Attachment as an Organizational Framework for Research on Close Relationships

Cindy Hazan; Phillip R. Shaver

Psychological Inquiry, Vol. 5, No. 1. (1994), pp. 1-22.

Stable URL: http://links.jstor.org/sici?sici=1047-840X%281994%295%3A1%3C1%3AAAAOFF%3E2.0.CO%3B2-S

Psychological Inquiry is currently published by Lawrence Erlbaum Associates, Inc..

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/leb.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is an independent not-for-profit organization dedicated to and preserving a digital archive of scholarly journals. For more information regarding JSTOR, please contact [email protected].

http://www.jstor.org Wed Feb 14 14:09:04 2007 Psychological Inquiry Copyright 1994 by 1994, Vol. 5, No. 1, 1-22 Lawrence Erlbaum Associates, Inc.

TARGET ARTICLE

Attachment as an Organizational Framework for Research on Close Relationships

Cindy Hazan Cornell University Phillip R. Shaver University of California, Davis

Recent years have witnessed a proliferation of research on close relationships and the emergence ofa new relationship subdiscipline within the social sciences. To date, the new science ofrelationships has been dominated by data. This article is based on the conviction that progress now hinges on the development oftheory to organize and interpret extant findings and to guide future investigations. Through a selective but extensive review of the major bodies of empirical literature, we attempt to show that attachment theory can incorporate a broad range offindings on adult relationships. In addition, attachment theory addresses an impressive array of research questions concerning the functions, emotional dynamics, evolutionary origins, and developmen­ tal pathways ofhuman affectional bonds. We conclude that a comprehensive theory of close relationships is both desirable and, with the integration of existing theories and concepts, currently achievable.

In 1958, Harry Harlow wrote: (Bowlby, 1973, 1980, 1969/1982, 1988) to propose a generative theoretical framework for organizing extant Our assigned mission as psychologists is to analyze all data and for making predictions beyond what can be facets of human and animal behavior into their com­ currently documented. What we seek is a comprehens­ ponent variables. So far as love or affection is con­ ive theory of close relationships, and what we propose cerned, psychologists have failed in their mission. The here is the outline of such a theory. little we know about love does not transcend simple observation and the little we write about it has been In searching for a theory of relationships, we are led written better by poets and novelists. (Harlow, 1958, to ask what, ideally, one would want from such a p. 673) theory. What questions should it answer? We believe that, at a minimum, it should address the following: Today, more than 30 years later, there is an inter­ What makes a potential relationship partner appealing? national network of researchers whose investigative How is a relationship formed, and how does it develop? efforts are devoted entirely to the study of personal How are relationships maintained, and what makes relationships. Have we made any progress in under­ them satisfying or enduring? Why and how are relation­ standing love and affection? Most definitely, we ships dissolved? What are the reactions to relationship have. breakup? Many have argued that what is needed in the field of Beyond these basics, a comprehensive theory of personal relationships is what every new science re­ relationships should offer models of both normative quires, a descriptive base of knowledge from which to and individual-difference phenomena and account derive principles and construct theory (e.g., Hinde, for the role of relationships in a person's overall 1979; Rubin, 1984). This article is based on the convic­ adaptation and functioning throughout life. It should tion that, after more than three decades of research on be able to explain the universal human tendency to close relationships, there is ample data to justify a form close relationships and the similarities and dif­ search for meaningful patterns and organizing princi­ ferences in affectional bonds from infancy through ples. In what follows, we draw on attachment theory adulthood. Perhaps most important, a theory of close HAZAN & SHAVER relationships must be consistent and compatible with tation followed his publication (Bowlby, 1944) of an existing empirical findings. And, as with any theory article entitled "Forty-Four Juvenile Thieves: Their that earns its keep, it must be parsimonious, testable, Characters and Home Life," which reported a strong and generative. association between early maternal separations and The challenge awaiting a comprehensive theory of subsequent delinquency among boys. The WHO report close relationships is formidable. Consider the diver­ (Bowlby, 1951) asserted that maternal deprivation, es­ sity of findings and phenomena to be incorporated: pecially during the first 3 years of life, puts children at Frequency of mutual gazing is a good indicator of the increased risk for physical and mental illness. Although amount of love that partners feel for each other influential and well received, the report was deficient (Rubin, 1973). We tend to fall in love with people in one important respect: It failed to explain why or how who seem especially responsive to our needs (A. P. early maternal deprivation has such deleterious effects. Aron, Dutton, E. N. Aron, & Iverson, 1989; Bowlby had been trained in the psychoanalytic Berscheid, 1984). Inadequate care during infancy is tradition but, almost as soon as he began to practice predictive of later troubled relations with peers child therapy, he found himself troubled by what he (Arend, Gove, & Sroufe, 1979). Children who are perceived to be inadequacies in psychoanalytic the­ anxiety prone are more inclined to develop romantic ory. Most objectionable was its exclusive focus on crushes (Hatfield & Rapson, 1987). Within the con­ fantasy and the internal life; a child's real-life expe­ text of laboratory marital interaction, physiological riences were of little interest relative to intrapsychic arousal predicts the eventual demise of the marriage events. Bowlby was also troubled by inconsistencies (Levenson & Gottman, 1985). Fear of intimacy is between psychoanalytic theory and his own observa­ associated with workaholism (Hazan & Shaver, tions. That institutionalized children suffered ex­ 1990). Adolescents who idealize their parents are treme distress and even sometimes failed to thrive judged as hostile by their peers (Kobak & Sceery, despite being fed and cared for by staff did not follow 1988). The typical initial reaction to the ending of a from psychoanalytic notions that children love their close relationship is anxiety (Weiss, 1988). Obsta­ mother simply because they associate her with the cles tend to enhance romantic passion (Driscoll, K. satisfaction of a hunger drive. Bowlby's growing W. Davis, & Lipetz, 1972). Parental divorce during dissatisfaction with psychoanalytic theory helped childhood is associated with chronic loneliness in launch a search for answers to questions raised in his adulthood (Shaver & Rubenstein, 1980). People tend WHO report. to overlook the faults and limitations of a new partner The search took him through the literatures of several (Tennov, 1979). Relationship satisfaction typically disciplines, but it was in ethology that he found what declines in the early years (Spanier, Lewis, & Cole, he believed to be an important part of the answer. 1975). And so on. Research on the bonding behavior of birds and mam­ Is it possible to detect laws of relationship structure mals was particularly influential in his thinking, grad­ and process that lend coherence to these and other ually leading to the insight that maternal deprivation is relationship facts? We believe that a parsimonious ex­ developmentally harmful because it thwarts the satis­ planation of much of the existing data is both desirable faction of an inborn need. Further, the work of etholo­ and possible, and we think attachment theory is a good gists like Lorenz and Tinbergen provided dramatic place to start. We are not ready to claim that, in its examples of how even inborn, instinctual tendencies current form, attachment theory tells us all we would can become distorted or fail to develop in non-optimal ever want or need to know in order to understand close environments. Almost 20 years later, Bowlby pub­ relationships. Some shortcomings of attachment theory lished the first of three major volumes on attachment can be handled nicely by interdependence theory theory. For obvious reasons, we cannot review all the (Kelley et al., 1983; Thibaut & Kelley, 1959) and supporting evidence on which the theory is based; we theories that stress the mutual construction of relation­ will settle instead for an explication of the theory itself. ship narratives (e.g., Duck, 1991). At the end of this (For a more extensive review of the evidence, see article, we say more about how we see the alternative Bowlby, 1973, 1980, 1969/1982.) theories fitting together.

The Attachment System Attachment Theory A basic assumption of attachment theory is that, In 1950, John Bowlby was invited by the World because of their extreme immaturity at birth, human Health Organization (WHO) to report on the mental infants can survive only if an adult is willing to provide health of London's many homeless children. The invi- protection and care. As a result of selection pressures,

2 ATIACHMENT AND CLOSE RELATIONSIDPS infants evolve behaviors that function to maintain prox­ affiliation, and sexual mating. Each system serves imity to a protector/caregiver. Adult caregiving is reg­ unique functions and responds to different environmen­ ulated by a complementary behavioral system. Babies tal cues. At least during infancy and childhood, attach­ smile, and parents find the smiles rewarding. Babies ment is the preeminent system, and its full activation cry, and parents are motivated to soothe them. Parents precludes activation of other systems. As long as the move away, and babies follow visually or physically. child experiences "felt security" (Sroufe & Waters, These two systems conspire to create the kind of rela­ 1977), the attachment system is quiescent, and other tionship that fosters the infant's survival. behavioral systems become available for activation, A behavioral system consists of a set of behaviors although periodic checking of the attachment figure's that serve the same function, although they may be availability continues to occur. In evolutionary terms, morphologically dissimilar (e.g., crying, smiling, it has been adaptive for human young to feel safe following). The attachment system is similar in some enough to engage in play and exploration only as long respects to the physiological systems that regulate as a famili:tr protector is available to respond if needed. body temperature, blood pressure, and the like. Any The safest reaction to any threat or strong uncertainty real or perceived obstacle to proximity maintenance is to devote all energy and attention to reestablishing results in anxiety, which in turn triggers attachment proximity. (See Figure 1 for a model of the attachment behaviors designed to reestablish proximity. Such system.) behaviors persist until the "set goal" for proximity has been achieved. The degree of proximity required to keep anxiety at bay is related to a variety of Attachment Formation endogenous and exogenous factors, including the child's age, emotional and physical state, and per­ Children could conceivably direct their attachment ceived environmental threat. The establishment and behaviors to any available person. In reality, however, maintenance of proximity engender feelings of secu­ by the sixth or seventh month of life, all normal infants rity and love, whereas disruptions in the relationship selectively direct these behaviors to one person, with typically beget anxiety and sometimes anger or sad­ whom they also seek proximity and from whom they ness (depending on particular appraisals). Hence, object to being separated (Schaffer & Emerson, 1964). Bowlby argued that an attachment is an emotional (See also Mizukami, Kobayashi, Ishii, & Iwata, 1990, bond. for evidence of selective attachment by 2 to 4 months Attachment is one of several distinct but interlocking of age.) How is this person "selected?" Of definite behavioral systems, including exploration, caregiving, significance to infants is who usually responds to their

playful, less inhibited, smiling, exploration­ oriented, sociable

Hierarchy of attach­ ment behaviors: maintenance of 1. visual checking proximity while 2. signaling to avoiding close reestablish contact, contact calling, pleading 3. moving to reestab­ lish contact, clinging

Figure 1. The attachment behavioral system. In the diamond is the test question (G. A. Miller, Galanter, & Pribram, 1960). The circles represent emotions triggered by the appraisals that answer the test question, and the boxes represent behaviors that follow from the appraisals and emotions.

3 HAZAN & SHAVER signals of distress. Also of significance is the qUJJlity of tionship and verbally communicate about and coordi­ the response. Thus, familiarity and responsiveness dic­ nate their respective goals. tate preferences and influence the selection of an at­ tachment figure. The functions of attachment are apparent in the Attachment Disruption and Dissolution infant's behavior. Proximity to the attachment figure is especially likely to be sought when the infant is fearful Bowlby's investigations of attachment stemmed or distressed for any reason. The caregiver serves as a from his interest in the effects of maternal deprivation. haven of safety to which the infant can retreat for He observed infants and young children who were comfort and reassurance during such times. In addition, being housed in residential nurseries, separated from the caregiver serves as a base from which to engage in their familiar caregivers for extended periods of time­ nonattachment behaviors, such as exploration. Accord­ in some cases, forever. Bowlby was struck by two ing to Bowlby,proximity maintenance (including prox­ aspects of the children's responses. First, there was a imity seeking and separation protest), safe haven, and remarkable degree of similarity across children in the secure base are the three defining features of attach­ way they responded to the separations, gradually re­ ment and the functions of an attachment relationship vealing a predictable and invariant sequence of emo­ (see Figure 2). tional reactions. The first was protest, which involves The process of attachment formation takes an aver­ crying, active searching, and resistance to others' age of 2 or 3 years (Bowlby, 1979, 1969/1982). The soothing efforts. This is followed by despair, charac­ endpoint of the process is referred to as a goal-cor­ terized by passivity and obvious sadness. The third and rected partnership, in which the goal of proximity final phase is emotional detachment. maintenance is adjusted for the child's ability to The second striking aspect of the children's re­ delay gratification and to mentally represent the sponses was that even short-term separations seemed caregiver's availability. At this point, caregiver and to have prolonged effects. Children who were re­ child can begin to negotiate the terms of their rela- united with their caregivers while still in the protest

Proximity Maintenance Staying near and resisting separations from

Attachment

Figure 2. The defining features ofattachment.

4 ATIACHMENT AND CLOSE RELATIONSlllPS phase exhibited heightened anxiety over abandonment are determined in part by the responsiveness of the and an excessive need for physical contact and reassur­ caregiving environment to individual needs for comfort ance. This insecurity continued, in some cases, for and security (Cassidy, 1988). According to Bowlby, months after the separation had ended. Those re­ these models guide thoughts, feelings, and behavior in united after passing through the phase of despair subsequent close relationships. tended initially to avoid contact with their moth­ ers, as if they had coped with the separation by emotionally detaching. However, in time they re­ Individual Differences sumed seeking contact and comfort. According to Bowlby, reactions of anxiety and pro­ Theoretically and logically speaking, there is no test, even detachment, are highly adaptive responses to limit to the amount and kind of variability that could separation from one's primary protector. A child ex­ exist in models of the caregiving environment. In presses distress because it usually brings the caregiver reality, however, infants parse the flow of informa­ around. If, however, there appears to be no hope of tion about caregiver behaviors into a limited number reestablishing proximity, continued expressions of dis­ of categories corresponding to responses to the fol­ tress not only risk attracting the attention of predators lowing question: "Can I count on my attachment (a very real threat in earlier periods of human evolution) figure to be available and responsive when needed?" but also physically exhaust the child. The characteristic There are three possible answers to this question: inactivity ofthe despair phase keeps the child quiet and yes, no, and maybe. That is, as concerns the internal still, allowing for recuperation. Detachment makes pos­ working model, a caregiver is consistently respon­ sible the resumption of normal activity, possibly even sive, consistently unresponsive, or inconsistent. In the search for a new attachment figure. Just as with fact, these three types of caregiver responsiveness routine proximity maintenance, reactions to prolonged have been empirically linked to three major patterns separation reflect the functioning of the attachment of infant-caregiver attachment (Ainsworth, Blehar, system. Waters, & Wall, 1978). The procedure that Ainsworth developed for as­ Internal Working Models sessing attachment quality-the Strange Situation­ was designed to activate an infant's attachment The attachment system is an organism-level system system through repeated separations from the car­ that is organized and regulated by social input, specif­ egiver in an unfamiliar environment. It was also ically by primary caregiver responsiveness to distress meant to activate the exploration system through the signals. On the basis of repeated interactions with the availability of attractive toys. Ainsworth was espe­ caregiver, infants learn what to expect, and they adjust cially interested in whether and when infants sought their behavior accordingly. These expectations form proximity and contact, to what degree they accepted the basis of mental representations (or, to use Bowlby's and were comforted by such contact, and whether term, internal working models) that can be used to their exploratory behavior was facilitated by the forecast caregiver availability and responsiveness and caregiver's presence. In other words, she was inter­ that include interrelated models of self and attachment ested in observing proximity-maintenance, safe­ figure. haven, and secure-base behaviors. Infant behavior in the laboratory setting was assumed to reflect expec­ Confidence that an attachment figure is, apart from tations (internal working models) based on the being accessible, likely to be responsive can be seen to caregiver's past responsiveness to the infant's bids turn on at least two variables: (a) whether or not the for contact and comfort. The models themselves attachment figure is judged to be the sort of person who were assumed to be founded on a history of actual in general responds to calls for support and protection; [and] (b) whether or not the self is judged to be the sort interactions. Home observations confirmed the link of person towards whom anyone, and the attachment between daily caregiver responsiveness and infant figure in particular, is likely to respond in a helpful laboratory behavior. The following paragraphs, way. Logically these variables are independent. In based on passages in Ainsworth et al. (1978), de­ practice they are apt to be confounded. As a result, the scribe the three major patterns. model of the attachment figure and the model of the self are likely to develop so as to be complementary and mutually confirming. (Bowlby, 1973, p. 238) Secure. The behavior of securely attached infants matched Bowlby's conception of Attachment theory thus implies that beliefs and feelings nature's prototype, in terms of proximity main­ about the self, especially social and global self-esteem, tenance, comfort seeking, and the ability to use

5 HAZAN & SHAVER

the caregiver as a secure base for exploration. In caregivers, and kept their attention directed to­ the laboratory, the typical securely attached in­ ward the toys (although with less apparent inter­ fant was distressed when the mother left the est and enthusiasm than the securely attached room, was comforted by her return, and engaged infants). On average, about 25% of American in active exploration as long as she was present. infants are classified as avoidantly attached During home observations made before the lab­ (Campos et al., 1983). oratory visits, caregivers were judged to be con­ sistently available and responsive. This is the Note that the effects of psychological availability most commonly observed pattern, averaging of the caregiver are remarkably similar to the effects about 60% in American samples (Campos, Bar­ observed for physical availability. Inconsistent re­ rett, Lamb, Goldsmith, & Stenberg, 1983). sponsiveness is functionally equivalent to short-term Anxious/Ambivalent. The typical caregiver of separations, and the corresponding attachment pat­ an anxious/ambivalently attached infant, ob­ tern-anxious/ambivalence-is characterized by the served in the home, exhibited inconsistent re­ same protest behaviors. Consistent caregiver unre­ sponsiveness to the infant's signals, being sponsiveness, like long-term absence, results in sometimes unavailable or unresponsive and at avoidance and apparent emotional detachment. other times intrusive. In the laboratory, anx­ Again, the behaviors are traceable to the organization ious/ambivalently attached infants appeared both and functioning of the attachment behavioral system anxious and angry and were preoccupied with (see Figure 3). their caregivers to such a degree that it precluded In recent years, researchers have identified a exploration. This is the most uncommon pattern, fourth pattern, disorganized/disoriented attachment averaging about 15% in American samples (Main & Solomon, 1990), that is distinguishable by (Campos et al., 1983). the absence of a coherent strategy for managing Anxious/Avoidant. At home, caregivers of anxiety and that is manifested in a mixture of avoid­ avoidantly attached infants consistently rebuffed ant and ambivalent behaviors. Research suggests that or deflected their infants' bids for comfort, espe­ this pattern arises in infancy, when the infant's pri­ cially for close bodily contact. In the laboratory mary caregiver is depressed, disturbed, or abusive in setting, these infants appeared not to be dis­ some way (e.g., Crittenden, 1988; Main & Hesse, tressed by separations, avoided contact with their 1990).

Secure Type

Anxious/Ambivalent --:-~ (or Preoccupied) playful, 1 Type less inhibited, I smiling, ).- exploration- 1 oriented, I sociable I ------~ L------~ ~-- l'L------, 1 Hierarchy of attach- I maintenance of ment behaviors: I I proximity while 1 1. visual checking I avoiding close 2. signaling to I contact, defensive reestablish contact, I exploration I calling, pleading I 1 3. moving to reestab­ ____ _j lish contact, clinging I I L ______j Avo1dant.t Typc

I<'igure 3. Three major patterns ofattachment as they correspond to various aspects ofnormative attachment-system dynamics.

6 ATIACHMENT AND CLOSE RELATIONSIDPS

Source and Stability oflndividual models, change can be difficult. It seems more sensible Differences to ask not whether working models are stable or unsta­ ble but under what conditions they are most and least Perhaps the two most controversial claims of attach­ likely to change. Bowlby has suggested several possi­ ment theory are that caregiver responsiveness largely ble routes to change, including (a) the capacity to think determines the quality of the attachment relationship about and reflect upon one's own working models and and that working models of attachment tend to be (b) "corrective" relationship experience. Both come stable. Many writers have criticized attachment theory into play in good therapeutic relationships (Bowlby, for its apparent failure to acknowledge the importance 1988). In light of what is known about the dynamics of of infant characteristics, especially temperament (e.g., the attachment system and the inborn need for security, Campos et al., 1983; Lamb, Thompson, Gardner, change would seem more likely to be in the direction Charnov, & Estes, 1984), and the theory's seemingly of secure than insecure attachment. The attachment deterministic view of development, which some see as system's "primary strategy" (Main, 1990) should al­ allowing little room for growth and change. The debate ways be to seek security if security is perceived as over temperament is by no means settled, but there is possible. Consistent with this, secure attachments have mounting evidence that both temperament and care­ been found to be more stable (Egeland & Farber, 1984). giver responsiveness are important influences on at­ Further, moving from an avoidant to a secure model of tachment quality. For example, one temperamental relationships would involve acknowledging long-re­ characteristi~istress proneness--has been linked to pressed insecurities and, as such, may necessitate an anxious/ambivalent attachment (Goldsmith & Al­ intermediate phase of anxiety and ambivalence. In fact, ansky, 1987). On the other hand, responsiveness train­ Heinicke and Westheimer (1966) observed that chil­ ing for the caregivers of distress-prone infants appears dren reunited with caregivers after prolonged separa­ to override the risks for later insecure attachment (van tions initially exhibited avoidance behavior, which was den Boom, 1990). At least one study (Crockenberg, followed by a period of dinginess and preoccupation 1981) has suggested that the relationship between tem­ (i.e., anxious/ambivalence) before normal (presepara­ perament and attachment may be mediated by maternal tion) behavior was eventually resumed. social support. To date, no investigation has shown Attachment patterns set in infancy need not be temperament to be better than caregiver responsiveness fixed for life. As with cognitive structures in general, at predicting attachment classification, and no reported internal working models of attachment face the sta­ findings would lead to the conclusion that the consis­ bility-plasticity dilemma (Grossberg, 1980). Mental tency and quality of caregiver responsiveness are not models strive for stability but have to remain plastic important determinants of infant attachment behavior. if they are to continue to be adapative and useful. To (See Colin, 1991, for a comprehensive review.) date, longitudinal studies have indicated significant In some ways, the debate over whether individual but not perfect continuity of attachment patterns over differences are stable is even murkier. Research subse­ the first several months and years of life (Cassidy, quent to Ainsworth's identification of the three patterns 1988; Owen, Easterbrooks, Chase-Lansdale, & (e.g., K. E. Grossmann & K. Grossmann, 1991; Main, Goldberg, 1984; Waters, 1978). Ultimately, the de­ Kaplan, & Cassidy, 1985; Sroufe, 1983; Waters, 1978) gree of continuity from infancy through adulthood has indicated that the patterns are generally stable over and the circumstances under which change is facili­ the first several years of life if family conditions are tated must be determined empirically. stable but can change if a child's social circumstances change (Egeland & Farber, 1984; Vaughn, Egeland, Sroufe, & Waters, 1979). Attachment theory does not Attachment Beyond Infancy dictate absolute stability of individual differences in­ duced during infancy. Nevertheless, as with any cogni­ Attachment is an integral part of human behavior tive construction, internal working models are resistant "from the cradle to the grave" (Bowlby, 1979). The to change, in part because they tend to be overlearned functions and dynamics of the attachment behavioral and operate out of awareness, and in part because the system are hypothesized to be virtually the same across default strategy for processing incoming information is the life span. Presumably, this is because the neural to assimilate it to existing schemes rather than modify foundation of the attachment system remains largely the schemes to accommodate the information (Fiske & unchanged. As Kanner (1982) put it, "The evolution of Taylor, 1991; Piaget, 1952). the brain would have to be considered unparsimonious When defensive emotional processes are intertwined if it were not able to draw upon the same basic capaci­ with overlearned cognitive and behavioral patterns, as ties of emotion and action in the various settings where is hypothesized to be the case with insecure-attachment strong attachment is called for" (p. 298).

7 HAZAN & SHAVER

Despite some basic similarities, adult attachment Related to this are differences in what motivates differs from infant attachment in important ways proximity seeking. Anxiety and distress appear to be (Weiss, 1982). First, childhood attachments are typi­ primary motivators in people of all ages. However, cally complementary. An attachment figure provides adult proximity-seeking can also result from a desire to but does not receive care; an infant or child seeks but protect or offer comfort (caregiving) or to engage in does not normally provide security. In contrast, adult sexual activity (sexual mating). Attachments are hy­ attachment relationships are typically reciprocal, with pothesized to form in the context of physical closeness, each partner being both a provider and a recipient of but the forces promoting such closeness may change care. In addition, the attachment relationship moves with development. from the level of external, observable interactions to If adult peers begin to serve similar functions and internally represented beliefs and expectations (Main satisfy the same needs for emotional support and et al., 1985). Whereas infants and young children may security for which parents are primarily responsible require physical contact with an attachment figure to during infancy and childhood, then at some point feel completely secure, older children and adults are attachment will be transferred from parents to peers. often able to derive comfort from the mere knowledge The timing ofthis transfer and the processes involved that their attachment figures can be contacted if needed are not specified within the theory. However, from (although the need for physical-contact comfort prob­ both the theory and the empirical literature on par­ ably never disappears entirely). What matters is "felt ent-child and peer relations (e.g., Furman & security" (Sroufe & Waters, 1977), and adults have Buhmester, 1985; Gottman, 1983; Hartup, 1983; G. more options for achieving it than do infants. Levinger & A. C. Levinger, 1986; Rubin, 1980; Another difference is that a child's primary attach­ Steinberg & Silverberg, 1986), a process model can ment figure is usually a parent, whereas an adult's be derived. The model (Hazan, Hutt, Sturgeon, & primary attachment figure is most commonly a peer, Bricker, 1991) is based on the assumption that, rather usually a sexual partner. Prototypical adult attachment than being shifted in concert, all attachment func­ relationships thus involve the integration of three be­ tions are gradually transferred, one by one. Accord­ havioral systems-attachment, caregiving, and sexual ing to this model, attachments begin with proximity mating (Shaver, Hazan, & Bradshaw, 1988; Weiss, seeking (motivated by security needs when directed 1982). (See Figure 4.) toward parents and by exploratory and affiliative

Prototypical Pair Bond

Figure 4. The components of a prototypical pair bond.

8 AITACHMENT AND CLOSE REIATIONSlllPS

TARGET OFAITACHMENT BEHAVIORS

DEVELOPMENTAL PHASE Parents Peers Infancy proximity maintenance safe haven secure base

Early Childhood proximity maintenance safe haven secure base

Late Childhood/ proximity maintenance Early Adolescence safe haven secure base

Adulthood proximity maintenance safe haven secure base

FigureS. A model of attachment transfer processes. In the course of normal development, parents are gradually relinquished as primary attachment figures as attachment behaviors are, one by one, redirected to peers. needs when directed toward peers). In late childhood An Attachment Perspective on and early adolescence, close proximity provides the Relationship Data context that eventually fosters support-seeking (i.e., safe-haven) behavior. Repeated interactions in which Having outlined attachment theory, we now briefly comfort is sought and provided or distress is expressed review some of the major bodies of data on close and alleviated may lead to reliance on the responder as relationships in order to show how each could be or­ a base of security. Parents are never completely relin­ ganized and explained by the theory. We have tried to quished as attachment figures, but their place in a preserve the fundamental theoretical constructs and hierarchy of attachment figures, relative to that of the processes, but we have not hesitated to extend or mold place of peers, naturally changes by adulthood. The the theory wherever warranted. In our opinion, the real model, still being developed and tested, is presented in strength of attachment theory lies in its ability to ex­ Figure 5. plain findings derived from other approaches and other theories in terms of the same limited number of proxi­ mal processes and evolved tendencies as it uses to Summary of Attachment Theory explain infant-caregiver bonds. The degree to which attachment theory and we succeed in this endeavor As a result of selection pressures over the course of reflects on the generative power of the theory. It is evolution, humans are naturally predisposed to form conceivable that each new finding in the field of per­ close relationships. Further, certain basic needs, of sonal relationships would necessitate the addition of a which the need for security is the most fundamental, are new theoretical construct. However, we attempt to best satisfied within social relationships. Behavioral show that attachment theory is able to incorporate a vast systems have developed to promote the satisfaction of range of such findings without significant alteration or all needs that have been important for survival and addition. And, when additions are required, they follow reproduction. The dynamics of close relationships­ naturally from the principles established for relation­ their formation, maintenance, and dissolution--can be ship functioning during infancy. understood in terms of the functioning of these systems. Our focus here is on close relationships of the attach­ Individual differences in thoughts, feelings, and behav­ ment variety-which adults prototypically form with a iors in close relationships are largely determined by the romantic or sexual partner-but much of what we say social environments to which individuals have had to is also applicable to other types of close relationships. adapt. Finally, these individual differences are main­ Our review of the literature is necessarily selective, but tained by mental models constructed out of actual rela­ it is extensive enough to illustrate both the diversity of tionship experience. available findings and the integrative capacity of at-

9 HAZAN & SHAVER tachment theory. The review is organized around what promoting potential more salient. Within attachment traditionally have been considered fundamental ques­ theory, anxiety is a signal to get closer. tions in the study of close relationships. However, attachment is not the only behavioral sys­ tem motivating interpersonal attraction. As stated ear­ lier, adult Jove can be conceptualized as a joint function What Makes a Potential Relationship of the attachment, caregiving, and sexual mating sys­ Partner Appealing? tems (Shaver et al., 1988), so, theoretically, attraction can result from one person's seeing the possibility of From an attachment perspective, humans possess another person's meeting attachment, caregiving, or basic needs that are naturally satisfied by social rela­ sexual needs. Another's attactiveness is thus deter­ tionships, such as the needs for emotional support, care, mined by the type of relationship that is sought and the and sexual gratification. Theoretically, each need is kinds of needs or desires that are likely to be satisfied. regulated by a distinct behavioral system designed to From an attachment perspective, the multiform features respond to specific social cues. We should, therefore, of attractiveness are reducible to a small number of be attracted to people who display these cues. Given conceptual categories that correspond to the behavioral that the most basic need (for felt security) is regulated systems relevant to close relationships, each of which by the attachment system, and given that this system is is triggered by different and distinct cues. assumed to function similarly across the life span, The caregiving system, for example, responds to among the most important characteristics of a potential babyish features, distress, and vulnerability-which in partner should be the very characteristics shown to be the case of adults may include self-disclosures of fear centrally important in the selection of an attachment or weakness, or the vulnerability inherent in the Jetting figure during infancy and childhood--namely, famil­ down of defenses. A person who wants to provide care iarity and responsiveness. should logically be attracted to someone who seems to Interpersonal attraction is not only the first stage of need such care. Cues associated with targets of care giv­ many relationships, it is also the topic with which the ing constitute one interpersonal-attraction category that field of close relationships began, in the pioneering has been understudied, perhaps due to the lack of a work of Berscheid and Hatfield (Berscheid & E. Wals­ theory that would encourage such investigations. Also, ter, 1974) and Rubin (1973). The Jist of features and many studies of interpersonal attraction involve adoles­ characteristics that have been shown to be determinants cents and young adults, for whom care giving may be a of attraction is long and heterogeneous (for reviews, see relatively undeveloped or Jess salient need. There are, Aronson, 1988; Berscheid, 1984). For example, we are in fact, age-related changes in the relevance of such attracted to people whose values, attitudes, opinions, cues. One study (Fullard & Reiling, 1976) documented and even physical features are similar to our own a developmental shift in preference for infant relative (Hinsz, 1989; Rubin, 1973). Unless we dislike someone to adult faces. Before puberty, both males and females to begin with, increased exposure is associated with preferred photographs of adult faces; after puberty, they increased liking (Zajonc, 1968). Socially responsive spent more time looking at infant faces. children are more attractive to their peers (Rubin, A person who seeks gratification of sexual needs 1980). Among all the people with whom we could should be attracted to someone who displays cues of socialize, we typically choose those who live or work sexual availability and value. In sexual relationships, nearby (Festinger, Schachter, & Back, 1950; New­ an attractive physical appearance can be remarkably comb, 1961). We have a predilection for people who important (e.g., E. Walster, Aronson, Abrahams, & smile frequently, who are good-humored, and who Rottmann, 1966). Also, there appears to be consider­ make us laugh (Folkes & Sears, 1977). We are espe­ able cross-cultural and cross-age consistency in what is cially attracted to people we know find us appealing (A. viewed as physically attractive, including characteris­ P. Aron et al., 1989). Anxiety tends to enhance attrac­ tics associated with youth and/or health-smooth skin, tion (Dutton & A. P. Aron, 1974; Hatfield & Rapson, clear eyes, white teeth, a lively gait. Also desirable in 1987) and the desire to affiliate (Schachter, 1959). a potential sexual partner-especially in a male being These data indicate that attraction to another person evaluated by a female-is evidence of social status and tends to increase with familiarity and the likelihood of resources, such as popularity, material wealth, physical a positive response. Anyone who is like us (in appear­ strength, intelligence, and wit (Buss, 1985). ance, attitude, etc.) or who likes us, is smiling, or makes Because romantic Jove includes elements of attach­ us smile is more likely than an unfamiliar or unrespon­ ment, caregiving, and sexuality, the self-attribution that sive person to be viewed as safe, approachable, and thus one is "in love" may result from noticing that another attractive. Anxiety apparently intensifies the need to be person is especially responsive, needs to be taken care near another person or makes that person's security- of, or is a sexual "turn-on." This could easily result in

10 ATIACHMENT AND CLOSE RELATIONSHIPS misunderstandings and confusion: Responsiveness Even though the motivation to seek closeness is need not mean sexual interest, sexual interest need not hypothesized to be somewhat different and more imply more general responsiveness, and the wish to be complex for adults than for infants, the first phase in taken care of carries no guarantee of reciprocal care. what may eventually become an attachment relation­ People can mistakenly believe that their entire proto­ ship is remarkably similar (Shaver et al., 1988). Like type of romantic love (Fehr, 1988) is about to be real­ an adult in love, the infant is preoccupied with and ized when in fact only one component is actually notably vigilant for signs of the target person's re­ present. sponsiveness. Further, for infants as well as adults, To say that we are attracted to someone is to say that one's emotional state hinges on the target person's we would like to be physically ancl/or psychologically behavior, with responsiveness generally leading to close to that person. Within an attachment framework, feelings of security and joy and unresponsiveness what motivates such proximity seeking depends on evoking anxiety and distress (Ainsworth et al., 1978; which social-behavioral system is activated. Conceptu­ Tennov, 1979). In both kinds of relationships-be­ alizing interpersonal attraction in this way makes it tween infant and caregiver and between adult lov­ possible to reduce the many interpersonal-attraction ers--strong forces attract the individuals to each factors to a few conceptually meaningful categories. other and, in some cases, hold them together long This in itself advances our understanding of attraction enough for an emotional bond to develop. In both, beyond the mere documention of factors. In our view, the formation of such a bond is facilitated by close however, the real strength of attachment theory in this physical contact. Only between parents and children domain is its ability to go beyond classification to an or between adult lovers is prolonged bodily contact explanation of when and why particular characteristics considered normal. It is noteworthy that laypersons of a person should be attractive. judge these two types of relationships to be the "clos­ est" (Berscheid & Graziano, 1979). For good or ill, the intensity of the need for close contact eventually How Is a Relationship Formed, and diminishes (Fisher, 1992; Traupmann & Hatfield, How Does It Develop? 1981)-an important fact that needs to be explained. Logically, the safe-haven component of attachment Some attractions develop into relationships, and would be expected to develop within the context of some do not. Also, it is generally assumed and well closeness. Consistent with this, researchers have found documented that relationships, after being formed, tend that mutual attraction and sexual passion are most to change over time (e.g., Berscheid & E. Walster, important early in a relationship, but the degree to 1974; Huston, Surra, Fitzgerald, & Cate, 1981; G. which a partner provides comfort and emotional sup­ Levinger, 1983; G. Levinger & Snoek, 1972; Lewis & port becomes increasingly important over time (Reedy, Spanier, 1979; Taylor & Altman, 1987; E. Walster & Birren, & Schaie, 1981). In attachment terms, what G. W. Walster, 1978). A nascent relationship is obvi­ eventually comes to matter most is whether the partner ously different from an established relationship, and serves as a reliable haven of safety. Mutual attraction discovering just what changes and why is an important and sexual interest can get couples together, but, if task for relationship researchers and has been the focus partners fail to satisfy each other's needs for comfort of much study (Altman & Taylor, 1973; Kerckhoff & and security, dissatisfaction will likely result. Kotler K. E. Davis, 1962; G. Levinger, 1983; Lewis, 1973; (1985) found that sensitive and responsive care, not Murstein, 1976; Reiss, 1960). In general, efforts to sexual attraction, was the most accurate predictor of identify a uniform sequence of stages for all close marital strength. relationships have met with great difficulty (e.g., When does a relationship become a base of security? Huston et al., 1981). It is relatively safe to assume that parents are committed From an attachment perspective, the formation of a to their offspring for life. This commitment is not close relationship between two individuals of any age typically questioned or broken. Commitments between typically forms in the context of close physical proxim­ peers, however, tend to be less robust and more suscep­ ity. In other words, for an attachment to form there must tible to both internal and external influences (G. Levin­ be a strong force promoting closeness. In infancy, ger, 1976). It is likely that, only after an extended period proximity is regulated by the attachment system and the of time and/or after an explicit commitment has been infant's need for security. In adult romantic relation­ made, a peer relationship can serve as a secure base ships, the sexual mating system (sexual attraction) is with a degree of certainty approaching that of the base hypothesized to be another primary instigator for the provided by parents. Marriage, for example, is usually proximity seeking that is the first step toward attach­ accompanied by a legally binding, public promise to ment formation. care for the partner until death.

11 HAZAN & SHAVER

In terms of the three defining attachment features predictions concerning both the nature and timing of described earlier (see Figure 2), the process of attach­ important milestones and transitions in developing re­ ment formation, at any age, is hypothesized to in­ lationships. The three components of a prototypical pair volve the same sequence: proximity seeking bond, each corresponding to a separate behavioral sys­ followed by safe-haven behavior followed by the tem-attachment, caregiving, and sexual mating­ establishment of a secure base. In some cases, of gradually become integrated. The hypothesized course, the process will not be completed. As sug­ developmental course of this integration, which is sim­ gested earlier, the major difference between infant­ ilar to Sternberg's (1986) proposal concerning the tem­ caregiver and adult-pair bonds is in the motivation poral development of various components of love, is for seeking closeness in the first place. The hypoth­ represented in Figure 7. esized process is summarized in Figure 6. An attachment perspective adds to our understanding of how close relationships develop and change over What Makes Relationships Satisfying time and helps explain some of the phenomena that and/or Enduring? have been repeatedly documented and described. The attachment view of relationship development helps ac­ As the wording of this question implies and as the count for some important characteristics of the time data confirm, an enduring relationship may not be a course of close relationships. It also makes implicit satisfying one. Therefore, it is essential that a theory of

RELATIONSHIP PHASE ATIACHMENT BEHAVIORS

Initial attraction *proximity maintenance

Established relationship proximity maintenance safe haven

Goal-corrected partnership, proximity maintenance long-term or committed relationship safe haven secure base

Figure 6. A model ofattachment formation processes. Attachment formation begins with proximity seeking, which may be motivated by a desire for security, sexual gratification, or the provision of care. In subsequent phases, the partner may gradually become a safe haven and then a secure base. *Motivations for proximity seeking include security (attachment), nurturance (caregiving), and sexual attraction (sexual mating).

7 attachment 6 caregiving 5 sexual mating Importance, 4 intensity 3

2

2 3 4 5 6 7

Time in Years

Figure 7. The developmental course of a prototypical adult attachment relationship in terms of three behavioral systems.

12 ATTACHMENT AND CLOSE RELATIONSHIPS close relationships specify and explain the factors that tion of basic needs for comfort, care, and sexual grati­ predict each. fication. Trust in a partner's willingness and ability to According to attachment theory, a relationship is meet needs is determined in part by the partner's actual satisfying to the extent that it meets basic needs. At any behavior and in part by the expectations or comparison age, attachment quality turns in large part on the answer levels (Kelley, 1983; Rusbult, 1983) each person brings to the question, "Can I trust my partner to be available to the relationship. A history of close relationships and responsive to my needs?" Trust promotes self-dis­ lacking in trust, for example, might be expected to closure and the development of intimacy (Reis & result in the kind of minimal expectations that could Shaver, 1988). Trust is also associated with open com­ lead one to stay in an unsatisfying relationship. Thus, munication about and the "voicing" of needs (Holmes relationship longevity may be influenced by relation­ & Rempel, 1989; Rusbult & Zembrodt, 1983). Satisfy­ ship history. After an emotional bond has developed, it ing relationships are not conflict free, but they involve can act as a psychological tether that provides some the kind of trust that allows couples to argue construc­ security and holds two people together regardless of tively (Rands, G. Levinger, & Mellinger, 1981) and to whether they still enjoy being together. Anxiety result­ engage in effective problem-solving behaviors (Kobak ing from contemplation of or attempted separation can & Hazan, 1991). The "hidden agendas" that interfere activate attachment behaviors that lead one back to the with successful conflict resolution are often about relationship, unless there is an available and willing unmet needs (Gottman, Notarius, Gonso, & Markman, alternative. 1976). Attachment theory tells us what the needs are likely to be and explains why trust in a partner's respon­ siveness to these needs is critical. What Are the Precursors and Many unsatisfying relationships endure, and these Reactions to Relationship Dissolution? are among the most challenging for researchers. What makes a partner decide to leave? The belief that import­ Just as research has uncovered multiple correlates of ant needs could be better met in another relationshi~ relationship satisfaction, scores of studies have been what has been called comparison level for directed at revealing the causes of breakup (e.g., alternatives-is influential in decisions to continue or Felmlee, Sprecher, & Bassin, 1990; Gottman, Mark­ end a relationship(Kelley, 1983; Rusbult, 1980, 1983). man,&Notarius, 1977;G.Levinger, 1966;Lund, 1985; In defining relationship commitment, it is helpful to Simpson, 1990). At the risk of sounding glib, one can distinguish between the desire to continue a relation­ conclude from this research that failure to engage in ship because it is satisfying and the tendency to stay behaviors that enhance relationship satisfaction and simply because the constraints against breakup seem longevity renders dissolution more likely. If partners too great to overcome (Stanley, 1986). External con­ lack trust in each other and, as a result, do not openly straints to break up include joint ownership of property, and clearly communicate their thoughts and feelings, poverty, children, and a relative absence of alternatives do not engage in effective strategies of conflict resolu­ (G. Levinger, 1976). tion, or are not committed to remaining together, the Attachment theory suggests an additional factor probability that the relationship will endure is lessened. that may contribute to the maintenance of an unsat­ Structural factors of the sort mentioned earlier-<:hil­ isfying relationshi~the emotional bond of attach­ dren, limited financial resources, and religious or soci­ ment. Recall that an attachment is typically formed etal prohibitions--also figure prominently in a couple's in the context of close proximity. Weiss (1982) ar­ decision to end or maintain their relationship. As long gued that proximity alone can maintain the bond. as each member of a pair has confidence (i.e., trust) in Often couples are unaware of the bond between them the other's ability and willingness to supply essential until it is disrupted or threatened in some way relational provisions, each will be motivated to main­ (Berscheid, 1983). Even a burdensome and unsatis­ tain the relationship. In G. Levinger's (1976) terms, fying relationship can contribute to one's sense of internal attractions will be so strong that external bar­ security. Perhaps the best evidence for the security­ riers and alternative attractions will be of little practical promoting function of an unhappy relationship is the importance. intense anxiety that typically accompanies separa­ The desire to leave a relationship necessarily comes tion (Weiss, 1975). Bowlby (1973) theorized that after the relationship has been formed. The question, separation from an attachment figure is one of many then, is what changes between the time a person decides natural cues to potential danger and, as such, triggers to enter a relationship and the time a decision is made a fear response and, in turn, attachment behaviors. to leave. The answer, we believe, lies in a process To summarize, from an attachment perspective, re­ model of attachment formation. Relationship satisfac­ lationship satisfaction depends largely on the satisfac- tion always reduces to whether needs are being satisfied

13 HAZAN & SHAVER or not (Shaver & Hazan, 1984). The problem is that the When the attachment system is activated-for ex­ relative importance of various needs changes over time. ample, by the sudden unavailability of the primary In fact, what gets two people into a relationship may be attachment-the natural response is to seek proxim­ what matters least in the long run. If sexual passion is ity to the attachment figure, and this seems to happen indeed the initial motivating force in the formation of whether or not establishing proximity is possible or many adult pair bonds, a decline in satisfaction is even rationally desirable. That an estranged partner inevitable unless the relationship meets other needs may still be attainable can foster not only hope but after they have become important. Unfortunately, peo­ also protracted protest (Weiss, 1975, 1988). In some ple in the throes of romantic passion may give relatively cases, chronic activation of the attachment system in little thought to whether the people to whom they are the absence of the former partner may result in pre­ attracted will make reliable long-term providers of care mature attachment to another person. Likely candi­ and support-which in time will come to dominate dates would include individuals providing support their feelings about the relationship. According to the and care during the difficult postseparation period. hypothesized time course presented in Figure 7, dissat­ By the time a couple decides to separate, all former isfaction would be expected to peak around the time fondness and affection may have eroded. For this that intense attraction has faded and the partner's com­ reason, many newly separated individuals, especially petence as a haven of safety and secure base assumes the ones who initiate separation, are surprised when relatively greater importance. It should be at this point they begin to experience a compulsion to be near the in a relationship that expectations, alternatives, and former partner. Weiss (1975) argued that this very constraints come into play in a major way. common feeling is due to the persistence of attach­ Attachment theory has much to say about how peo­ ment and suggested that an attachment bond can be ple respond when a relationship ends. Response to broken only by an extended period of separation. The separation and loss, after all, was the topic with which same sequence of responses would not be expected Bowlby began his inquiries. It is difficult to discuss the unless an attachment had been fully formed, which function of attachment without considering disruptions may take several years. Weiss (1988) noted that and separations, because the system that regulates at­ responses to divorce are distinctly different in indi­ tachment feelings and behaviors includes "built-in" viduals married for fewer than 2 years. responses to disruption. Even though there is tremen­ Most of what attachment theory has to say about dous cultural variation in associated rituals and cus­ separation and loss applies equally well to loneliness toms, the human response to the breaking of an (Peplau & Perlman, 1982), which seems to take two attachment bond hardly varies (Gorer, 1973; Marris, major form~motional and social isolation (Weiss, 1958; S. I. Miller & Schoenfeld, 1973; Palgi, 1973). 1973). Emotional isolation is the kind of loneliness Moreover, the way in which adults respond to attach­ associated with the lack of an intimate companion, ment disruption is not essentially different from the way whereas feelings of social isolation result from the lack infants and children respond (Bowlby, 1980; Hazan & of a social network or sense of community. The two Shaver, 1992; Heinicke & Westheimer, 1966; Parkes forms of loneliness correspond well to Bowlby's dis­ & Weiss, 1983). tinction between the attachment and affiliation behav­ The first reaction to the disruption of an attachment ioral systems, which are thought to have different relationship, whether due to death or voluntary separa­ functions. In addition, the two types of loneliness have tion, is intense separation-protest behavior. Individuals different symptoms, causes, and cures (Rubenstein & report feeling agitated, anxious, and preoccupied with Shaver, 1982). Weiss (1973) viewed loneliness as an thoughts of the lost partner, coupled with a compulsion adaptive if uncomfortable emotional state because it to search for him or her, as though trying to undo the serves as a reminder that important social needs are not loss even if it is consciously known to be irreversible. being met, and it tends to continue until corrective Eventually, with the realization that the loss cannot be action is taken. Again, reactions to the loss or absence recovered, there comes a period of deep sadness, during of an emotional bond are hypothesized to be the direct which intense activity and rumination give way to result of the functioning of the attachment behavioral depression and despair. Many individuals experience system. an unusual and marked lack of concern about or interest in life and other people. Gradually, the sadness sub­ sides, and most people achieve an adaptive degree of What Is the Role of Relationships in emotional detachment from the lost partner and return OveraU Functioning? to ordinary living. In adults especially, constructing a causal account of the loss helps bring about acceptance Close relationships are important in the lives of most and detachment (Harvey, Orbuch, & Weber, 1990). people and are among the greatest sources of subjective

14 ATIACHMENT AND CLOSE RELATIONSIDPS well-being (Freedman, 1978; Veroff, Douvan, & How and Why Do Individuals Differ in Kukla, 1981). As further evidence of the significance the Way They Think, Feel, and Behave of close relationships for healthy functioning, the dis­ in Relationships? ruption or loss of a relationship, especially through divorce, makes one more susceptible to everything Up to this point, we have emphasized normative from automobile accidents to alcohol abuse to admis­ aspects of close relationships. However, one cannot sion into a psychiatric facility (Bloom, Asher, & White, ignore the immense variability in the ways people 1978). In addition, the grieving and lonely are vulner­ relate to one another. Some people fear intimacy, able to disease, as studies of their hearts and immune whereas others embrace it (Hatfield, 1984). Some systems have demonstrated (e.g., Kiecolt-Glaser, Gar­ self-disclose to an excessive degree, whereas others ner et al., 1984; Kiecolt-Glaser, Ricker et al., 1984; disclose little or not at all (Altman & Taylor, 1973). Lynch, 1977), and are at greater risk of death from Attitudes about romantic involvements range from cancer (Goodwin, Hurt, Key, & Sarret, 1987). Social game playing to pragmatism (C. Hendrick & S. deficiencies and losses jeopardize not only health and Hendrick, 1986). Relationships can involve commit­ happiness but job performance and achievement as well ment without passion or passion without intimacy (Baruch, Barnett, & Rivers, 1983; Lee & Kanungo, (Sternberg, 1986). Partners may attribute each 1984; Vaillant, 1977). other's problematic behavior to character flaws or The diverse and numerous links between relation­ explain it in terms of situational factors (Fincham, ships and overall functioning are well documented Beach, & Nelson, 1987). In response to conflict, but poorly understood. Why should the absence of partners may withdraw or accommodate (Rusbult, socioemotional connections have such profound psy­ Zembrodt, & Gunn, 1982). Communication may be chological, physical, and even occupational conse­ open and coherent or defensive and disorganized quences? We believe that Bowlby's insights supply (Bretherton, 1990). Relationships appear to be as the missing link: Social deprivation is harmful be­ diverse as the individuals involved. cause it thwarts the satisfaction of inborn needs. Ideally, the multitude of differences could be re­ Powerful emotional cues signal when these crucial duced to a more manageable number of conceptual needs are not being met, and the result is a subjective categories. If we accept that confidence and trust in the state of anxiety and discomfort. Stated differently, responsiveness of others are central issues in close social deprivation is stressful and, as with any pro­ relationships, then a few broad categories of individual longed stressor, can be both psychologically and differences naturally follow. As Main et al. (1985) saw physically harmful. it, the social environment can be perceived as consis­ Because of the interrelations among behavioral tently responsive, inconsistently responsive, or consis­ systems, malfunctioning in one system can cause tently unresponsive to an individual's attempts to dysfunction in another. Recall that, in studies with establish security-promoting closeness. Given that this infants, attachment and exploratory behavior are important issue appears to be the same in infancy and closely intertwined. Smooth functioning of the ex­ adulthood, we assume that the important individual­ ploratory system requires a quiescent (satisfied) at­ difference categories will also be essentially the same. tachment system, which itself depends on the One would not expect to see the exact same behaviors availability and responsiveness of an attachment fig­ but rather the same basic strategies for maintaining felt ure. In adulthood, the balance between relationships security. In emphasizing normative attachment, we and work, between emotional connectedness and in­ have actually been describing security or the secure dependent activity, is in important respects similar to attachment type. In this section on individual differ­ the attachment-exploration balance marking healthy ences, we focus instead on the other end of the dimen­ functioning in early life (Hazan & Shaver, 1990). In sion-insecurity. Bowlby's (1979) words, The strategy associated with inconsistent respon­ siveness--anxious/ambivalent (preoccupied) attach­ Human beings of all ages are happiest and able to ment-is characterized by a lack of confidence in the deploy their talents to best advantage when they are reliable responsiveness of others. The proximal goal of confident that, standing behind them, there are one or all attachment behavior is to achieve a state of felt more trusted persons who will come to their aid should security. In the case of anxious/ambivalent attachment, difficulties arise. (pp. 103-104) this is attempted and at times accomplished by devoting immense mental energy and behavioral effort to keep­ Thus, attachment theory explains why and how close ing others close by and engaged. It is manifested in relationships play a central and critical role in overall intensified expressions of distress and anger and dimin­ feelings about and adjustment to life. ished exploratory activity.

15 HAZAN & SHAVER

In studies of adult attachment, anxious/ambivalent findings on avoidant attachment, see Shaver & Hazan, attachment is associated with obsessive preoccupation 1993.) with a romantic partner's responsiveness; falling in Theoretically, these two major patterns of insecurity love easily; being extremely jealous; being subject to are based on internal working models constructed from fear, anxiety, and loneliness (even when involved in a actual attachment experience, beginning with parents. relationship); having low self-esteem (Collins & Read, The anxious/ambivalent strategy is ''logical" in the 1990; Feeney & Noller, 1990); and experiencing a sense that it reflects a history of inconsistent respon­ higher rate of relationship dissolution (Hazan & siveness. Expecting close relationship partners to be Shaver, 1987). The anxious/ambivalently attached also somewhat unreliable may lead to heightened vigilance tend to view partners as reluctant to commit and as and fears of abandonment and neglect, both of which inadequate or insufficiently attentive caregivers can interfere with nonattachment activities. The avoid­ (Kunce & Shaver, 1991). They engage in indiscrimin­ ant strategy is equally logical in light of a history of ant and overly intimate self-disclosure (Mikulincer & frequent rejection or inhibitions on physical affection Nachshon, 1991) and assert their own feelings and and intimate emotional expression. Such experiences needs without adequate regard for the partner's feelings can lead to an avoidance of closeness, extreme self-re­ and needs (Daniels & Shaver, 1991). In laboratory liance, and a habit of regulating anxiety by keeping problem-solving tasks with their partners, anxious/am­ oneself distracted. In line with social psychological bivalent subjects tended to express dysfunctional anger research showing that actions often follow from beliefs (Kobak & Hazan, 1991). They also reported more and interpersonal schemas in a way that encourages physical and psychological symptoms (Fiala, 1991; repeated confirmation (e.g., Snyder & Swann, 1978), Hazan & Shaver, 1990) and had greater difficulty mak­ internal working models may have self-fulfilling ef­ ing friends in a new setting (Hazan & Hutt, 1991b). At fects on social behavior and social-information pro­ work, anxious/ambivalence is associated with distrac­ cessing. For example, there is evidence that mates are tion, procrastination, and suboptimal performance selected for their ability to confirm attachment-related (Hazan & Shaver, 1990) and, in discussions of attach­ expectations, even if the expectations are negative ment history, with overly effusive and poorly organized (Kirkpatrick & K. E. Davis, in press; Swann, Hixon, & discourse (Main et al., 1985). (For a review of addi­ De La Ronde, 1992). We believe that individual differ­ tional correlates of anxious/ambivalent attachment, see ences in attachment, mediated by internal working Shaver & Hazan, 1993.) models, may underlie many of the interpersonal differ­ In contrast, avoidant attachment is believed to result ences that have been discovered by researchers work­ from consistent unresponsiveness. The avoidant strat­ ing from other theoretical bases. Further, insecure egy for maintaining felt security involves avoidance of attachment might be at the root of many dysfunctional intimate social contact, especially in stressful or dis­ behaviors contributing to relationship dissatisfaction tressing circumstances, and compensatory engagement and dissolution. in nonsocial activities. According to research on adult Despite forces favoring the stability of individual attachment, avoidance is manifested in fear of intimacy differences in attachment, change is always possible. and a tendency to maintain distance in "close" relation­ For example, the experience of just one important rela­ ships, with pessimistic views of relationships and a tionship that disconfirms insecure expectations of un­ relatively high rate of relationship dissolution (Hazan reliability or rejection increases the likelihood of & Shaver, 1987). When answering interview questions forming a secure attachment in adulthood (Hazan & concerning childhood relationships with parents, Hutt, 1991a). In most cases, these disconfirming rela­ avoidantly attached adults (a) use idealized descriptors tionship experiences were formed with nonparental but are unable to provide supporting examples (Main adults (e.g., teachers, relatives) during childhood or et al., 1985) and (b) show spikes in skin conductance with romantic partners during late adolescence or early when probed for such examples (Dozier & Kobak, in adulthood. Consistent with studies of change in infancy press). They avoid self-disclosure and experience dis­ and childhood, secure attachment is the most stable comfort with relationship partners who do self-disclose pattern. (Mikulincer & Nachshon, 1991). They are more sus­ The distribution of adults across attachment catego­ ceptible to sudden religious conversion (Kirkpatrick & ries-55% secure, 25% avoidant, and 20% anx­ Shaver, 1990). They are judged by their peers to be ious/ambivalent-has been replicated in many studies hostile (Kobak & Sceery, 1988). They tend to use work in several different countries (e.g., Feeney & Noller, to avoid social interaction (Hazan & Shaver, 1990) and 1990; Hazan & Shaver, 1987; Mikulincer, Florian, & are prone to engaging in uncommitted sexual relations Tolmacz, 1990). The proportions are similar to those and using alcohol and other substances to reduce ten­ found in infant studies using Ainsworth's Strange Sit­ sion (Brennan, Shaver, & Tobey, 1991 ). (For additional uation procedure (Campos et al., 1983). To date, there

16 ATIACHMENT AND CLOSE RELATIONSIHPS have been no reliable gender differences reported in the we have proposed will be able to explain all important distribution of subjects across the three categories. relationship phenomena. Each theory has its bound­ aries, and attachment theory is no exception. In fairness to Bowlby, be was not attempting to explain every Gender Differences aspect or type of close relationship. His aim was simply to explain the structure and functions of attachment, No treatment of close relationships would be com­ and it took him three volumes to do it. Even in begin­ plete without some discussion of gender differences. ning to extend his theory to adult attachments, we have Males and females differ significantly in their styles of taken great liberties. communication (fannen, 1990), in their tendency to At the outset, we said that we see attachment and engage in extra-relationship affairs (Skolnick, 1978), interdependence as largely complementary theories. and in their skill at reading nonverbal cues (Hall, 1978). We asked what a good theory of close relationships They differ as well in their likelihood of initiating should be able to do. A reasonable response is that it breakups (Hill, Rubin, & Peplau, 1976) and in their should offer a precise and operational definition of subsequent adjustment (Wallerstein & Kelly, 1980). close and should emphasize relationships more than They also differ in the degree to which they value individual relationship partners. Interdependence the­ physical attractiveness (Berscheid, Dian, E. Walster, & ory is superior to attachment theory in both regards, G. W. Walster, 1971) and material resources (Buss & and, in its models of interaction and types of transfor­ Barnes, 1986) in a potential mate. mations that occur over the course of an interaction, The measure we developed for assessing individual interdependence theory is unparalleled. Attachment differences in adult attachment (Hazan & Shaver, theory, in contrast, emphasizes what of psychological 1987) might have been expected to show gender differ­ value individuals will try to accomplish in their inter­ ences. The anxious/ambivalent pattern sounds very actions. In other words, it supplies much of the psycho­ much like the clingy, dependent aspects of the female logical substance of what Kelley (1983) called the stereotype, and the avoidant pattern strongly resembles given matrix. Attachment is a motivated model, and the the stereotypical intimacy-evading male. That males motives on which it focuses are based in biology. We and females do not fall disproportionately into either are active, biological organisms that build internal category lends support to Bowlby's claim that all structures. As with other behavioristic models in psy­ human beings have an inborn need for felt security. chology, the strength of interdependence theory rests Informal examination of the data suggests that gen­ with its operational definitions and measurement of der differences lie primarily in the domains of caregiv­ observables. However, it would be a mistake to believe ing and sexuality, rather than attachment. In general, that all important relationship phenomena can be re­ females are more oriented toward caregiving, and duced to observable ones. males are more oriented toward sex. Some of the major The theories that emphasize socially constructed, differences between lesbian and gay male relationships shared meanings and narratives also add something seem to support this view (Peplau & Gordon, 1983). important to attachment theory, which, because of its Buss (1985) developed a sociobiological theory of roots in ethology, does not stress the unique properties human mate selection that is also consistent with this of human verbal communication. Nevertheless, attach­ conceptualization of gender differences. However, ment theory may provide some of the themes of and within an attachment framework, gender differences do constraints on interpersonal story construction, not entail a biological explanation. Because the care­ whereas communication-oriented theories add to at­ giving and sexual mating systems develop later than the tachment theory's conceptualization of the ways in attachment system, it is likely that they are more subject which internal working models of self, relationships, to sex-role socialization pressures. The extent of bio­ and relationship partners get constructed. A complete logical versus social causation is still a legitimate mat­ integration of attachment, interdependence, and com­ ter for dispute and further research. munication theories would require more space and expertise than we have at present, but we firmly believe that such a union would be fruitful. Concluding Remarks It is common to hear attachment theory described as a theory about three (or maybe four) types of babies. Our review of the literature has necessarily been Such emphasis on individual differences does not re­ selective, and there are many additional facts that we flect the main thrust of attachment theory, which is first have not attempted to integrate, although many seem and foremost a normative theory. An added strength is amenable to explanation within an attachment frame­ that it can also account for the nature and form of work. Even so, we do not believe that the framework individual differences. Ainsworth's creation of an in-

17 HAZAN & SHAVER novative paradigm for assessing attachment quality in Cindy Hazan and by National Science Foundation infancy focused attention on individual differences. By Grant BSN-8808736 to Cindy Hazan and Phillip R. creating a simple self-report measure of adult attach­ Shaver. ment styles, we inadvertently helped extend the indi­ We thank Rick Canfield, Robert Turgeon, and Jo­ vidual-differences approach to the study of adult anne Sturgeon for their helpful comments on drafts of relationships. The normative implications of attach­ this target article. ment theory have rarely been spelled out or tested. Cindy Hazan, Department of Human Development The prevailing problem with the overemphasis on and Family Studies, Van Rensselaer Hall, Cornell Uni­ individual differences has been a lack of consensus versity, Ithaca, NY 14853-4401. concerning how to measure these differences beyond infancy. Four or five different methods are currently used to assess attachment in toddlers and children References (Greenberg, Cicchetti, & Cummings, 1990); for adults, there are two highly similar interviews (Bartholomew, Ainsworth, M.D. S., Blehar, M. C., WateiS, E., & Wall, S. (1978). 1990; Main, 1991) and several self-report measures that Patterns of attachment: A psychological study of the Strange are derivatives of our 1987 measure (e.g., Collins & Situation. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Altman, 1., & Taylor, D. A (1973). Social penetration: The develop­ Read, 1990). Current debate includes questions about ment of interpersonal relationships. New York: Holt, Rinehart whether measures of adult attachment should concen­ &Winston. trate on parent-child dyads or adult-pair bonds, how the Arend, R., Gove, F., & Sroufe, LA. (1979). Continuity ofindividual various methods (e.g., interview, self-report) compare, adaptation from infancy to kindergarten: A predictive study of ego resiliency and curiosity in preschooleiS. ChildDevelopment, and how to conceptualize the differences (i.e., in terms 50, 950-959. of dimensions or types). There is even ongoing debate Aron, A. P., Dutton, D. G., Aron, E. N., & IveiSOn, A. (1989). about how many major types of attachment exist in Experiences of falling in love. Journal ofSocial and Personal adulthood (Bartholomew & Horowitz, 1991). We think Relationships, 6, 243-257. Aronson, E. (1988). The social animal (5th ed. ). New York: Freeman. it is important to point out that attachment theory does Bartholomew, K. (1990). Avoidance of intimacy: An attachment not stand or fall with the success of any individual-dif­ perspective. Journal of Social and Personal Relationships, 7, ference measure. Regardless of which measure is used, 147-178. the results invariably show that secure attachment is the Bartholomew, K., & Horowitz, L M. (1991). Attachment styles among young adults: A test of a four-category model. Journal norm, so we would do well to devote additional re­ ofPersonality and Social Psychology, 61, 226-244. search effort to investigating normative (i.e., secure) Baruch, G., Barnett, R., & RiveiS, C. (1983). Lifeprints: New panerns attachment phenomena. ofloveandworkfortoday'swomen. New York: McGraw-Hill. We imagine that Harlow would be pleased to see that BeiScheid, E. (1983). Emotion. In H. H. Kelley, E. Berscheid, A. Christensen, J. H. Harvey, T. L Huston, G. Levinger, E. researchers are finally grappling with love and affec­ McClintock, L A. Peplau, & D. R. PeteiSOn (Eds.), Close tion and are analyzing these important facets of human relationships (pp. 110--168). New York: Freeman. behavior into their component variables. Such a bot­ Berscheid, E. (1984). lnterpeiSOnal attraction. In G. Lindzey & E. tom-up approach and the construction of a descriptive Aronson (Eds.), Handbook of social psychology (3rd ed., pp. 413-484). Reading, MA: Addison-Wesley. base constitute a useful first phase in the development BeiScheid, E., Dion, K., Walster, E., & Walster, G. W. (1971). of our science, but it is not the ultimate goal. Eventually, Physical attractiveness and dating choice: A test of the match­ the accumulation of facts in the absence of theory ing hypothesis. Journal of Experimental Social Psychology, becomes inefficient. Without theory as a guide, re­ 7, 173-189. BeiScheid, E., & Graziano, W. (1979). The initiation of social rela­ search is difficult to plan, and findings are difficult to tionships and social attraction. In R. L Burgess & T. L Huston interpret. Our goal here has been to persuade the reader (Eds. ), Social exchange in developing relationships. New York: that a theoretical integration of research findings on Academic. close relationships is neither premature nor impossible BeiScheid, E., & Walster, E. (1974). A little bit about love. InT. L. Huston (Ed.), Foutulotionsofinterpersonal anraction (pp. 355- and that attachment theory can provide the core con­ 381). New York: Academic. structs of such an integrative framework. If we have Bloom, B. L, Asher, S. J., & White, S. W. (1978). Marital disruption failed to be persuasive, we hope we have at least been as a stressor: A review and analysis. Psychological Bulletin, 85, provocative enough to inspire the kind of integrative 867-894. Bowlby, J. (1944). Forty-four juvenile thieves: Their characteiS and thought and conceptual debate that will help advance home life.lnternationalJournal ofPsycho-Analysis, 25, 19-52, the science of close relationships. 107-127. Bowlby, J. (1951 ). Maternal care and mental health. Geneva: World Notes Health Organization. Bowlby, J. (1973). Attachment and loss: Vol. 2. Separation: Anxiety and anger. New York: Basic. Preparation of this target article was facilitated by a Bowlby, J. (1979). The malcing and breaking of affectional bonds. Special Projects Grant from Cornell University to London: Tavistock.

18 ATTACHMENT AND CLOSE RELATIONSHIPS

Bowlby, J. (1980). Attachment and loss: VoL 3. Loss: Sadness and York: Harper. depression. New York: Basic. Fiala, K. B. (1991). Attachment and psychosocial functioning of Bowlby, J. (1982).Attachment and loss: Vol. ].Attachment (2nd ed. ). depressed, remitted depressed, and nondepressed women and New York: Basic. (Original work published 1969) their partners. Unpublished doctoral dissertation, University of Bowlby, J. (1988). A secure base: Parent-child attachment and Massachusetts, Amherst. healthy human development. New York: Basic. Fincham, F. D., Beach, S., & Nelson, G. (1987). Attribution processes Brennan, K., Shaver, P.R., & Tobey, A. E. (1991 ). Attachment styles, in distressed and non-distressed couples: Self-partner attribution gender, and parental problem drinking. Journal of Social and for spouse behavior. Cognitive Therapy and Research, 11, 71-86. Personal Relationships, 8, 451-466. Fisher, H. E. (1992).Anatomy of love. New York: Norton. Bretherton, I. (1990). Open communication and internal working Fiske, S. T., & Taylor, S. E. (1991). Social cognitiorL New York: models: Their role in the development of attachment relation­ Random House. ships. In R. A. Thompson (Ed.), Nebraska Symposium on Moti­ Folkes, V. S., & Sears, D. 0. (1977). Does everybody like a liker? vation: VoL 36. Socioemotional development (pp. 57-113). Journal a/Experimental Social Psychology, 13, 505-519. lincoln: University of Nebraska Press. Freedman, J. (1978). Happy people: What happiness is, who has it, Buss, D. M. (1985). Human mate selection. American Scientist, 73, and why. New York: Harcourt Brace Jovanovich. 47-54. Fullard, W., & Reiling, A.M. (1976). An investigation of Lorenz's Buss, D. M., & Barnes, M. (1986). Preferences in human mate "babyness." Child Development, 47, 1191-1193. selection. Journal of Personality and Social Psychology, 50, Furman, W., & Buhmester, D. (1985). Children's perceptions of the 559-570. personal relationships in their social networks. Developmental Campos, J. J., Barrett, K., Lamb, M. E., Goldsmith, H. H., & Psychology, 21, 1016--1024. Stenberg, C. (1983). Socioemotional development. In P. H. Goldsmith, H. H., & Alansky, J. A. (1987). Maternal and infant Mussen (Ed.), Handbook of child psychology: Vol. 2. Infancy temperamental predictors of attachment: A meta-analytic re­ and developmental psychobiology (pp. 783-915). New York: view. Journal of Consulting and Clinical Psychology, 55, 805- Wiley. 816. Cassidy, J. (1988). Child-mother attachment and the self in six-year­ Goodwin, J. S., Hurt, W. C., Key, C. R., & Sarret, J. M. (1987). The olds. Child Development, 59, 121-134. effect of marital status on stage, treatment and survival of cancer Colin, V. L. (1991 ). Human attachment: What we know now. Chevy patients. Journal of the American Medical Association, 258, Chase, MD: U.S. Department of Health and Human Services. 3125-3130. Collins, N. L., & Read, S. J. (1990). Adult attachment, working Gorer, G. (1973). Death, grief and mourning in Britain. In E. J. models, and relationship quality in dating couples. Journal of Anthony & C. Koupernik (Eds.), The child in his family: The Personality and Social Psychology, 58, 644-663. impact of disease and death. New York: Wiley. Crittenden, P. M. (1988). Relationships at risk. In J. Belsky & T. Gottman, J. M. (1983). How children become friends. Monographs Nezworski (Eds. ), Clinical implications ofattachment (pp. 136-- of the Society for Research in Child Development, 48(3, Serial 174). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. No. 201). Crockenberg, S. B. (1981 ). Infant irritability, mother responsiveness, Gottman, J ., Markman, H., & Notarius, C. (1977). The topography of and social support influences on the security of infant-mother marital conflict: A sequential analysis of verbal and nonverbal attachment. Child Development, 52, 857-869. behavior. Journal of Marriage and the Family, 39, 461-477. Daniels, T., & Shaver, P. R. (1991). Attachment styles and power Gottman, J., Notarius, C., Gonso, J., & Markman, H. J. (1976). A strategies in romantic relationships. Unpublished manu­ couple's guide to communicatiorL Champaign, IL: Research script, State University of New York at Buffalo, Department Press. of Psychology. Greenberg, M., Cicchetti, D., & Cummings, M. (Eds.). (1990). At­ Dozier, M., & Kobak, R. R. (in press). Psychophysiology and ado­ tachment in the preschool years: Theory, research, and inter­ lescent attachment interviews: Converging evidence for repress­ vention. Chicago: University of Chicago Press. ing strategies. Child Development. Grossberg, S. (1980). How does a brain build a cognitive code? Driscoll, R., Davis, K. W., & lipetz, M. E. (1972). Parental interfer­ Psychological Review, 87, 1-51. ence and romantic love. Journal of Personality and Social Grossmann, K. E., & Grossmann, K. (1991). Attachment quality as Psychology, 24, 1-10. an organizer of emotional and behavioral responses in a longi­ Duck, S. W. (1991, May). New lamps for old: A new theory of tudinal perspective. In C. M. Parkes, J. Stevenson-Hinde, & P. relationships and a fresh look at some old research. Paper Marris (Eds.), Attachment across the life cycle (pp. 93-114). presented at the meeting of the International Network on Per­ London: Tavistock/Routledge. sonal Relationships, Normal, IL. Hall, J. A. (1978). Gender effects in decoding nonverbal cues. Psy­ Dutton, D. G., & Aron, A. P. (1974). Some evidence for heightened chological Bulletin, 85, 845--857. sexual attraction under conditions of high anxiety. Journal of Harlow, H. (1958). The nature of love. American Psychologist, 13, Personality and Social Psychology, 30, 510-517. 673--685. Egeland, B., & Farber, E. A. (1984). Infant-mother attachment: Hartup, W. (1983). Peer relations. In E. M. Hetherington (Vol. Ed.) Factors related to its development and changes over time. Child & P. H. Mussen (Series Ed.), Handbook of child psychology: Development, 55, 753-771. VoL 4. Socialization, personality, and social development (pp. Feeney, J., & Noller, P. (1990). Attachment style as a predictor of 103-196). New York: Wiley. adult romantic relationships. Journal ofPersonality and Social Harvey, J. H., Orbuch, T. L., & Weber, A. L. (1990). A social Psychology, 58, 281-291. psychological model of account-making: In response to severe Fehr, B. (1988). Prototype analysis of the concepts of love and stress. Journal ofLanguage and Social Psychology, 9, 191-207. commitment. Journal ofPersonality and Social Psychology, 55, Hatfield, E. (1984). The dangers of intimacy. In V. J. Derlega (Ed.), 557-579. Communication, intimacy and close relationships (pp. 207- Felmlee, D., Sprecher, S., & Bassin, E. (1990). The dissolution of 220). New York: Academic. intimate relationships: A hazard model. Social Psychology Hatfield, E., & Rapson, R. L. (1987). Passionate love: New directions Quarterly, 52, 13-30. in research. In W. H. Jones & D. Perlman (Eds.), Advances in Festinger, L., Schachter, S., & Back, K. W. (1950). Social pressures personal relationships (Vol. 1, pp. 109-140). Greenwich, CT: in informal groups: A study of human factors in housing. New JAI.

19 HAZAN & SHAVER

Hazan, C., & Hull, M. J. (1991a). Continuily and change in internal models of attachment, affect regulation, and perceptions of self working models of attachment. Unpublished manuscript, Cor­ and others. Child Development, 88, 135-146. nell University, Department of Human Development and Fam­ Konner, M. (1982). The tangled wing: Biological constraints on the ily Studies. human spirit. New York: Holt, Rinehart & Winston. Hazan, C., &Hutt, M.1. (1991b).Fromparentstopeers: Transitions Kotler, T. (1985). Security and autonomy within marriage. Human in attachment. Unpublished manuscript, Cornell University, Relations, 38, 299-321. Department of Human Development and Family Studies. Kunce, L. 1., & Shaver, P. R. (1991). An attachment-theoretical Hazan, C., Hutt, M.1., Sturgeon, 1., & Bricker, T. (1991, April). The approach to caregiving in romantic relationships. Unpublished process of relinquishing parents as attachment figures. Paper manuscript, State University of New York at Buffalo, Depart­ presented at the meeting of the Society for Research in Child ment of Psychology. Development, Seattle. Lamb, M. E., Thompson, R. A., Gardner, W. P., Charnov, E. L., & Hazan, C., & Shaver, P.R. (1987). Romantic love conceptualized as Estes, D. (1984). Security of infantile attachment as assessed in an attachment process. Journal of Personalily and Social Psy­ the Strange Situation: Its study and biological interpretation. chology, 52,511-524. Behavioral and Brain Sciences, 7, 127-171. Hazan, C., & Shaver, P.R. (1990). Love and work: An attachment­ Lee, M. D., & Kanungo, R. N. (Eds.). (1984). Management of work theoretical perspective. Journal of Personalily and Social Psy­ and personal life: Problems and opportunities. New York: chology, 59, 27~280. Praeger. Hazan, C., & Shaver, P. R. (1992). Broken attachments. InT. L. Levenson, R. W., & Gottman, 1. M. (1985). Six physiological and Orbuch (Ed.), Close relationship loss: Theoretical approaches affective predictors of change in relationship satisfaction. Jour­ (pp. ~ 108). Hillsdale, N1: Lawrence Erlbaum Associates, Inc. nal ofPersonality and Social Psychology, 49, 85-94. Heinicke, C., & Westheimer, I. (1966). Briefseparations. New York: Levinger, G. (1966). Systematic distortion in spouses' reports of International Universities Press. preferred and actual sexual behavior. Sociometry, 29, 291-299. Hendrick, C., & Hendrick, S. (1986). A theory and method of love. Levinger, G. (1976). A social psychological perspective on marital Journal ofPersonality and Social Psychology, 50, 392-402. dissolution. Journal ofSocial Issues, 32(1), 21-47. Hill, C. T., Rubin, Z., & Peplau, L. A. (1976). Breakups before Levinger, G. (1983). Development and change. In H. H. Kelley, E. marriage: The end of 103 affairs. Journal ofSocial Issues, 32, Berscheid, A. Christensen, 1. H. Harvey, T. L. Huston, G. 147-168. Levinger, E. McClintock, L.A. Peplau, & D. R. Peterson (Eds.), Hinde, R. (1979). Towards understanding relationships. London: Close relationships (pp. 315-359). New York: Freeman. Academic. Levinger, G., & Levinger, A. C. (1986). The temporal course of close Hinsz, V. B. (1989). Facial resemblance in engaged and married relationships: Some thoughts about the development of couples. Journal of Social and Personal Relationships, 6, children's ties. In W. W. Hartup & Z. Rubin (Eds.), Relation­ 223-229. ships and development (pp. 111-134). Hillsdale, N1: Lawrence Holmes, 1. G., & Rempel, 1. K. (1989). Trust in close relationships. Erlbaum Associates, Inc. In C. Hendrick (Ed.), Close relationships (pp. 187-220). New­ Levinger, G., & Snoek, 1. D. (1972).Attraction in relationship:A new bury Park, CA: Sage. look at interpersonal allractioTL Morristown, N1: General Huston, T. L., Surra, C. A, Fitzgerald, N. M., & Cate, R. M. (1981). Learning Press. From courtship to marriage: Mate selection as an interpersonal Lewis, R. A. (1973). A longitudinal test of a developmental frame­ process. In S. Duck & R. Gilmour (Eds.), Personal relation­ work for premarital dyadic formation. Journal ofMarriage and ships: VoL 2. Developing personal relationships (pp. 53-88). the Family, 35, 16-25. New York: Academic. Lewis, R. A., & Spanier, G. B. (1979). Theorizing about the quality Kelley, H. H. (1983). Love and commitment. In H. H. Kelley, E. and stability of marriage. In W. R. Burr, R. Hill, F. I. Nye, & Berscheid, A. Christensen, 1. H. Harvey, T. L. Huston, G. I. L. Reiss (Eds.), Contemporary theories about the family Levinger, E. McClintock, L.A. Peplau, & D. R. Peterson (Eds.), (Vol. 1). New York: Free Press. Close relationships (pp. 265-314). New York: Freeman. Lund, M. (1985). The development of investment and commitment Kelley, H. H., Berscheid, E., Christensen, A, Harvey, J. H., Huston, scales for predicting continuity of personal relationships. Jour­ T. L., Levinger, G., McClintock, E., Peplau, L. A., & Peterson, nal ofSocial and Personal Relationships, 2, 3-23. D. R. (Eds.). (1983). Close relationships. New York: Freeman. Lynch, 1. 1. (1977). The broken heart: The medical consequences of Kerckhoff, A. C., & Davis, K. E. (1962). Value consensus and need loneliness. New York: Basic. complementarity in mate selection. American Sociological Re­ Main, M. (1990). Parental aversion to infant-initiated contact is view, 27, 295-303. correlated with the parent's own rejection during childhood: The Kiecolt-Glaser, 1. K., Gamer, W., Speicher, C., Penn, G. M., Holli­ effects of experience on signals of security with respect to day, 1., & Glaser, R. (1984). Psychological modifiers of im­ attachment. InK. E. Barnard & T. B. Brazelton (Eds.), Touch: munocompetence in medical students. Psychosomatic The foum:lation of experience (pp. 461-495). Madison, CT: Medicine, 46, 7-14. International Universities Press. Kiecolt-Giaser, 1. K., Ricker, D., George, 1., Messick, G., Speicher, Main, M. (1991). Metacognitive knowledge, metacognitive monitor­ G. E., Garner, W., & Glaser, R. (1984). Urinary cortisol levels, ing, and singular (coherent) vs. multiple (incoherent) model of cellular immunocompetence, and loneliness in psychiatric inpa­ attachment: Findings and directions for future research. In C. M. tients. Psychosomatic Medicine, 46, 15-23. Parkes, 1. Stevenson-Hinde, & P. Marris (Eds.), Attachment Kirkpatrick, L. A., & Davis, K. E. (in press). Attachment style, across the life cycle (pp. 127-159). London: Tavistock/ gender, and relationship stability: A longitudinal analysis. Jour­ Routledge. nal ofPersonality and Social Psychology. Main, M., & Hesse, E. (1990). Parents' unresolved traumatic experi­ Kirkpatrick, L., & Shaver, P. R. (1990). Attachment theory and ences are related to infant disorganized status: Is frightened religion: Childhood attachments, religious beliefs, and conver­ and/or frightening parental behavior the linking mechanism? In sion. Journal for the Scientific Study ofReligion, 29, 315-334. M. T. Greenberg, D. Cicchetti, & E. M. Cummings (Eds.), Kobak, R. R., & Hazan, C. (1991). Attachment in marriage: Effects Attachment in the preschool years (pp. 161-184). Chicago: of security and accuracy of working models. Journal ofPerson­ University of Chicago Press. ality and Social Psychology, 60, 861-869. Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, Kobak, R. R., & Sceery, A. (1988). The transition to college: Working childhood, and adulthood: A move to the level of representation.

20 ATIACHMENT AND CLOSE RELATIONSHIPS

Monographs of the Society for Research in Child Development, Rusbult, C. E. (1983). A longitudinal test of the investment model: 50(1-2, Serial No. 209), 66-104. The development (and deterioration) of satisfaction and com­ Main, M., & Solomon, J. (1990). Procedures for identifying infants mitment in heterosexual involvements. Journal of Personality as disoganized/disoriented during the Ainsworth Strange Situa­ and Social Psychology, 45, 101-117. tion. In M. T. Greenberg, D. Cicchetti, & E. M. Cummings Rusbult, C. E., & Zembrodt, I. M. (1983). Responses to dissatisfac­ (Eds.), Attachment in the preschool years (pp. 121-160). Chi­ tion in romantic involvement: A multidimensional scaling anal­ cago: University of Chicago Press. ysis. Journal ofExperimental Social Psychology, 19, 274-293. Marris, P. (1958). Widows and their families. London: Routledge & Rusbult, C. E., Zembrodt, I. M., & Gunn, L. K. (1982). Exit, voice, KeganPaul. loyalty, and neglect: Responses to dissatisfaction in romantic Mikulincer, M., Florian, V., & Tolmacz, R. (1990). Attachment involvements. Journal of Personality and Social Psychology, styles and fear of personal death: A case study of affect 43, 1230-1242. regulation. Journal ofPersonality and Social Psychology, 58, Schachter, S. (1959). The psychology of affiliation: Experimental 273-280. studies ofthe sources ofgregariousness. Stanford, CA: Stanford Mikulincer, M., & Nachshon, 0. (1991). Attachment styles and University Press. patterns of self-disclosure. Journal of Personality and Social Schaffer, H. R., & Emerson, P. E. (1964). The development of social Psychology, 61, 321-331. attachments in infancy. Monographs ofthe Society for Research Miller, G. A., Galanter, E., & Pribram, K. H. (1960). Plans and the in Child Development, 29(3, Serial No. 94). structure of behavior. New York: Holt, Rinehart & Winston. Shaver, P.R., & Hazan, C. (1984). Incompatibility, loneliness, and Miller, S. 1., & Schoenfeld, L. (1973). Grief in the Navajo: Psycho­ "limerence." In W. Ickes (Ed.), Compatible and incompatible dynamics and culture. International Journal ofPsychiatry, 19, relationships(pp. 163--184). New York: Springer-Verlag. 187-191. Shaver, P. R., & Hazan, C. (1993). Adult romantic attachment: Mizukami, K., Kobayashi, N., Ishii, T., & Iwata, H. (1990). First Theory and evidence. In D. Perlman & W. Jones (Eds.), Ad­ selective attachment begins in early infancy: A study using vances in personal relationships (Vol. 4, pp. 29--70). Green­ telethermography. Infant Behavior and Development, 13, wich, CT: JAI. 257-271. Shaver, P.R., Hazan, C., & Bradshaw, D. (1988). Love as attachment: Murstein, B. I. (1976). Who will marry whom? New York: Springer. The integration of three behavioral systems. In R. J. Sternberg Newcomb, T. M. (1961). The acquaintance process. New York: Holt, & M. L. Barnes (Eds. ), The psychology oflove (pp. 68--99). New Rinehart & Winston. Haven, CT: Yale University Press. Owen, M. T., Easterbrooks, M.A., Chase-Lansdale, L., & Goldberg, Shaver, P.R., & Rubenstein, C. (1980). Childhood attachment expe­ W. A. (1984). The relation between maternal employment status rience and adult loneliness. In L. Wheeler (Ed.), Review of and the stability of attachments to mother and father. Child personality and social psychology (Vol. 1, pp. 42-73). Beverly Development, 55, 1894-1901. Hills, CA: Sage. Palgi, P. (1973). The socio-cultural expressions and implications of Simpson, J. A. (1990). The influence of attachment styles on romantic death, mourning and bereavement arising out of the war situa­ relationships. Journal ofPersonality and Social Psychology, 59, tion in Israel. Israel Annals ofPsychiatry, 11,301-329. 971-980. Parkes, C. M., & Weiss, R. S. (1983). Recovery from bereavement. Skolnick, A. (1978). The intimate environment: Exploring marriage New York: Basic. and the family (2nd ed. ). Boston: little, Brown. Peplau, L. A., & Gordon, S. L. (1983). The intimate relationships of Snyder, M., & Swann, W. B., Jr. (1978). Behavioral confirmation in lesbians and gay men. In E. R. Allgeier & N. B. McCormick social interaction: From social perception to social reality.Jour­ (Eds.), The changing boundaries: Gender roles and sexual nal ofExperimental Social Psychology, 14, 148--162. behavior (pp. 226--244). Palo Alto, CA: Mayfield Spanier, G. B., Lewis, R. A., & Cole, C. L. (1975). Marital adjustment Peplau, L. A., & Perlman, D. (Eds.). (1982). Loneliness: A over the family life cycle: The issue of curvilinearity. Journal sourcebook ofcurrent theory, research and therapy. New York: ofMarriage and the Family, 37, 263--275. Wiley-Interscience. Sroufe, L. A. (1983). Infant-caregiver attachment and patterns of Piaget, J. (1952). The origins of intelligence in children. New York: adaptation in preschool: The roots of maladaptation and compe­ International Universities Press. tence. In M. Perlmutter (Ed.), Minnesota Symposium on Child Rands, M., Levinger, G., & Mellinger, G. D. (1981). Patterns of Psychology (Vol. 16, pp. 41--83). Hillsdale, NJ: Lawrence conflict resolution and marital satisfaction. Journal of Family Erlbaum Associates, Inc. Issues, 2, 297-321. Sroufe, L.A., & Waters, E. (1977). Attachment as an organizational Reedy, M. N., Birren, J. E., & Schaie, K. W. (1981). Age and sex construct. Child Development, 48, 1184-1199. differences in satisfying love relationships across the adult life Stanley, S. M. (1986). Commitment and the maintenance and en­ span. Human Development, 24, 52~. hancement of relationships. Unpublished doctoral dissertation, Reis, H. T., & Shaver, P. R. (1988). Intimacy as an interpersonal University of Denver. process. In S. Duck (Ed.), Handbook of research in personal Steinberg, L., & Silverberg, S. B. (1986). The vicissitudes of auton­ relationships(pp. 367-389). London: Wiley. omy in early adolescence. Child Development, 57, 841--851. Reiss, I. L. (1960). Toward a sociology of the heterosexual love Sternberg, R. J. (1986). A triangular theory of love. Psychological relationship. Marriage and Family Living, 22, 139--145. Review, 93, 119-135. Rubenstein, C., & Shaver, P.R. (1982). In search of intimacy. New Swann, W. B., Jr., Hixon, J. G., & De La Ronde, C. (1992). Embrac­ York: Delacorte. ing the bitter "truth": Negative self-concepts and marital com­ Rubin, Z. (1973). Liking and loving. New York: Holt, Rinehart & mitment. Psychological Science, 3, 118--121. Winston. Tannen, D. (1990). You just don't understand: Women and men in Rubin, Z. (1980). Children'sfriendships. Cambridge: Harvard Uni­ conversation. New York: Morrow. versity Press. Taylor, D. A., & Altman, I. (1987). Communication in interpersonal Rubin, Z. (1984). Toward a science of relationships. Contemporary relationships: Social penetration processes. In M. Roloff & G. Psychology, 29, 856--858. Miller (Eds.), Interpersonal processes: New directions in com­ Rusbult, C. E. (1980). Commitment and satisfaction in romantic munication research (pp. 257-277). Newbury Park, CA: Sage. associations: A test of the investment model. Journal ofExper­ Tennov, D. (1979). Love and limerence: The experience of being in imental Social Psychology, 16, 172-186. love. New York: Stein & Day.

21 HAZAN & SHAVER

Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of Wallerstein, J. B., & Kelly, J. B. (1980). Surviving the brtllllcup: How groups. New York: Wiley. children and parents cope with divorce. New York: Basic. Traupmann, J., & Hatfield, E. (1981). Love: Its effects on mental and Walster, E., Aronson, E., Abrahams, D., & Rottmann, L. (1966). The physical health. In J. March, S. Kiesler, R. Fogel, E. Hatfield, & importance of physical attractiveness in dating behavior. Jour­ E. Shana (Eds.),Aging: Stability and chonge in the family. New nal ofPersonality and Social Psychology, 4, 508-516. York: Academic. Walster, E., & Walster, G. W. (1978). A new look at love. Reading, Vaillant, G. E. (1977). Adaptation to life: How the best and the MA: Addison-Wesley. brightest came ofage. Boston: Uttle, Brown. Waters, E. (1978). The reliability and stability of individual differ­ van den Boom, D. (1990). Preventive intervention and the quality of ences in infant-mother attachment. Child Development, 49, mother-infant interaction and infant exploration in irritable 483-494. infants. In W. Koops, H. J. G. Soppe, J. L van der Unden, P. C. Weiss, R. S. (1973). Loneliness: The experience of emotional and M. Molenaar, & J. J. F. Schroots (Eds.), Developmental psy­ social isolation. Cambridge, MA: MIT Press. chology behind the dikes: An outline of developmental psy­ Weiss, R. S. (1975). Marital separation. New York: Basic. chology research in The Netherlands. The Netherlands: Weiss, R. S. (1982). Attachment in adults. In C. M. Parkes & J. Uitgeverij Eburon. Stevenson-Hinde (Eds.), The place of attachment in human Vaughn, B., Egeland, B., Sroufe, L A., & Waters, E. (1979). Indi­ behavior (pp. 171-184). New York: Basic. vidual differences in infant-mother attachment at 12 and 18 Weiss, R. S. (1988). Loss and recovery. Journal ofSocial1ssues, 44, months: Stability and change in families under stress. Child 37-52. Development, 50,971-975. Zajonc, R. B. (1968). Attitudinal effects of mere exposure. Journal Veroff, J., Douvan, E., & Kukla, R. A. (1981). The inner American: ofPersonality and Social Psychology Monograph Supplement, A self-portrait/rom 1957 to 1976. New York: Basic. 9(2, Pt. 2), 1-27.

22