s barbed penis causes damage to the female vagina, the pain of which effectively deters her from further copulation. The males of one species of South American butterfly introduce an anti-aphrodisiac into the female which repels other males for weeks at a time (Gilbert 1976). And if you ever believed that sperm was a benign gift to females, Rice’s work certainly makes you think twice. The male housefly’s seminal fluid proteins enter not only the reproductive tract but also the female’s bloodstream and bind to receptors in the brain, altering her behaviour. They trigger a series of responses that are clearly beneficial to males; they decrease her sexual appetite (making her less likely to re-mate with a rival), they increase her ovulation rate and they control sperm competition with alien sperm in her reproductive tract. As we have noted, they also slowly poison her so that the more she copulates, the sooner she dies. What is happening is a conflict of literal life and death. From the female’s perspective, her best bet is to mate wisely (select a mate with motile sperm and good genes) but not too frequently. The male, however, benefits from ensuring she does not re-mate but invests heavily in his current zygote. A second male that now appears on the scene will have his own agenda—he aims to entice the female to mate again (she must trade-off the quality of the new male’s genes against the toxicity of a double sperm dosage) and to displace or immobilize the previous male’s sperm. This tripartite conflict of interest generates gene warfare between all the parties. And notice now that we have left the biological effects of genes and begun to consider their behavioural effects. When genes code for behaviours they are just as (and perhaps more) important in the process of what William Rice and Brett Holland (1997) call interlocus contest evolution (ICE). Males adopt ploys that benefit their own reproductive interests while females develop counterploys to stop them. The familiar maxim of sexual selection is that females specialize in reproductive quality while males specialize in copulatory frequency. Males who can mate with many females hold a significant advantage in terms of reproductive success. So it pays

Mhoc08.fm Page 236 Friday, December 14, 2001 10:18 AM


    

males to attempt to seduce females as often as they can and it pays females to slow down their rate of mating—another Red Queen contest. Consider poecilid fish which come in two varieties. One called Xiphophorus have long swords while Priapella do not. The longer the sword, the more attractive the male is to Xiphophorus females. It operates as a persuasive seduction device; it encourage females to mate more than they would usually do. Indeed it is so effective that Priapella females also find it attractive when experimenters attach artificial swords to the normally swordless males of their own species. But if excessive mating (while good for the male) is bad for the female then we would expect that over time females would develop a resistance to its attractiveness. Sure enough, the Xiphophorus females have begun to resist its charms as evidenced in the fact that they are attracted significantly less to the sword than their sister Priapella who have not had the opportunity to develop an adaptive resistance. The theory is still speculative but it may be that frivolous seductive ornaments have a limited shelf-life. Useful male qualities such as resources and commitment in monogamous species cannot afford to be ignored as they actively favour females who make the correct choice. But gimmicky male features that seduce women may not work forever. Like fashion, what seemed stylish today looks ridiculous tomorrow.

Intragenomic conflict The battle between the sexes is sometimes played out in the offspring that they jointly create and it can happen on any of the chromosomes not just the X and Y. The conflict stems at bottom from male’s ancient wanderlust and deep attachment to promiscuity. Because a man comes and goes (so to speak), there is a good chance that the embryo created in a single act of intercourse may be the only one that he ever shares with this particular woman. This means that her long-term reproductive health is of only minor interest to him. Provided that she carries, bears and raises his child, why should the father care if she endures a difficult pregnancy or never conceives again? Polygyny is a heartless strategy from a female’s point of view. So when females undertake almost all of the paternal investment, they must have an eye to the future—a woman must consider all the future children that she will bear. Each one will be unambiguously a bearer of her genes despite uncertainty about whose the father’s might be. She must look out for her long-term reproductive interests for it is very clear that her current mate will not. The male’s genes exploit the female’s body in the interests of his offspring, the female’s genes counter them in any way that they can. This principle has been demonstrated by the creation of individuals (mice, in fact) who have a single parent, either male or female. How can this be done? Both sperm and egg carry their genes in their nucleus. When a sperm first penetrates an egg, the two pronucei are separate briefly before they fuse together. This moment is the opportunity for human intervention. If the sperm nucleus is removed and

Mhoc08.fm Page 237 Friday, December 14, 2001 10:18 AM

   :     replaced with a second egg nucleus, a zygote with no father is created. Similarly, the removal of the egg nucleus and the replacement with a second sperm nucleus results in a motherless zygote. But the effect on development is very different in each case. A zygote composed of maternal genes looks fairly normal but the placenta and yolk sac are grossly undersized. Father-only genes create a healthy placenta but instead of an embryo there is a disorganized aggregation of cells. Male genes build the placenta while female genes build the embryo’s body and especially its head. It was David Haig who suggested the reason: the placenta favours the foetus even at the expense of the mother—it parasites the mother’s body (Haig and Westoby 1989). Of course, the mother has her own reasons for accommodating the infant with a blood supply but she must also watch out for herself and her own survival. After this pregnancy, there will be others. She favours a placenta large enough to support the current foetus but not so large as to place excess demands on her body. Faced with this lack of female cooperation, the father ensures his own genetic interests over those of the mother by constructing the placenta himself. (It has recently been suggested that the growth of the placenta may serve to sequester and deny resources to any unrelated siblings rather than to enhance foetal growth; Hurst and McVean 1997.) As Ridley (1999, p. 209) succinctly describes it: ‘the father’s genes do not trust the mother to make a sufficiently invasive placentas; so they do the job themselves.’ The male is less concerned with the long-term welfare of the mother than with exploiting her as an accommodating vehicle for his offspring. (Oddly, the foetus should be more concerned about the mother than the father is. This child will be 50 per cent related to any further children she produces and therefore his inclusive fitness is related to hers.) The mechanism behind these exploitative genes is a phenomenon called genomic imprinting. Until recently it was believed that genes from the mother and father had equal chances of being expressed in the offspring. Every generation, genes are shuffled in the process of meiosis and crossover so that the 23 chromosomes that a woman and a man contribute to the total of 46 needed to build a baby are a random set taken equally from each partner. But we know now that some genes are ‘tagged’ with the sex of the parent who donated them and that tag determines whether or not they are expressed. The ‘imprinted’ gene is silenced while its rival allele is switched on and this imprinting can occur on both the mother’s and the father’s genes. Many of the phenotypic effects we see probably occur via the impact of imprinting on other genes that regulate cell behaviour. It is paternally active genes that build big placentas and maternally active genes that build the body and brain. The conflict between what suits men and what suits women must reach far back in our history and is a sufficiently powerful dynamic to drive genomic imprinting—the tug-of-war between maternal genes devoted to embryonic growth and paternal genes devoted to uterus exploitation. In an optimistic frame of mind, we might be tempted to think that the genes are not in conflict at all but rather are cooperating in the production of a viable placenta

Mhoc08.fm Page 238 Friday, December 14, 2001 10:18 AM

      and a viable embryo (Badcock 2000). But if this were so, then why not have both parents contribute equally and at half dosage to both? What we see is the end-point of a Red Queen struggle in which reproductively successful men have shown a casual genetic disregard for the mother’s long-term survival and women have developed counterploys of their own. One such counterploy is the discovery of genomic imprinting of a protein called insulin-like growth factor II (IGF II). This protein is essential to foetal growth. It promotes the transition of metabolites across the placenta and so is vital to acquiring resources from the mother. The paternal IGF II gene is expressed while the mother’s version is silenced. But the mother has not allowed this potential exploitation to go unchallenged. IGF IIR is a receptor whose purpose seems to be to mop up IGF II and so modulate the delivery of resources across the placenta. The maternal gene coding for this receptor is expressed. And the necessity of this female counterploy is evident in individuals who lack the maternal receptor gene—they are 16 per cent larger than normal (Haig and Graham 1991). In cases where an individual inherits two copies of the paternal IGF II, they develop enlarged hearts and livers and are prone to embryonic tumours. Like two wrestlers straining against one another, the opposing maternal and paternal effects keep everything upright. When one lets go, the imbalance is disastrous. But mothers and fathers also have conflicts about which parts of the embryo are most important. To study this, Keverne et al. (1996a,b) created mice where two different genomes were fused into the same body. They took advantage of the technique of egg and sperm pronucleus replacement that we have already described. But such manipulations are not viable—they rarely survive the eleventh day of gestation. To build a chimerical baby that could actually survive, they fused a normal embryo with an embryo made from two egg pronuclei. The result was a mouse with a very large head. But when they fused the normal embryo with one derived from an embryo made from two sperm pronuclei, they grew a mouse with a big body and small head. Indeed the body of the paternal chimeras grew so large that they had to be delivered by Caesarean section. So fathers contribute more to bodies and muscle while mothers specialize in brains. But not just brains—particular parts of them. By biochemically marking the maternal and paternal cells, the researchers were able to see where they ended up (Keverne et al. 1996b). The input of the paternal cells to brain construction was to the hypothalamus, amygdala and preoptic area—the areas that control emotion and evolutionarily critical ‘automatic’ behaviours such as sex, reproductive behaviour, aggression and fear. The mother’s cells migrated and proliferated in the cortex, striatum and hippocampus—areas implicated in reasoning, thought and behavioural inhibition. (If you are now blaming your father for inheriting his emotional tempo remember that half the hypothalamus-building genes that he bequeathed to you came from his mother. They may not have been expressed in him but they were inherited from his mother, retagged in him as paternal instead of maternal and then passed on

Mhoc08.fm Page 239 Friday, December 14, 2001 10:18 AM

   :     to you.) Why should females be so committed to constructing the cortex? What is in it for them? We don’t yet know but there are several fascinating suggestions. Christopher Badcock (2000) has taken perhaps the most daring and speculative viewpoint to date by suggesting that the conflict between the male-constructed limbic system and the female-constructed cortex parallels the psychoanalytic tension between the id and the ego. The paternally-built id makes egocentric, infantile and constant demands upon the mother, while the maternally-controlled cortex represses them. In this view, the cortex is essential for children’s autonomy and the mother’s ability to reinvest in new offspring. Badcock sees the genomically imprinted two-part brain as biological support for Freud’s psychoanalytic model. According to Haig’s theory, one of the chief functions of the hypothalamus is growth promotion and so it makes sense that, just as the father’s genes try to maximize foetal resources, they also try to maximize growth beyond the womb. Lactation requires a high calorie intake and the pressure on the mother to find and consume resources is even greater in the face of a hungrier infant. But again, why should the father care about this extra female effort? He will be long gone by then and his farewell gift is a fast-growing, highly demanding infant that will push the mother’s foraging ability to its limit. Another way of looking at this is as a special case of the Red Queen contest. Rice has shown that one result of ICE is very fast evolutionary pace. The enemy of male adaptation is always snapping at a woman’s heels and she must run fast to outpace him. The frontal cortex, with its ability to promote planning and foresight, is in a unique position to override the emotional tyranny of the limbic system. As males built bigger limbic systems, females fought back by building an ever-larger cortex and the result is the massive expansion of the human brain (Keverne et al. 1996b) that seems to be constrained only by the sheer impossibility of getting an infant with a bigger brain through the female pelvis. If the male–female contest gave us bigger brains, it was the female contribution that steadily increased our powers of reason and rationality. And over evolutionary time, the cortex has expanded in size much more than the lower brain, suggesting that women have been winning this particular race. From my own viewpoint, it is hard to resist the speculation that females have always had more to gain than males by the processes of emotional and behavioural inhibition that are controlled by the frontal cortex (see Chapter 3). An impulsive approach to aggression and sex may be advantageous to males in their short-term, all-out competition to vanquish other males and inseminate females, but the benefits to females are less evident. Males try to build offspring with plenty of derring-do motivation (and let the females take the consequences) while females counter by trying to slow the offspring down with longer-term reasoning (and extend their life expectancy). Fathers, it seems, wanted a little more ‘gut feeling’ in their children while females opted for more planning and restraint. If males had got their way unhindered, the human race would have a lot more r-strategists (breed plentifully and

Mhoc08.fm Page 240 Friday, December 14, 2001 10:18 AM


    

hope that some survive) while if females had triumphed we would be a K-selected species (have few children but invest massively in each). The battle of the sexes at the brain-building level points to tension between men’s desire for the drama of fast-track sex without personal consequences and women’s longer view incorporating protracted parental commitment. But brains are structures that run on chemicals and recent work suggests that humans, with their slow-maturing offspring, may have evolved a brain chemistry that works on women’s side.

The chemistry of bonding Prairie voles are sociable and deeply monogamous creatures. Together they support one another and cooperate in raising their young. The male not only defends his mate against rival male intruders but prefers her over any other available female. If their mate is removed, 80 per cent of them refuse to re-mate and spend the rest of their lives alone. The bond is normally established at the time of mating and seems to involve two peptides that are especially relevant to females and males, respectively: oxytocin and vasopressin (Insel et al. 1998). These peptides are remarkably similar in structure, differing only in the position of two amino acids. They have been linked to sexual, maternal, affiliative and aggressive behaviour and (importantly for monogamy) to social memory (Young et al. 1998). The distribution of their receptors varies markedly across even closely related species but, wherever species share a marked behavioural characteristics such as monogamy, they have a very similar pattern of receptor distribution in the brain. Both peptides are manufactured in the limbic system and one site that contains dense oxytocin receptors is the nucleus accumbens—part of the pleasure circuit of the brain. Partner preference has also been linked to increases in the release of dopamine (the neurotransmitter associated with wanting and liking) in this same part of the brain. In the wild, a female vole copulates between 15 and 30 times in a 24-hour period and the vaginal stimulation triggers the release of oxytocin. It seems likely that it is the presence of dopamine, resulting from sexual stimulation, that activates the oxytocin and generates the pair bond (Gingrich et al. 2000). In monogamous species, blocking the uptake of dopamine after mating eradicates the female’s partner preference and, by using chemicals that enhance dopamine’s effects, experimenters can generate a preference for a male even without copulation. Just the same pattern has been found for oxytocin. In the laboratory, oxytocin facilitates her preference for a familiar partner even without mating and an oxytocin antagonist, given before she mates, blocks the formation of a partner preference (Insel et al. 1998). Non-monogamous voles have very few oxytocin receptors in the nucleus accumbens and perhaps this is why they form no association between sexual pleasure and their partner. In the male prairie vole, our monogamous hero, administration of vasopressin, but not oxytocin, directly into the brain also induces pair bonding and male parental

Mhoc08.fm Page 241 Friday, December 14, 2001 10:18 AM

   :     care without mating (Wang et al. 1998). The administration of an antagonist (that blocks the binding of the peptide) leads to a decrease in these behaviours even among males who have mated repeatedly with the female. In short, by artificially administering vasopressin it is possible to cause males to bond without sex, and by blocking its action, to have sex without longer-term commitment. Scientists have discovered a long DNA sequence in the promoter region of the vasopressin receptor gene that seems to be important in determining when and where the gene is activated. Young et al. (1999) created transgenic mice by inserting the genetic sequence that codes for the vasopressin receptor gene from a monogamous prairie vole into a polygamous and less social mouse. The ‘monogamice’, as they were christened, showed the same pattern of receptors in the brain as intact prairie voles and although they did not become magically monogamous, when given vasopressin, they did show a significant and atypical increase in social contact with the female. We should avoid the trap of believing that there is something inevitable and incontrovertible about the fact that sexes regulate their behaviour by different substances. There is nothing about vasopressin that genetically or hormonally marks it out as a ‘male’ peptide. In plainfin midshipman fish, males come in two types— regular courting males and ‘sneak’ fertilizers who surreptitiously fertilize eggs and depart leaving the cuckolded courting males to raise them. The two males can be distinguished by the fact that sneak males are unable to produce the elaborate female-luring song of the courting males. Instead, like the females, they emit only a grunting noise. Courting males’ song is affected by vasopressin while the female’s grunt is triggered by oxytocin. The grunt of the sneak males, like females, responds only to oxytocin not vasopressin. Yet there is nothing female about them—in fact, they have enlarged reproductive organs. Sensitivity to these peptides is not an inevitable consequence of an organism’s sex (Goodson and Bass 2000). And although among prairie voles the endogenous action of the two peptides is sexually dimorphic, pair bonding can be achieved by administering very high doses of vasopressin to females and oxytocin to males (Cho et al. 1999). Prairie voles are all very well but does oxytocin work for women too? It is too early to be sure. But Turner et al. (1999) investigated whether oxytocin release is responsive to positive and negative romantic recollections, as well as to massage which (like orgasm) has been shown to trigger oxytocin release. Massage did indeed raise oxytocin levels while negative recollections of abandonment decreased them. Alas, positive romantic recollections had no effect. (But then, thinking about someone is a very far cry from having intimate, tender sex with them.) One interesting finding was that women who were currently in happy relationships or who had a history of secure relationships showed a much greater release of oxytocin in response to romantic thoughts than did women whose love-lives were less satisfactory. Perhaps background levels of relationship security and happiness prime us to respond to new ones with equal commitment. Could this be the neurochemical realization of attachment theory?

Mhoc08.fm Page 242 Friday, December 14, 2001 10:18 AM

     

The riddle of monogamy Evolutionary psychologists have stressed, perhaps even overstressed, the philandering nature of men. In their defence, they often cite the fact that 84 per cent of world societies condone polygyny. (The implicit or explicit contrast is with the tiny prevalence of polyandry (0.5 per cent) which is often used to document the greater sexual appetite and sexual proprietariness of males compared with females.) But counting societies is not the same as counting people. The vast majority of the world’s population inhabit societies that prohibit polygyny and, even where it is legal, only about 10 per cent of men take multiple wives. A central question about human behaviour therefore is why men, who can gain such reproductive advantages from promiscuity and polygyny, in the main elect to be monogamous? As yet there is no definitive answer, but there is no shortage of contenders. Some believe that monogamy was driven by women, some by men but most agree that it was a compromise that benefited both sexes. What does monogamy offer females? One strong possibility is protection. Northern elephant seals provide a classic example because there is such a huge difference in the size and weight of males and females. Males are usually about four times bigger but can be as much as eleven times as large. A male can crush a female when he lets his body weight fall on her during intromission and she can be injured by his lessthan-affectionate biting during copulation. But if injury from a partner is a hazard, things get worse when she attempts to leave the colony and return to the sea. She must run the gauntlet of dozens of subdominant males who attempt to intercept and mate with her. During the thirty minutes it takes her to get to the safety of the water, she receives on average twenty times more blows, mounts and copulations than when she was in the harem (Mesnick 1997). Among primates too, females receive more threats, attacks and wounds when they are in oestrus than at other times. Rhesus macaque females who choose to approach low-ranking males are punished by the dominant males, resulting in between three and six attacks every day when she is in oestrus. In chimpanzees, males try to persuade females to join them in consortships (trips away from the body of the group for the purposes of mating that can last for hours or days). Sometimes the female is willing but, if she is not, the male performs aggressive displays to intimidate her and, if that fails, he simply attacks her. What can a female do to protect herself? A strong alliance between females could make males think twice about an attack. In primate species where females remain in their natal group, the female matrilines are powerful and male behaviour seems to take account of the possible consequences of a massed female counterattack. But in other species where females leave their kin and do not forge strong bonds with one another, females are in a much weaker position. As we have seen, unrelated bonobo females do bond strongly and they respond vigorously and unanimously to the rare attacks that males dare to launch. Among humans, the time that women are able to spend with unrelated women and the

Mhoc08.fm Page 243 Friday, December 14, 2001 10:18 AM

   :     friendships they forge may also be important sources of protection from male aggression (see Chapter 5). Where the aggression is sexual, one avenue is for the female to simply give way to it in order to protect her own life—a depressing capitulation called ‘convenience polyandry’. Chimpanzee females do it. So do the poor elephant seals as they desperately try to reach the freedom of the sea (Mesnick 1997). Eighty-eight per cent are mounted by at least one male and 60 per cent of the females lay quietly and even speed up the unwanted copulation by spreading their hindflippers. Those who do, manage to decrease the number of blows they receive by 8 per cent. And it seems their strategy has another pay-off. After the copulation, 78 per cent of the male rapists escort the female to the sea (doubtless to protect his own seminal investment rather than out of any gallantry) and aggressively drive away other males. Here we can see vividly the dynamics of what has been called the ‘bodyguard’ hypothesis of monogamy. A male can effectively protect the female from the aggression of other males. (As Brownmiller (1975) has pointed out, it only takes a minority of males to be rapists to induce a female to seek protection from the remainder.) But there is a price. Although male baboons have been found to protect female ‘friends’ from male aggression, it is very rare indeed that the female does not mate with him subsequently (Mesnick 1997). The proposal to the female is simple—sex for safety. Some human facts seem to support such a theory. Women place a higher value than men do on physical strength in a prospective mate (Buss 1989). Rape is especially prevalent during war when men are absent, dead or unable to protect women (Brownmiller 1975). At all ages, married and cohabiting women are significantly less likely to have experienced sexual aggression during the last year than are women who are single, separated, divorced or widowed (Wilson and Mesnick 1997). Barbara Smuts (1992, 1995) offers an account of how monogamy could have arisen under the bodyguard hypothesis. Human males are remarkable in capacity to form alliances. The importance of these alliances probably lay in increased betweengroup competition or in response to the need for cooperative hunting or joint power-seeking within the group. To hold these alliances together and prevent the despotic rule of a single male, men may have extended respect (in the sense of forgoing rape) to one another’s long-term mates. Once this became common, it would be increasingly hard for a female to buck the trend. A promiscuous female would have no steady partner and consequently become the object of unrestrained sexual interest by males and she would have no defender to protect her from such attacks. But there is sting in the tail of this notion of chivalrous men: human marriage can be considered a means by which cooperating males agree about mating rights, respect (in principle) one another’s ‘possession’ of particular females, protect their mates and their mates’ children from aggression by other men, and gain rights to coerce their own females with reduced interference by other men. (Smuts 1992, p. 11)

Mhoc08.fm Page 244 Friday, December 14, 2001 10:18 AM

      Although pair bonds may have formed originally to cement male intragroup solidarity, the persuasive pay-off for female cooperation in the enterprise was protection and safety. But what good is protection from other males if the real threat comes from your mate? Today a woman is more at risk of homicide or physical abuse from a male partner (or ex-partner) than from a stranger. The most common form of rape is ‘date rape’—sexual aggression from a partner not from a stranger. Either the bodyguard hypothesis has never worked in a female’s interest or, if it did in the Pleistocene, it no longer does. But perhaps another kind of protection was more central—protection for the children not for the mother. Infanticide by strange males is a real and lethal threat in primates, accounting for more than a third of infant deaths (Smuts 1992). The explanation of monogamy in baboons seems to lie in protection against infanticide. Van Schaik and Dunbar (1990) were able to dismiss the commonly held belief that monogamy was a masculine default option—the result of a male’s inability to monopolize sexual access to more than one female. They argued that the advantages of monogamy must lie in the fact that monogamous males could produce surviving children at a rate which exceeded that which could be achieved by the inseminations they had to forgo. Close scrutiny of the baboons’ behaviour showed that the appeal of monogamy did not lie in the reduced risk of predation that males offered, nor in the males ability to defend a foraging area for the female. What made monogamy successful was the dramatic decrease in infanticide among monogamously fathered infants. Among humans too the protection of a father is a distinct advantage. Nowhere is the coldness and instrumentality of male infanticide demonstrated as starkly as in inter-tribal warfare. Hrdy (1999, p. 242) relates one woman’s account of the calculated violence of Yanamamo warriors against children: the men began to kill the children; little ones, bigger ones, they killed many of them. They tried to run away, but [the Karawetari raiders] caught them, and threw them to the ground, and stuck them with bows, which went through their bodies and rooted them to the ground. Taking the smallest by the feet, they beat them against the trees and the rocks. The children’s eyes trembled. Then the men took the dead bodies and threw them among the rocks, saying ‘Stay there, so that your father can find you and eat you.’

If a male is to risk his life to defend an infant against such a vicious attack, it pays him to make very sure that it his own—hence the exchange of female fidelity for male protection of infants. But a quite different spin has been put on the relationship between infanticide and mating patterns in primates by Sarah Hrdy (1979). Not only was she the first primatologist to document the systematic use of infanticide by strange males but she also noticed that langur males never attacked an infant born to a female with whom they had mated. They seemed to err on the conservative side of paternity certainty, showing reluctance to risk the possible killing of one of their own children. This being so, she reasoned, a useful female strategy would be to copulate with as many

Mhoc08.fm Page 245 Friday, December 14, 2001 10:18 AM

   :     males as possible in the hope that each one would have reason to believe any subsequent child was his own and spare them. Perhaps that is why females show a preference for subadult males on their way to the top and for strange males who are not yet part of the group. In a promiscuous mating system, Hrdy argues, males should avoid infanticide because of the uncertainty that surrounds paternity. So we have two conflicting scenarios. In one, females trade sexual fidelity for their mate’s protection from ‘dangerous males’ (that is males with whom the female has not mated). In the other, the female forgoes the protection of a mate and protects herself by promiscuously confusing paternity. Now this latter strategy can only work if the male’s latitude of tolerance is very high—if they are willing to believe that it just might be their infant. But among our closest relative, chimpanzees, such tolerance is not much in evidence. Male chimpanzees will kill an infant simply because the mother’s home range is too close to a neighbouring group and hence she may have mated with an outgroup male (Hiraiwa-Hasegawa and Hasegawa 1994). In other words, they will kill an infant if they believe that it just might not be fathered by them. Men are notable among primate species in the degree of involvement that they exhibit in provisioning and protecting their own offspring. But that assistance is most likely to be forthcoming when paternity is certain. Like chimps, males tend to be acutely suspicious of females (Buss 2000). Although women might have evolved to express their sexuality in a more promiscuous way, their need for help with child care precluded it—they needed male assistance too much. The paternal investment that goes with pair bonding has traditionally, in one way or another, been associated with hunting and food sharing (Foley 1987; Lovejoy 1981). Infants were born increasingly dependent and they relied upon their mother for extended periods of time when they were carried and cared for. Protracted childhood meant that the mother was often caring for more than one child at a time (Lancaster and Lancaster 1983). And this created a brake on the number of children that could be produced in a woman’s lifetime. She needed assistance to help her current children survive and to shorten the interval between births. Although a woman could provision herself and her infant by gathering in the local vicinity, men could travel more widely. They could hunt and scavenge and the extra calories they provided were needed to fuel the growth of our expanding brains. A human brain consumes 22 per cent of our metabolic energy compared to only 8 per cent in the chimpanzee. Meat is a very useful source of protein. Among hunter-gatherer societies the arrival of meat is greeted with delight, the taste and smell of roasted fat is a positive treat. And most men are unable to secure enough meat to support more than one family (Draper and Harpending 1988). But meat from the hunt is given to the wife rather than the child, suggesting that it might be connected more to mating effort than to direct child care. Studies of hunter-gatherer societies suggest that although men do share meat with their mates, they also share it beyond the immediate family. Indeed it seems that men channel

Mhoc08.fm Page 246 Friday, December 14, 2001 10:18 AM


    

the fruits of their hunting labour preferentially towards attractive women in the hope of securing sex with them (Hawkes 1991). Nevertheless, fathers do seem to be an asset (Geary 1998). The fact that children without fathers die more frequently is in part because they are less protected from aggressive males but it may also be related to the adequacy of their diet. The children of skilled hunters die less frequently than those whose fathers are poor hunters (Blurton Jones et al. 1997; Hill and Kaplan 1988). In part this derives from the fact that these children are accorded more care than others—including a greater likelihood that other adults will share food with them in times of scarcity. Hunting prowess affects a man’s social standing and even among pre-industrial peoples this is associated with infant survival. In current Western societies, a family’s socio-economic status predicts the health of family members and in most families the primary wage-earner is male. (The impact of divorce on children is often mediated by the sudden drop in income and associated lifestyle changes; Tucker et al. 1997.) In the United States, children in divorced families have a one-third greater risk of death (Friedman et al. 1995). As Geary (1998, p. 114) sums it up: ‘paternal investment that increases the social competitiveness of children likely provides a buffer—reducing mortality and morbidity risks—against unforeseen large-scale social stressors.’ But among many hunter-gatherer societies, women can forage successfully without depending upon male provisioning. And this raises the question of how and why females came to depend so heavily upon males. Gowaty (1997) sees mating patterns as the outcome of a Red Queen dynamic between the sexes which begins with males’ attempts to control females directly. Such direct control could vary from physical coercion, rape and mate guarding to more attractive tactics including nuptial gifts, defending females’ foraging areas and protection from predation (Smuts and Smuts 1993). But when these tactics do not work in the female’s best reproductive interests, female counter-strategies will evolve. In response, males move to the next stage of control—they ‘broker’ females’ access to the resources that they depend on to reproduce. Here males focus their acquisitive attention, not directly on females, but rather on the environmental resources on which females depend. Gowaty sees males and females as in competition to control the resources that are vital to female reproduction. Why is it that males have so often won? In the case of humans, Barbara Smuts (1995) identifies three key steps by which females lost the battle: loss of female kin bonds, the evolution of male alliances and the shift to agriculture. When mankind moved from a hunter-gatherer lifestyle, women lost the ability to forage self-sufficiently. Under agriculture and animal husbandry, women found themselves restricted to a permanent plot of land used for growing or grazing. They were now under the permanent surveillance of men and their work was no longer independent; they assisted in farming whenever their child-care duties allowed. This dependence on male provisioning weakened their bargaining power and forced them into an entirely new kind of dependency. For men, agriculture meant that surplus resources could be accumulated and traded,

Mhoc08.fm Page 247 Friday, December 14, 2001 10:18 AM

   :     creating wealth differences that, through nepotism, could be perpetuated through generations. Women now had to compete with one another for access to the wealthiest men who, holding all the power, could enforce requirements for chastity and faithfulness which women were in no economic position to resist. In the past ten thousand years, women, unlike their primate sisters, became economically dependent upon males. The survival of their children depended increasingly upon maintaining a strong bond with a provisioning man. But males would only provide on this grand scale if they could be certain that each child was their own. The bargain was straightforward: exclusive sexual rights for the male and longterm paternal assistance for the female. Women gained the benefit of paternal investment but lost the advantages of multiple mates. One lost advantage may have been decreased risk of aggression to the child—if potential sires erred on the side of caution. Another may have been increased provision and protection for the child— studies of partible paternity suggest the optimum number of fathers to ensure an infant’s long and healthy life is two. Non-monogamous women also gained the advantage of genetic diversity in their offspring and ‘serial monogamy’ allowed women the opportunity to discard an unsuitable mate and select a new one (Fisher 1989). Perhaps most importantly a woman made her choice based purely on genetic quality—an attractive man who promised ‘good genes’ might be an unwilling and unreliable long-term mate but an eager sperm donator. Whatever the advantages, multiple sexual partners must have been part of female history for how else can we account for the size of human testes, the huge number of ‘killer’ and ‘blocker’ sperm that men produce in each ejaculate and the evolution of mate guarding and intense sexual jealousy in men? If males have traded promiscuity for monogamy, they have not done so as a favour to women but because it benefited them. The fitness profits for monogamy must have offset the costs. The arithmetic amounts to paternal certainty plus the improvement in child survival earned by his presence minus the cost of lost mating opportunities elsewhere plus the probability that these women could raise the child without him. In short, the choice was between being a Dad or a Cad. Dads in the main won because of the benefits that they achieved in raising their children successfully rather than relying on the success of a woman acting alone. But monogamy may have been a route that had a special appeal to those men who were lowest on the social hierarchy. Women want it all but often have to choose between good phenotypic and genetic quality and long-term assistance with child care. Men with the best genes are the ones that are most reluctant to become monogamous fathers—they have too many mating opportunities to give up (Gangestad and Simpson 2000; Waynforth 1999). If less desirable males cannot compete on those terms, they can compete by giving females what they want in the form of extended paternal care. But men with plenty of money had another mating strategy—polygynous marriage. In egalitarian societies where there is little variation in male resources, polygyny is

Mhoc08.fm Page 248 Friday, December 14, 2001 10:18 AM


    

usually not an economic possibility. Wives are just too expensive. But as societies become more complex and imbalanced, polygyny becomes more common. Betzig (1982) took measures of how much a society departed from egalitarianism: group size, the degree of hierarchy, asymmetry in conflict resolution benefiting the powerful and the use of fines, fees and bribes by the powerful. All four of these variables correlated higher than 0.80 with the number of wives taken by those at the top of the social ladder. Polygyny ensures an excessive number of children to any man who can afford it. But while he may benefit, most others do not. Was polygyny good for women? The notion of the ‘polygyny threshold’ was developed from the study of bird species where the males arrive at the breeding site ahead of the females. The male carves out a territory for himself by driving other males away and when the female arrives she inspects what the various males have to offer. In some cases, it is more advantageous for a female to take up residence in the territory of an already-mated male than to opt for one with only a poor site (resource defence polygyny). This idea was applied to women too by anthropologists who suggested that women may find it in their own interests to become the second or third wife of a rich man, rather than the sole wife of a poor man (Borgerhoff Mulder 1990). Robert Wright (1994) argues that polygyny is a positive benefit to the majority of women and gives a nice demonstration of why. Imagine a group of 100 women and 100 men each ranked according to how desirable they are on the marriage market. Under monogamy each woman will end up with a man who shares her ranking. (This is assortative mating. Although men and women would both prefer to have a mate who ranks higher than them, simple market forces make gross disparities between mates unlikely.) But now imagine we legalize polygyny. The first woman to take advantage of it rises from rank 80 to rank 20 (admittedly sharing the wealth with her co-wife). But the effect on every other woman is to move them up one place so that they each marry a man who is one rank higher than their last partner. Polygyny, Wright argues, redistributes male wealth in a much more egalitarian way between women. So, even if a given woman refuses to share a husband, she is still better off under polygyny because the fact that some other women would, benefits her by freeing up males who rank higher than her current mate. The losers are top-ranking women (who have nowhere to advance to and now have to share what they previously had to themselves) and the ‘bottom-ranked men’ (who miss out on marriage altogether). But there are numerous costs to women that affect not just the cream of the crop. Second and later wives have lower fertility than monogamous women (Daly and Wilson 1983; Sichona 1993) and their children are less likely to survive (Dorjahn 1958; Strassmann 1997). As the percentage of polygynous marriages increases in society, women’s ability to control the fruits of men’s labour decreases and women become less likely to inherit property. Most importantly, women just do not seem happy in these marriages—we have seen in Chapter 6 that the co-wife relationship is the epitome of jealousy, unhappiness and conflict.

Mhoc08.fm Page 249 Friday, December 14, 2001 10:18 AM

   :     Given that ruling males benefited from polygyny, why did they give it up? One answer is to increase social harmony among men further down the status ladder. When a few men monopolize more than their fair share of women, other men must go without. This creates a cadre of unmated and alienated males who can be socially dangerous; male-on-male male aggression is higher in polygynous societies (Betzig 1986). Unmated males are more likely to kill one another (Daly and Wilson 1988) than their married same-aged counterparts. Indeed unmarried men are over-represented in a range of criminal activities from robbery to rape. This kind of societal infighting worked against strong male alliances which were necessary to maintain and expand territory. As Smuts (1995, p. 14) puts it: ‘males developed the ability to control male–male competition within the group, presumably in order to improve their ability to compete against males from other groups.’ Socially imposed monogamy (rather than ecological monogamy that results from equality of male resources) may be an institutionalized response designed to control and contain male aggression. And if monogamy was to be imposed as a sop to the male masses, ruling males had to forgo polygyny as well—those at the top could hardly claim for themselves what they denied to others. In this view, monogamy was imposed to increase within-group solidarity against external threats (Alexander 1979) and because increasing specialization of labour meant that the value of the proletariat to the aristocracy lay not just in their sheer numbers and manpower but in their skilled knowledge that gave them a new worth and bargaining power (Betzig 1986). But if rulers were unwilling to give up their own sexual indulgences, the advent of democracy saw to it that they must. When all men could vote, they voted in their own interests and the majority of men could neither find nor support multiple wives. As Wright (1994) points out, those at the top still got the most desirable women but they at least had to restrict themselves to one. In the past one hundred years, the increased acceptance of divorce means that we have a kind of de facto polygyny since men who divorce are more likely to remarry and produce children from second marriages than are women. So the most desirable men (often the richest) can still sequester and enjoy a number of attractive women provided that they do not marry them simultaneously. Let us pause and take stock of monogamy. Let us tally the advantages, not forgetting the costs. For men, monogamy buys relief from the bachelor conflict, enhancement of male alliances, the opportunity of at least modest reproductive success (rather than none at all), sexual intercourse ad libidum, increased certainty of paternity, monopoly of a woman’s entire reproductive career and enhanced offspring survival. For all this he incurs some costs: the lost opportunity for massive reproductive success, the price of mate guarding, protection and provisioning, and the possibility that—if he has been cuckolded—all these costs were paid in vain. For women, monogamy means greater time and energy to invest in her young, extra calories, protection from sexual aggression and infant harassment by other males, energy saved by servicing a single partner rather than several and an increase

Mhoc08.fm Page 250 Friday, December 14, 2001 10:18 AM

      in offspring survival. On the debit side, she lowers the diversity of her children’s genetic inheritance, forgoes the ‘sexy sons’ of higher-quality genotypic males, exposes herself to the risk of jealousy-motivated violence, and incurs energy and time costs in guarding him from takeovers by unpartnered, often younger women. Humans are capable of a living with a variety of different mating systems. But some are more difficult for us than others. Polyandry is rare because males find it hard to share a wife and to entertain uncertainty of paternity (when it does occur it is most often between brothers and the younger brother aims to leave and set up his own monogamous home when he is economically able). Indiscriminate promiscuity is hard for us too because men and women build specific, exclusive bonds with one another and experience jealousy when they see their partner with another person (Buss 2000). Polygyny is a system that has been and is practised in many societies but it is one that most women have resisted (it generates strong conflicts of interest between co-wives) and, while men may have envied polygynists, democracy has meant that their envy has ultimately overthrown the system. Monogamy seems to have won for now. Nowhere have the advantages of monogamy been so elegantly shown as by Holland and Rice (1999) working with the humble housefly. Normally promiscuous, some of these flies were randomly selected to have monogamy forced upon them. The experimenters acted as marriage brokers, teaming up and housing together individual males and females over 32 generations. Monogamy means not only that the reproductive success of males and females is identical but also that their reproductive interests should converge. While polygyny means that males can exploit females quite ruthlessly without suffering themselves, monogamy means that anything that hurts a female (prevents her from achieving her reproductive potential) hurts her male partner just as much. After several generations of monogamy, Holland and Rice performed the key tests. First, they introduced non-experimental, traditional females to mate with the monogamous ‘new males’. They found that monogamy had led to a decrease in the toxicity of the male’s seminal fluid and also to a reduction in male courtship—an activity that is harmful to females. Then they looked at the effect of monogamy on the experimental females. During monogamy, these females’ male partners had behaved in a less exploitative way towards them and so the monogamous females had not needed to evolve counter-strategies of resistance. As expected, when these monogamous females mated with normal males, a larger proportion died than among traditional females who had been allowed to co-evolve with male polygyny. The monogamous flies also produced more offspring than the promiscuous flies, having been freed from the extra ‘load’ of sexual selection (developing ploys and counterploys takes time and energy as well as being potentially fatal). Non-monogamous arrangements can have advantages for females (such as getting the best genes) but they also have costs (males evolving to ‘dupe’ females into behaviour that is counter to their own best interests). Freed from this antagonistic tussle, males become more benign and females produce more offspring and suffer lower mortality.

Mhoc08.fm Page 251 Friday, December 14, 2001 10:18 AM

   :     In the end monogamy is a compromise between all parties—and one with definite pay-offs. But never forget that official monogamy does not rule out temporary bouts of polygyny or polyandry in the form of infidelity. And although affairs may require one member of each sex, the effects are not gender-blind.

Marital troubles Given the inherent conflict in the reproductive strategies of men and women, we should expect marriage to be a relatively fragile institution, vulnerable to collapse when the cost–benefit equation for either the man or the woman becomes too imbalanced. One way to examine the kinds of costs that trigger marital strife is to look at the causes of divorce. Divorce has become increasingly easy and in the United States and Europe between 30 and 50 per cent of first marriages end in divorce (Hatfield and Rapson 1996). Of these first divorcees, about 75 per cent will marry again and about half of these second marriages will also fail. The number of divorces in a society is, of course, a function of how easy divorce is to obtain but nonetheless the reasons that drive people to dissolve their unions are likely to be the same sources of dissatisfaction that, in less liberal eras, couples simply had to endure. Marriages fail because of a clash of personalities—at least that is how the parties involved often explain it. They made the wrong choice—over the years it became clear that they just did not approach life in the same way, share the same values or ‘gel’ as a functioning partnership. But behind these kinds of general statements, we can look for the specific areas of disagreement that seem to give rise to couples’ realization that they were not ‘meant for each other’. Betzig (1989) used ethnographic data on causes of divorce from 160 cultures. She was able to identify the leading causes on a worldwide basis by simply tabulating the percentage of societies that recognized divorce as a reasonable response to each of a number of possible ‘wrongdoings’. The most common cause was wifely infidelity: ‘If marriage qualifies as near universal, so must the double standard. Almost every one of the causes of conjugal dissolution that might be related to infidelity is ascribed significantly more often to one sex than to the other’ (Betzig 1989, p. 658). Second on the list comes infertility and again it is the wife’s sterility that is recognized more often than the husband’s. In polygynous societies, a wife’s failure to produce a child constitutes grounds for the husband taking a second wife. However, the reverse is never the case—no culture endorses a woman taking a second husband if her first should prove sterile. Failure to produce children may be a less frequent ground for divorce in the West but nonetheless it draws attention to the importance of children in sustaining marriages as we shall see shortly. The third leading cause is personality differences but as Betzig remarks, these grounds ‘often seem tantamount to no grounds at all’. From an evolutionary viewpoint, they cannot help us to identify exactly what constitutes the focal grounds for disagreement. Next in line is failure of

Mhoc08.fm Page 252 Friday, December 14, 2001 10:18 AM

      economic support and here finally we see a reversal of the double standard—in all but one case it is failure of provision by the husband that is grounds for divorce. One of the most comprehensive studies of contemporary reasons for divorce was undertaken by Kitson (1992) who interviewed men and women divorcees in Cleveland, Ohio. Men were more likely to be unsure about what had caused the divorce and blamed external events such as his overcommitment to his job, a death in the family or interference by in-laws. Women were significantly more likely to mention their husband’s involvement in extramarital sex, his untrustworthy nature, too much time spent ‘out with the boys’, his excessive drinking and money problems. Men seem out of touch with the social and emotional troubles that have beset their marriage whereas women focus clearly on the failure of a man’s romantic and economic commitment to her as evidenced by his duplicitous and selfish behaviour. But the problem with the majority of studies of divorce is that they collect their data retrospectively. Perceptions of the reasons for the divorce not only vary between husband and wife but they also change over time. After-the-fact accounts can involve redefining what had previously been acceptable behaviour in order to rationalize and justify the divorce (Rasmussen and Ferraro 1979). Amato and Rogers (1997) avoided this problem by undertaking a prospective study. In 1980, they contacted a random sample of 2033 married persons and asked about sources of conflict in their marriage and whether the problem was caused chiefly by the husband or the wife. They managed to recontact 78 per cent of their sample again 3 years later, 71 per cent 8 years later and 61 per cent after 12 years. They could then examine the problems that had preceded the divorce in real time. Their results confirmed Betzig’s and many other researchers’ (Kitson et al. 1985) findings that the chief source of divorce is infidelity or suspected infidelity. Husbands who reported that their wife was unfaithful showed a 363 per cent increase in the likelihood of divorce while wives who believed their partner was unfaithful increased their chances of divorce by 299 per cent. Women’s infidelity (at least as suspected by the partner) is more closely associated with divorce than men’s. Male jealousy was also more likely to lead to divorce then female jealousy. Behind this, came a cluster of reasons that all seem to reflect Betzig’s category of failure to economically support the partner—spending money foolishly. Men’s profligacy (as viewed by their wives) increased the odds of divorce by 187 per cent while women’s foolish use of money increased divorce odds to a much lesser extent (77 per cent) suggesting that it is men’s failure to provide economically that matters more than women’s. The reasons for divorce also vary as a function of whether men and women are asked about their own behaviour or their spouse’s. Men were more likely than their wives to report that their own criticism, moodiness, hurt feelings and absence from home were problematic. The wives, on the other hand, were more critical of their husbands’ jealousy and irritating habits. Although the two parties emphasized different problem behaviours by the husband, they were both equally aware of the extent to which his misbehaviour was a problem in the marriage. As for the wives,

Mhoc08.fm Page 253 Friday, December 14, 2001 10:18 AM

   :     they more often blamed their own anger, hurt feelings, criticism, moodiness and refusal to talk as the causes of the marital problems compared to the husbands who thought their wife’s jealousy was the real problem. Husbands reported significantly fewer marital problems caused by the wife than did women. Men, in short, seem to be relatively under-aware of the contribution their wife is making to the collapse of the marriage. But both parties agree that it is the jealous behaviour of their partner that creates problems. And sometimes that jealousy may be founded in reality.

Fidelity—who cares? In countries around the world until the nineteenth century, the notion that a wife was a man’s possession was encoded in law, that is in laws which were made by men and reflected male interests. A wife who had intercourse with a man other than her husband was guilty of a crime (Daly and Wilson 1988). The wronged husband had the right to damages, to divorce and even to violent revenge. In every traditional society studied, there was a clear double standard—adultery by the husband was never subject to the same extreme sanctions. Only recently and especially in Western societies have laws altered to become ‘gender-blind’ in their treatment of fornication by men and women. Not only is such a grave offence grounds for divorce, it may also be a justification for extreme violence. The state of Texas resisted until 1974 changing a law which stated that homicide committed by a husband upon the wife’s lover—as long as it was committed before the adulterous couple separated—was justifiable and hence not a crime. Even today, murders committed as a result of finding the wife in flagrante delicto are reduced to manslaughter in most states in America and may be treated even less severely when prosecutors feel that they have little chance of persuading a jury to imprison a ‘wronged’ husband. From an evolutionary viewpoint there is no real mystery as to the peculiar importance attached to wifely fidelity—a woman who has sex outside marriage places her husband in danger of raising and investing in another man’s child. Daly and Wilson (1988) undertook a comprehensive study of 80 homicides committed in Detroit in 1972 where the offender and victim were legal or common-law spouses. Twenty-three were caused by jealousy and in 16 of the cases it was the husband who committed the murder. Male spousal homicide also seems to be associated with strong proprietary feelings about the wife. The wife is his property and he owns her. Hence if she threatens to leave him (and, worse, start a new relationship with another man) some men believe that she must be killed. A distinct male pattern involves the offender not only killing his wife (and sometimes the children too) but taking his own life also. As Daly and Wilson (1988, p. 215) succinctly put it: ‘The prospect of losing his family through death apparently strikes the desperate familicidal father as no more

Mhoc08.fm Page 254 Friday, December 14, 2001 10:18 AM

      disastrous than the prospect of losing them through desertion!’ Central to male killing patterns is his determination that ‘If I can’t have her no-one can’. Separated and divorced women are between five and seven times as likely to be killed by their partner than those who are still residing with him (Wilson et al. 1993) and over 90 per cent of these killings occur in the first year of separation. This pattern has also been found in Australia, Chicago and Canada (Wilson and Daly 1993). Inspection of homicide data leaves little doubt that a lethal response to infidelity is a very male reaction. Wilson and Daly (1993) describe male sexual jealousy as an evolved psychological module. A man who was a conscientious mate guarder, alert to the possibility of cuckoldry, left a greater proportion of his genes in the subsequent generation’s gene pool than one who took a laissez-faire attitude to his wife’s desire for sexual novelty. But mate guarding can become paranoid and extreme; the wife finds that she is prevented from leaving the house unescorted, punishment ensues for tardiness, she is barred from seeing even relatives or female friends, she is subject to phone calls throughout the day to check on her whereabouts. This pattern of obsessive concern can erupt into violent acts aimed at deterring her from pursuing—not just affairs— but even a normal degree of autonomy. Spending too long out of the husband’s sight at a party can result in accusations of sexual betrayal and the use of physical violence. The majority of women who seek refuge in shelters report these patterns of excessive control, which at the start of the relationship the woman often interprets as a testament to his devotion to her. And there may be an element of truth in these women’s belief. Extremes of male mate guarding cause appalling suffering and even physical injury, but David Buss (2000) has argued that a husband who shows no concern about his partner’s interest in other men is a man who may be less than fully committed to her. He calls this the ‘jealousy paradox’. A degree of jealousy is an inevitable part of love. For better or worse, human pair-bonding depends upon a sense of exclusivity. Experientially, we feel that our chosen partner is uniquely suited to us, he is the ‘other half’ of ourselves. Like a lock in a key, we feel ourselves to have been predestined to coalesce in marriage. And that sense of the partner as ‘specially designed’ for us is utterly broken by romantic and sexual betrayal. We become just one of many that could have been their partner. Worse still, we have been chosen and then found wanting. Such rejection prompts not only depression but anger too for, as the old saying has it, the real opposite of love is not hate but indifference. A degree of jealousy may be an index of the intensity of love. A study that gathered data on dating couples and followed them up over time found that those pairs who expressed greater jealousy about their partner were more likely to marry them subsequently (Mathes 1986). Jealousy reveals something about a person’s estimation of the value of their mate. The probability of infidelity in men and women is associated with their attractiveness relative to their spouse. Individuals who rate themselves (and who are rated by others) as the more attractive of the pair are more likely to have affairs and to have

Mhoc08.fm Page 255 Friday, December 14, 2001 10:18 AM

   :     them earlier in the marriage (Berscheid et al. 1973; Buss and Shackelford 1997; Prins et al. 1993). Attractive and desirable partners invite more mate guarding than their more homely peers. On a darker note, studies of spousal homicide also confirm that women are most likely to be killed by their husbands when they are younger and hence more attractive to rival men (Daly and Wilson 1988; Willbanks 1984). The path of jealousy that leads to divorce has an evolutionary logic (Buss 2000). Infidelity must be concealed and the tell-tale signs are far from evident. To detect infidelity, a person must become highly attuned to even subtle variations in their partner’s behaviour—the credit card statement that goes missing, the new dress that seems an unusually fashionable choice, the race to answer the phone which often turns out to be a wrong number. Once a sign has been picked up, the husband lowers his threshold for detecting further signs. A wife who senses her husband’s suspicion must in turn raise her level of deception. A co-evolutionary spiral develops where the mate’s ability to see through deception results in increasing skill by the partner in concealing their true intentions. This process, spread out over evolutionary time rather than in the microscopic time scale of a human marriage, may have made all of us more proficient at both deceiving and detecting. Jealousy seems to be the result of selection for strong sensitivity to signals of sexual infidelity. And such sensitivity can develop when the costs of failing to detect a true signal are greater than the costs of a false alarm. Such a finely tuned detection system can lead to unwarranted (we often call it pathological) jealousy. And jealousy as well as infidelity can ruin marriages as the prospective study of divorce has shown (Amato and Rogers 1997). The problem is that it is almost impossible to know whether the jealousy attributed to a spouse was in fact a legitimate response to actual infidelity. But does this mean that women don’t experience jealousy as intensely as men? They certainly do. We gain a very different and much less sex-biased view of jealousy from research that does not measure it by such an extreme act as murder. Women, as I have been at pains to argue, steer clear of lethal violence in general, so it comes as no surprise that they are less likely then men to risk lethal retaliation against their straying spouses. A number of studies have asked men and women to report how distressed they would feel in jealousy-eliciting situations such as watching their partner flirting, dancing, hugging, kissing or having sex with somebody else. One of the most extensive studies investigated jealousy in over 2000 people in seven countries (Buunk and Hupka 1987) and found no sex differences. Many other studies confirm this finding as David Buss (2000, p. 51) concludes in a wide-ranging review: ‘The conclusion is clear: women and men alike can be plagued by jealousy, both in its everyday manifestations and in its most florid clinical expressions.’ But what are women so jealous about? A man’s infidelity carries no parallel threat of maternal uncertainty. Evolutionary theory suggests that in a biparental species such as our own, women’s jealousy should centre chiefly upon the potential loss of a

Mhoc08.fm Page 256 Friday, December 14, 2001 10:18 AM


    

providing male. Males enjoy sex with a variety of partners (if they are able) and can do so with startling lack of discrimination and emotional involvement. Seventy-six per cent of men (but only 37 per cent of women) answered ‘Yes’ when asked ‘Have you ever continued to have sex on a regular basis with someone you did not want to get emotionally involved with?’ (Townsend 1995). Since desertion is more likely to follow if the male falls in love with another woman, women ought to be especially sensitized to the possibility that the man is becoming strongly emotionally attached to another woman. In an ingenious study (Buss et al. 1992), men and women were asked to form a strong visual impression of their partner either having sex or falling in love with somebody else. They measured changes in heart rate, sweating and facial expressions. As expected, men showed a stronger response to sexual infidelity than the women did. But most studies have used self-reports in which men and women are forced to make a choice. They were asked to select either sexual involvement or emotional involvement with another person as being the stronger elicitor of jealousy. The prediction that males would opt for sex and women for emotion was borne out (Bailey et al. 1994; Buss et al. 1992; Buunk et al. 1996; DeSteno and Salovey 1996; Geary et al. 1995; Harris and Christenfeld 1996; Shackelford and Buss 1996; Teismann and Mosher 1978; Wiederman and Kendall 1999). Typically slightly over a half of male subjects selected sexual infidelity as most upsetting. The sex difference was largely driven by females’ greater selection of emotional betrayal. Some have objected that these self-report studies are logically flawed (DeSteno and Salovey 1996). Suppose that most women, quite reasonably, believe that men are eminently capable of sex with a variety of women without emotional involvement but that sometimes a man may begin to form a strong attachment to one of those women. If a woman is told that her partner is emotionally involved with another woman, she naturally assumes that he has already had sex with her. Hence the women tend to choose this scenario as the most upsetting because it represents a ‘double shot’ of jealousy. Men, however, believe that women only have sex with men that they are already emotionally involved with—hence sexual infidelity implies the existence of prior emotional infidelity—the male ‘double shot’. Buss et al. (1999) addressed this issue by varying the instructions in their next studies. This time, participants were asked which would upset them more and the two choices were ‘Imagining your partner forming a deep emotional (but not sexual) relationship with another person’ or ‘Imagining your partner enjoying a sexual (but not emotional) relationship with that person’. The results were the same as the original study—men opted for sex (even without emotional attachment) and the women did the reverse. In a further study, they asked men and women to imagine that their partner was both emotionally and sexually involved with another person and they had to indicate which component of the relationship troubled them more. Once again, the original predictions held. However, in all these studies subjects are forced to make a choice. When men and women are simply presented with both scenarios and asked to rate each one

Mhoc08.fm Page 257 Friday, December 14, 2001 10:18 AM

   :     independently in terms of how upsetting it would be, no sex differences appear. Both sexes score between six and seven on a seven-point scale for each scenario (Buss 1989). In the real world, and in the world of our imagination, sexual and emotional betrayal are so tightly interwoven as to be effectively equal in their impact. Neither women nor men are happy about infidelity of either kind, yet homicide data reveal that men are far more likely to kill their partner because of it. Of 214 jealousymotivated partner homicides in Canada, the husband was the killer in 195 (Daly and Wilson 1988). Why should it be that equal jealousy results in such an unequal reaction? The usual argument is that more is at stake for men than for women. Infidelity by a wife entails the possibility of misattributed paternity and the waste of a substantial portion of the man’s resources on a child with whom he has no genetic relationship. A wronged wife bears no such cost. But there is nonetheless a very substantial cost in losing a supporting husband and the costs are felt not just by the deserted wife but by her children as well. Divorce, as we saw in Chapter 2, causes psychological damage to the children and can have an immediate and disastrous effect on the family’s financial situation. (In the United States deserting husbands are more likely to keep up their car repayments than their child support.) The costs are extreme and women without husbands have suffered marked disadvantage (to say nothing of demonization) throughout history (see Low 2000). But even if the reproductive costs are less disparate than we have thought, for men there may be associated social costs that women are free from. Could it be that a man’s reputation is more damaged by an unfaithful partner than is a woman’s? Certainly cases of pathological jealousy indicate that impotent men or those who are concerned about their inability to satisfy their wife are especially prone to believing that she is having an affair. If similar thoughts run through the minds of neighbours and friends, men may feel that adultery is a more serious affront to his reputation in the community than a woman might (Buunk 1986). For a wife, men’s proclivity for variety in sexual partners means that an affair can as easily be attributed to her husband’s excessive sexual desire than to her particular failure as a sexual partner. On the other hand, a study of five pathologically jealous women indicated that common themes in their lives were their belief that others knew of his affairs and their distress at the loss of status they experienced in the face of others’ sympathy (Buss 2000). Another study which asked participants about the expected loss of self-esteem that they would suffer if their partner was unfaithful showed no significant differences between men and women (Paul et al. 1993). Could the difference in jealousy-motivated homicides stem from a difference in how easily the two sexes are provoked to anger? To examine this we have to narrow our focus to rating studies that have explicitly asked about anger in reaction to infidelity. Most studies find that women report as much (Hansen 1982; McIntosh 1989; Paul et al. 1996; Pines and Aronson 1983; Pines and Friedman 1998; White 1981; Wiederman and Allgeier 1993; Yarab et al. 1999) or more anger (Bryson 1991; Buss

Mhoc08.fm Page 258 Friday, December 14, 2001 10:18 AM

      1989; Buunk 1981, 1982; de Weerth and Kalma 1993; Geary et al. 1995 (US sample only); Hansen 1985; Paul et al. 1993) in response to betrayal than men. These studies, like others that have investigated anger in response to less traumatic events, suggest that it is not the emotional drive to aggression that differs between the sexes. Indeed, jealous women seem to engage in a good deal of fantasy about revenge. When asked to report on their likely behavioural reaction to infidelity, women scored higher than men in their reported likelihood of physically and verbally abusing their partner (de Weerth and Kalma 1993) and in their likelihood of showing their anger and destroying a sentimental gift (Paul and Galloway 1994). But women seem less inclined to act on their violent impulses—they are better at inhibiting their anger as we might expect if a woman’s reproductive success hinges on ensuring her own survival. Homicide is an extreme response to jealousy and one that is fortunately rare. The convergence of extreme male intolerance of possible cuckoldry together with their less inhibited approach to aggression places wives in a far more vulnerable position than husbands. But even when divorce rather than murder is the end-point of an infidelity-plagued marriage, it is the wife’s not the husband’s betrayal that is critical. This certainly reflects the strength of evolved male jealously but the gender disparity is also augmented by the economic self-sufficiency of fathers compared to mothers. Where the wife and children depend upon the husband’s economic support, women may tolerate sexual dalliances provided that emotion and resources are not diverted away from her household. Just as women accept but do not embrace polygyny, they may live with infidelity if the costs of leaving exceed those of staying behind.

The pros and cons of children Marriage is intimately tied to the production of children. Even today where cohabitation is common, couples will often marry in anticipation of or in response to pregnancy. Children are ‘the crown of marriage’ (Proverbs 17:6) according to the Bible. In India, Andaman Islanders believe that the marriage is not consummated until a child is produced. In rural Japan, marriages are not entered in the family register by the village officer until the first child is born. Among the Tiv of Nigeria, the husband pays only a small bride-price for his ‘trial’ wife with further instalments made on the birth of each subsequent child. In other societies, the husband may demand his money back or the provision of a replacement woman if his wife fails to give him children (Daly and Wilson 1988). Betzig’s cross-cultural analysis found that a variety of factors linked to childlessness all constitute grounds for divorce: the death of existing children, too few children, and failure to produce sons. In addition, factors which mitigate against pregnancy are also legitimate grounds for divorce: refusal to have sex and—in the case of women only—old age. The production of children (at least where paternity is not in doubt) creates a mutual genetic investment between two otherwise unrelated individuals. The child represents half of each parent’s genes and ensuring its survival is of equal importance

Mhoc08.fm Page 259 Friday, December 14, 2001 10:18 AM

   :     to both. Children, from the point of view of evolutionary theory, provide the glue that holds two people together despite their lack of kinship. We should expect that this bond would be particularly strong during the time when the child is most vulnerable—in its early years. As a topic, marriage has been the traditional preserue of sociology and economics. Sociological theories propose that marriage is based upon the sexual division of labour (Durkheim 1933). This enhances the ties of mutual dependence between the marital partners and sex-linked specialization traditionally reaches its peak when children arrive—the mother taking charge of the home and children and the father assuming primary responsibility for financial provision. This division of labour is especially intense in the preschool years because infants require more time and attention than older children. As children grow up, they attend school, and siblings, friends and fathers (rather than the mother alone) occupy an increasing part of their leisure time. The child’s age has implications for divorce. As children grow older, divorce means that the non-custodial parent gives up less current and future time with the child. Younger children are seen as more vulnerable than older children and the loss of a parent through divorce is believed to have more harmful effects. Divorce often propels the mother back into the full-time labour force and parents feel that this may be harmful when the child is young. Although phrased in terms that omit consideration of the protection of genetic investment, and highlighting the immediate circumstantial (rather than the evolved adaptational) causes, the theory makes predictions that are quite compatible with those of evolutionary theory: small children cement marriages. Economic theories focus on investment in the present marriage and the availability and attractiveness of alternatives (Becker 1981). Children, in this view, are seen as investments that are likely to deter divorce. Although phrased in market-place terms, there is no real contradiction between economic and evolutionary theories as Daly and Wilson (1988, p. 188) point out: A pair of beavers maintain their dams and domicile through the winter. We can call these unions, with their division of labour and exchanges of benefits, ‘economic’ if it pleases us to do so. All animals accumulate and allocate resources, ‘economic’ activity if you like, but economic activity that has evolved because of its reproductive consequences.

Economic theory makes no formal predictions about the effects of children’s age on the probability of divorce. But it does argue that the greater the investment, the less the probability of ending the union and this implies that divorce should be most likely when children are younger because the investment-to-date is lower. While the Concorde effect—the logically flawed belief that having invested so much in the past one cannot ‘pull out’ of the deal now—may represent faulty reasoning, it seems to be a cognitive error to which humans are uniquely vulnerable (Arkes and Ayton 1999). As for the second part of the economic equation—the availability of alternatives— men have the upper hand. The chances of a mother remarrying are lower than those

Mhoc08.fm Page 260 Friday, December 14, 2001 10:18 AM


    

of comparably aged, childless women, so in general a mother has fewer outside alternatives than a father. Remarriages with stepchildren—usually the woman’s— are more likely to fail and stepchildren are a source of tension and discord in second marriages (Teachman and Heckert 1985; White and Booth 1985). Nevertheless, a younger mother is better placed than an older one—although age decreases remarriage market value, the effect is much more marked in women than in men. This suggests that a young woman (with young children) is likely to have more alternatives than an older woman (with older children). Again the prediction would be that divorce should be more common among those with younger children. Given these competing hypotheses, what do the data tell us about the effect of children on marriage? Data from modern societies around the world show that the chances of divorce decrease with increasing number of children (Betzig 1989; Fisher 1989). But the problem is that the duration of marriage and the number of children produced are obviously correlated and the longer a couple remain together (with or without children), the lower is the probability of divorce. But studies that control for the length of the marriage still show that couples with no children divorce more often than those with one child who, in turn, divorce more often than those with three or more children (Betzig 1989; Jacobson 1959). But the children that are counted in these studies may not necessarily reside with the parents, they may be adults who have left home and stepchildren may be included (whom, as we have seen, are likely to increase the probability of divorce). And studying the effect of children as a whole does not distinguish possible different effects of number, age and spacing of children. Cherlin (1977) examined the effects of preschool children, older children, total number of children and premarital childrearing. Only the presence of preschool children had an effect on the probability of divorce—it lowered it. This was replicated by Morgan and Rindfuss (1985) who found that first-births decreased the probability of divorce in all marital cohorts and at all marital durations. A study by Waite and Lillard (1991) clarified the relationship further using national data from 4400 American first marriages (of which 1424 ended in separation) followed from 1968 to 1985. They found that childless couples face a much higher likelihood of divorce in their first year together but the likelihood falls gradually over the years. For those who had children, the first birth dramatically lowered the risk of divorce until the child reached the age of 6 and each later birth had a similar though smaller effect. Couples with children under 5 had a much lower risk of divorce than other couples. But as the child grows older, the marriage-protecting effects decrease and by the time the child is 13 the chance of its parents divorcing actually increases. In short, parents have a very much lower risk of divorce when their first child is young and the risks remain lower (though less dramatically so) until their children become teenagers. As the authors explain: ‘The early protective effects of children are largely offset by the later disruptive effects, making the overall effect of

Mhoc08.fm Page 261 Friday, December 14, 2001 10:18 AM

   :     children on marital stability quite modest’ (Waite and Lillard 1991, p. 948). The overall number of children that a couple has simply changes the timing of the marital disruption but has little effect on its long run possibility. What is going on here? Research indicates that couples with children do not have happier marriages than childless couples and they are not happier with their lives in general (Glenn and McLanahan 1982). For many less-than-happy childless couples the marriage ends quickly, most frequently during its first year. However, had these same couples had a child during that time, there is a much greater chance that they would have stayed together. But this would not be because they were any happier. It may well be to do with the high dependency and demandingness of a young child which ‘glues’ the parents together. The chance of divorce remains lower until the last or only child reaches the age of 13. At this point, the presence of the child actually increases the probability of divorcing. Why? It might be that the strain of living with a turbulent adolescent is simply too much for a less-than-perfect marriage to bear. If this were the case we would expect that, after the child leaves home, there would be a significant reduction in the likelihood of divorce—simply removing the troublesome adolescent should ameliorate the situation. So far, this remains an untested proposition. Another explanation is that the birth of a child locks together couples that would have otherwise separated— they hold their discontent in check until they believe that the youngest child is sufficiently mature to no longer need their joint attention. This suggests that the bond between spouses depends heavily on their mutual investment in dependent children. When that investment has been realized, little remains to unite them. The pressure to remain together is strongest while the children are young, weakening every year until the last child reaches the age of 13, and finally reversing. Could it be that lifetime monogamy was not a part of the environment of evolutionary adaptation? Could it be that men and women were designed to remain together only as long as it took for their children to become self-sufficient? Could the seven-year itch have a basis in fact? Helen Fisher (1989) believes the itch has been exaggerated by about three years— humans evolved to strike four-year bonding agreements. Her argument has three strands. First and most obvious, the distribution of divorce plotted on duration of marriage: marriages worldwide are most likely to end in the fourth year. These data, however, are complicated by the fact that we lack information on time spent in premarital courtship, engagement, kin negotiations and wedding preparations. Nor do we know how long the marriage formally continued despite being effectively over or the time taken to obtain a divorce (which can vary from minutes to years). In short, human monogamous relationships probably both start and end earlier than official records show. In addition, the modal ‘peak’ of four years is not strongly marked—more marriages end in the second and third years combined that in the fourth year. The second strand of her argument is that the modal age for divorce is between 25 and 29 for men and between 20 and 29 for women. This is young and

Mhoc08.fm Page 262 Friday, December 14, 2001 10:18 AM


    

clearly occurs during the peak reproductive years. Taken together with the fact that 80 per cent of divorced men and 75 per cent of divorced women remarry, Fisher believes that there are grounds for suggesting systematic serial monogamy. Her third argument bears upon the impact of children. The fact that young children suppress divorce and that typical birth spacing among hunter-gatherers is four years (identical to the modal length of marriage) suggests that monogamous unions were created primarily to support the raising of one child before the couple separated and each moved on to a new partner. Research on the physiology of sexual attraction suggests that it typically lasts between two and three years; just the right length of time to fall in love, produce a child and get it to the beginning of an independent life before starting the whole process again. What might be the benefits of such a strategy? Fisher suggests that it enabled females to escape from pair bonds that proved unsatisfactory, it meant that males could select younger and more fertile mates and both sexes could benefit from increasing the genetic diversity of their lineages. But there are problems with this argument. At age 4, the child does not disappear and it is the woman who continues to have the major responsibility for its well-being. Why should her next male partner voluntarily take on the additional burden of a stepchild? (Indeed, depending upon the age of the mother, he may be taking on responsibility not just for one but for several children to whom he is genetically unrelated.) Given everything we know about discriminative parental solicitude, why would a woman voluntarily want to expose her child to the indifference and possible cruelty of a stepfather? Given the deleterious effects of stepchildren on subsequent bonds, why would a woman expect to form a successful union with her next partner? As she ages, reaching the farther reaches of her fecundity and gathering more and more children, a woman becomes a less and less attractive sexual partner. A husbandswapping strategy seems to serve her badly when she is competing against younger, attractive and unencumbered nuliparous competitors. Fisher’s reply is that by age 5, children in hunter-gatherer societies spend a considerable amount of their time with age-mates and other members of the village so that they make minimal demands upon the parents. While this may be true relative to the extensive demands made by Western children, it is also true that there is considerable community reluctance to provision children whose parents do not make a significant contribution to subsistence. Children cannot be sloughed off at 5 leaving the mother free to pursue her subsequent genetic interests unimpeded. Children’s parental demands may lessen with time but they do not disappear. This is not to say that humans are perfectly evolved for contented monogamy. There is simply too much adultery (and too much behavioural mate guarding and physiological sperm competition) to imagine that this is true. People seem to keep a running tabulation of alternative possibilities to their current union (Udry 1981). The risk of divorce rises where either wives or husbands encounter an abundance of spousal alternatives. Where the operational sex ratio is favourable, it increases people’s willingness to abandon their marriage. Where there is an excess of women, men are more likely to

Mhoc08.fm Page 263 Friday, December 14, 2001 10:18 AM

   :     seek a divorce and the presence of a disproportionate number of unmarried, employed women is especially associated with this trend (South and Lloyd 1995). Nonetheless, although people may be acutely aware of their value in the marriage market, the presence of young children makes them less likely to take advantage of this knowledge. At the same time, adolescent children seem to have the reverse effect of increasing parents’ willingness to separate. To understand this we have to consider once again Trivers’ insights about parental investment. Reproductive investment is a limited resource. The time and effort expended on one child preclude expenditure on other activities so that parents (especially mothers) have to be selective about the amount of care devoted to any one child. Matters look rather different from the child’s perspective. It is in her best interests to sequester as much of that investment as possible. While children have been designed to extract whatever resources they can, parents have been designed to withhold resources after the child reaches a point of independence. During the last century, enormous social changes have altered the tempo of the life cycle. Nutritional improvements mean that children reach puberty earlier than ever; many girls now reach menarche while still at primary school. These physical signals are those that would normally signal the imminent reproductive independence of the child and parents continue to read these changes as indicators that the child is now less reliant upon them. The same nutritional benefits (together with stateprovided education) have raised the mean IQ considerably over the last century but there is little evidence that there has been any major acceleration in children’s acquisition of emotional autonomy. Teenagers respond with considerable distress to divorce which from their viewpoint looks like a halving of parental investment and may even auger parents’ new investment in half-siblings. Although the age of children’s sexual maturity has lowered, contemporary society has simultaneously extended the period of care required from parents by the introduction of compulsory secondary schooling, the encouragement of tertiary education and the debarring of children under 16 from the labour or marriage market. Parents are being asked to invest in their children for longer than ever before and it seems that many, despite their best efforts, reach breaking point when the youngest child reaches sexual maturity.

With all my worldly goods Monogamy offered ancestral women male assistance with provisioning especially during the years when her children are young and demanding of her time and energy. Note the word ‘assistance’. There is a real danger of imagining a primeval scenario in which the woman sits by her hut singing lullabies to her children, waiting patiently for her mate to return with food for them all. This picture corresponds quite closely to a particularly unusual time and place for the division of labour—the Western world in the postwar era. So short is our historical perspective that sociologists are tempted to see women’s entry into the labour force in the late twentieth century as something of a novelty. But this is a serious misunderstanding.

Mhoc08.fm Page 264 Friday, December 14, 2001 10:18 AM


    

If contemporary hunter-gatherers are a useful proxy for the lifestyle that we enjoyed about 100,000 years ago, then women played a very active role in provisioning the family. Two or three times a week a woman would set off carrying her youngest child and perhaps with an older child in tow into the surrounding territory to search for food—wild onions, mongongo nuts, plums, baobab fruit, bird’s eggs, tortoises and honey. Men might hunt several days a week but on average they made a kill on every fourth attempt—a woman who depended solely on men would have starved. The shift away from nomadic hunting and gathering life came with the swidden farmers—sometimes called ‘slash and burn’ agriculturists (Jolly 1999). In the Amazon forest, they cleared trees, created fields and grew plants before moving on to begin the cycle anew in a different spot. Eventually they would return after sufficient time had passed for the land to become enriched again during its fallow period. The earth took carbon and minerals from the burned trees, new tree shoots turned the soil over, weeds died out as the trees grew above them and the crop pests were forced out. Population density meant that this pattern of continuous relocation could not go on for ever. As the number of people increased, the fallow time shortened. Soon the areas to which they returned had not had time to regenerate into full forests, they were still grass meadows. Eventually, movement to new spots became impossible and the only way to sow seed was to plough. Helen Fisher (1993) believes that the plough has much to answer for. Ploughing is a task too heavy and timeconsuming to be taken on by a woman with children. And a man needed a woman to sow, weed, pick, prepare and store the vegetables and crops he produced. Men and women became dependent on each other. Agriculture also meant that there was a joint investment in the family’s parcel of land. If one partner wanted to move away, they would lose one-half of their lifetime’s work and have no capital to buy another. As Fisher (1993, p. 106) puts it: ‘Farming women and men were tied to the soil, to each other, and to an elaborate network of stationary kin. Under these ecological circumstances, divorce was not a practical alternative.’ The greater the economic dependence, the greater the expressed commitment to (though not necessarily happiness with) the marriage (Nock 1995). The lowest rates of divorce have been found in pre-industrial Europe and in India and China—all cultures where ploughing and agriculture were the chief means of subsistence. Then came the industrial revolution. Men and women left the land to exchange their labour directly for money. Money is paid to an individual, not to a couple. Even mutual wealth can be divided and carried away. The divorce rate began to climb in the early twentieth century as women began to earn their own wages. In the postwar years there was a drastic but temporary interruption of this pattern as men returned from the war and women vacated their jobs to make way for them. Women were warned of the dangers of ‘failing their children’ by leaving the home and a full-time wife at home became a status badge for men. But women were not content to be relegated to the kitchen permanently. With the arrival of the liberation movement, women entered the waged economy in increasing numbers and at the

Mhoc08.fm Page 265 Friday, December 14, 2001 10:18 AM

   :     same time the divorce rate began to soar, doubling between 1966 and 1976 (Cherlin 1981). The moral of the story is straightforward. Men and women remain married when their fates are economically interlinked. When women revert to the kind of independence that predated agriculture (about 10,000 years ago), they are less likely to remain married. Out of 331 marriages documented by one anthropologist among the Kung San, 134 ended in divorce (Howell 1979). Among the Hadza of Tanzania, the divorce rate is five times higher than in the United States (Friedl 1975). But waged work in contemporary western society presents mothers with problems that they simply did not have to face in pre-industrial society. How can a woman be in two places at the same time? How can she earn some financial autonomy and simultaneously earn her child’s love? It is a new and cruel bind. Farm work allows child care to take place alongside milking or weeding. Foraging can be done with children as well as alone. But work in the industrial or service economy takes place at regulated times and places at which infants are not welcomed. To make matters worse, the increasing mobility of the labour force means that there is good chance that she is far away from supportive and babysitting kin. She may not even remain in one place long enough to establish strong bonds with neighbours. If she elects to work, she must place the safety and happiness of her child (as well as the lion’s share of her pay packet) in the hands of a stranger. To buy the right to work, working women must rely on a second income from their spouse (Herzog et al. 1983). Perhaps then it should come as no surprise that, not just among societies living on the edge of starvation but in ‘liberated’ ones also, a leading cause of divorce is a man’s failure to provide for his family. If marriage is an agreement in which the woman is responsible for child-rearing (even if she also holds a waged job) and the man provides economic support, it seems reasonable that failure to hold up his end of the bargain might well lead to trouble. From this we can make predictions. Marriage should be less frequent and divorce more common under two particular conditions. First, when men’s economic situation makes them poor providers. Secondly, when women are able to earn sufficient resources to obviate the need for male assistance. Black women are reluctant to marry men who have little to offer (Cherlin 1981). The high proportion of births to single mothers (which rose from 42 to 70 per cent between 1960 and 1989; US Bureau of the Census 1992) is chiefly the result of a declining rate of marriage rather than an increase in the birth rate to younger women (Weinraub and Gringlas 1995). With economically supportive males thin on the ground, young women are likely to seek support from female kin and this arrangement has knock-on effects on their fertility patterns. Black women die earlier than their white counterparts and are more prone to illnesses that make child care difficult or impossible. In black urban communities, a teenager who bears a child has a 75 per cent probability that her mother will still be alive and able to assist in child-rearing when her child is 5 years old. A woman who waits to the age of 20 has

Mhoc08.fm Page 266 Friday, December 14, 2001 10:18 AM


    

only a 40 per cent chance of maternal assistance (Low 2000). When men cannot provide resources, women adapt and adjust their reproductive strategy. A husband who does not contribute is not merely economically neutral. The wife is effectively paying for the privilege of having him around and, given the still modest contribution of men to household duties, she is likely to receive little in return. Ethnographic studies of black ‘cultures of poverty’ in the United States have documented again and again the male response to such situations. Unable to effectively contribute in the home and viewing it as a female preserve, the men head for the streets to pass their days talking, drinking and attempting to salvage some pride—if only in their ability to attract women despite their conspicuous absence of resources (Liebow 1967). Divorce rates among black Americans are higher than among other ethnic groups (Norton and Glick 1979). However, when the effects of income, differences in family size and home ownership are controlled, black families are less likely to separate than whites (Hampton 1975). This underlines the central importance of economic rather than cultural factors in divorce. Black men earn less, are unemployed more often and occupy low-status positions—all of which make them economically dispensable (Becker 1981). Across all ethnic groups there is a clear relationship between socio-economic status and divorce (Kitson et al. 1985). Divorce is higher among the less educated and those with the most restricted access to high earnings. A husband’s level of education remains a protective factor for divorce, even controlling for his income, which suggests that women who are married to men with the educational potential for high earnings are less likely to divorce. Divorce rates rise as the husband’s job status and income decline and as his unemployment and instability of work and income increase (Cherlin 1979; Hoffman and Duncan 1995). Regardless of how much the wife earns for herself, women married to richer men are more likely to stay married to them. But the divorce-protective effect of being married to a wealthy man is most pronounced among couples where the wife has no income of her own and is reliant on her husband for support (Ono 1998). Our second prediction is that increasing economic success of women will also make marriage less appealing and divorce more likely. The increase in cohabitation without marriage began to soar in the 1970s at the same time that women began to gain financial autonomy and had less financial need to tie themselves to one man (Fisher 1993). The steepest increase in single-birth rates has been among white, employed, college-educated women (Bachu 1993). The percentage of collegeeducated women having a child out of marriage rose from 5.5 per cent in 1982 to 11.3 per cent in 1992 (Weinraub and Gringlas 1995). As for divorce, the anthropological record attests to the fact that in societies where women have more economic power, the divorce rate is higher. In the Mayan peninsula, the Caribbean, New Guinea, southern Africa, Polynesia, Alaska, Brazil and among North American native people, there is a consistent relationship between female autonomy and divorce (Fisher 1993). Among the Yoruba of western

Mhoc08.fm Page 267 Friday, December 14, 2001 10:18 AM

   :     Africa, women take principal responsibility for growing crops and selling them at an all-female market. The women have their own income and nearly half the marriages end in divorce. Sociologists have generally supported the ‘independence effect’ of working wives; the wife’s income increases the likelihood of divorce (Ross and Sawhill 1975; Spitze and Spence 1985). But a mirror-image argument has also been proposed called the ‘wife’s income effect’ according to which the wife’s earnings can help to stabilize a marriage by reducing economic strife. A study by Ono (1998) helps to clarify the apparent contradiction between the two positions. He found that divorce was more likely when the wife had no earnings than when she earned a small or moderate amount (up to US$18,000 per annum). This seems to be a ‘wife’s income effect’ where the extra money she contributes seems to stabilize the marriage. Her salary is not enough to live on independently but may be enough to remove some of the economic stress that contributes to marital disruption. However, when her earnings rise above US$18,000 the odds of divorce increase significantly—in line with her ability to support herself independently. Women with the highest levels of education (graduate degrees) and lucrative jobs are also more likely to divorce. But perhaps more significant is the disparity in earning power between a wife and her husband—women who earn more than their spouses are more likely to divorce (Fisher 1993). But there is a twist. Heckert et al. (1998) examined over 2000 couples and divided them into four categories; the traditionals (the wife earns 0 to 25 per cent of the income), new traditionals (the wife earns 25 to 50 per cent of the income), non-traditionals (the wife earns 50 to 75 per cent of the income) and reverse traditionals (he wife earns 75 to 100 per cent of the income). They followed the sample over three one-year periods starting in 1986. Taking traditional couples as their baseline for the likelihood of separation, they asked the question: How do these other couples compare in terms of their likelihood of divorce? The only significant effect, in line with our prediction, was for the ‘nontraditional’ group (where the wife earned the bulk of the couple’s income) who were over twice as likely to divorce. But what then of our brave ‘new women’ who earned virtually all of the couple’s income? Why did they not simply leave the marriage? To understand, we have to radically rethink our natural stereotype of these woman— they were far from being hypersuccessful career women with the world as their oyster. These women earned far less than the other two groups of working women (about US$14,000 p.a. in fact), only half of the husbands worked and the chief reason for this was that they were permanently or temporarily disabled. From the point of view of evolutionary theory, something has got in the way of our clean prediction that these wives should leave the marriage. And it seems to be their heart. Emotions, not rational calculation, drive human behaviour. Love, an emotion understudied by mainstream psychology, is as much an adaptation as fear or anger. It holds us in pair bonds (mostly) and creates a stable world for children to grow in. Once it takes hold, it is impossible to voluntarily switch off.

Mhoc08.fm Page 268 Friday, December 14, 2001 10:18 AM


    

But women (and men) vary in the way that they respond to this as to every other emotion. It can impel us to jealousy or murder and also to extremes of self-sacrificing care. Some of these differences are a product of circumstance but even in the very same situation, people differ. Faced with the same philandering husbands, some women stay and some go. Some smile through and others crack up. How can we explain these differences between women? How can evolution, with its emphasis on the universality of human nature, cope with the evident fact of our individuality? It is to this final challenge that we now turn.

Mhoc09.fm Page 269 Friday, December 14, 2001 10:18 AM

 


Counting the ways: The unique woman

Women are everywhere the same and yet everywhere different. How can we reconcile the idea that natural selection winnows down the variability and hones in on an improved (but never perfect—environments change, arms races continue) female prototype with the manifest evidence that each woman is unique? How can there be a universal female nature when every baby that is born carries her uniqueness not in only her fingerprints but in her shyness, determination or musical ability? There can be few other areas of evolutionary psychology that raise serious critics’ hackles more than its apparent failure to acknowledge the manifest differences between people. And it has taken time for evolutionary psychologists to seriously address this issue. Their principal quest has been to describe the universal aspects of human behaviour, to find the fundamental commonalities that typify our species. This task has been driven by the need to challenge wholly environmental theories that argue that the human mind contains only the capacity for learning and little else. The tactical emphasis that has been placed upon a species-typical psychology has obscured consideration of individual as well as cultural differences. At a more pragmatic level, the emphasis upon human universals has also provided a bulwark against those who have tried to use evolutionary premises to argue for innate racial differences. The greatest enemy of evolutionary psychologists is the ghost of Social Darwinism that threatens to taint the scientific research programme. Some believe that opening the door to individual differences might also open it to group differences and to the exploitation of Darwinian theory for racist politics. Increasingly, however, evolutionary psychology is trying to create a socio-biological framework in which individual differences can be understood. The task has only just begun and as we shall see that there are opposing positions and, to date, more theoretical speculation than corroborating data. The principle from which most theorists start is that there is a single fundamental human design. The basic components of mind—long-term memory, language acquisition, the ability to recover depth from a two-dimensional retinal image, the use of electrical and chemical signals to pass information along and between nerve cells—everywhere share the same fundamental architecture. When we speak of individual differences we speak of modifications to this blueprint rather than

Mhoc09.fm Page 270 Friday, December 14, 2001 10:18 AM


    

differences in kind. Individuals may vary in their threshold of response to certain stimuli or in their sensitivity to certain neurotransmitters but the modules of the brain are built and organized in much the same way for everyone. The difference between an extravert and an introvert does not depend upon them being born with a different kind of mind—although it may depend upon their being born with a different rate of synthesis or uptake of certain neurotransmitters. Brains have been built by millions of years of evolution and nobody seriously supposes that natural selection operated on two spatially distinct ancestral species—extraverts and introverts—converging by sheer chance to produce two brains that were indistinguishable except for a difference in their sociability. The emergence of a unique never-to-be-replicated person is a result of environments and genes and their interaction with one another. To find our way through these tangled sources, let us start on the outside with the impact of different environments and work in towards the more theoretically challenging issue of why genetic variability persists before we consider in what ways the two elements can and do affect each other. As we do so, it is useful to bear in mind a distinction between alternate and conditional strategies for I shall be discussing both. Alternate strategies are consistent and stable paths taken by different individuals. They correspond roughly to what most of us call ‘lifestyles’ and psychologists refer to as ‘traits’. They are enduring differences between people but their stability over time, as we shall see, does not necessarily require that they be of genetic origin. Conditional strategies are ‘if . . . then’ rules which tailor behaviour to the immediate circumstances. They correspond to what we might call tactics and what psychologists might call ‘situational effects’. Their flexibility, as we shall see, need not necessarily imply that they are of purely environmental origin.

A single human nature meets a variable environment Let us imagine a population of individuals who do not differ genetically in any way from one another. They are effectively genetic clones—identical twins on a massive scale. Under such circumstances we might still expect to see differences in their behaviour as a result of the interaction of exactly the same genes with different environments. The same mental module if given different inputs will produce different outputs. Before we continue with this thought experiment bear in mind that when we speak of environmental effects we are not saying that genes do not matter. Holding genes constant is not the same as removing their effect. It is our genes that allow the environment to act upon us and then only in certain ways. As Tooby and Cosmides (1990, p. 21) explain: The environment per se is powerless to act on the psyche of an animal except in ways specified by the developmental programs and psychological mechanisms that already happen to exist in the animal at a given time. These procedures take environmental

Mhoc09.fm Page 271 Friday, December 14, 2001 10:18 AM

  :   


information as input and generate behaviour or psychological change as output. The actual relationship between environment and behavior is created solely by and entirely by the nature of the information-processing mechanisms that happen to exist in the animal . . . The smell of excrement may be repulsive to us but it is attractive to dung flies.

In our population of clones, differences would first arise as a result of whether or not a mental module was activated at all. Sexual jealousy is a good example—a mental module designed by selection to ensure mate fidelity thus increasing paternal certainty and paternal investment. There are conceivably people who have never experienced this psychological state. In past times, when the media was not available to create widespread familiarity with this emotion, religious celibates may have gone to their graves quite unaware that such an emotion existed. Although the module is present in everyone, it is not necessarily activated in everyone—that depends upon environmental factors such as attraction to a desirable partner, forming an attachment to them and becoming sensitized to cues that threaten the relationship. But the environment can also contribute not just to the temporary state of being jealous but to the enduring trait of jealousy as well (Buss 1991). Lifelong vulnerability to jealousy might turn out to have a genetic component but that is not the only possibility. A state can become a trait if the situation that generated it in the first place persists over time. A woman may continue to experience jealousy over several months or years if her husband continues with his infidelity for that long. Or the environment might recalibrate her threshold of jealousy to a lower value. The wife may become aware that her value on the marriage market is lower than her husband’s. A chance remark that she ‘made a good catch’ or ‘got lucky’ is enough to plant such an idea. From then on, even a relatively innocuous event (a glance held too long, an enthusiastic kiss in greeting another woman) can trigger apparently excessive bouts of jealousy despite continued reassurance from her husband. Chance events that alter her attractiveness may manifest themselves as an enduring change in her level of jealousy. A facially disfiguring scar or a successful diet can lower or raise her assessment of her own attractiveness and lead to a lifelong change in her level of jealousy. As another example, long-term changes in a man’s financial situation can alter his ability to attract women and hence to pursue the malepreferred strategy of short-term relationships. If a man of only average good looks accumulates wealth over the course of his career, he may be able to jettison his longstanding wife in favour of a series of younger women (as long as the money lasts). Chance events such as a lottery win or a risky but successful stock-market investment can have the same effect. Temporary or local changes in the sex ratio can create market forces that allow one sex to cash in their preferred mating strategy, causing what we might conclude was a change in their personality. The arrival of US army personnel in Britain during World War II was greeted with delight by the many indigenous women whose menfolk were stationed abroad. The increase in fatherless Anglo-American children during the 1950s was perhaps a predictable result of a sex ratio imbalance combined with the influx of chocolate-laden, stocking-rich men.

Mhoc09.fm Page 272 Friday, December 14, 2001 10:18 AM


    

The change in American men’s behaviour when they reached British shores might have looked like a personality change but was a response to a favourable changed sex ratio. In childhood, the environment provides important cues about the social world to which the child must adapt their behaviour. Early experience can alter personality over the long term in one of two ways. It can recalibrate a mental organ or trigger a distinct switch in life history strategy. Abused children are at greater risk than others of growing into aggressive adults—a puzzle that has exercised psychologists for several years. Surely the dreadful experience of this kind of treatment would be expected to cause feelings of repugnance for violence? But violence witnessed at an early age may act as useful data telling the child about the degree of competitiveness in the prevailing social climate. This could cause the threshold for violence to be reset to a lower level in order to increase that individual’s chance of survival in a hostile environment. The longitudinal stability of aggressiveness from an early age suggests that just such a mechanism may be operating. Other childhood experiences may shunt the child into a different life history trajectory. Children from father-absent homes are more likely to display a suite of behaviours that can be characterized as a preference for a short-term mating strategy. Children from such families, as we have discussed, show a pattern of insecure attachment to parents, reach menarche and have their first sexual intercourse earlier, have a larger number if sexual partners and are more likely to have unstable partnerships (Draper and Belsky 1990). The early experience of an unpredictable family and a transient father signal the future environment that the child faces. Under such conditions, making the best of a bad job entails early and frequent reproduction with lower investment in each child. All individuals have the potential for both strategies but the one that is activated is dependent upon environmental input. The developmental programme is triggered by an ‘if . . . then’ switching device. And a given strategy is not simply the result of selecting one behaviour rather than another (for example have sex now not later) but entails the activation of a whole suite of changes—psychological and biological—that together constitute a coherent reproductive alternative. If the environment is unpredictable and resources scant, then invest in mating and hope that some of your many offspring survive. If the environment is more predictable and benign, invest in parenting and forgo quantity of offspring in favour of quality of rearing. These examples are not without interpretative problems. The link between parent and child may be a genetic rather than environmental one. Studies suggest that aggressiveness is heritable with genetic effects accounting for approximately 40 per cent of the population variance (Rushton et al. 1986). A study which replicated the relationship between childhood stress and pubertal timing also found that a genetic model provided the best fit to the data (Moffitt et al. 1992). A relationship was found between the mother’s and daughter’s age at menarche. If mothers can transmit developmental timing, it is more than likely that they can also transmit a suite of

Mhoc09.fm Page 273 Friday, December 14, 2001 10:18 AM

  :   


personality characteristics that are associated with an unstable temperament that breaks both the mother’s and the daughter’s partnerships and interrupts their ability to forge strong bonds with their children. Indeed MacDonald (1997) has argued that differences in degree of parental investment are themselves heritable. But the key theoretical question concerns the mechanism by which environmental factors lead to different phenotypic effects. Here we will consider two possibilities— one that sees a biological chain of events leading from external variables to changes in behaviour and another more cognitive suggestion that emphasizes the role of information processing and strategy selection.

How environments change gene activation Although each individual receives their full genetic complement at birth, the activation of genes responds to the environment. The idea of environment affecting genes has a way of appearing mystical until we understand a little about how such a transaction might be physically realized. Genes have two major functions. The first is self-duplication, through meiosis and mitosis. The second function of DNA is the transcription of the genetic code. This code, held in the 23 chromosomes of each cell of our bodies, is translated into messenger RNA which moves from the cell nucleus to the cell body where it is translated by ribosomes into amino acid sequences that are the building blocks of proteins and enzymes. These are called structural genes and their activity is largely unaffected by the environment. These are the genes that we most often think of— the Mendelian structures that give us blue eyes or long legs. But most genes are not structural but regulator genes that code for products that can bind with DNA and so control the action of other genes. Regulator genes are able to interface with the environment. An early description of how such an effect could occur won the Nobel Prize (see Plomin 1994). The single-celled bacterium Escherichia coli depends upon lactose for nutrition and so needs to be able to digest it. Next to the structural gene that codes for the enzyme that E. coli needs to metabolize lactose is an operon that switches the gene off and on. But how does the operon ‘know’ when the gene should be activated? Far away on the genome, there is a regulator gene that produces an amino acid product—a repressor—that is able to bind both to the operon and to lactose. When there is no lactose available to consume, the repressor binds to the operon and, in doing so, switches it off. When lactose is present, the repressor binds to the lactose instead and this turns the operon on. In this way, the gene which enables lactose metabolism is only activated when lactose is present. On a grander scale, it is clear that embryonic development itself (including neural development) is not the straightforward transcription of genetic information but is acutely responsive to the internal environment of the foetus. Neurotrophins penetrate cells, selecting which genes should be activated and so instructing neurons into

Mhoc09.fm Page 274 Friday, December 14, 2001 10:18 AM


    

specific differentiated pathways. They can be shunted structurally into sympathetic nervous system cells or sensory neurons. And they can be shunted functionally into different chemical languages by switching on genes that synthesize and recycle particular neurotransmitters. This process may be at work in the behavioural problems that are seen in the children of malnourished women (Lukas and Campbell 2000). Maternal malnourishment at the time of conception reduces the level of nutrients that supply the foetus. In response, the foetus transfers whatever energy it can obtain to its most critical organ, a phenomenon called ‘brain sparing’. Malnutrition in the foetus leads to a reduction in competition between amino acids to cross the blood–brain barrier that favours the precursors of the neurotransmitters serotonin and dopamine. The high levels of these neurotransmitters force the brain to compensate which it does by reducing receptor sensitivity. The balance between the inhibitory effects of serotonin and the excitatory effects of dopamine are lost, resulting in patterns of hyperactivity, poor behavioural restraint and behavioural problems. It is likely that the downregulating of transmitter receptors is controlled by regulator genes responding to detected ‘overdoses’. In the human brain, genes are responsive to the external as well as the internal environment. Communication in the brain occurs chiefly between neurons but neuronal activity itself depends upon information obtained from the outside world (by the retina, for example) which is coded into the internal language of the brain. Neuronal cells carry incoming information electrically from their dendrites (which pick up as many as 150,000 inputs from adjoining cells) to the cell body, synthesizing their output and sending it along the axon to form part of the incoming information of their neighbouring cell. Between adjacent cells chemical messengers— neurotransmitters—convey the information across the synaptic cleft. And these neurotransmitters can communicate with the very nucleus of the cell, the home of the genes. The neurotransmitter locks on to a receptor on the cell membrane and this alerts second messengers within the cell which communicate with intermediate switching proteins that are able to penetrate the cell nucleus. Inside the nucleus, they communicate with transcription factors that control the expression of regulator genes. These regulator genes are responsible for managing the sensitivity of receptors and the production of enzymes which create neurotransmitters (Comb et al. 1987). Hormones have an even simpler route because they themselves are transcription factors that can operate directly on the genes. Earlier we noted that males and females differ because the Y chromosome carries a genetic switch that shunts development down a male pathway. In the absence of the Y chromosome or if the SRY gene were not activated, the foetus would develop as a female. In this case, the switch is a gene donated in the father’s sperm to the offspring. But in other animals, the switch that canalizes sexual development can be the ambient external temperature or the prevailing sex ratio—variables that exist unambiguously beyond the organism. All the animal needs is the ability to ‘read’ the correct external signals that activate the switch. As a species, we do not determine

Mhoc09.fm Page 275 Friday, December 14, 2001 10:18 AM

  :   


our sex this way but that does not preclude the possibility that sources of variability between individuals may work just like this (Wilson 1994). The activation or deactivation of a variety of genes as a result of experience could create differences between people who began life with exactly the same genome. Take post-traumatic stress disorder (PTSD), a condition that has been studied in American war veterans. Exposure to a fear-provoking situation, as we have seen, causes increases in norepinephrine implicated in processing threatening stimuli and linked to the activation of the ‘fight-or-flight’ sympathetic nervous system. Under stress, neurons in the noradrenergic pathways of the brain fire at up to ten times their resting rate resulting in a massive increase in norepinephrine—synthesis of the transmitter is increased and enzymes that break it down work harder to free the synapse from the excess. And the effect is also felt by the genes contained in the nucleus of the brain cells. The c-fos gene is switched on coding for DNA transcription factors whose job it is to regulate the future behaviour of receptors, enzymes and proteins. untoward events do not just upset neuronal communication in the present, but can remodel the brain’s early warning system, fine-tuning sensitivity to match the relative safety of the surrounding environment. . . . Acute responses—firing rates, transmitter release—gradually slow and stabilize, and levels of norepinephrine drift back to baseline as stores are replenished. But underneath the calm demeanor, noradrenergic neurons have actually shifted into over drive. Manufacturing capacity is quietly enhanced, as levels of norepinephrine-synthesizing enzymes are ratcheted upwards. Baseline firing rates are reset to a new, higher value, while the number of post-synaptic receptors—perhaps in an attempt to keep the lid on the system—declines. The result is a security system that tolerates familiar hazards but responds more vigorously than ever to the unknown . . . Persistent threat has kindled, or sensitized norepinephrine-containing neurons until they overreact to every provocation. (Niehoff 1999, p. 121)

We tend to think of PTSD as a diagnostic category—you have it or you don’t. But it is probably more accurate to think of it as the far end of a continuum with emotional underreaction at one end and overreaction at the other. This raises the issue of whether a genetic switch that controls a binary choice could also create the whole spectrum of individual differences we see along a personality dimension. If you plot the distribution of individuals on any trait you like (jealousy, extraversion, intelligence, aggressiveness), you will almost always find a normal distribution— plenty of people in the mid-range and ever fewer as you approach the two extremes. Let us suppose that every member of the population has the same genome (an unlikely scenario as we shall see) and only the environment can switch genes on and off. Let us also suppose (and this is far more likely) that a given phenotypic trait depends upon many genes. If each gene contributes additively and if the number of regulator genes that are activated depends upon the number of appropriate environmental triggers that are encountered, then it is unlikely that people will never encounter a single trigger and equally unlikely that they will encounter the number needed to

Mhoc09.fm Page 276 Friday, December 14, 2001 10:18 AM


    

switch on every relevant gene. Most people will lie somewhere in the middle. For example, let us suppose that an individual’s level of aggression depends upon how many times they have been exposed to aggression during development. In this way, individuals could tailor their levels to the social environment around them. Most people will have been on the receiving end of a certain amount of verbal and (occasionally) physical aggression. Some will never have been picked on at all and happily even less will have been the consistent victims of attack. The effect of a single genome interacting with a variety of environmental experiences will tend to create a normal distribution if the quantity of those external triggers is itself normally distributed.

Do humans use environmental cues to choose strategies? But there is more to human uniqueness than genetic switches. We need to look at how environments can provide information to the brain that allows the active choice of strategies. Humans are extraordinarily clever and creative strategists as a study by David Buss (1988) shows. He was curious to know what tactics people employ to retain their partners and to prevent them seeking new relationships with others. He decided the obvious first step was to ask them. He distributed questionnaires to 105 undergraduate students asking them to list up to five specific behaviours that males and females use to retain their partners. After eliminating duplicated answers, he was still left with 104 different acts that he classified as acts either directed towards one’s mate or towards one’s prospective rival. The first group were further subdivided into direct guarding (e.g. concealing mates from friends, monopolizing their time), negative inducements (e.g. threatening one’s own infidelity or punishing their partner’s, derogating competitors) and positive inducements (e.g. enhancing one’s appearance, displaying resources). Tactics directed at rivals were broken down into public signals of ownership (e.g. possessive ornamentation, verbal signals of possession) and negative inducements (e.g. derogation of mate to rivals, threats of violence). Now what is remarkable about this is the sheer variety of tactics used. When we look at animals’ range of strategies we often find only two or three. For example, among one species of sunfish, males come in three types: ‘parentals’ (large males who defend the nest), ‘sneakers’ (who are smaller) and ‘mimics’ (who resemble the female). The latter two stealthily fertilize the females’ eggs while the parentals are copulating (Wilson 1994). But among Buss’s undergraduates there were in excess of a hundred different alternatives. And nobody would seriously argue that their availability depends upon long-term genetic changes as we saw in the case of PTSD. These undergraduates were able to produce these various conditional tactics without undergoing any of the physiological change associated with actual jealousy. Evolutionary psychologists write much about ‘facultative strategies’ and ‘behavioural flexibility’ but we need to take a closer look at what they mean by these terms.

Mhoc09.fm Page 277 Friday, December 14, 2001 10:18 AM

  :     When used to describe animal behaviour, the metaphor for flexibility is often a computer program with a set of ‘If . . . then’ rules. Question: Should I fight or flee? Answer: If the antagonist is large, then withdraw. If the antagonist has already been beaten by you (or another animal that you have beaten), proceed. This metaphor does not explicitly raise the issue of the origins of these rules because the implication is that they have been placed there by the process of natural selection. This kind of analysis does a reasonable job of explaining simple phenomenon such as why gulls will peck at a beak-like shape with a red dot on it but not at a beak lacking such a dot. Sometimes the unstated supposition that contingent rules are ‘automatically’ available (i.e. hard-wired) is applied to humans also. In a famous analogy, Tooby and Cosmides (1992) describe a jukebox armed with navigational equipment that is programmed to play certain records as a function of its location, time and date. Although each jukebox in the world might be playing a different tune, there is a fundamental lawfulness that determines which one it is. Here we have human strategy choice conceived of as a set of universal rules that are both situational-sensitive and ‘given’. For Tooby and Cosmides, the tricky issue is how humans actually read the navigation device—how humans recognize different situations as equivalent (and deploy the same strategy) or as distinct (requiring different strategies). The strategy needed to find a mate is obviously different from that needed to find food. In short, the problem is how humans achieve what psychologists call ‘stimulus equivalence’ within classes. Tooby and Cosmides argue that current psychological theories will ‘have to be replaced with theories positing a far more elaborate motivational architecture, equipped with an extensive set of evolved information-processing algorithms that are contingently sensitive to a long list of situational contents and contexts’ (Tooby and Cosmides 1992, p. 99). In fact the list may not have to be that long if people are capable of recognizing underlying similarities between different surface structures. If humans recognize that their partner’s extended eye contact with their rival has the same meaning as their partners over-close slow dancing with a rival then the same jealousy module would be activated. Perceptual, cognitive and linguistic psychologists have long recognized the stimulus equivalence problem. Perceptually, we see a rectangular coffee table as remaining the same shape even though, as we walk around it, the geometry of the retinal image alters. Conceptually, we see a coffee table and a dining table as both being exemplars of the category ‘table’ despite their differences in colour, size and shape. Linguistically, we understand that ‘The table is in front of the sofa’ is identical in meaning to ‘The sofa is behind the table’. There is no reason to doubt (though the precise mechanisms may not yet be fully understood) that humans can and do recognize the identity of deep structure despite manifest variation in surface structure. But once the situation is diagnosed, from where does the appropriate strategy come? Tooby and Cosmides’ argument become less clear at this point, and it invokes three principles which move successively away from the notion of an ‘automatic’ if–then menu supplied by natural selection.

Mhoc09.fm Page 278 Friday, December 14, 2001 10:18 AM


    

The first is domain specificity. The more specialized the class of events to which a dedicated mechanism responds, the more closely tailored can be the response. The frog can have a simple ‘bug detector’ precisely because insects share features with one another that are not shared by many members of more inclusive classes such as ‘animals’ or ‘objects’. . . . At the limit of perfect generality, a problem-solving system can know nothing except that which is always true of every situation in any conceivable universe and, therefore, can apply no techniques except those that are applicable to all imaginable situations. In short, it has abandoned virtually anything that could lead to a solution. (Tooby and Cosmides 1992, p. 104)

In this view, as in the jukebox analogy, the identification of a problem domain elicits a perfectly tailored response (that is one that has been honed by natural selection because of the fit between the problem and the adaptiveness of the response to it). But humans do not have a bug detector device that reliably triggers a protruding sticky tongue—or anything like this bar a few preprogrammed responses such as the Babinski response or the sucking reflex. What is remarkable about humans is the sheer diversity of available tactics (both within and between individuals) when faced with identical situations. Tooby and Cosmides acknowledge that other mental algorithms are at our disposal and these seem to be of two kinds. The first is a small repertoire of domain general abilities—such as the ability to reject propositions that are contradicted, to form associations and to fine-tune behaviour as a result of experience—that can operate across specific domains. The second mechanism that enhances flexibility is the aggregation of a number of domain-specific modules. When we add together, for example, a semantic inference module, a sexual attraction module and a theory of mind module we begin to approximate a human mind. Yet Tooby and Cosmides are far from clear about how such modules summate or interact. Can principles from one module (e.g. reciprocal altruism) be extended to others (e.g. mate selection)? If so, how similar must the problems or solutions be for this to occur and with what degree of generality do they carry over? Does the simultaneous mental stimulation of multiple modules create a new emergent property that was not initially part of either one alone? Most importantly, is the synergy of multiple modules the same thing that they also call a domain-general problem solver? The third device that Tooby and Cosmides propose is one that chimes more clearly with those discussed by others (e.g. Barkow 1989; Symons 1992). It is the notion that the main impact of natural selection has been upon goal setting. Rather than a finite set of solutions that have been genetically fixed, humans experience emotional responses that indicate a problem situation. Jealousy indicates potential mate loss, anger indicates competition, lust indicates the need for sexual activity. These troubleshooting emotions generate mental activity designed to achieve a goal that alters the emotional state. In this view we still retain domain specificity in that natural selection specifies both the diagnostic emotion and the goal that will alter it.

Mhoc09.fm Page 279 Friday, December 14, 2001 10:18 AM

  :   


Food satisfies and removes hunger, sex satisfies and removes lust, aggression ends anger and removes competition. The organism needs to define an adaptive target and it needs a ‘guidance system’ that recognizes when the target has been met. It is this guidance system that has been constructed by evolution. what kind of guidance system can propel computational systems sufficiently often toward the small scattered islands of successful outcomes in the endless expanse of alternative possibilities? . . . In order to perform tasks successfully more often than chance, the architecture must be able to discriminate successful performance from unsuccessful performance. Because a domain general architecture by definition has no built-in content-specific rules for judging what counts as error or success on different tasks, it must have a general rule . . . The only unifying element in discriminating success from failure is whether an act promotes fitness (design-propagation). But the relative fitness contribution of a given decision cannot be used a criterion for learning and making choices because it is inherently unobservable by the individual. . . . Consequently our evolved psychological architecture needs substantial built-in content-specific structure to discriminate adaptive success from failure. (Tooby and Cosmides 1992, pp. 102, 111)

The quote above is important because of evolutionary psychology’s emphatic rejection of fitness maximization as a goal. Darwinian anthropologists have been subject to criticism because, implicit in their empirical work, is what evolutionary psychologists take to be an important misreading of Darwin. Darwinian anthropologists attribute to humans the ability to devise strategies that are aimed at increasing their inclusive fitness. Many of their studies are aimed at showing that a particular strategy (polyandry, cannibalism) is adaptive in the sense of increasing the number of descendants of those that practise them. But to evolutionary psychologists this is a fundamental mistake—adaptedness refers to the past not the present. Genes that control specific psychological mechanisms are retained if the phenotypic behavioural result benefits the bearers over thousands of generations. It is natural and sexual selection that decides whether something is adaptive, not a human decision maker. And, for reasons we have discussed, evolutionary psychologists are committed to a modular view of mind. Selection does not select for fitness maximization as a general principle but for the psychological mechanism that produces behaviour that solves discrete problems which may have little in common (e.g. seeing depth, learning language, having a lower threshold for recognizing fear rather than joy). If this view is correct, then evolution has provided us with a cue that a problem needs to be solved and mechanisms for knowing when that end-point has been achieved. This mechanism is domain specific (satiety after hunger, orgasm after sexual deprivation, sleep after tiredness) and corresponds to what we describe in the vernacular as satisfaction or pleasure or relief. They correspond to what psychologists used to call primary reinforcers—states that ‘automatically’ feel pleasurable to such a degree that they can be used as rewards for other kinds of behaviours. As Symons (1992, pp. 138, 139) puts it: ‘Human behavior is flexible, of course, but this flexibility is of means, not ends, and the basic experiential goals that motivate humans are

Mhoc09.fm Page 280 Friday, December 14, 2001 10:18 AM


    

both inflexible and specific . . . these (psychological) mechanisms merely make possible novel behavioural means to the same old specific ends.’ So how do we stop our mate from roaming? By a thousand means that achieve the desired end—threaten them, charm them, intimidate rivals. The alternatives are not given by our genes. These are not the simple instantiation of an if–then routine (or if they are, it is a deeply complex one with as many as 104 conditional alternative statements). We can generate dozens of strategies and the first test they must pass is their ability to solve the specific problem at hand. From the subset that could potentially achieve this end, we sieve out those that are not available or appropriate to this particular circumstance. I could retain my mate by increasing my attractiveness or by warning off rivals. My choice will depend upon factors such as how attractive my partner is, how many and how competitive my rivals are, the extent to which physical attractiveness is a critical criterion in my partner’s decision and what possibilities exist for improving my appearance. What makes humans special is their ability to run thought experiments. You don’t have to actually execute every possible strategy until you hit on a success (dye your hair, warn off the barmaid, etc.). Representational intelligence is the capacity to manipulate symbolic representations of the world inside your head. This capacity has evolved as part of human intelligence. It is, in Tooby and Cosmides’s sense, a domain-general strategy. Representation intelligence lets you add up 4 + 5 without having to use beads, lets you imagine what would happen if you drove your car into a wall, lets you contemplate the likely response of your partner to an unprovoked attack on the barmaid. It evolved because humans who could do this had a distinct advantage over those who could not. From the simplest creatures whose response to environmental events was genetically programmed (where a mistake meant death) came organisms capable of behavioural plasticity controlled by rewards and punishers (where a bad move was unlikely to be repeated) and then came what Dennett (1995) calls Popperian creatures whose ability to run thought experiments meant that their hypotheses, rather than their bodies, were the casualties. Every ‘new’ organism, even with the rudiments of the evolving representational system, held an evolutionary advantage over its peers and the chief advantage was successful flexibility. Building brains is costly and selection favours a net profit so the advantages of representational thought for adaptive plasticity must have been great indeed. (For a discussion of what propelled humans down this costly but advantageous avenue, see Dunbar 1992; Miller 2000; Mithen 1996; Wills 1995.) Jerome Barkow (1989, p. 103) makes the intriguing suggestion that ‘an organism with a model of external reality that includes a representation of itself will experience consciousness’. Humans are most emphatically conscious in this sense for we are able to include ourselves in our own thought experiments. We are aware, to a greater or lesser degree of accuracy, of our own strengths, abilities and potentials. You can use knowledge about yourself (recollections of your past behaviour, the way in which others respond to you) to evaluate the likely success of behavioural alternatives in

Mhoc09.fm Page 281 Friday, December 14, 2001 10:18 AM

  :     your thought experiments. The ability to represent one’s own assets and liabilities as well as those of others improves dramatically the likelihood of locking onto a winning strategy. Representational thought allows us to move from problem to solution without leaving our armchairs. But this special form of intelligence does not operate independently of our evolved architecture—it relies on emotions to detect problems and upon a domain-specific guidance system to tell us the best alternatives. Representational intelligence allows us to build a picture of human behaviour that does better justice to the creativity and insightfulness of humans’ adjustment to the environment.

The problem of genetic variance So far we have considered how different environments could produce different phenotypes even if the population was genetically identical. But behaviour genetic studies indicate that this convenient assumption is wrong. Because such results, though now well replicated, are still treated with scepticism, let us pause briefly to consider how behavioural geneticists reach these conclusions. The two main methods are twin and adoption studies. In twin studies measurements of a trait, such as risk-taking, are obtained from both members of pairs of monozygotic and dizygotic twins. Because identical twins share all their genes while non-identical pairs share only half, we would expect that if genes have an effect on risk-taking then monozygotes should be more similar than dizygotes. We have to qualify this prediction in two ways. First, because monozygotic (MZ) twins are genetically interchangeable they also share the same patterns of dominance (for each heterozygous allele) and epistasis (the interaction between genes at different loci). This is what geneticists mean when they say that they are matched for broad, as well as narrow sense, heritability. This means that they would be slightly more than twice as similar as dizygotes if risk-taking depends on these non-additive sources of variance. The second critical assumption is that both types of twin share an equally similar environment. In other words, MZ twins are not treated more similarly than dizygotic (DZ) twins. This has been the assumption that has been most strongly questioned by critics. Yet studies which have included measures of parental treatment find that similarly-treated MZ twins are no more alike on intelligence and personality than are MZ twins whose parents make conscious efforts to raise them as distinctive individuals (Plomin et al. 2001a). Parents who mistakenly believe that their twins are monozygotic produce children who are no more similar than other DZ twins. Monozygotic twins who are so physically similar that they are frequently mistaken for each other are not more psychologically similar than MZ twins who look less alike (Loehlin and Nichols 1976). Perhaps most importantly, the similarity on personality traits between identical twins who have been reared apart is comparable to the similarity found for MZ twins raised in the same household. Raised as individuals (not twins) and by different parents, their similar personalities cannot be the result of parents responding to their physical resemblance.

Mhoc09.fm Page 282 Friday, December 14, 2001 10:18 AM

      The second kind of study looks at adopted children. For any given trait, to the extent that children resemble their biological mother, the effect can only be genetic. To the extent that they resemble their adoptive mother, the effect can only be environmental. This follows if we can be sure that there is no important effect of selective placement (Rowe 1994). If a child were placed with an adoptive mother who was genetically similar to the biological mother, the child’s phenotypic similarity to the adoptive mother would come not just from their shared environment but also from shared (but non-inherited) genes. Behaviour geneticists can establish the extent of selective placement by computing the correlation between the biological and adoptive mother on the trait that they are examining in the child. This effect can then be statistically removed from their estimates. In practice, correlations that are found tend to be for social class and IQ—they do not present a serious threat to studies of personality traits. Convergence of results from different adoption studies with varying degrees of selective placement also increase confidence in the conclusions. (It is ironic that the ‘selective placement’ objection is often made by strong environmentalists who dispute the idea of a biological basis to personality. The fact that some agencies attempt a degree of selective placement is precisely because they implicitly accept the idea of a biological basis to the child’s character.) The most powerful design blends these two methods by comparing identical and non-identical twins, some of whom have been raised together and some of whom have been separated and adopted at birth (Bouchard et al. 1990). Here the effects are unambiguous—if MZ twins are just as similar when raised apart as when raised together then there is clearly a powerful impact of genetics on the trait in question. For 18 scales measuring different facets of personality from an instrument called the California Personality Inventory, the average correlation between identical twins reared together is 0.49. The correlation for identical twins who have lived in different homes with different parents is 0.48. The results from all the different methods converge on the same pattern. Between 50 and 70 per cent of variability in intelligence is genetic. For traits such as extraversion and other personality factors, heritability accounts for approximately 50 per cent of the phenotypic variance. Significant heritability has even been found characteristics that have always been assumed to be the indisputable products of rearing, such as social attitudes (Tesser 1993), time spent watching television (Plomin et al. 1990) and divorce (McGue and Lykken 1992). Heritability is a measure of the amount of variance in the distribution of a trait that is explicable by genetic factors. Heritability is a ratio of the genetic variance to total phenotypic variance. It is not an absolute measure that holds true for all people and places. Because genetic and environmental factors must always sum to 1.00, the ratio can fluctuate depending upon how much variability there is in the genes and in the environment. As environments become more similar for all members of a population, the environmental variance decreases and consequently the heritability estimate must increase. For example, as opportunities for success become more

Mhoc09.fm Page 283 Friday, December 14, 2001 10:18 AM

  :   


widely spread through a society, the heritability of ‘social success’ will become larger. By the same logic, when individuals differ very little in the relevant genes governing a trait, the environmental component tends to increase. The fact that important traits show genetic variation is problematic for evolutionary psychology because when a trait is essential for survival it is strongly selected and goes to fixation. That is why most human morphological structures are universal. People who lacked the genes to build hearts, brains or two legs simply did not survive to reproduce. It is impossible to perform a meaningful analysis of the genetic contribution to ‘leg building’ precisely because there is no variance that can be partitioned into genetic or environmental origin. Counterintuitive though it seems, if such an analysis was attempted, it would show a heritability of near zero. People who have only one leg have overwhelmingly lost one through trauma rather than being born with only one. Those characteristics that are indispensable to a species such as the ability to construct two legs would appear to have no heritability. Yet nobody in their right mind would say that legs are not built by genetic instruction. To keep matters clear it is helpful to make a distinction between ‘inherited’ (genetically instructed species-typical attributes) and ‘heritable’ (characteristics that vary between people where that variability is genetic in origin). The issue that has sparked much debate among evolutionary psychologists is the explanation of such genetic heritability. If a particular value of a personality trait was optimal for survival and reproduction, then why is there not uniformity among all members of the species? The fact is that brain size and intelligence have grown dramatically during our evolutionary past so that, as a population, Homo sapiens is smarter than Australopithecus. (We may also be more or less extraverted as well— although we will probably never know how outgoing our forefathers were.) But the point is that we show genetic variability now and almost certainly always have. Why are we not identical in our willingness to take risks or in our levels of anxiety or aggression? There seem to be three possible answers. Individual differences are the incidental by-product of some other trait that was selected. Or selection failed to act on individual differences because they were irrelevant to our survival. Or natural selection actively selected for diversity because it was advantageous.

Inherited individual differences as by-products Tooby and Cosmides (1990) presented a persuasive argument that personality differences are by-products of selection for something quite different. They compare the human mind to a car. Both are complex machines that are composed of many subparts that have to function together as whole. Because all the parts have to coordinate with one another for the car to work effectively, cars, like brains, have a ‘monomorphism of integrated functional design’ (p. 27). Changing one component, for example fitting a radiator from a Jaguar into a Toyota, will cause the car to fail, but changing the colour of the radiator from silver to black will have no important

Mhoc09.fm Page 284 Friday, December 14, 2001 10:18 AM


    

effects on the functioning. In the same way, the human brain cannot tolerate radical alteration in one of its components (e.g. a short-term memory system that generates products that cannot be stored in long-term memory) but can function quite well if the variation is so superficial that it does not affect the interdependence of the parts (a preference for opera rather than rock music). Tooby and Cosmides decompose the kinds of evolutionary and genetic changes that can occur into four types. The first is superficial variation (such as radiator colour), analogous, as they see it, to personality differences. At a slightly more radical level comes limited functional variation describing small incremental changes in a single part that either improves or degrades the overall performance. These are the kinds of small random mutations which provide the basic fodder of natural selection—improvements which by definition increase the reproductive success of the individual are retained while detrimental mutations tend over time to be selected out. (We shall return to the question of whether this kind of variation would be an adequate account of personality differences.) Thirdly, there is disruptive variation; critical mutational change that causes the car to malfunction. Mutations most often have negative effects and genetic effects on the synthesis of some proteins can be devastating. Phenylketonuria (PKU) results from an error in the production of amino acid so that the individual cannot convert phenylalanine into tyrosine. Without intervention, this results in mental retardation. Finally, there is the possibility of radical but coordinated functional variation. In this case, there is a series of major and interrelated changes that produce a whole new variety of functioning car. Often referred to as ‘hopeful monsters’, this type of genetic change is extremely rare. Probability theory suggests the vanishingly small likelihood that a set of gene mutations will be, by chance, perfectly coordinated to produce an integrated design. It would be analogous to selecting hundreds of car parts at random from a junk heap piled with different makes of vehicle, putting them together and finding that the car actually works. Tooby and Cosmides believe that personality differences are variations that have no significant effect on the overall functioning of the brain. They argue that it is only at the level of specific enzymatic pathways that qualitative differences between people appear. They believe that at this microbiological level of analysis there are compelling advantages to genetic variability which hinge on a link between sexual reproduction and genetic polymorphism (Tooby 1982). Why have sex? Many animals reproduce asexually and thrive. Sex has significant costs—the most major being that sex dilutes an organism’s contribution to the next generation by half. But this latter factor becomes a positive virtue in Tooby and Cosmides’ argument. Increasingly, biologists are finding that sexual reproduction has a significant but non-obvious benefit. The greatest threat to our species is disease. Disease-carrying microorganisms survive as parasites on larger animals, including humans, and the two of us are in an antagonistic arms race. To survive, humans must evolve defences against them and they in turn must counter-evolve ways around our defences. But they have a very significant advantage over us—their generation

Mhoc09.fm Page 285 Friday, December 14, 2001 10:18 AM

  :   


time is short. This means that they can evolve at a much faster rate than we can. During a human lifetime, a pathogen can have one million generations in which to find a way round our immune system (Tooby 1982). Imagine a genetically identical population again but this time from a pathogen’s viewpoint. Once it has overcome one individual’s defence, it has overcome every individual’s defence. But sexual reproduction mixes genes from two genomes creating a unique individual who is also a unique challenge to pathogens. Genes code for proteins—the environmental niche of the pathogen. The more alternative alleles that exist at different loci on the genome, the more unique each individual is and the harder is the pathogen’s job. In this way, genetic variability is actively favoured not because it is good for us (in any social sense) but because it is bad for pathogens. Enzymatic variability, coded for by genes, foils our smallest (and yet biggest) enemy. For many years a group of biologist called neutralists argued that this variability was selectively neutral because, although the proteins looked different structurally, they were functionally equivalent and therefore their variability was irrelevant to selection. But this combination of functional equivalence and structural difference is precisely the most advantageous combination from our point of view. The proteins do the same job in our bodies and brains but are chemically distinct. Pathogens choose their microenvironments not according to their function but according to their chemistry. Enzymatic variability gives us the twin advantages of disrupting the pathogen while not disrupting the functioning of the body part that it controls. As Tooby and Cosmides (1990, p. 34) put it: ‘the resolution of this conflict is to produce variation which is significant from the point of view of the pathogen’s life cycle, but superficial from the point of view of the ultimate functional design of the organ system.’ This analysis also throws light on the distribution of human genetic diversity. Eightyfive per cent of human variation occurs within groups, 8 per cent occurs between nations and 7 per cent occurs between races. This is exactly what we would expect if the function of allelic diversity is protection from disease. We catch diseases from those with whom we come into daily contact and any protective device has to be geared to them not to distant groups that we will never meet. But what does this genetically based diversity of proteins have to do with personality? Tooby and Cosmides (1990, p. 49) explain: Pathogens select for protein diversity introducing the maximum tolerable quantitative variation and noise into the human system . . . Given the intricate design complexity of the nervous system (as well as other organ systems), this protein variation gives rise to a wealth of quantitative variation in nearly every manifest feature of the psyche: Tastes, reflexes, perceptual abilities, talents, deficits, thresholds of activation, motor skills, verbal skills, activity level, abilities to remember different kinds of things, and so on—all vary from individual to individual in a quantitative way.

And just as there are differences in reflex speed, there are differences in anxiety, aggression and risk-taking. Personality is not disembodied and mystical. It is genes

Mhoc09.fm Page 286 Friday, December 14, 2001 10:18 AM


    

that, via proteins, create receptors and neurotransmitter sensitivity. The production, uptake and clearance in the synaptic cleft of neurotransmitters can result in different gradients of excitement-seeking, fear or contentment. Our production of and sensitivity to hormones can affect our mental state, sometimes transiently and sometimes throughout our lives. But if the pathogen-resistance theory is right, these are by-products that carry no real significance in and of themselves. However pleasing we may find human diversity, whatever rewards it brings to our social and emotional life, these are just bonuses that natural selection neither saw nor chose.

Inherited individual differences as genetic junk Extraverts and introverts do not differ in critical mental modules—they both acquire language using the same device and share a common sexual jealousy mechanism. Just as all humans have one stomach, one heart and two lungs so we all share the same basic architecture of the mind. However, no two stomachs are exactly the same size, nor do they produce identical amounts of acid. But this is mere superficial variation and as such it is simply too small a target for selection to hit. Selection can only act on organisms who vary in their survival and reproductive success—if being one standard deviation above or below the mean on extraversion has no discernible effect on either of these two outcomes then selection will not weed it out. So just as having a larger-than-average stomach has no significant effect on survival, neither do minor fluctuations in personality. According to this model, extraverts have not, in our evolutionary past, had more children than introverts and the distinction between them is therefore evolutionarily irrelevant. Although it was Thiessen (1972) who coined the term ‘genetic junk’, Charles Darwin himself subscribed to the idea that variability suggested the fitness-irrelevance of a trait: ‘The preservation of favourable variations and the rejection of injurious variations I call natural selection. Variations neither useful nor injurious would not be affected by natural selection and would be left a fluctuating element, as perhaps we see in the species we call polymorphic’ (quoted in Majerus et al. 1996, p. 82). But what of Tooby and Cosmides’s limited functional variation—small random mutations that provide the basic fodder of natural selection? Could that not explain personality differences? The whole process of evolution critically depends upon the continuous supply of variation for that is the menu of possibilities from which nature selects. Just as evolution removes additive genetic variance, making us more and more alike, mutation introduces new forms to replenish the supply of alternatives. Evolution is composed of two processes—natural selection and random genetic drift. We have spoken mainly of the former because this is the process that drives directional selection over time. But genes can also vary in frequency by chance effects. When an individual dies before reproduction they take their genes with them and this is as true when they die as a result of a chance effect (being struck by lightening) as when there is a genetic vulnerability (they are not resistant to flu).

Mhoc09.fm Page 287 Friday, December 14, 2001 10:18 AM

  :   


Chance deaths can cause fluctuations in allele frequency. Imagine a bag holding six white balls and six black balls (corresponding to two different alleles at a genetic locus). Put in your hand and randomly select six balls to go into the next generation of the game. You pull out four white and two black. The population is now reset at a new value for allelic diversity. Next time around, you select two white and two black; the population drifts back to equilibrium. (There is a chance that on second round you might have pulled out nothing but white balls—the white allele would have gone to fixation but as a result of chance not selection.) The process is random and provides a continuous backdrop of genetic noise fluctuating up and down behind the directional force of natural selection. Random genetic drift is usually associated with characteristics that do not affect fitness. Levels of such neutral genetic variability— traits that neither increase nor decrease fitness—are larger in bigger populations. In a small population, mutations will not occur very often but when they do they will quickly go to fixation—that is, all members of the population will inherit them. In a bigger population, mutations will be more frequent but because of the population size it will take much longer for the novel trait to spread through the whole population and become species-typical (Majerus et al. 1996). This means that large populations, such as human beings, at any one time have many mutations that are drifting towards either elimination or fixation. This might result in genetic variance in fitness-irrelevant traits.

Inherited individual differences as adaptations There are two circumstances where natural selection does not use up variability in the population and these warrant special consideration as candidates for explaining personality differences. The first of these is frequency-dependent selection. In this process, the fitness of a genotype is related to the number of others in the population who also carry it. The most often cited example is the case of Batesian mimicry. Some animals, notably butterflies, depend for their survival on mimicking other related species that are toxic to predators. The success of the mimic strategy depends upon the relative frequencies of mimics and the ‘real things’. When there are few mimics, predators are likely to meet the real thing on their first encounter with the prey and will henceforth avoid it—and of course the mimic who is indistinguishable from it. But as more and more mimics enter the population, the chances increase that the prey’s first experience will be with a mimic who, lacking toxicity, will provide a tasty meal. The success of mimicry depends on a low ratio of mimics to models. The same kind of dynamic can be responsible for the generation of new genotypes. Many birds form a search image of their prey (see Majerus et al. 1996). If the bird lives in an area where yellow snails are common, this will be incorporated into its search image and the bird will be on the lookout for yellow targets. If a mutation arises that is pink, it will be ignored by predatory birds and so the pink

Mhoc09.fm Page 288 Friday, December 14, 2001 10:18 AM


    

mutation will increase in frequency. But there is such a thing as too much success. Soon pink snails will be in the majority and yellow snails will be hard for the bird to find. At this point, the bird will alter its search image to pink, eventually removing the threat to yellow snails and allowing them to become more numerous. This process need not stop at only two varieties. As Majerus et al. (1996, p. 65) explain: ‘The simple balance between being common and eaten, or rare and ignored, will always favour new forms.’ The amount of variability around equilibrium depends upon how quickly the bird changes its search image. If it changes swiftly in response to even slight changes in yellow–pink frequency, there will be very little slack in the system but if there is a longer latency to change, the balance of yellow and pink may at times be quite extreme. Although the swiftness of the correction factor may vary, frequency-dependent selection can maintain allelic diversity indefinitely. Frequency-dependent selection may operate in our own species too. Mealey (1995) has provided just such an analysis of primary psychopathy. This clinical label describes a syndrome that includes a deficit in the ability to experience empathy, a lack of guilt, a willingness to exploit others, low levels of fear and superficial charm. These individuals make their way through life exploiting the goodwill of others— they con others into giving them material goods, loyalty and affection without feeling any need to reciprocate. They are social cheaters who take but fail to repay in life’s game of tit-for-tat. This strategy can be extremely successful—they receive sex without commitment, money without work, shelter without rent and friendship without emotional engagement. Before the victim knows what has happened, the psychopath is off in search of his next target. As a strategy, psychopathy can only work if it is used by a minority of the population. Too many psychopaths and the rest of the population starts to alter its response to strangers. They become more wary, hold back from generosity, wait to see if this person can be trusted. The more wary they become, the lower the success of the cheater. And psychopaths can only thrive in large populations where their anonymity and speedy exit can be guaranteed. It is hard to use this strategy successfully in a small, remote village because the first con game will quickly result in a reputation that alerts others and there is nowhere to find a stock of fresh victims. And primary psychopathy does seem to have a strong biological component. Deficits have been identified in the functioning of the amygdala (Blair et al. 1997), the orbitofrontal cortex and the autonomic nervous system (Raine et al. 2000). Female short- and long-term mating strategies (Gangestad and Simpson 1990) may be another candidate for this kind of frequency-dependent analysis (Buss and Greiling 1999). Some women adopt a restricted sexual practice, taking time to evaluate the likelihood that the male will make a committed and investing longterm partner and parent. Others, whose goal is gene quality rather than long-term investment, will be prepared to have intercourse after only a short delay. The dynamic that holds the strategies in equilibrium is their relative frequencies. As the number of unrestricted women rises so do the number of ‘sexy sons’ that they produce and

Mhoc09.fm Page 289 Friday, December 14, 2001 10:18 AM

  :   


hence the value of women selecting for good genes diminishes. As the number of restricted women begins to rise in response, the competition among them increases and advantages begin to accrue to women who do not waste time and effort searching for providing fathers. So far we have spoken of a finite number of competing ‘strategies’ but everything we know about human personality tells us that we are distributed along a continuum, rather than ‘chunked’ into typologically discrete groups. Could a continuously distributed trait such as extraversion also be a frequency-dependent strategy that depends for its success upon the number of introverts that are around? Before we consider this, bear in mind that the majority of genes are homozygous— that is the parents supply the same alleles to the offspring. You and I share 99.9 per cent of our genes (the same amount we also share with the ‘reference sequence’ of the Human Genome Project). That 0.1 per cent that makes us different amounts to 3 million variants in the 3.2 billion base-pairs that compose the genome (Wellcome Trust 2001). For features or traits that are common to us all, both parents provide the same instructions at each genetic loci. But two different ‘strategies’ such as extraversion and introversion will be coded for by different alleles at many genetic loci. And sexual recombination creates a problem for maintaining the coordination of these multiple-allele genes that must function together to create the two types. To get extraversion the whole suite of alleles and their activation must be passed intact to the offspring. When two different individuals mate (especially if one is an introvert and the other an extravert) the combinations will be broken apart, resulting in a mix of both types. This is probably what gives rise to the normal distribution that we associate with personality traits. As we saw before, if each of ten genes makes an independent ‘choice’ of activating or not then the probability is that few individuals will switch on all ten and equally few will switch on none. A coin flipped ten times will usually give about five heads and five tails creating a roughly normal distribution. But there are mechanisms that can link the fate of multiple genes so that they are carried through generations together without being broken apart. One is called linkage disequilibrium. It happens most often when particular combinations of alleles are particularly advantageous. Imagine two genetic loci 1 and 2. At each one of these loci there are four possible allelic combinations that can be inherited: AA, Ab, BB, Ba. If AA and BB (which, if real-world genetics were simpler, might be introversion and extraversion) are the fittest then there will be a selective advantage to these combinations that will keep their frequencies higher than the other alternatives. Linkage between genes depends on their physical location—linkage is more likely when the critical loci are close to one another on the same chromosome. This is because recombination is less likely to split them apart. One factor that enhances linkage is assortative mating. Extraverts who mate with one another are more likely to produce extraverted children because alleles are identical. Although the sheer number of genes that would have to be in linkage to generate the complexities of a human personality dimension are thought by some to be simply too great to

Mhoc09.fm Page 290 Friday, December 14, 2001 10:18 AM

      withstand the forces of recombination (Tooby and Cosmides 1990), it might be the case that alteration to a single gene or relatively few tightly-linked genes (a supergene) is all that is needed for us to see a whole suite of changes, such as those associated with extraversion. A change in one locus can have far-reaching consequences. Recall that the SRY gene switches on a series of biological events that cause production of high levels of testosterone and consequent masculinization. This hormonal change manifests itself in many different forms both morphological (muscle strength, height) and psychological (visio-spatial ability, sex drive). Here a single gene has considerable power to alter the many components of the final phenotype. It may be the case that single-gene loci alter production or responsiveness to a particular neurotransmitter resulting in extraversion or introversion. Another means by which coherent suites of changes can be achieved economically is through supergenes. The butterfly Papilio memnon survives by mimicry but in this case it may mimic one of two different species. To succeed it must resemble one of the two models in a variety of traits including colour, morphology and pattern. All the different genes that control similarity to one or other of the species have to be passed on together because there is little to be gained by half-mimicking one species and half-mimicking another. In the P. memnon, scientists have found a tightly linked supergene composed of six loci responsible for passing on the alternative forms intact (Clarke and Sheppard 1971). Much more research is required before we can even approach an answer to the coordination and control of personality— the human genome project is a critical first step. So selection for suites of coordinated changes that constitute alternative strategies may be biologically possible. But to prove frequency dependency we would have to go much further (Buss and Greiling 1999). We would have to show that these suites of changes are coordinated with one another to solve specific problems. Next we would have to demonstrate that the different types are truly in a frequency-dependent relation one another—that as the number of one type increases its pay-offs diminish—and we would have to posit a homeostatic mechanisms that maintains the stable equilibrium. The criteria for demonstrating a human genetically-based frequency-dependent strategy are stringent. My guess is that they will rarely be met in full. Here’s why. If the differences between humans correspond to different strategic approaches to the same problem then, when we plot the distribution of individuals on a given trait, we would expect to see a bimodal (or perhaps multimodal) distribution. For example, we would expect to see introverts and extraverts as two distinct types, clumped at either end of the personality continuum with rather few individuals in the middle (since there is no selective advantage in falling between two strategic stools). What we see is just the reverse. Now this may be because we have been looking in the wrong places. Perhaps we should be exploring theoretically-driven, evolutionarily-relevant dimensions (such as mate selection or reproductive strategy or parenting style),

Mhoc09.fm Page 291 Friday, December 14, 2001 10:18 AM

  :   


rather than atheoretical correlated attributes such as the empirical coalescing of impulsivity and sociability into extraversion. But that said, the normal distribution does seem remarkably ubiquitous—as true for personality as for height or weight. And it is clearly suggestive of additive genetic variance—a dosage effect of increasing numbers of relevant genes—rather than of genetic linkage—the intact inheritance of coordinated gene complexes. Humans do not seem to be easily classifiable as discrete numbers of types representing different strategies—rather we vary quantitatively and, to make matters worse, we do so simultaneously on a whole range of personality dimensions. The second mechanism that maintains diversity in a gene pool is heterozygotic advantage. Suppose that there are two alleles (variants) of a given gene (A1 and A2). Neither one will go to fixation at the expense of the other if the best possible combination that an individual can inherit is one of each. So the frequencies that the two alleles finally arrive at depend upon the relative advantage that the heterozygote (A1A2) has over the homozygotes (A1A1 and A2A2). The classic example in humans is sickle-cell anaemia. Most Europeans carry the two identical alleles for the normal oxygen-carrying haemoglobin gene. Their blood oxygen levels are healthy but one negative side-effect is that they are prone to malaria. Other individuals are also homozygous but for a mutant form of this gene—the sickle allele—that produces abnormal sickle-shaped red blood cells which are poor oxygen carriers and result in serious anaemia. But there is one positive side-effect—their red blood cells are strongly resistant to malaria. In regions where malaria is common, the best possible combination to have is one each of the two alleles. Although it creates mild anaemia it protects the individual from malaria. Two homozygous alleles are likely to result in death either from malaria or anaemia. When the heterozygote is more advantageous than either of the homozygotes, a situation called heterozygote advantage, the population remains polymorphic—it retains more than one form. Early biologists took polymorphism as evidence that the trait in question must be irrelevant to natural selection (genetic junk) but others argued that it showed that selection forces are strong but balanced (Majerus et al. 1996). Such genetic variability would change only very slowly, in line with the shifting balance of advantages and disadvantages. When the different forms are perfectly balanced in fitness, variability can be long-lasting and geographically widespread. The beneficial and detrimental effects of heterozygosity can alter at different ages, times and places. Some genes can be beneficial in youth and a liability in old age— one explanation of senescence depends upon exactly this principle. As another example, in red deer two alleles (f and s) have been found at one genetic locus (Pemberton et al. 1991). Deer born with the ff combination are much more likely to die in the first year of life. On the other hand, the fs combination causes females to reproduce earlier and more numerously. The f allele remains because its advantages outweigh its costs. Pleiotropy might favour two life history strategies which differ in

Mhoc09.fm Page 292 Friday, December 14, 2001 10:18 AM

      whether they allocate initial effort to reproduction or to growth. One might be grow quickly, breed early and be prolific (even at the risk of dying young). Another might be grow slowly, breed late and ensure your survival to take care of off-spring (even if you have fewer of them). Some will take the first route and some the second. Each has costs and benefits. But overall both strategies are successful and both remain in the population. Temporal variation can also maintain heterozygosity. If the environment changes according to the seasons, there may be an advantage to having two different alleles. Spatially distinct environments may also favour different genotypes—where an animal moves frequently between difference niches then heterozygotes have the advantage of retaining two different strategies. As long ago as 1930, Fisher (Fisher 1930) suggested the possibility of modifier genes that through selection would actively favour the positive functions of each allele creating a heterozygote with the best of both worlds. It may be that over evolutionary time and place different values of a trait have been optimal. Imagine a time and place where food is scarce and predators few. Risk-taking might have been a strongly favoured trait if it resulted in securing meat while not exposing oneself to extreme danger. Imagine another place where the single biggest killer was disease. In such a place those who associated least with other people might have fared best, resulting in selection for introversion. Individuals with a genotype that had greater ability to store fat may have been advantaged in times of famine but risk obesity in times of plenty (Bailey 1998). Given thousands of such variable environments across evolutionary space and time, the optimal level of different traits has probably been at every conceivable value. What we see in the population today may be the result of these thousands of instances of varying selection pressure.

Variable genes meet variable environments We have considered why, from an evolutionary perspective, genes might code for differences in personality and ability. But at birth, that unique genotype comes face to face with the environment. For many years, when standard social models dominated our thinking in psychology and genetic variability was ignored, it was assumed that parents created adults from the raw and invariant clay of infancy. Parents took credit for children who ‘did well’ and more frequently took the blame for children who did not. Responsibility for autism, anorexia, homosexuality, schizophrenia and delinquency were all laid at the parents’ door. Yet the same studies that showed us the importance of heritability also showed us the importance of environment—but not in the way we had imagined. In addition to genetic sources, variance is decomposed into two other sources. Shared environment captures the similarity that results from experiencing the same treatment—from living in the same home in the same neighbourhood with the same parents. Unshared environment captures the impact of differential experience by siblings. Studies clearly show that,

Mhoc09.fm Page 293 Friday, December 14, 2001 10:18 AM

  :     to the extent that children raised by the same parents show similarities, they are not the result of being exposed to the same parental treatment (Plomin and Daniels 1987). Monozygotic twins raised by the same parents are no more alike than those raised by different parents. The largest portion of the environmental variance is unshared—our own particular experiences that we do not share with our brothers and sisters. The maths teacher that we had, the chickenpox we contracted, the time we nearly drowned on holiday. This finding shook the developmental community but it also challenged behaviour geneticists too. If such important effects were random, how could they ever be studied? Happily, not all unshared environmental effects are random (though many probably are). To pursue any sensible line of research we have to think of some areas where siblings have systematically different experiences. One such area might be the way they are each treated by their parents. If your parents elected you to be the black sheep of the family and your sister was the angel in the house, then it is likely that the two of you would be treated quite distinctively which could then cause the two of you to become increasingly different from one another over time. The harder question—lurking in the wings—is what makes parents decide to treat their children differently? One of the first forays into the area of differential parental treatment was by David Rowe (1981) who performed a modest study of 89 pairs of adolescent twins in which he asked them to rate their perception of their parents behaviour towards them. He was interested in two main dimensions of parenting—discipline (the firmness of control and degree of autonomy accorded to the child) and warmth (the affection towards and acceptance of the child). For the time being we will concentrate only on warmth. Monozygotic twin pairs’ ratings were much more highly correlated than DZ pairs. He replicated the study using a different set of twins and a different measuring instrument (Rowe 1983). Again, MZ twins were much more concordant for perception of parental warmth. What could this mean? The most obvious hypothesis was that MZ twins are more genetically similar in their temperament and this causes them to view their parents with the same rose-tinted (or blue-tinted) glasses. They behold the world with the same eyes. A new study was undertaken (Plomin et al. 1988) with 386 twin pairs, some of whom had been separated at birth, and once again the same results were found. Here the ‘common perception’ effects would have to be strong indeed because many of the twins were rating completely different families. There was another possibility; the MZ twins rated parental warmth more similarly because the parents were responding to genotypically identical children with the same emotional response. If we ask parents to rate their own behaviour to their children (instead of asking the children) we can rule out the possibility that the similarity of MZ twin ratings derives from similar perceptual processes. Another technique bypasses verbal reports and goes directly to observation. A number of studies which have calculated heritability in different ways suggest that half the variance in maternal closeness, which traditionally has been viewed as the cornerstone

Mhoc09.fm Page 294 Friday, December 14, 2001 10:18 AM

      of successful child development, is attributable to the child’s genotype (Plomin 1994). Recall that Rowe also measured discipline. Here he found little evidence of genetic effects. Parents do not differ in the standards they set or the sanctions that they employ towards genetically identical twins as compared to non-identical twins. In case you are wondering why the parental reactive effects are so marked for warmth but not control, there is one plausible reason: parents do not set out on a purposeful program to create and direct attachment to them, and acceptance is also largely a matter of the child’s disposition and relationship with the parent, rather than the parent’s conscious choice. On the other hand, most parents do engage in a conscious, purpose-driven program of shaping and controlling the child’s behavior—hence it is their dispositions that count here. (Lytton 1991, p. 399)

So studies show that parental response varies as a function of the child’s genotype, but can we go on to show that the way the child is treated affects the child’s later behaviour? Another step forward in understanding the enigma of unshared variance has come from explicitly relating differences in siblings’ experiences to differences in later outcomes. Negative parental behaviour, such as conflict and shouting, directed to a child is correlated with that child’s antisocial behaviour. Many previous studies have shown this before using the standard design of comparing children raised in different households. But new studies (Pike et al. 1996; Plomin et al. 2001b) showed that the effect held up even within the same family—differential treatment by parents was associated with differential antisocial behaviour among siblings. These researchers then asked a further and crucial question—is parenting style a cause or an effect of the child’s misbehaviour? Because they had included twins, full sibs, half sibs and unrelated sibs in the sample, they were able to examine possible genetic effects. Recall that in most studies a single variable is decomposed into genetic and environmental sources. Statistical techniques can also partition the covariance between variables. To do this, a cross-correlation is first computed (the correlation between parental negativity for Sib 1 and antisocial behaviour for Sib 2) and a genetic effect is implicated to the extent that correlations are higher for more genetically related individuals. The results showed that most of the associations were indeed mediated by genetic effects. Unshared environment (the different ways in which the siblings were treated) was driven by genetic differences between them. As Plomin et al. (2001b) ruefully conclude: ‘The . . . quest for nonshared environment led to genotype-environment correlation, that is children select, modify, construct and re-construct their experiences in part on the basis of their genetic propensities.’ The particular genotype–environment (G–E) correlation at which Plomin unexpectedly arrived is one among three possible types (Scarr and McCartney 1983). Plomin’s work suggests a ‘reactive’ G–E correlation—this is where the environment responds to the child in such a way as to augment a particular genetic predisposition.

Mhoc09.fm Page 295 Friday, December 14, 2001 10:18 AM

  :     But the reactive environment need not be confined to the family. Studies of highly aggressive boys suggest that their peers respond to them by social exclusion which in turn sensitizes the boys to their rejection. They start to interpret even neutral social events (an accidental ball to the back of his head) as hostile and intended which triggers further aggression in a vicious cycle (Dodge and Coie 1987). Shy children may evoke gentle, quiet responses from others which allow them to avoid the cutand-thrust of normal childish interaction. Happy children are easier to love and the security that this fosters adds to their benign easy-going nature. Earlier in life, passive G–E correlations are more likely to be at work. Here parents provide the child not only with genes but also with an environment that is correlated with them. Literary parents not only supply ‘literary genes’ but also shelves groaning with books. Athletic parents not only give athletic genes but take their child with them to sports meetings. Extraverted parents may have genetically extraverted children but they also have a huge circle of friends inhabiting their house during the child’s early years. Both genes and environment work together to canalize the child’s personality in a particular trajectory but we call the effect ‘passive’ because the environmental component is given to the child without regard to their genotype. This type of effect is thought to be most prominent in infancy. Passive G–E correlations seem to be the easiest to locate. For example, parents who produce easy-going infants are also parents who create a family environment that is higher in expressiveness and cohesiveness (Plomin and DeFries 1985). For 12 other correlations between measures of parental behaviour and infant development, the values were significantly higher in biological than in adoptive families. This suggests that many genetic links between the parents’ behaviour and infant development are due to passive G–E correlations. However, most work so far has been conducted on young children and we would expect the other forms of G–E correlation to become stronger as the child grows up. And the methods that investigators use also play a role in the findings, as Robert Plomin (1994, p. 154) notes: ‘It is possible that passive GE correlations will loom large, even in adolescence, as long as we continue to assess the environment using a passive model in which the child is a mere receptacle for the environment—for example by assessing what parents do to children.’ More marked in adolescence and adulthood are active G–E correlations. The individual selects social and occupational niches that are congenial to her own genotype. Introverts are more likely to be librarians than actresses. People select friends and partners who are similar to themselves in personality. People with good visio-spatial skills are more likely to study mathematics at university, while excellent language skills make for better linguists. Niche-picking may also explain an unexpected finding from genetic studies of IQ (Loehlin et al. 1989). In childhood, adopted children show some similarity to their adoptive and biologically unrelated siblings. Their IQ correlates about 0.26. But by the age of 18, the correlation has disappeared. Special programmes that offer children enriched environments also raise IQ while the child is attending them but the effects evaporate a few years later. Reciprocally,

Mhoc09.fm Page 296 Friday, December 14, 2001 10:18 AM


    

MZ twins show greater similarity in IQ as the years pass. What this suggests is that an educationally enhancing environment can have an effect on a child’s performance on intelligence tests as long as the child remains within that environment. But as a child grows up and selects her own hobbies, friends and vocations, she gravitates to those that are most congenial to her genotype. You can lead a child to literacy but you can’t make her enjoy books. Children do not choose their parents but they choose their peers and are chosen by them. Peer choice reflects active (selection by the child) and reactive (acceptance by peers) effects. By adolescence, measures of peers’ characteristics show even higher heritability than measures of parenting. In other words, the child’s own genotype is highly predictive of the kinds of friends that the child selects and is accepted by (Plomin 1994). The child’s genotype is even more predictive of the type of friends she has than of her parents style of interaction with her. Such a niche-picking notion lies at the heart of Wilson’s (1994) proposal of the adaptiveness of individual differences. He suggests that human populations are a mix of generalists (who can match their behaviour to prevailing conditions) and specialists (who cannot). When specialists are in their preferred niches, they always do better than generalists. But generalists have the advantage of being able to move between niches when the competition becomes too great in one spot. These opposing forces keep both types in the population. Children vary in how shy or bold they are and in a longitudinal study Kagan et al. (1988) measured this trait at 14 months and again at 48 months of age. They found that children at the extremes of the distribution were remarkably stable over time but those in the mid-range were not. The extreme scorers were specialists, Wilson suggests, who would function well as long as they could gain access to their most congenial social niches. The mid-rangers were generalists able to modify their behaviour depending on which niche they chose to be in or were put in by others. What looked like a continuum of personality, he suggests, actually conceals the possibility of a higher-order typology of specialists and generalists. Niche picking may be more important to specialists than to generalists. Psychopathy may also be a niche-dependent phenomenon—at least to some extent. Mealey (1995) distinguishes between primary and secondary psychopaths. The first group, as we have discussed, conform to the classical clinical definition of a personality constellation characterized by a lack of empathy, remorse and fear. Heritability estimates for this form of personality disorder are substantial and societies everywhere are likely to carry a small proportion of such individuals whose life strategy is to some extent already biased towards the exploitation of others at birth. However, societies vary markedly in the prevalence of secondary psychopaths. These are individuals less extreme on the genetic distribution who adopt an exploitative lifestyle as a function of the ecological niche in which they find themselves. Adverse developmental and environmental factors trigger the use of a ‘cheating’ strategy to others. These individuals come from families marked by inconsistent and harsh discipline, an absence of rewards for pro-social behaviour and academic achievement,

Mhoc09.fm Page 297 Friday, December 14, 2001 10:18 AM

  :   


high levels of family discord and father absence. They experience impoverished environments with high competition for scarce resources. Although they experience the normal ranges of fear and empathy, they adopt an exploitative response to others because they are unable to compete in the legitimate economic market-place as a result of poor social and academic skills. Secondary psychopathy is a far more niche-dependent phenomenon amenable to social policy intervention since it represents a response to family stress, competitive handicapping and inadequate economic resources. Although not yet investigated empirically, it is a theoretical possibility that negative relationships might hold for all the three types of gene–environment correlations that we have discussed. Negative passive correlations are the most difficult to conceive of but might be seen if, for example, parents with no musical aptitude provided music lessons to their son despite his equal lack of natural aptitude. Negative reactive correlations are more intuitively obvious and could occur where parents or schools seek to compensate for deficits that become obvious during development by, for example, encouraging acting classes for children lacking in self-confidence or encouraging sport among children who lack physical coordination. Negative active correlations sound perverse but could include the child’s own attempt to compensate for what she regards as her liabilities such as emotionally labile children seeking out friends who are calm enough to control and contain her outbursts. Gene–environment interactions also remain to be examined; the impact of different levels of an environmental variable may have a different effect on different genotypes. For example, strict parenting may effectively curb antisocial behaviour for introverts but not for extraverts. Girls’ misbehaviour may alter in response to its negative impact on loved ones, while boys’ does not. This brief survey of the field indicates the challenge of understanding the gene– environment interface even for a single trait. But the stunning diversity of personality that we see around us results from G–E relations (correlations and interactions) operating simultaneously on a range of different traits. What we do know is that these G–E relations have the effect of increasing phenotypic diversity beyond the independent effects of genes and environment alone (Plomin 1994). They help to guarantee that we will never meet the same individual twice.

Variable genes meet variable memes So far we have taken a rather asocial view of human individuality. Differences between people have been described as arising from different genotypes interacting with different environments. Other people form an important part of that environment, of course, but they are much more than mere stimuli to which we react. They are also information transmitters. It is time to consider head-on the fact that people live in social groups that share a common culture—a culture that exerts a powerful impact upon their experience.

Mhoc09.fm Page 298 Friday, December 14, 2001 10:18 AM

      Evolutionary psychology’s first goal was to reject the traditional notion that the entire content of our mental life derives from culture. In this ‘standard social science model’, the baby is an empty vessel equipped only with the power to learn. What it learned is given entirely by culture. The invocation of culture became the universal glue and explanatory variable that held social science explanations together. Why do parents care for their children? It is part of the social role that culture assigns to them. Why are Syrian husbands jealous? Their culture tied their status to their wives’ honor. Why are people sometimes aggressive? They learn to be because their culture socialises them to be violent. Why are there more murders in America than in Switzerland? Americans have a more individualistic culture. Why do women want to look younger? Youthful appearance is valued in our culture. And so on. (Tooby and Cosmides 1992, p. 41)

Given evolutionary psychology’s evident contempt for this ubiquitous and poorly defined concept, it was some time before they were willing to recognize that a full account of human psychology would have to face it in the end. Culture has been ascribed a bewildering array of over 150 meanings by anthropologists whose stock-in-trade has been its study (Kroeber and Kluckhohn 1952). For evolutionary psychologists, there is broad consensus that culture simply means socially transmitted knowledge—knowledge that has not been individually discovered but has been acquired by observation of or interaction with other people. The advantages of acquiring information in this way are not hard to see. For an individual, it is safer and faster than learning by trial and error. For a society, it allows for the cumulative assembly and transfer of knowledge over generations. Through culture we can learn how to speak specific languages (English, Swahili), how to decode meaning from actions (thumbs-up signs), procedural knowledge (how to drive a car or cook spaghetti), history and myths, styles of art, gossip and rumour—the list is well-nigh endless. The very scope of culture’s influence means that it is helpful to have a word that can be used to refer to any one of its many manifestations. Richard Dawkins has suggested the term ‘meme’ and its similarity to gene is not coincidental. Although there is some debate as to whether cultural information can be atomized into discrete units in this way (is Beethoven’s Fifth Symphony a meme or only the first four notes?), it nonetheless solves a referential problem so I will employ it. Dawkins (1989) proposed the meme to describe a unit of cultural transmission that operates analogously to genetic evolution. A meme is above all a replicator but instead of passing from body to body across generations like genes, it employs the vehicle of the human brain and can move horizontally and at a much higher speed. The success of a meme depends, as with genes, upon longevity (how long a meme lasts through the generations—religion and Shakespearian plays have survived remarkably well), fecundity (how fast it can make copies of itself in different people’s heads—its ‘catchiness’ or popularity) and its copying fidelity (its capacity to maintain its integrity, despite some degree of personalization, when transmitted).

Mhoc09.fm Page 299 Friday, December 14, 2001 10:18 AM

  :     Dawkins idea has been taken up and expanded by many, generating a new subdiscipline known as memetics (Blackmore 1999; Brodie 1996; Dennett 1995; Lynch 1996). In what follows I will restrict myself chiefly to the implications of the meme for an understanding of individual differences. Dawkins’ thinking was concentrated upon memes themselves and much less upon their human vehicles. His emphasis was upon memes as parasites actively striving for replication in the human brain. In his writing there is little sense of a host brain doing the choosing but a strong emphasis upon memes evolving in such a way as to make their spread more successful. Other writers such as Blackmore (1999) have explicitly argued that there is no ‘chooser’ there at all since ‘we’ are simply the sum total of the memes that have sprouted like weeds in our brains. But, to the extent that Dawkins admits a selection factor on memes, it is in the form of coadapted meme complexes. This describes the tendency of individual memes to coalesce into coherent higher-order complexes that in turn are capable of influencing the uptake of new memes. Religion is a favourite example. But the same argument might be applied to political ideology—a philosophy that endorses the primacy of the individual over the state can more easily accommodate memes for unfettered capitalism, low taxes, severe criminal penalties for expropriative crimes and private health care. Dawkins’s (1989, p. 199) ‘meme’s eye view’ describes the process thus: ‘Selection favours memes that exploit their cultural environment to their own advantage. This cultural environment consists of other memes which are also being selected. The meme pool therefore comes to have the attributes of an evolutionary stable set, which new memes find it hard to invade.’ Another way of putting it is that people more readily accept memes that mesh with already installed memes—they tend to avoid the cognitive dissonance that might cause an overhaul of the entire system. People tend to absorb opinions and facts that confirm their own world-view. This same view has been taken by William Durham (1991) although in his writing it is a part of a much larger picture of the relationship between memes and genes. While Dawkins sees memes as pursuing their own interests, unconcerned about the welfare of the biological vehicle that they inhabit, Durham argues that memes and genes are intimately linked with one another. Memes can constrain genes. For example, the human ability to digest lactose is restricted to those cultures that have subsisted by dairying—in non-dairying cultures a child’s ability to digest milk ends at weaning. Dairying culture selects for lactose-digesting genes. In the western world, the deve