Review of General Psychology Copyright 2000 by the Educational Publishing Foundation 2000, Vol. 4, No. 2, 186-204 1089-268O/O0/$5.0O DOI: 10.1O37//1089-2680.4.2.186

The Place of Attachment in Human Mating

Cindy Hazan and Lisa M. Diamond Cornell University

Application of the principles of evolution and natural selection to the phenomena of human mating does not lead inevitably to a single theoretical model. According to the standard evolutionary model, formally known as sexual strategies theory (D. M. Buss & D. P. Schmitt, 1993), biologically based sex differences in parental investment have resulted in hard-wired sex differences in mate preferences and mating strategies. A critical analysis of the logical and empirical foundations of the theory reveals several weaknesses and limitations. This article demonstrates how attachment theory (J. Bowlby, 1969/1982, 1973, 1979, 1980, 1988) can be used to integrate a diverse set of ideas and research findings and provide a more grounded account of human mating.

In the past decade, there has been a resur- theory, which remains one of the most robust in gence of interest among social scientists in Dar- modern science, has anything of value to con- win's theory of evolution. In the field of psy- tribute to the understanding of mating in our chology, this is exemplified by the work of species. Although all theories about the ances- David Buss and colleagues, who have applied tral nature of human mating are necessarily the theory to human mating. One measure of the speculative, in what follows we hope to dem- success of this effort is the fact that their model onstrate that the available evidence supports a is widely viewed, at least among psychologists, very different evolutionary perspective on how as the definitive evolutionary perspective on the Homo sapiens go about the business of mating. topic. Put differently, many in our field seem to think that application of the principles of natural selection to human mating phenomena inevita- The Standard Evolutionary Model bly results in the Bussian framework. An unfor- tunate result is that those who find fault with The standard evolutionary model is formally this particular framework tend to reject outright known as sexual strategies theory (Buss & the possibility that Darwin's grand and elegant Schmitt, 1993). It is essentially an extension of the theory of parental investment, proposed by Trivers in 1972. He argued that in any species in which differences exist in what it costs mem- Cindy Hazan and Lisa M. Diamond, Department of Hu- bers of each sex to reproduce their genes, there man Development, Cornell University. Lisa M. Diamond is now at the Department of Psychol- will be corresponding sex differences in mating ogy, University of Utah. behavior. The biological reality in humans is We are grateful to Richard L. Canfield for his thoughtful that males can reproduce their genes with the critique of an earlier version of this article, and to Emily minimal investment of a few minutes and a few Allen, Jeffrey Ellens, Samantha Goldman, Samara Guzman, Janienne Kondrich, and Shelby Semino for help with the sperm, whereas the cost to females is usually literature review. Our work also benefited from the support years of investment in the form of gestation, of a National Science Foundation grant (SJBR-9320364). lactation, and offspring care. In theory, such Many of the ideas and arguments contained herein were asymmetry has resulted in hard-wired sex-spe- previously presented by Cindy Hazan in conference papers at the 1997 (Toronto, Ontario, Canada) and 1998 (Lexing- cific strategies for achieving reproductive suc- ton, Kentucky) meetings of the Society for Experimental cess. Males "naturally" seek out and take ad- Social Psychology and the 1999 (Denver, Colorado) meet- vantage of opportunities to copulate with as ing of the American Psychological Society. many different females as possible, especially Correspondence concerning this article should be ad- ones who display the fertility markers of youth dressed to Cindy Hazan, Department of Human Develop- and beauty. The female is "by nature" more ment, Cornell University, Ithaca, New York 14853-4401. Electronic mail may be sent to [email protected]. sexually cautious, preferring one male who has

186 SPECIAL ISSUE: ATTACHMENT AND MATING 187 resources and appears willing to share them Putting Mate Selection Theories to with her and the offspring she will have to the Test nurture. Although sexual strategies theorists ac- knowledge that the mating behavior of men and The most straightforward way to identify the women can be similar in some respects under criteria that people use to select actual as op- certain ecological conditions, above all they posed to hypothetical mates would be to study a emphasize sex differences. The inescapable group of already-mated individuals and deter- conclusion from their writings is that differ- mine the basis on which they chose their part- ences between the sexes represent the hallmark; ners. But simply asking them has obvious lim- of human mating. itations. For one, respondents might be reluctant Scores of studies have been conducted to test to answer honestly, especially if their selection criteria included qualities that could be judged sexual strategies theory, including one massive superficial, such as money or looks. However, survey of more than 10,000 individuals in 37 concerns about this possible limitation can be different cultures (Buss, 1989). The findings easily dismissed in light of the fact that large indicate that males worldwide assign greater numbers of respondents in Buss et al.'s studies importance than females to the appearance of have voluntarily mentioned such qualities on potential mates, preferring those who are young their lists of mate preferences. A more trouble- and physically attractive. In contrast, females some possibility is that the average person lacks generally report caring more than males about awareness of the factors that truly influence his the social status, ambition, and earning power of or her mating decisions (for a similar point, see potential mates. In sum, the sex differences in Pietromonaco & Feldman Barrett, 2000). mate preferences that sexual strategies theory All theories that regard mate choice as a predicts appear to be both statistically reliable strategic process presume that choices are sys- and culturally universal. tematically guided by characteristics of the in- As the name implies, a central tenet of sexual dividuals making them. That is, whether a re- strategies theory is that human mating is inher- searcher adopts an assortative mating perspec- ently strategic. In theory, mating behavior is tive (one prefers those who are similar), an guided by evolved psychological mechanisms idiographic view (one's preferences are learned compelling men and women not only to desire and idiosyncratic), or a sexual strategies model certain qualities but to select mates on the basis (one's preferences are determined by biological of these innate desires. However, the methods sex), the same prediction would hold: Mate used in tests of the theory do not and cannot choice is driven primarily by qualities of those directly test the selection part of the prediction. doing the choosing. A corollary is that individ- Participants (often undergraduate students) are uals who have similar qualities should also have asked to generate a list of the qualities on which similar mate preferences. This makes possible they believe their eventual mate selections will an elegant "natural" test of these theories using be based or, sometimes, simply asked to rank or identical twins. Monozygotic (MZ) twins are as rate the importance of a list of traits provided by similar as two people can be. Not surprisingly, researchers. Lists typically include the full they tend to make similar choices in a wide range of desirable qualities: a winning person- range of domains, including jobs, clothing, ality, above-average intelligence, great financial friends, cars, hobbies, and vacations, to name just a few. If mate selection is lawfully guided prospects, high social status, and good looks. by the characteristics of choosers, MZ twins Although the findings are generally consistent should prefer and select mates who are, them- with theoretical predictions—at least as they selves, more alike than the mates of less similar pertain to sex differences in the estimated rela- pairs of individuals. tive future value of certain qualities—they re- veal nothing about whether reported selection Lykken and Tellegen (1993) recently tested criteria translate into actual mate choices. this hypothesis using the Minnesota Twin Reg- Clearly, the validity of sexual strategies theory istry, which contains a wealth of information, hinges on its ability to predict real-world mate including personality, achievement, IQ, atti- tude, occupation, and physical appearance data, selection behavior. on more than 1,000 twin pairs. Even with this 188 HAZAN AND DIAMOND large sample size and correspondingly strong Overview analytic power, no evidence was found to sup- port any strategic model of mate selection. The If mates are not chosen on the basis of the unexpected and hugely significant finding was sex-specific criteria that the standard evolution- that the mates of MZ twins are no more similar ary model posits or according to the criteria of to each other than are the mates of randomly other strategic mate selection theories, then selected pairs. what determines who ends up with whom? In At first, one might think that this finding the following sections, we outline an alternative deals a bigger blow to theories of assortative or evolutionary perspective on human mating that idiographic models than to sexual strategies the- proposes an answer to this burning question. ory, given that the former attribute mate choice We begin with an overview of mating relation- ships as seen through the lens of attachment to the chooser's personal characteristics (inde- theory (Bowlby, 1969/1982, 1973, 1979, 1980, pendent of sex), whereas sexual strategies the- 1988). It is argued that pair bonding is the norm ory attributes mate choice to the chooser's sex in our species, that this normative mating pat- (independent of personal characteristics). But tern is reflected in male as well as female mate personal characteristics play just as important a preferences, and that it is advantageous in re- role in sexual strategies theory via their link to productive fitness terms. Propinquity, familiar- mate value. Mate value represents an individu- ity, and romantic infatuation are proposed as al's capacity to attract high-quality mates. In the mechanisms that explain how actual mate other words, the closer one is to the ideal for "choices" are made and how mating relation- one's sex, the better able one is to attract part- ships are established. We conclude by address- ners who are ideal for their sex. ing some of the additional empirical and logical The personal characteristics that factor into limitations of sexual strategies theory and by these ideals (most notably, social status for men explicating some of the reasons that pair-bond and physical attractiveness for women) are attachment is an essential component of any more similar in twins than in randomly paired theory purporting to offer a comprehensive individuals. Thus, the genetic similarity be- model of human mating. tween twins equalizes their mate value, giving them the same range of "attractable" mates. As a result, twins should end up with mates that are Human Mating From an Attachment more similar to one another in physical attrac- Theory Perspective tiveness and social status than the mates of If one wants to know the most reproductively randomly selected pairs. But it is an empirical advantageous mating strategy for any species, a fact that they do not. safe bet is simply to observe what the members Previous tests of sexual strategies theory have of that species normally do. In species in which relied exclusively on self-reports from individ- conception is all that is needed to achieve re- uals who are asked about the qualities they productive success, sexual partners part ways as believe will influence their eventual mating de- soon as a viable pregnancy has been achieved, cisions. The twin findings strongly suggest that sometimes after a single copulatory sequence. the criteria people think they will use to select Yet, this is not the case for our species. Instead, their mates are not the factors that ultimately most humans opt to remain with their reproduc- determine their actual mate choices. Equally tive partners for an extended period of time important is the fact that the twin data challenge (Fisher, 1989, 1992; Lancaster & Kaplan, 1994; the fundamental assumption that mate selection Van den Berghe, 1979). This runs counter to the is a strategic process. It is perhaps ironic that impression conveyed by sexual strategies the- social scientists have taken the premise of stra- ory, namely that an enduring relationship be- tegic selection at face value and devoted much tween human mates is difficult to achieve and research attention to identifying the criteria on therefore relatively rare as a result of males' which such selections are based. We concur evolved inclination to forgo long-term mating with the twin researchers' conclusion that hu- associations in favor of multiple inseminations. man mating may be more "adventitious" than is The reality, however, is that human reproduc- generally assumed. tive partners typically remain together for a SPECIAL ISSUE: ATTACHMENT AND MATING 189 minimum of several years, a fact that any theory Defining Components of human mating must take into account. If short-term mating is so advantageous for Bowlby took care to define the specific type males, why should this long-term mating pat- of relationship to which his theory of attach- tern ever have evolved? It is thought that a trend ment applied and to distinguish this special toward extended associations between human bond from other kinds of social ties. Attach- mates evolved in response to a birthing crisis in ments have four defining features, all of which which the infant's large head could not easily are evident in the overt behaviors directed to- pass through the birth canal of our bipedal fe- ward an attachment figure: seeking and main- male ancestors (Trevathan, 1987; Washbum, taining physical proximity (proximity mainte- 1960). Infants who were born prematurely, with nance), seeking comfort or aid when needed less developed brains and correspondingly (safe haven), experiencing distress on unex- smaller heads, were more likely to survive, as pected or prolonged separations (separation were their mothers. Immaturity at birth offered distress), and relying on the attachment figure as a base of security from which to engage in the additional advantage of a longer period of exploratory and other nonattachment activities postnatal neural plasticity and potential for (secure base). learning, which is important for a highly social species. However, the benefits of premature In infancy, these behaviors are directed al- birth would have posed new adaptive chal- most exclusively toward primary caregivers. lenges. Specifically, the effort required to ade- Babies often seek and enjoy contact with many quately nurture exceptionally dependent off- individuals and even look for and accept com- fort from them. Yet, usually a much smaller spring during a protracted period of immaturity number of people, not uncommonly one, qual- made paternal investment an advantage, if not a ify as an attachment figure by being the target of necessity (Mellen, 1981; Small, 1993). Helpless all four behaviors. Infants are especially dis- and vulnerable offspring would have had little criminating in the expression of separation dis- chance of surviving to reproductive age or de- tress. This particular component of attachment veloping the skills needed for their own even- typically emerges between 6 and 8 months of tual mating and parenting roles unless fathers age, and its appearance is the accepted marker participated in their protection, care, and social- that an attachment bond is in place (Ainsworth, ization. Thus, a new adaptive problem arose: Blehar, Waters, & Wall, 1978). This time how to keep males around and involved in the course is universal and, within normal ranges, care of their progeny. independent of rearing environment (Kagan, As it happened, our species already had avail- 1984). able a specialized but flexible mechanism for An explicit prediction of attachment theory is fostering an enduring bond between two indi- that, in adulthood, these same behaviors will be viduals. The mechanism was attachment, which redirected toward a mate. Two empirical inves- had helped to ensure a survival-enhancing tie tigations have confirmed this prediction (Fraley between infant and mother. Given the generally & Davis, 1997; Hazan & Zeifman, 1994). Al- conservative tendencies of evolution and natu- though adults, like infants, seek and enjoy con- ral selection, it is parsimonious to suppose that tact with a variety of people and sometimes turn this mechanism was exploited for the (new) to them for comfort or reassurance, most of purpose of cementing a bond between reproduc- these relationships do not qualify as attachment tive partners. This co-opting of an evolved bonds. Those that do, by virtue of containing all structure for a novel purpose is fairly common; four defining components, are almost exclu- such structures are called "exaptations" (Gould sively formed with romantic partners. Thus, & Vrba, 1982). To suggest that the attachment by this standard, mate relationships are attach- system was exploited to keep adult mates to- ments in the technical sense of the term. gether is to claim that the tie between estab- lished pairs is an attachment bond in the strict Reactions to Separation and Loss and specific sense of the term. There is diverse and abundant evidence that this is so (Hazan & Additional evidence that attachment is an in- Zeifman, 1999). tegral part of mate relationships comes from the 190 HAZAN AND DIAMOND literature on bereavement. The original inspira- attachment figure or if an established bond is tion for attachment theory was Bowlby's obser- disrupted (Harlow & Harlow, 1965; Kraemer, vations of infants and children separated from 1997; Robertson, 1953; Spitz, 1946; Suomi, their primary caregivers. He found it remark- 1997). Although adults are clearly less depen- able that the separations were so distressing dent on an attachment figure for basic survival, even in situations in which nutritional and hy- there is ample evidence that they incur health gienic needs were being met quite adequately benefits from having one and are at significantly by surrogates. More striking was the similarity increased risk for numerous physical and mental across children in how they responded. Bowlby health problems if they do not. For example, identified what appeared to be a universal pat- divorce increases the likelihood of admission tern of reactions that he labeled the protest- to a psychiatric facility, alcoholism and other despair-detachment sequence. The initial reac- forms of substance abuse, accidents, suicide, tion is characterized by behavioral agitation, and impaired functioning of the cardiac, endo- hyperactivity, crying, resistance to others' of- crine, and immune systems (e.g., Bloom, Asher, fers of comfort, and extreme anxiety, often to & White, 1978; Goodwin, Hurt, Key, & Sarret, the point of panic. Eventually, active protest 1987; Lynch, 1977; Uchino, Cacioppo, & subsides and is followed by a period of lethargy, Kiecolt-Glaser, 1996). Among the most com- inactivity, depressed affect, disrupted sleep, and mon life stressors, attachment-related losses reduced appetite. In time, a degree of emotional cause the most subjective distress. Death of detachment from the lost attachment figure fa- a spouse is rated as the most stressful event, cilitates the resumption of normal, presepara- followed by divorce and marital separation tion activities and functioning. (Holmes & Rahe, 1967). Other losses can be Several studies have documented essentially quite painful, but they generally have not been the same sequence in adults grieving the loss of found to jeopardize physical or psychological a long-term partner: initial panic and anxiety well-being to the same degree as disruption of a followed by lethargy and depression and, even- pair-bond attachment. If separation distress is tually, recovery through emotional detachment the marker of attachment, then bonds between (Fraley & Shaver, 1999; Hazan & Shaver, 1992; mates clearly qualify. Parkes & Weiss, 1983; Weiss, 1975). This se- quence of reactions is not limited to situations Nature of Physical Contact of permanent loss. Even brief, routine separa- tions are enough to trigger a less intense but Like caregivers and their infants, romantic essentially identical pattern of responses in mar- partners (at least initially) spend much time ital partners (Vormbrock, 1993). engaged in mutual gazing, cuddling, nuzzling, It is important to note that the protest-de- sucking, and kissing in the context of prolonged spair-detachment sequence is observed almost face-to-face, skin-to-skin, ventro-ventral con- exclusively in two social relational contexts: tact. These most intimate of human interper- infant-caregiver relationships and adult pair sonal exchanges are universally typical of only bonds. The fact that this three-stage reaction is two types of relationships: those between in- the norm among adults separated from their fants and caregivers and those between mates mates, and not the normal reaction to the loss of (Eibl-Eibesfeldt, 1975). Not coincidentally, this other kinds of social ties, is another indication type of physical contact is known to foster that the attachment mechanism is operative in attachment. pair bonds. And it is another fact of human There is even evidence that the chemical ba- mating that belongs in theories of what such sis for the effects of close physical contact may relationships are like. be the same for lovers and mother-infant pairs (for a review, see Insel, 2000). Oxytocin is an Health Effects endogenous hormone that triggers labor in preg- nant women and milk letdown in nursing moth- It is well established that human infants, as ers. It is thought to aid infant attachment and well as the young of other primate species, maternal caregiving by inducing a state of calm suffer long-term negative health effects if they and contentment and by stimulating a desire for are not given an opportunity to bond with an continued close bodily contact. Oxytocin is also SPECIAL ISSUE: ATTACHMENT AND MATING 191 present during sexual exchanges between adult the erect human penis is 13 cm, as compared lovers. It builds with sexual stimulation and with approximately 3 cm for the gorilla, a much excitement, is released at climax—in both larger animal in terms of overall body size. The males and females—and has been implicated in exceptional length of the human penis, in addi- the cuddling or "afterplay" that often follows tion to its unique thickness and flexibility rela- sexual intercourse (Carter, 1992, 1998). Cud- tive to that of all the great apes (Eberhard, 1985, dling or contact comfort, as was famously dem- 1991), made possible a wide variety of copula- onstrated by Harlow, is crucial for the establish- tory positions, including more intimate face- ment of attachment bonds. Thus, if sexual to-face, mutually ventral (i.e., bond-promoting) orgasm triggers release of a hormone that stim- positions. Distinctive features of the human pe- ulates a desire for bond-promoting contact, it nis also make female orgasm more likely, which effectively increases the chances that a mating in turn increases the probability of conception pair will become emotionally attached. by causing her cervix to dip rhythmically into the semen pool and her uterus to contract in a Sexual Anatomy and Physiology manner that helps move sperm toward egg (Baker & Bellis, 1995). In addition, it may have Many unique features of human sexuality served to heighten her readiness for engaging in foster and help maintain an enduring bond be- sexual activity, thereby strengthening the bond tween reproductive partners. The most striking with her mate (G. F. Miller, 1998). difference in our reproductive physiology in It deserves noting that penis size alone is not comparison with that of other mammalian spe- an accurate predictor of monogamous versus cies is the absence of outward signs of estrus. polygamous mating patterns among primates, Most mammals mate only during the short es- nor is our species unique in such reproductive trus periods of the female, but human sexual characteristics as hidden ovulation, female or- desire and activity are not so restricted. Women gasm, or face-to-face copulation (Blaffer-Hrdy, can be sexually receptive during any phase of 1988). We do not claim that the presence of any their reproductive cycle, even though concep- of these characteristics represents proof that hu- tion is possible only during a small fraction of mans are hardwired for monogamy. As many it. This physiological adaptation facilitates a anthropologists have argued quite convincingly continuous tie between partners on the basis of (e.g., Small, 1993), facultative flexibility in sexual reward. Among older couples, in which mating strategies has been a key component of the female's reproductive potential may be our survival as a species. Nevertheless, multiple practically nonexistent, nearly half continue to features of human sexual anatomy and physiol- have sex an average of once per week (Lau- ogy support the view that we evolved to bond mann, Gagnon, Michael, & Michaels, 1994). with our reproductive partners. In addition, the All of this suggests that sex in our species decrease in sexual dimorphism over the course serves more than a reproductive function. of human evolution is another indication of Hidden ovulation may also serve to diminish movement away from opportunistic mating in the benefits of straying. Males of many diverse the direction of more enduring pair bonds, not species guard their mates during periods of sex- as a culturally prescribed arrangement but as a ual receptivity so as to ensure paternity. When pathway to reproductive success for a species the fertile period has ended, he can safely move whose young fare best with two investing par- on to another receptive partner. However, if ents. If one gives this capacity for bonding and ovulation is hidden, making it impossible for the reproductive implications of reduced sexual the male to determine just when fertilization dimorphism their due, one is led to a rather will be possible, his optimal strategy may shift different evolutionary perspective on human toward guarding and remaining with the same mating. sexual partner for longer periods of time (Al- cock, 1989). Mate Preferences Genital differences between humans and their closest primate relatives also suggest the If natural selection pressures on human mat- important role of sex in maintaining the human ing operated in a manner that is consistent with pair bond. For example, the average length of our pair-bond attachment hypothesis, then at- 192 HAZAN AND DIAMOND tachment considerations ought to be reflected in But do people really consider a trait such as mate preferences. Moreover, there is no theo- kindness more important than other qualities retical basis for predicting that males and fe- when choosing a mate? To address this ques- males would differ in this respect. The qualities tipn, Graziano and colleagues (Graziano, that make one a good attachment figure—being Jensen-Campbell, Todd, & Finch, 1997) con- kind, warm, responsive, and competent—do not ducted a series of experiments in which hetero- vary as a function of sex. In contrast, sexual sexual women rated the attractiveness of men strategies theory makes the explicit prediction whose hypothetical traits of dominance and that males and females will look for different agreeableness were systematically manipulated. qualities in a potential mate owing to differ- The experimental conditions represented the ences in parental investment that are present various combinations of these traits that women even before conception (Trivers, 1972). Earlier might encounter in their real-world dating ex- we sketched the basic premises of this argu- periences. Some men are really nice but not ment; here we revisit it in greater depth. very dominant, whereas others may have high From puberty through late adulthood, males status in the dominance hierarchy but be some- produce approximately 500 million tiny sperm what lacking in agreeableness. cells per day (Zimmerman, Maude, & Mold- From a sexual strategies perspective, domi- awar, 1965). Females, on the other hand, pro- nance should triumph. After all, agreeableness duce an average of one large egg per month in a mate is a luxury for females facing years of from puberty until middle adulthood. For males, parental investment. They should favor the male whose contribution to their progeny can be as who will reliably beat out other males in the little as a few cheap sperm, the most effective competition for status and resources. This is strategy for achieving reproductive success may not what Graziano et al. (1997) found. When be to take advantage of all opportunities for sex women were forced to choose, they consistently with fertile female partners. The female, for preferred the nice guy. Apparently, our chim- whom every sexual encounter is potentially panzee cousins do too (Sapolsky, 1998). In fact, quite costly, is predicted to be far choosier in agreeableness was three times as powerful as accepting or encouraging copulations. Once her dominance in predicting women's attractions. egg is fertilized, she has to forgo other repro- Further evidence that mate preferences are ductive opportunities for a relatively long pe- not always reducible to sex differences in pa- riod of time. Thus, her most effective strategy is rental investment comes from studies of male to limit her sexual encounters to males who and female facial attractiveness. In a series of possess and appear willing to share valuable detailed analyses involving facial-metric meth- resources with her and her offspring. ods and cross-cultural samples, Cunningham Previously we cited the 37 cultures study and his collaborators (Cunningham, Druen, & (Buss, 1989) in which it was found that males Barbee, 1997) have sought to identify the fea- and females universally differ in the qualities tures that make potential mates of both sexes they seek in a mate. Men assign greater weight most appealing. The findings vary somewhat as to physical attractiveness, and women assign a function of gender, but the overall combina- greater weight to social status. But what gets tion of features judged to be most attractive is little press or theoretical attention is the fact that essentially the same for men and women. The neither physical appearance nor social status is winning configuration combines three types of given top billing by either sex. When men and facial features: expressive, neotenous, and sex- women the world over make their mate wish ual maturational. Expressive features (e.g., size lists, there are other qualities that they rank as of smile area) serve as cues of warmth and more important. The top-ranked qualities are sensitivity; neotenous features (e.g., large eyes) identical for men and women. What are these signal vulnerability; and sexual maturational universally coveted mate traits? The answer is features (e.g., prominent cheekbones) function kindness, understanding, and intelligence (Buss as cues of reproductive capability. In other & Barnes, 1986), roughly the same qualities that words, judgments about attractiveness do not appeal to infants in potential attachment figures boil down to the kinds of sex differences that (Bowlby, 1969/1982). sexual strategies theory would predict. Instead, SPECIAL ISSUE: ATTACHMENT AND MATING 193 attachment qualities figure prominently in the Evidence That Pair-Bond Attachment Is judgments of both sexes. Reproductively Advantageous Such findings demonstrate that human mate preferences are not governed exclusively by a In our species, reproductive success requires self-contained psychological mechanism cus- negotiation of at least three adaptive challenges: tom designed by evolution to propel individuals surviving to reproductive age, acquiring and toward the singular, circumscribed goal of sex- retaining a mate, and providing adequate care to ual intercourse. Not only is such a system in- offspring so that they also survive to reproduce. consistent with the data, we think it makes little It was argued previously that the emergence of sense in light of the survival problems faced by bipedalism favored premature birth and that the our ancestors in the environment of evolution- extreme immaturity of human neonates created ary adaptation (EEA). Just as Bowlby theorized, a situation in the EEA in which survival de- human mating involves multiple evolved mech- pended not only on infants forming a strong anisms. Whereas attachments between infants bond to a protector but also on the joint invest- and caregivers are inherently asymmetric—with ment of parents. This necessitated a mechanism infants seeking the emotional security, protec- that would hold reproductive partners together for an extended period of time. We proposed tion, and care that parents provide—attach- that attachment, which had evolved to ensure an ments between romantic partners are symmet- enduring bond between infants and caregivers, rical in that both partners mutually seek and was exploited for this new purpose (for a dif- provide the same social provisions. In the course ferent view of the adaptive value of pair bond- of normative development, these evolved psycho- ing, see Insel, 2000). In what follows, we argue logical mechanisms (attachment and parental that the advantages of pair bonding include en- caregiving) become integrated with the sexual hanced survival and reproductive fitness for mating system (Hazan & Shaver, 1994a, 1994b; mates as well as their offspring. Hazan & Zeifman, 1994; Shaver & Hazan, 1993; Whether opportunistic, short-term mating Shaver, Hazan, & Bradshaw, 1988). From an at- strategies are inferior to stable long-term ap- tachment perspective, it thus comes as no surprise proaches is the source of much debate (e.g., that the faces judged most appealing by men and Belsky, 1999; Buss, 1997; Chisholm, 1996), but women alike are those that combine the signal it is important to remember that these are not stimuli of all three mechanisms. mutually exclusive strategies. Pair-bond attach- The differences between a sexual strategies ment is not synonymous with sexual exclusiv- perspective on mating and an attachment ap- ity. In fact, genetic analyses of offspring pro- proach should now be apparent. For sexual vide objective evidence that extrapair copula- strategies theory, mating is about the reproduc- tions are common even in species that by all tion of genes via sexual activity. Emotional ethological criteria qualify as monogamous (Carter et al., 1997; Mendoza & Mason, 1997). bonds are relevant only insofar as they help or The relative value of short-term versus long- hinder this process. From an attachment per- term mating strategies will vary depending on spective, these bonds belong at the center, rather environmental contingencies. According to life than the periphery, of evolutionary theories of history theory (Stearns, 1992), organisms pos- human mating. However, to advance this argu- sess a finite amount of resources that must ment requires more than simply citing evidence be allocated across various evolutionary chal- that pair bonding is our species-typical mating lenges, including survival, growth, mating, and pattern, that the attachment mechanism is oper- parental investment. Local circumstances deter- ative in pair bonds, that numerous features of mine the balance of time and energy an indi- human sexual anatomy and physiology foster vidual devotes to each. Thus, both long- and mate attachment, or that attachment consider- short-term strategies can be viewed as reason- ations consistently outweigh physical appear- able and comparably adaptive responses, given ance and social status in male and female mate a particular ecology. preferences. A case must be made that attach- It is clearly advantageous for humans to be ment bonds between reproductive partners carry capable of adapting their mating strategies to advantages in terms of reproductive fitness. local ecological conditions (Daly & Wilson, 194 HAZAN AND DIAMOND

1988). This does not necessarily mean that spring of stable pairs are better equipped to short- and long-term mating strategies are dif- attract and retain their own mates. Adolescents ferent but essentially equal solutions to the same from father-absent homes show precocious sex- evolutionary challenges (but see Belsky, 1999, ual interest, relatively early sexual maturation, for an alternative viewpoint). Adjustments to more negative attitudes toward potential mates, nonoptimal circumstances are sometimes nec- and less interest in long-term relationships than essary, but they may not produce optimal re- do their counterparts in father-present homes sults. Take feeding behavior, for example. Sur- (Draper & Belsky, 1990; Draper & Harpending, vival depends on the regular intake of food, and, 1982; Surbey, 1990). In other words, when re- if hungry enough, humans will consume almost productive partners do not maintain a long-term anything to stay alive. But garbage is unlikely bond, their children are more likely to adopt to have the same nutritional value as a well- approaches to mating that emphasize quantity rounded meal, nor would it be expected to sup- over quality. Parental divorce has also been port physical development equally well. found to affect offspring mating behavior. Fe- Likewise, quick and frequent copulations male children of divorce tend to fear intimacy coupled with an avoidance of parental invest- and have difficulty establishing committed rela- ment may be the best available strategy in some tionships, whereas the effects for males are ev- circumstances, but it hardly stands as the best idenced in a lack of achievement orientation route to reproductive success. The clearest ob- (Wallerstein, 1994) and lower eventual socio- jective evidence of this fact is that infant mor- economic status (Lillard & Gerner, in press). tality rates are significantly higher among chil- Thus, the failure of reproductive partners to dren without an investing father (Hill & Hur- maintain a long-term bond can negatively affect tado, 1995). Even putting aside issues of the mating appeal and success of their offspring. offspring survival, there is other evidence that Attachment plays an important role in the long-term bonds between partners directly en- three adaptive challenges that our species must hance their own reproductive success. For ex- negotiate to achieve reproductive success: sur- ample, women ovulate more often and more viving to reproductive age, mating, and nurtur- regularly if they are in a stable sexual relation- ing offspring to reproductive age. Why, then, is ship (Cutler, Garcia, Huggins, & Preti, 1986; attachment nowhere to be found in the predom- Veith, Buck, Getzlaf, Van Dalfsen, & Slade, inant evolutionary model of human mating? 1983). They also tend to continue ovulating The (over)emphasis on sex differences has dis- longer and reach menopause significantly later. tracted us from the reality that men and women Earlier we cited evidence that partners in long- are basically similar in what they seek in a mate, term relationships enjoy more robust physical the processes by which they become attached to and mental heath. The more fit an individual, a mate, and the benefits that accrue to them as a the better able he or she is to function in all of result of being in a stable pair bond (Zeifman & the various roles adults are required to fill, in- Hazan, 1997). cluding those of mate, parent, and grandparent. A healthy member of any social group is more A Mechanism to Foster Mate Attachment valued, more valuable, and more capable of protecting the self as well as loved ones. A How does one get from attraction to pair stable bond with a trusted and reliable compan- bonding? An enduring emotional bond between ion also promotes the kind of exploration and reproductive partners takes time to develop. The engagement in constructive activity on which attachment mechanism can maintain the bond, family welfare depends (Hazan & Shaver, but it would have required a different kind of 1990). Like children, adults benefit from having psychological mechanism to hold the pair to- someone to look out for and keep track of them, gether long enough for an attachment to form. someone to initiate a search if they fail to show Ideally, this mechanism would engender a sin- up at the expected time, to care for them when gle-minded focus on the partner at hand, to the they are sick, dress their wounds, and help de- exclusion of other potential mates. In addition, fend them against external threats. it would need to operate at least as strongly, if Other evidence that pair bonding is a superior not more so, in males as females. Moreover, it reproductive strategy is the finding that the off- would have to promote and sustain a strong SPECIAL ISSUE: ATTACHMENT AND MATING 195 desire for the type of close physical contact dized if infants rejected any protector who known to foster attachment. failed to match some ideal of the perfect care- In fact, such a mechanism exists; it is called giver. And, in actuality, the attachment mecha- romantic infatuation. In the largest and most nism in human infants can be engaged by al- systematic investigation of romantic infatua- most any conspecific. Although babies are hap- tion, Tennov (1979) conducted a content anal- piest and develop optimally when caregivers are ysis of the questionnaire and interview re- consistently warm and responsive (Ainsworth sponses of hundreds of men and women. From et al., 1978), they become fully attached to this database, she was able to identify the com- abusive caregivers (Crittenden, 1995) and even mon features of this highly common phenome- other children if adult figures are not available non. In addition to a reduction in perceived need (Freud & Dann, 1951). Both Harlow and Lorenz for or interest in food and sleep and a paradox- demonstrated rather dramatically the flexibility ical increase in energy, the symptoms include of the attachment mechanism in other species. It mental preoccupation with and idealization of is worth seriously considering the possibility the target of infatuation and an intense longing that the search image for human mating is also for intimate physical contact. The overwhelm- inherently flexible. ing majority of male and female participants in As noted previously, sexual strategies re- Tennov's study had experienced this constella- search on mate selection has tended to focus on tion of feelings. Leibowitz (1983) proposed that qualities of the ideal mate, usually assessed by the physiological arousal and idealization that asking people to describe the kind of individual typify romantic infatuation are mediated by they would most like to have as their partner. phenylethylamine, an endogenous amphetamine Evolutionary theorists assign special impor- that also has mild hallucinogenic effects. tance to the finding that males and females Whatever the source of these symptoms, the reliably differ in their relative rankings of phys- available empirical evidence indicates that they ical appearance and status-resources. However, strike men and women with equal frequency rankings and ratings of hypothetical partner and intensity. If sexual strategies theory were qualities, no matter how consistent within gen- correct in its postulation of a powerful evolved der or across cultures, represent only weak ev- male preference for short-term mating, why idence that any of these qualities figure into or would men be so susceptible to falling in love influence mate selection. with the female targets of their attraction? It In the twin study described earlier, selection makes perfect sense within an attachment model criteria were examined not by asking people of mating but is more difficult to explain from a what they desired in a mate or what they sexual strategies theory perspective. Perhaps thought influenced their decisions but by ex- this is because sexual strategies theory is a ploring the outcomes of actual matings. The rational model of mating, and everyone knows findings are not consistent with those from stud- that there is nothing rational about romantic ies based on evaluations of hypothetical part- infatuation! ners, nor do they support predictions derived from sexual strategies theory. In fact, the pattern How Specific Mates Are "Chosen" of pairings observed in this large sample of couples led the authors to conclude that mating Although romantic infatuation explains why is not lawfully guided by any of the factors two potential mates would engage in the kinds that relationship researchers have proposed as of intimate interaction that might result in their important. becoming attached to each other, it begs the The authors raised the intriguing possibility question of how and why one particular indi- that human mating may be a largely "adventi- vidual becomes the sole focus of attention and tious" process. After all, if our hominid ances- passion. The sexual strategies model implies tors lived in relatively small and isolated bands, specific triggers for infatuation. In our view, it as many anthropologists contend, there may not would be maladaptive for the triggers to be very have been many choices available to them. If specific. From an attachment perspective, the choice was not an adaptive problem faced by mental image of a suitable figure is only sche- our predecessors in the EEA, then there would matic. Imagine how survival would be jeopar- have been no need for—and thus no pressure 196 KAZAN AND DIAMOND for—the development of a mate choice mecha- easy enough to recognize to satisfy the first nism. This possibility questions the core requirement; the innate disgust response could premise underlying decades of research on mate serve to satisfy the second. selection. Have the countless hours spent theo- There are three important points to be made rizing about and investigating the criteria by here. First, if nearly every postpubescent young which humans choose their mates been for person is fertile and there is a "natural" aversion naught because, in the end, there is little or no to sexual contact with those who display age- or active choosing going on? disease-related cues of infertility, it is doubtful There are data to support a less extreme con- that natural selection would have needed to clusion. After all, logic dictates that mating exert pressure for the development of a fertility- decisions cannot be completely random. Our detecting mechanism. Second, if our mate se- species would have expired long ago had we not lection mechanisms evolved in an environment succeeded in choosing mates who were capable characterized by small and relatively isolated of reproducing. Thus, at a minimum, there must groups, the average person would not have had be some species-typical mechanism for elimi- the wide range of mate options that typify the nating infertile partners from the pool of eligi- contemporary undergraduate on whose mate bles. An important foundation of sexual strate- preferences sexual strategies theory is largely gies theory is the claim that this was a problem based. Thus, it is quite possible that, in the EEA, faced by men but not women because post- there was never strong selection pressure for the pubescent males are reproductively capable development of any highly specific mate choice throughout their lives. This is not actually the mechanism. Third, given these considerations, case. Male fertility and reproductive capability evolution may have equipped us with mate re- also decline with age, albeit not as precipitously jection as opposed to mate selection criteria. as for females. For example, the proportion of That is, our mating system may be designed to morphologically abnormal sperm increases steer us away from fatal choices rather than starting at about 35 years of age (Schwartz et toward ideal ones. Everyone can describe his or al., 1983), and the ability to achieve full erec- her ideal mate, but even those who qualify as tion decreases steadily from middle adulthood high in mate value often have to compromise. through old age (Rowland, Greenleaf, Dorfman, Yet, there are undoubtedly limits to the com- & Davidson, 1993). Needless to say, sperm are promises that any individual is willing to make. of little use in reproductive fitness terms if the To be in the running, a potential mate must delivery system is nonfunctional. The fact that simply surpass some threshold of acceptability. life expectancy for males and females during Does this mean that mating decisions (within the Pleistocene probably did not extend far the pool of acceptable partners) are random? beyond the age of menopause (Mellen, 1981) No, but neither are they necessarily strategic. may render the whole debate somewhat moot We agree with the twin researchers that there anyway. may be an adventitious or chance factor driving Putting aside issues of sex differences for the mate choice, and we propose that propinquity is moment, it clearly would have been important this factor. In other words, the true pool of for members of our species, like all others, to eligibles is not defined as those individuals who choose reproductively capable mates. The ques- are the best match to one's ideals (even control- tion is whether this was, as sexual strategies ling for one's own mate value), nor does it theory posits, a problem to be solved. If one include every fertile person under the sun. examines fertility in contemporary hunter-gath- Rather, it consists of those individuals who both erer societies, the usual comparison group for exceed one's rejection criteria and happen to be speculations about life in the EEA, it does not living or working or playing nearby. This may appear that it was. In such societies, more than seem so commonsensical as to be trivial, but we 90% of postpubescent young people are fertile view it as a critically important point that is (Symons, 1979). Thus, it may have sufficed for routinely overlooked. A cost-benefit evaluation our ancestors to (a) avoid mating with conspe- would favor mating with one who is adequately cifics who had not yet reached puberty and (b) attractive and readily accessible. It might be avoid mating with those who showed signs of possible to find a marginally more attractive or aging or disease. The markers of puberty are higher status mate with a longer and wider SPECIAL ISSUE: ATTACHMENT AND MATING 197 search, but a small incremental change on either they would ideally like in a mate, the traits most dimension would probably make little differ- valued by both males and females include kind- ence in the final (reproductive fitness) analysis. ness and understanding. In other words, they Furthermore, familiarity generally breeds at- say they want a mate who will respond posi- traction (Rubin, 1973). Familiarity effects are tively to and treat them well. Reciprocal liking themselves thought to reflect an evolved ten- can be interpreted as a signal that one has found dency to respond favorably to individuals who just such a person, and the fact that this person have passed the fhend-versus-foe test. Propin- is acceptably attractive and possesses all of the quity affords not only opportunities for mating "right" body parts may be enough to send most but the kind of prolonged association that in- people head over heels. For a significant num- creases familiarity and, in turn, enhances the ber, this alone is enough to capture and hold mutual attractiveness of potential mates. As ev- their interest until a more enduring bond devel- ery zookeeper knows, a nearly sure-fire way to ops. How long does that take? In adult pairs, it get two members of any species to mate is requires approximately 2 years, give or take 6 simply to house them in the same cage. Why months (Hazan & Zeifman, 1994). Not coinci- must it be different for Homo sapiens? dentally, this time frame corresponds to the The evidence suggests that it is not. Most average duration of romantic infatuation (Ten- humans end up with a mate who lives within nov, 1979). walking distance (Eckland, 1968). Yet, of Further evidence that infatuation is more course, "lives within walking distance" would powerful than strategic choice comes from re- never appear on anyone's wish list of ideal mate search detailing its effects on social perception. characteristics. This is important for what it For example, although physical attractiveness reveals about the mate selection process in our can be defined objectively, how attractive one species. Specifically, the factors that exert the perceives a particular individual to be depends strongest influence on humans' mating behavior in large part on how one happens to feel about may be those that operate outside of their con- him or her. More than 20 years ago, Murstein scious awareness. (1976) conducted a study in which members of We have yet to address the question of why a married couples were asked to (secretly) rate particular pair ends up together. Propinquity the physical attractiveness of their respective and familiarity, along with individual qualities spouses. Eighty-five percent of the husbands and preferences, obviously narrow the pool con- rated their wives as above average on looks, siderably, but what determines the pairing of whereas fewer than 25% were judged to be specific mates? Stated differently, what triggers above average in physical attractiveness by a romantic infatuation? What is responsible for panel of judges using the same rating scale. It is the transition from merely finding another per- often assumed that a woman is desired because son attractive to making him or her the sole she is attractive. This study demonstrates that focus of intense longing and preoccupation? In the causal arrow can run in the opposite one of only a small number of mating studies to direction. use individuals who were already paired, Aron Another example of the influence of romantic and his colleagues (Aron, Dutton, Aron, & Iver- infatuation on social perception comes from a son, 1989) discovered a possible answer. study conducted by Simpson and colleagues The approach they took was to ask partici- (Simpson, Gangestad, & Lerma, 1990). Hetero- pants to provide detailed accounts of their fall- sexual participants with steady dating partners ing-in-love experiences. According to the find- rated opposite-sex age-mates as less physically ings, the factor primarily responsible for the and sexually attractive than did participants shift from attraction to infatuation is reciprocal without steady partners. The two groups did not liking, or the perception that the person one is differ in their ratings of much younger or much interested in feels the same way. Whether ex- older opposite-sex individuals, suggesting that pressed in a warm smile or a prolonged gaze, the effect was specific to potential mates. Along the message is unmistakable: "It's safe to ap- the same lines, Murray and Holmes (1993, proach. I like you, too. I'll be nice. You're not 1994; Murray, Holmes, & Griffin, 1996) have, in danger of being rejected." Recall that when in numerous studies, documented the benefits of people are asked to make lists of the qualities mutual idealization in couples. Given that the 198 HAZAN AND DIAMOND availability of attractive alternative partners As in the study just cited, sometimes findings constitutes one of the greatest threats to the that appear to be supportive of sexual strategies stability of an existing relationship (Rusbult, theory have plausible alternative explanations. 1980, 1983), the power of idealization to simul- Another example is a study conducted by De- taneously inflate perceptions of a mate's appeal Steno and Salovey (1996). Sexual strategies and reduce the appeal of potential rivals under- theory predicts sex differences in reactions to a scores its importance in maintaining the integ- partner's infidelity depending on whether the rity of pair bonds. In short, romantic infatuation indiscretion is sexual or emotional. Specifically, leaves people with the feeling that they have men are predicted to show greater concern and found the ideal mate, their "one and only." upset over a female partner's sexual infidelity for the reason that it undermines paternity cer- tainty. Women are predicted to be less bothered More Limitations of Sexual Strategies by a male mate's sexual wanderings than by his Theory emotional involvement with another woman, because the latter raises the possibility of losing We believe that the empirical evidence re- his valued resources. In a study that involved viewed thus far presents a serious challenge to multiple indexes of distress, including heart sexual strategies theory and makes a compelling rate, perspiration, and frowning, these theory- case for a very different evolutionary model of based predictions were confirmed (Buss, Lar- human mating. But an obvious question at this sen, Westen, & Semmelroth, 1992). point is, If the evidence for sexual strategies DeSteno and Salovey took issue with the theory is so weak, why does it appear to be so authors' interpretation, arguing that the results strong? In the remaining sections, we offer a could be explained in terms of participants' possible explanation, beliefs concerning the covariation between sex- As previously noted, a major limitation of the ual and emotional infidelity. Leaving aside for support for sexual strategies theory is a reliance now questions about whether sex differences in on self-reports and hypothetical mating scenar- sexual behavior reflect evolved mating strate- ios. But there are some behavioral data that gies, it is nonetheless true that males are more Buss and Schmitt (1993) cited to bolster their likely than females to engage in casual sex. theory. For example, in a study by Clark and More important, males and females are aware of Hatfield (1989), an attractive male or female this sex difference. DeSteno and Salovey rea- research confederate approached opposite-sex soned that participants' knowledge of this dif- strangers on a college campus and, after an- ference could explain the findings. Specifically, nouncing that he or she found the stranger at- women might reasonably assume that male tractive, posed one of three randomly selected emotional infidelities are strongly suggestive of invitations: to go for dinner, to go to the con- sexual infidelity as well, and men can safely federate's apartment, or to have sex. The big- assume that female sexual unfaithfulness also gest sex difference was found for the third con- involves emotional infidelity. In other words, dition. Fully 75% of the men accepted the in- maybe Buss et al.'s participants were showing vitation for sex, whereas none of the women the greatest distress not to a single type of consented. The results have been questioned on infidelity but instead to the type of infidelity that the grounds that the risks associated with ac- implied both: what DeSteno and Salovey called cepting an invitation to visit the apartment of or the "double-shot" hypothesis. And when partic- have sex with a complete stranger are quite ipant beliefs about the covariation between sex- different for men and women, and thus have ual and emotional infidelity were assessed, be- ambiguous implications for their respective in- liefs were more powerful in predicting distress terest in sex per se (Small, 1995). Nevertheless, than was biological sex. this finding and others like it are offered as evidence that males have evolved qualitatively different mating tactics than females and that Do Men "Naturally" Prefer Short-Term their "best" strategy for achieving reproductive Mating ? success is to have sex with as many fertile There is probably no issue on which sexual females as they can. strategies theory depends more, and thus no SPECIAL ISSUE: ATTACHMENT AND MATING 199 issue on which the merits of the theory can be this hard-wired preference not evident in the more critically evaluated, than the sex differ- self-reports of men who are at their sexual peak ence in casual sex, or what has been termed and living in the midst of extraordinary numbers sociosexuality (Simpson & Gangestad, 1992). of young and beautiful potential mates? Could Accordingly, in their original comprehensive the answer be that they have been designed by exposition of the theory, Buss and Schmitt evolution for pair bonding? As for those few (1993) devoted three separate figures to demon- men who skewed the distribution by reporting a strating sex differences in sexual promiscuity. desire for 100 or more future sexual partners, Participants were asked about the degree to Miller and Fishkin found that most were inse- which they seek short-term (one-night-stand) as curely attached, especially to their fathers, A opposed to long-term (marriage) partners, the general theory of human mating must be based estimated likelihood of sexual intercourse af- on the behavior of people in general, not a ter acquaintanceships or relationships ranging minority of atypical and poorly adjusted young from 1 hr to 5 years, and the ideal number of men. sexual partners over periods of time extending Another problem is that in much of the re- from 1 month to 30 years. By all measures, search on short- and long-term mating, the two males on average were more interested in and strategies are presented in a forced-choice more willing and eager to pursue casual sexual format (for an exception, see Kenrick, Groth, liaisons. Trost, & Sadalla, 1993). This approach sets up a Arithmetic averages are sometimes informa- false dichotomy. Although rates of sexual infi- tive and sometimes misleading. L. C. Miller and delity vary somewhat across samples and his- Fishkin (1997) set out to replicate one of these torical periods, a significant number of long- studies. They asked a large group of undergrad- term maters of both sexes engage in short-term uates "Ideally, how many sex partners would extrapair copulations (Fisher, 1989). And de- you like to have in the next 30 years?" The spite the fact that males consistently report results were consistent with those reported by higher rates of infidelity than females, some previous investigators (e.g., Buss & Schmitt, members of both sexes pursue both strategies in 1993). The mean number of ideal future sexual tandem, even though the risks for females partners for women was 2; for men, it was 64. sometimes include death (Small, 1993), The findings appear to support sexual strategies This raises a related issue. In line with sexual theory in that reported male desires confirm strategies theory, one could argue that a man hypothesized sex differences in "best" repro- who has sex with 50 women could conceivably ductive strategies, with males preferring larger end up with 50 offspring, whereas a woman numbers of sex partners than females. However, who has sex with 50 men can still get pregnant another measure of central tendency paints a by only one. Theoretically, then, it never really qualitatively different picture. The median ideal pays for her to be sexually promiscuous. But, in number of future sexual partners reported by fact, it can be quite advantageous for women to women was one; for men, the number was also be sexually unfaithful to their partners. An es- one! timated 99% of sperm in each ejaculate func- The demographic characteristics of the sam- tions not to fertilize an egg but, rather, to fight ple are also informative. The typical male un- off other men's sperm (Baker, 1996). By having dergraduate is young, has yet to commit to a sex with multiple males, a female can help long-term relationship, and has access to what is ensure that her offspring are of the highest pos- arguably the largest pool of potential mates he sible genetic quality. Not coincidentally, human will ever encounter in his lifetime (almost cer- females are most likely to stray when they are tainly larger than his hominid ancestors had ovulating. access to in the EEA), and he says that what he Assume, for the sake of argument, that hu- would "ideally" like is one sexual partner over man males are "naturally" more sexually pro- the next 30 years. It is difficult to imagine an miscuous than human females. This alone empirical finding more contradictory to the pre- would not constitute proof that high male rela- dictions of sexual strategies theory. If human tive to female sociosexuality reflects an evolved males are programmed by evolution to desire sex difference in mating strategies. High levels and seek out multiple sexual partners, why is of androgens are necessary to produce sperm, 200 HAZAN AND DIAMOND and androgens are known to powerfully in- vestment in offspring is so much greater for fluence sexual desire in both sexes (Carani, females than males—years versus minutes— Granata, Fustini, & Marrama, 1996). Male sex- females should care more than males about a ual promiscuity could be simply a by-product of mate's social standing and financial prospects. hormonal activity as opposed to an adaptive, And worldwide, this appears to be the case selected-for mating strategy. (Buss, 1989). The finding that this sex differ- For both empirical and logical reasons, we ence is universal is taken as an indication that it question whether an exclusive short-term mat- is part of an evolved female mating strategy. ing strategy was ever adaptive for human males. There are, however, universal features of the It has been estimated that in contemporary hunt- EEA that support an equally plausible account er-gatherer societies—the model for specula- of this reliable cross-cultural sex difference. As tions about the EEA—the typical female is Bern (1993) has pointed out, early humans did pregnant or lactating (and therefore not ovulat- not have access to the modern conveniences of ing) during approximately 24 of the average 26 birth control or digestible substitutes for moth- years between puberty and menopause (Sy- er's milk enjoyed today. In addition, much of mons, 1979). As a result, she would be fertile on the work of daily survival depended on physical only about 80 of these 8,000 or so days. In other strength, which males (then and now) possess to words, only 1 of 100 random copulations could a greater degree than females. It is not improb- even potentially result in conception. Normally, able that all human groups would arrive at the it takes several months of unprotected sex to same sex-based division of labor: She would produce a viable pregnancy. In light of these stay home with the babies while he and his facts, it seems quite improbable mat moving friends took care of the business of hunting and from one-night stand to one-night stand would looking after the group. As human civilization have been an effective strategy for achieving evolved, males had a power and status advan- reproductive success or that such a strategy tage relative to females that continues today. would ever have been selected for. Even contemporary professional women cannot Why has so much store been placed in the sex anticipate financial rewards equal to those of difference in promiscuity? Perhaps the reason is their male counterparts. that it is so salient. What person, male or fe- If this environmental universal, rather than an male, has failed to observe that males are gen- evolved psychological mechanism, is at the erally more willing and eager than females to heart of female concern about a potential mate's have sex on a moment's notice with a total status and resources, then the degree to which stranger? It is a biological fact that men can women judge men on the basis of their social enjoy sex without concerns about becoming position and earning capacity should vary as pregnant, and it is a sociological fact that males, a function of the local financial standing of relative to females, are more often rewarded and women. In fact, it does. Eagly and Wood (1999) less often punished for their sexual exploits. reanalyzed the 37 cultures data collected by Surely these facts have some degree of signifi- Buss and found exactly this kind of covariation. cance. Whatever its origins, the sex difference It is also noteworthy that one female character- in casual sex could be what makes sexual strat- istic universally valued by males that was not egies theory so intuitively appealing and right- included in the original published report (Buss, sounding to some. In fact, it may well be the 1989) is that she be a good housekeeper and critical "hook" that leads many to swallow the cook. Was there selection pressure in the EEA entire theory, line and sinker. for female housekeeping skills (before people had houses) or culinary talents (before people Do Women "Naturally" Care More About did any cooking)? The possibility strikes us as Status and Resources? unlikely. But the findings are important for what they can reveal about human mate preferences. There is one additional issue on which sexual Some may reflect evolved tendencies and some strategies theory appears to have strong support, may simply mirror socially constructed gender and that is sex differences in the value of a roles, and the mere universality of their mention potential mate's status and resources. In theory, by research participants cannot settle disputes given the reality that the minimal requisite in- about which is which. SPECIAL ISSUE: ATTACHMENT AND MATING 201

Conclusion intimacy in turn triggers a release of hormones that boost desire for continued contact In time, A recent American Scientist article on the their neurobehavioral systems become mutually evolutionary psychology of human mating conditioned to the stimulus of the mate such that (Buss, 1994) includes a photograph of business- she or he comes to have a uniquely powerful man Donald Trump and his (then) wife Maria effect on physical and psychological well- Maples. According to the photo caption, they being. A pair bond is in place. The two are represent "species-typical mating preferences." attached. We think not. Admittedly, when mating deci- Admittedly, not all mating relationships turn sions are fully conscious and calculated, youth out this way. As many as half may dissolve and good looks can be traded for status and before an attachment bond is fully formed. Even resources. But this rule seems also to apply in so, most endure long enough to see at least one cases in which the one with resources is female. infant through the most vulnerable early years More celebrity examples, this time rich women (Fisher, 1992). with "boy toys," come to mind: Madonna, Cher, If the goal is to understand human mating, Roseanne, Jackie Collins. But they are no more one must go beyond self-reported ideals and typical of mating in our species than are Donald hypothetical scenarios to a thorough investiga- and Maria. The truly typical human being ends tion of the processes by which mating relation- up with a mate who is roughly his or her socio- ships are actually established and what such economic and physical attractiveness equal relationships are really like. There can be no (Buss, 1984). comprehensive evolutionary theory of human An alternative evolutionary model of the pro- mating without a full accounting of our species- cess might go something like this: The physio- typical mating pattern. Moreover, given the ir- logical and psychological changes that accom- refutable necessity for successful negotiation of pany puberty motivate individuals of both sexes the adaptive challenges associated with infant to begin the search for a mate. A natural aver- survival, mate acquisition, and offspring care— sion to sexual intimacy with those who are and the central place of attachment in all physically repulsive for any number of reasons three—we contend that an evolutionary theory (e.g., wrinkled skin, sagging body parts, open of human mating that does not include attach- sores, or gross asymmetries) helps ensure that ment is hollow at its core. Bowlby's grand and they avoid mating with anyone who has "bad" elegant theory complements Darwin's by pro- genes or is unlikely to be fertile. Most members viding the foundations of a new and more ac- of their local group who reached sexual matu- curate model of mating in our species. rity before them will already be paired; all re- maining postpubescent members of the desired sex who exceed their threshold of acceptability constitute the pool of eligibles. References Propinquity and familiarity further narrow Ainsworth, M. D. S., Blehar, M. C, Waters, E., & the pool. Potential mates who are encountered Wall, S. (1978). Patterns of attachment. Hillsdale, daily at the river's edge have an advantage over NJ: Erlbaum. those residing on the other side. Within this Alcock, J. (1989). Animal behavior: An evolutionary pool, they are vigilant for signs of reciprocal approach. Boston: Sinauer. interest, expressed in easily recognized flirta- Aron, A., Dutton, D. G., Aron, E. N., & Iverson, A. tion behaviors (the proceptive program; Eibl- (1989). Experiences of falling in love. Journal of Eibesfeldt, 1989). A slightly prolonged gaze, a Social and Personal Relationships, 6, 243-257. smile, or a subtle violation of personal space Baker, R. R. (1996). Sperm wars: The evolutionary may trigger romantic infatuation. If mutual, the logic of love and lust. London: Basic Books. psychological and neurochemical processes that Baker, R. R., & Bellis, M. A. (1995). Human sperm competition. London: Chapman & Hall. ensue make each person the sole focus of the Belsky, J. (1999). Modern evolutionary theory and other's attention and passion and render alter- patterns of attachment. In J. Cassidy & P. R. native potential mates less desirable. The same Shaver (Eds.), Handbook of attachment: Theory, processes stimulate a seemingly insatiable long- research, and clinical applications (pp. 141-161). ing for close physical contact. This physical New York: Guilford Press. 202 HAZAN AND DIAMOND

Bern, S. L. (1993). The lenses of gender. New Haven, Chisholm, J. S. (1996). The evolutionary ecology of CT: Yale University Press. attachment organization. Human Nature, 7, 1-38. Blaffer-Hrdy, S. (1988). The primate origins of hu- Clark, R. D., & Hatfield, E. (1989). Gender differ- man sexuality. In R. Bellig & G. Stevens (Eds.), ences in receptivity to sexual offers. Journal of The evolution of sex (pp. 101-131). San Francisco: Psychology and Human Sexuality, 2, 39-55. Harper & Row. Crittenden, P. M. (1995). Attachment and psychopa- Bloom, B. L., Asher, S. J., & White, S. W. (1978). thology. In S. Goldberg, R. Muir, & J. Kerr (Eds.), Marital disruption as a stressor: A review and Attachment theory: Social, developmental, and analysis. Psychological Bulletin, 85, 867-894. clinical perspectives (pp. 367-406). Hillsdale, NJ: Bowlby, J. (1973). Attachment and loss: Vol. 2. Sepa- Analytic Press. ration: Anxiety and anger. New York: Basic Books. Cunningham, M. R., Druen, P. B., & Barbee, A. P. Bowlby, J. (1979). The making and breaking of af- (1997). Angels, mentors, and friends: Trade-offs fectional bonds. London: Tavistock. among evolutionary, social, and individual vari- Bowlby, J. (1980). Attachment and loss: Vol. 3. Loss: ables in physical appearance. In J. A. Simpson & Sadness and depression. New York: Basic Books. D. T. Kenrick (Eds.), Evolutionary social psychol- Bowlby, J. (1982). Attachment and loss: Vol. 1. At- ogy (pp- 109-140). Mahwah, NJ: Erlbaum. tachment (2nd ed.). New York: Basic Books. Cutler, W. B., Garcia, C. R., Huggins, G. R., & Preti, (Original work published 1969) G. (1986). Sexual behavior and steroid levels Bowlby, J. (1988). A secure base: Parent-child at- among gynecologically mature premenopausal tachment and healthy human development. New women. Fertility and Sterility, 45, 496-502. York: Basic Books. Daly, M., & Wilson, M. (1988). Homicide. New Buss, D. M. (1984). Toward a psychology of person- York: Aldine de Gruyter. environment (PE) correlation: The role of spouse DeSteno, D. A., & Salovey, P. (1996). Genes, jeal- selection. Journal of Personality and Social Psy- ousy, and the replication of misspecifled models. chology, 47, 361-377. Psychological Science, 7, 376-377. Buss, D. M. (1989). Sex differences in human mate Draper, P., & Belsky, J. (1990). Personality develop- preferences: Evolutionary hypotheses tested in 37 ment in evolutionary perspective. Journal of Per- cultures. Behavioral and Brain Sciences, 12, 1-49. sonality, 58, 141-161. Draper, P., & Harpending, H. (1982). A sociobiologi- Buss, D. M. (1994). The strategies of human mating. cal perspective on the development of human re- American Scientist, 82, 238-249. productive strategies. In K. MacDonald (Ed.), So- Buss, D. M. (1997). The emergence of evolutionary ciobiological perspectives on human development social psychology. In J. A. Simpson & D. T. Ken- (pp. 340-372). New York: Springer-Verlag. rick (Eds.), Evolutionary social psychology (pp. Eagly, A. H., & Wood, W. (1999). The origins of sex 387-400). Mahwah, NJ: Erlbaum. differences in human behavior: Evolved dispositions Buss, D. M., & Barnes, M. (1986). Preferences in versus social roles. American Psychologist, 54, 408- human mate selection. Journal of Personality and 423. Social Psychology, 50, 559-570. Eberhard, W. G. (1985). Sexual selection and animal Buss, D. M., Larsen, R. J., Westen, D., & Semmel- genitalia. Cambridge, MA: Harvard University roth, J. (1992). Sex differences in jealousy: Evo- Press. lution, physiology, and psychology. Psychological Eberhard, W. G. (1991). Copulatory courtship and Science, 3, 251-255. cryptic female choice in insects. Biological Re- Buss, D. M, & Schmitt, D. P. (1993). Sexual strat- view, 66, 1-31. egies theory: An evolutionary perspective on hu- Eckland, B. K. (1968). Theories of mate selection. man mating. Psychological Review, 100, 204-232. Eugenics Quarterly, 15, 71-84. Carani, C, Granata, A. R., Fustini, M. F., & Mar- Eibl-Eibesfeldt, I. (1975). Ethology: The biology of rama, P. (1996). Prolactin and testosterone: Their behavior. New York: Holt, Rinehart & Winston. role in male sexual function. International Journal Eibl-Eibesfeldt, I. (1989). Human ethology. New ofAndrology, 19, 48-54. York: Aldine de Gruyter. Carter, C. S. (1992). Oxytocin and sexual behavior. Neu- Fisher, H. E. (1989). Evolution in human serial pair roscience and Bhbehavhral Reviews, 16, 131-144. bonding. American Journal of Physical Anthropol- Carter, C. S. (1998). Neuroendocrine perspectives on ogy, 73, 331-354. attachment and love. Psychoneuroendocrinol- Fisher, H. E. (1992). Anatomy of love. New York: ogy, 23, 779-818. Norton. Carter, C. S., DeVries, A. C, Taymans, S. E., Rob- Fraley, R. C, & Davis, K. E. (1997). Attachment erts, R. L., Williams, J. R., & Getz, L. L. (1997). formation and transfer in young adults' close Peptides, steroids, and pair bonding. Annals of the friendships and romantic relationships. Personal New York Academy of Sciences, 807, 260-272. Relationships, 4, 131-144. SPECIAL ISSUE: ATTACHMENT AND MATING 203

Fraley, R. C, & Shaver, P. R. (1999). Loss and Kenrick, D. T., Groth, G. E., Trost, M. R., & Sadalla, bereavement: Attachment theory and recent con- E, K. (1993). Integrating evolutionary and social troversies concerning "grief work" and the nature exchange perspectives on relationships: Effects of of detachment. In J. Cassidy & P. R. Shaver (Eds.), gender, self-appraisal, and involvement level on Handbook of attachment: Theory, research, and mate selection criteria. Journal of Personality and clinical applications (pp. 735-759). New York: Social Psychology, 64, 951-969. Guilford Press. Kraemer, G. W. (1997). Psychobiology of early so- Freud, A., & Dann, S. (1951). An experiment in cial attachment in rhesus monkeys. Annals of the group upbringing. In R. Eisler, A. Freud, H. Hart- New York Academy of Sciences, 807, 401-418. mann, & E. Kris (Eds.), The psychoanalytic study Lancaster, J. B., & Kaplan, H. (1994). Human mating of the child (Vol. 6). New York: International and family formation strategies: The effects of Universities Press. variability among males in quality and the alloca- Goodwin, J. S., Hurt, W. C, Key, C. R., & Sarret, J. M. tion of mating effort and parental investment. In T. (1987). The effect of marital status on stage, treat- Nishida, W. C. McGrew, P. Marler, M. Pickford, ment and survival of cancer patients. Journal of the & F. B. M. de Waal (Eds.), Topics in primatology, American Medical Association, 258, 3125-3130. vol. 1: Human origins (pp. 21-33). Tokyo: Uni- Gould, S. J., & Vrba, E. S. (1982). Exaptation—A versity of Tokyo Press. missing term in the science of form. Paleobiol- Laumann, E. O., Gagnon, J. H., Michael, R. T., & ogy, 8, 4-15. Michaels, S. (1994). The social organization of Graziano, W. G., Jensen-Campbell, L. A., Todd, M., & sexuality. Chicago: University of Chicago Press. Finch, J. F. (1997). Interpersonal attraction from an Leibowitz, M. (1983). The chemistry of love. New evolutionary psychology perspective: Women's re- York: Berkeley Books. actions to dominant and prosocial men. In J. A. Lillard, D., & Gerner, J. (in press). Getting to the Ivy Simpson & D. T. Kenrick (Eds.), Evolutionary social League: How family composition affects college psychology (pp. 141-167). Mahwah, NJ: Erlbaum. choice. Journal of Higher Education. Harlow, H., & Harlow, M. K. (1965). The affectional Lykken, D. T., & Tellegen, A. (1993). Is human systems. In A. M. Schrier, H. F. Harlow, & F. mating adventitious or the result of lawful choice? Stollnitz (Eds.), Behavior of nonhuman pritnates A twin study of mate selection. Journal of Person- (Vol. 2, pp. 287-334). New York: Academic Press. ality and Social Psychology, 65, 56-68. Hazan, C, & Shaver, P. R. (1990). Love and work: Lynch, J. J. (1977). The broken heart: The medical An attachment theoretical perspective. Journal of consequences of loneliness. New York: Basic Books. Personality and Social Psychology, 59, 270-280. Mellen, S. L. W. (1981). The evolution of love. Ox- Hazan, C, & Shaver, P. R. (1992). Broken attach- ford, England: Freeman. ments. In T. L. Orbuch (Ed.), Close relationship Mendoza, S. P., & Mason, W. A. (1997). Attachment loss: Theoretical approaches (pp. 90-108). Hills- relationships in new world primates. Annals of the dale, NJ: Erlbaum. New York Academy of Sciences, 807, 203-209. Hazan, C, & Shaver, P. R. (1994a). Attachment as an Miller, G. F. (1998). How mate choice shaped human organizational framework for research on close nature: A review of sexual selection and human evo- relationships. Psychological Inquiry, 5, 1-22. lution. In C. Crawford & D. L. Krebs (Eds.), Hand- Hazan, C, & Shaver, P. R. (1994b). Deeper into attach- book of evolutionary psychology: Ideas, issues, and ment theory. Psychological Inquiry, 5, 68-79. applications (pp. 87-129). Mahwah, NJ: Erlbaum. Hazan, C, & Zeifman, D. (1994). Sex and the psy- Miller, L. C, & Fishkin, S. A. (1997). On the dy- chological tether. Advances in Personal Relation- namics of human bonding and reproductive suc- ships, 5, 151-177. cess: Seeking windows on the adapted-for human- Hazan, C, & Zeifman, D. (1999). Pair bonds as environment interface. In J. A. Simpson & D. T. attachments: Evaluating the evidence. In J. Kenrick (Eds.), Evolutionary social psychology Cassidy & P. R. Shaver (Eds.)T Handbook of at- (pp. 197-235). Mahwah, NJ: Erlbaum. tachment: Theory, research, and clinical applica- Murray, S. L., & Holmes, J. G. (1993). Seeing virtues in tions (pp. 336-354). New York: Guilford Press. faults: Negativity and the transformation of interper- Hill, K., & Hurtado, M. (1995). Demographic/life his- sonal narratives in close relationships. Journal of tory of Ache foragers. New York: Aldine de Gruyter. Personality and Social Psychology, 65, 707-722. Holmes, T. H., & Rahe, R. H. (1967). The Social Murray, S. L., & Holmes, J. G. (1994). Story-telling Readjustment Rating Scale. Journal of Psychoso- in close relationships: The construction of confi- matic Research, II, 213-218. dence. Personality and Social Psychology Bulle- Insel, T. R. (2000). Toward a neurobiology of attach- tin, 20, 663-676. ment. Review of General Psychology, 4, 176-185. Murray, S. L., Holmes, J. G., & Griffin, D. W. Kagan, J. (1984). The nature of the child. New York: (1996). The benefits of positive illusions: Idealiza- Basic Books. tion and the construction of satisfaction in close 204 HAZAN AND DIAMOND

relationships. Journal of Personality and Social Spitz, R. A. (1946). Anaclitic depression. Psychoan- Psychology, 70, 79-98. alytic Study of the Child, 2, 313-342. Murstein, B. I. (1976). Who will marry whom? The- Stearns, S. (1992). The evolution of life histories. ories and research in marital choice. New York: New York: Oxford University Press. Springer. Suomi, S. J. (1997). Early determinants of behaviour: Parkes, C. M., & Weiss, R. S. (1983). Recovery from Evidence from primate studies. British Medical bereavement. New York: Basic Books. Bulletin, 53, 170-184. Pietromonaco, P. R., & Feldman Barrett, L. (2000). Surbey, M. (1990). Family composition, stress, and The internal working models concept: What do we human menarche. In F. Bercovitch & T. Zeigler really know about the self in relation to others? (Eds.), The socioendocrinology of primate repro- Review of General Psychology, 4, 155-175. duction (pp. 71-97). New York: Liss. Robertson, J. (1953). Some responses of young chil- Symons, D. (1979). The evolution of human sexual- dren to the loss of maternal care. Nursing ity. New York: Oxford University Press. Times, 49, 382-386. Tennov, D. (1979). Love and limerence: The experi- Rowland, D. L., Greenleaf, W. J., Dorrman, L. J., & ence of being in love. New York: Stein & Day. Davidson, J. M. (1993). Aging and sexual function Trevathan, W. (1987). Human birth. New York: Al- in men. Archives of Sexual Behavior, 22, 545-557. dine de Gruyter. Rubin, Z. (1973). Liking and loving. New York: Holt, Trivers, R. L. (1972). Parental investment and sexual Rinehart & Winston. selection. In B. Campbell (Ed.), Sexual selection Rusbult, C. E. (1980). A longitudinal test of the investment model: The development (and deterio- and the descent of man (pp. 1871-1971). Chicago: ration) of satisfaction and commitment in hetero- Aldine. sexual involvements. Journal of Personality and Uchino, B. N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. Social Psychology, 45, 101-117. (1996). The relationship between social support and Rusbult, C. E. (1983). Commitment and satisfaction physiological processes: A review with emphasis on in romantic associations: A test of the investment underlying mechanisms and implications for health. model. Journal of Experimental Social Psychol- Psychological Bulletin, 119, 488-531. ogy, 16, 172-186. Van den Berghe, P. L. (1979). Human family sys- Sapolsky, R. M. (1998, July-August). Sex and the tems: An evolutionary view. Westpott, CT: Green- single monkey. The Sciences, pp. 10-11. wood Press. Schwartz, D., Mayauz, M. J., Spira, A., Moscato, J. L., Veith, J. L., Buck, M., Getzlaf, S., Van Dalfsen, P., & Jouannet, P., DzygJik, F., & David, G. (1983). Semen Slade, S. (1983). Exposure to men influences the characteristics as a function of age in 833 fertile men. occurrence of ovulation in women. Physiology and Fertility and Sterility, 39, 530-535. Behavior, 31, 313-315. Shaver, P. R., & Hazan, C. (1993). Adult romantic Vormbrock, J. K. (1993). Attachment theory as ap- attachment: Theory and empirical evidence. In D. plied to war-time and job-related marital separa- Perlman & W. Jones (Eds.), Advances in personal tion. Psychological Bulletin, 114, 122-144. relationships (Vol. 4, pp. 29-70). Greenwich, CT: Wallerstein, J. S. (1994). Children after divorce: JAI Press. Wounds that don't heal. In L. Fenson & J. Fenson Shaver, P. R., Hazan, C, & Bradshaw, D. (1988). (Eds.), Human development 94/95 (pp. 160-165). Love as attachment: The integration of three be- Guilford, CT: Dushkin. havioral systems. In R. J. Sternberg & M. L. Bar- Washbum, S. L. (1960). Tools and human evolution. nes (Eds.), The psychology of love (pp. 68-99). Scientific American, 203, 3-15. New Haven, CT: Yale University Press. Weiss, R. S. (1975). Marital separation. New York: Simpson, J. A., & Gangestad, S. W. (1992). Socio- Basic Books. sexuality and romantic partner choice. Journal of Zeifinan, D., & Hazan, C. (1997). Attachment: The Personality, 60, 31-52. bond in pair-bonds. In J. A. Simpson & D. T. Simpson, J. A., Gangestad, S. W., & Lerma, M. Kenrick (Eds.), Evolutionary social psychology (1990). Perception of physical attractiveness: (pp. 237-263). Mahwah, NJ: Erlbaum. Mechanisms involved in the maintenance of ro- Zimmerman, S. J., Maude, M. B., & Moldawar, M. mantic relationships. Journal of Personality and (1965). Frequent ejaculation and total sperm count, Social Psychology, 59, 1192-1201. motility and forms in humans. Fertility and Steril- Small, M. F. (1993), Female choices; Sexual behav- ior of female primates. Ithaca, NY: Cornell Uni- ity, 16, 342-345. versity Press. Small, M. F. (1995). What's love got to do with it? Received March 18, 1999 The evolution of human mating. New York: An- Revision received September 8, 1999 chor Books. Accepted September 8, 1999 •