Capacity for Intimate Relationships: a Developmental Construction

Home , Girlfriend

Reprinted from: Furman, W., Brown, B. B., & Feiring, C. (Eds.) (1999), The development of romantic relationships in adolescence (pp. 125-147). New York: Cambridge University Press.

Capacity for Intimate Relationships: A Developmental Construction

W. Andrew Collins and L. Alan Sroufe Institute of Child Development, University of Minnesota

Abstract An enduring hypothesis in the history of psychology is the notion that early caregiving experiences provide the prototype for later significant relationships. Although this hypothesis often has been extended to adult love relationships, compelling models of developmental processes that would account for continuity in relationship characteristics over so long a period are still needed. We propose a developmental view of the development of a capacity for intimacy in which distinct relational experiences are linked across time. Drawing on experiences and findings from a twenty-year longitudinal study of relationships and development, we illustrate evidence for intervening links during childhood and adolescence that eventually may predict the quality and significance of romantic relationships among young adults. We give particular attention to the importance of intervening links between early caregiving relationships and developing friendship capacities throughout middle childhood and adolescence. We then speculate about the links between the implications of different patterns of childhood closeness for adolescents' orientations to dating, romantic relationships, and transitory sexual liaisons. Finally, we propose criteria for future research that would advance understanding of the development of capacities for intimate relationships.

The advent of romantic relationships is a hallmark parents or other authority figures or social conven- transition of adolescence in both popular perceptions to nullify or force or otherwise render invol- tions and theoretical formulations. Images of the untary the selection of a romantic partner are sudden onset of preoccupation with the other, shy- viewed as inimical to the essential nature of ro- ness and self-consciousness, awkwardness in inter- mance. Romantic relationships also involve de- actions, and sexual awakening suffuse popular pendency, which is reciprocal between the partners, treatments of the topic. In developmental perspec- unlike the more asymmetrical dependency of child tive, however, romantic relationships are embedded on caregiver; and the reciprocal dependency of ro- in fundamental human motivations to form and mantic partners is likely to be both greater and maintain close relationships (Baumeister & Leary, more extensive than the reliance of friends upon 1995; MacDonald, 1992) and in a meaningful pro- one another. Finally, romantic relationships are gression of relational forms across the life course marked by an amalgam of love, passion, and actual (Ainsworth, 1989; Feeney & Noller, 1996; Furman or anticipated sexual activity. Although friendship & Wehner, 1994). Early caregiver-child relation- may be caring and passionate (see Diamond, Dube, ships, peer relationships in preschool and middle & Savin-Williams, this volume), adolescents' ro- childhood, and close mutual friendships in adoles- mantic relationships are likely either to involve sex cence all potentially contribute to the behavioral or to be the kind of relationship in which shared patterns and emotional orientations that mark a re- feelings of love and passion make sex a likely and lationship as romantic. appropriate possibility at some future time. Romantic relationships also are distinctive from The common relational threads that eventually these forerunners. Romantic relationships are vol- form the fabric of romantic relationships are ex- untary and symmetrical, in contrast to the kinship periences that support the development of inti- or legal bonds that commonly circumscribe care- macy. During preadolescence and adolescence inti- giving relationships. In Western literary and popu- macy becomes increasingly central to social com- lar portrayals of adolescent love, the attempts of petence, because the salient developmental task of Capacity for Intimate Relationships Egeland & this period is forming close mutual relationships. A Developmental Viewpoint on Intimacy Indeed, Reis and Shaver (1988) have defined intimacy as: Relationships of all types in all periods of life have several features in common. First, relation- ... an interpersonal process within ships are not simply the sum of personal character- which two interaction partners experi- istics of each member of the dyad; rather, relation- ence and express feelings, communicate ships consist of the unique patterning and qualities verbally and nonverbally, satisfy social of dyadic interactions that endure over time (Hinde motives, augment or reduce social & Stevenson-Hinde, 1987; Sroufe & Fleeson, fears, talk and learn about themselves 1986). Second, individuals and relationships are and their unique characteristics, and reciprocally related, such that individuals are both become "close".... (pp. 387-388). the products and the architects of the relationships in which they participate (Baldwin, 1911; Mead, Reis and Shaver further note that emotionally 1934; Sroufe, 1989; Sroufe & Fleeson, 1986). close interactions at all ages derive their signifi- Third, relationships are integral to competence, or cance not only from mutually self-disclosing be- the ability "to make effective use of personal and haviors, but from the experiences of feeling under- environmental resources to achieve a good devel- stood, validated, and cared for as a result of them. opmental outcome" (Waters & Sroufe, 1983, p. The interactions that serve these functions change 81). Good developmental outcomes are those developmentally. With parents, closeness, as ex- which lead to healthy adaptations during later de- pressed by cuddling and extensive joint interac- velopmental phases or, at least, that do not limit or tions, declines as children mature, but conversa- foreclose on important developmental changes tions in which information is conveyed and feelings (Elicker, Englund, & Sroufe, 1992; Sroufe & are expressed increase. With peers, closeness fol- Fleeson, 1986). lows an age-related pattern in which close relationships are increasingly defined in terms of mutual According to attachment theorists (e.g., Ains- caring and commitment, rather than merely patterns worth, Blehar, Waters, & Wall, 1978; Bowlby, of shared activities, which suffice as markers of 1973; Sroufe & Fleeson, 1986), earlier and later friendship at earlier ages (Hartup, 1992). These ad- forms of closeness reflect processes linking distinct aptations are appropriate adaptations to the maturity relational experiences across time. Capacities for level and changing needs of individuals (Collins, intimacy emerge through a transactive process, in 1996). which expectations concerning self and relationships and patterns of arousal modulation character- In this chapter, we consider what is known and istic of early relationships lead to particular forms what still must be learned about the common fea- of engagement with persons and objects. Other per- tures of these various forms of relating and how the sons commonly react in a complementary way, distinctive qualities of each contribute uniquely to thus perpetuating the pattern, albeit in new forms the development of capacities for intimate romantic and in new contexts, across developmental periods relationships. Drawing on experiences and findings (Sroufe, Carlson, & Shulman, 1993). from a nineteen-year longitudinal study, we examine precursors of intimacy in early caregiver-child We regard the capacity for intimacy as a classic relationships and in relationships with peers in developmental phenomenon. By this we mean both childhood. We then consider how experiences in that it has "emergent properties"--not being fully these relationships might be manifested in the con- specified by capacities that precede it--and that it text of normative changes in relationships with par- nonetheless evolves in a logical manner from pre- ents and with peers and, eventually, with romantic cursors through a series of transformations. At each partners during adolescence. In the final sections of phase of life, beginning in infancy, foundations are the chapter, we speculate about how experiences laid down that support the capacity for intimacy during childhood and adolescence may serve as with peers in adolescence. Given the non-linear precursors to entry into, and growth-enhancing ex- nature of development, these foundations include periences in, romantic relationships, and we outline experiences with parents, as well as experiences needs for additional research to fill gaps in this lit- with peers. At the same time, the self-disclosure erature. and sexual intimacy of adolescence are qualitative

2 Capacity for Intimate Relationships Egeland & advances, non-obvious derivatives of earlier behav- the preschool and middle-childhood periods. We ior. Certainly, nothing one can see in infancy would give particular attention to alternative possibilities directly forecast such capacities. Only within an for "carrying forward" the experiences of early epigenetic view can links between infant or early caregiving relationships to later voluntary relation- peer experience and later adolescent intimacy be ships with age mates. discerned. Early Relationships and the Capacity for Closeness The multifaceted nature of adolescent intimacy Theoretically, relationships in infancy contrib- helps to underscore the importance of a develop- ute to three components of closeness and, ulti- mental viewpoint. Intimacy involves complex moti- mately, of intimacy. First, relationships with care- vational, emotional, and behavioral aspects. One givers, when based on a history of availability and must first be oriented toward closeness, to value it responsiveness, should lead to positive expectan- and to seek it. Closeness is an active, creative proc- cies about interactions with others. Because of the ess. Second, one must be able to tolerate, and even caregiver's key role in comforting and other aspects embrace, the intense emotions that are inextricably of affect regulation, the "other" will be valued and part of intimacy and be able to freely share such appraised in parallel to the caregiver's treatment of emotional experiences. Finally, one must be capa- the child. In Bowlby's (1973) terms, children de- ble of self disclosure, mutual reciprocity, sensitivity velop internal working models of self and others to the feelings of the other, and concern for the that guide them toward similar interactions with other's well-being. others. Second, such relationships provide a con- Although this totality is a unique achievement text for learning reciprocity, even though only the of adolescence, foundations for each of these three more mature (parental) partners can purposefully aspects of intimacy are to be found in earlier devel- fit their behavior to the child's actions. Further, par- opment. Some have argued that certain pre-requi- ticipating in a relationship with an empathic, re- site capacities are drawn primarily from experi- sponsive caregiver affords learning the very nature ences with parents, while others are drawn primar- of empathic relating ( Sroufe & Fleeson, 1986), ily from peers (e.g., Furman & Wehner, 1994). which can later be applied in more symmetrical There is persuasive logic underlying this analysis. relationships (e.g., with peers, romantic partners). Our own view emphasizes the integrative nature of Third, through a history of responsive care and family and peer experience. Closeness with friends, support for autonomy within the relationship, the and reciprocity and conflict resolution between child develops a sense of self-worth and efficacy. equals, must certainly be practiced and mastered Feelings that the self is worthy of respect and care within the peer group; but successful negotiation of underlie characteristics that are likely to be attrac- these issues is dependent upon a history of emo- tive to future partners (e.g., self-confidence, curios- tional closeness with caregivers and ongoing paren- ity, enthusiasm, and positive affect) (Elicker et al., tal support. The positive experiences with peers that 1992) and orient the child to expecting and accept- are so important in channeling individuals toward ing certain kinds of reactions. the capacity for intimacy are more likely for those Bowlby's (1973) notion of working models ex- with supportive family experiences. Although posi- plicitly includes the possibility that patterns of feel- tive peer experiences could be rehabilitative for ings, expectations, thoughts, and behaviors can those with unsupportive parenting histories, such change with experience. Researchers since Bowlby children are the very ones who are least likely to nevertheless have found impressive relations be- have them. tween early caregiver-child relationships and later Our view is epigenetic. The capacity for inti- key relationships and between both these earlier macy evolves through a series of phases, each and later patterns and friendships in which inti- building on the preceding ones. Consequently, our macy can occur in adolescence (Elicker et al., discussion focuses on early attachment relation- 1992; Shulman, Elicker, & Sroufe, 1994). In the ships and early peer experiences, as well as the rest of this section, we outline the evidence for friendships of middle childhood which are the im- these links, drawing from longitudinal findings mediate precursors of intimacy. In the following from our 19-year study of individual adaptation sections, we address how early relationships with and from other sources (e.g., Shulman et al., 1994; caregivers might provide a basis for establishing Sroufe & Egeland, 1991; Sroufe et al., 1993). developmentally appropriate close relationships in

3 Capacity for Intimate Relationships Egeland &

Attachment History and Emerging Closeness in Links to peer relationships. Children who mani- Peer Relationships fested these three contrasting patterns during infancy also later showed striking differences in peer Our view of developmental precursors to ro- relationships. During preschool, the salient issue is mantic relationships has been shaped by experi- to master challenges and problems using their own ences in the Minnesota Parent-Child Project, in resources. Because other persons are potentially which 190 first-born individuals have been studied valuable resources in meeting common challenges, since the third trimester of the mother's pregnancy. effectiveness in relating to adults, especially (See Egeland & Brunnquell, 1979, for an early re- non-caregivers, and with peers is fundamental to port.) Infants and/or mothers were seen seven times competence. In the Minnesota sample, children in the child's first year, twice in each of the next 3 with secure histories indeed demonstrated greater years, and yearly though grade 7. Assessments in- competence by participating more actively in the cluded neurological status, motor, cognitive, and peer group, manifesting more positive affect and intellectual development, maternal personality and less negative affect in their encounters than inse- IQ, parent-child interaction, temperament, peer re- curely attached children (Sroufe, 1983; Sroufe, lationships, personality development, and contex- Schork, Motti, Lawroski, & LaFreniere, 1984). Se- tual variables such as life stress and social support. cure children also were more popular. By contrast, Children were observed in home, laboratory, and children with anxious-avoidant attachment history school. During childhood and adolescence, all par- were not only significantly less competent in all of ticipants were seen at ages 13, 16, and 19. In addi- these respects, but were more aggressive in the tion, subsamples of these participants have been classroom. Those with anxious/resistant attachment studied intensively in a semester-long nurs- were easily frustrated and oriented toward their ery-school program, a four-week summer camp at teachers at the expense of engaging peers (Pancake, age 10, and observations in a weekend retreat at age 1985). 15. These subsamples were representative of the sample at large. Early attachment history also forecast differences in qualities of interpersonal relationships dur- The primary measure of early relationships was ing preschool, at times with extraordinary specific- the Strange Situation procedure (Ainsworth et al., ity. Children with early histories of secure attach- 1978). This 20-minute procedure involves a series ment displayed greater reciprocity and dealt more of episodes with primary caregiver and infant in a effectively with conflicts in interactions with pre- playroom. The two are joined by a stranger who school peers (Liberman, 1977; Suess, 1987). When ultimately initiates a brief interaction with the baby, we focused on specific pairs of children who followed by two brief separations and reunions with played together frequently, pairs containing at least the caregiver, with the infant left entirely alone dur- one avoidant member formed relationships that ing the second separation. Based on behavior in the were less deep (less characterized by mutuality, Strange Situation, Ainsworth identified secure at- responsiveness, and affective involvement) and tachment with use of the caregiver as a secure base more hostile than the other pairs (Pancake, 1985). for exploration and a pattern of emotional responses In addition, of 19 dyadic relationships in the sub- to separation and reunion that indicates confidence sample, five involved "victimization," a repetitive in the accessibility and responsiveness of the care- pattern of physical or verbal exploitation or abuse giver. In a second, contrasting pattern, termed by one child of the other (Troy & Sroufe, 1987). In "anxious/resistant attachment," infants have diffi- each case, the "exploiter" was a child with an culty exploring even when the caregiver is present; avoidant history, and the victim was another anx- moreover, they both become quite upset by the iously attached child (avoidant or resistant). Such a separation episodes and show great difficulty set- pattern was observed every time such a pairing oc- tling upon reunion, even when in contact with the curred. Children with secure histories were never caregiver. In the third pattern, called "anxious/ victimizers or victims. Thus, by preschool distinc- avoidant attachment," children commonly show tive relational patterns among children with differ- little distress during the separations, and upon reun- ent attachment histories clearly extend beyond the ion they ignore, turn or move away from, or show confines of the family to the voluntary affiliations abortive approaches to their caregivers. Earlier, we of classrooms and play arrangements. outlined several reasons why secure attachments likely would be more conducive to the subsequent The patterns extend to teachers as well as peers development of a capacity for intimacy than the (Sroufe & Fleeson, 1988). Teachers were judged to two insecure patterns.

4 Capacity for Intimate Relationships Egeland & be warm and straightforward in their engagement of were part of a larger group, but they also freely in- children with histories of secure attachment, to hold teracted with others. Other individuals and group out age-appropriate standards for them, and to ex- activities were not a threat to their relationship. pect self-direction and compliance with classroom Their friendship was apparent in group settings, yet rules. With children from the resistant group, teach- they freely interacted with others. In contrast, ers were unduly nurturant and caretaking, were con- avoidant-avoidant pairs exhibited very exclusive trolling, had low expectations for compliance, and relationships. They often were physically separate were quite tolerant of minor violations of classroom from the others, seldom participated in voluntary rules. With children with avoidant histories, teach- groups, rarely interchanged with other individuals, ers behaved in a controlling and, at rare times, even and showed jealousy regarding each other. When angry manner and displayed low expectations for either child was absent, the remaining partner had compliance, although they did not accept viola- difficulty participating socially. Yet a third pattern tions. For their part, avoidant children often en- characterized resistant-resistant pairs. These friends gaged in hostile or defiant behavior that would had difficulty sustaining their relationships, be- alienate teachers as well as other children. These cause one of two often was absorbed by the group, observations provide further evidence that the re- thus separating from the other. Children in these jecting relationships of the avoidant child's early pairs showed little loyalty to each other; conse- years were recapitulated in the preschool class- quently, they were unreliable as a resource for each room. other. Their relationship was neither the detriment to group functioning that avoidant-avoidant rela- In late middle childhood (roughly ages 10-11), tionships were nor a base for effectiveness in the children face additional demands in relationships group provided by secure-secure relationships. with peers. Normatively, the prospect of future intimacy is enhanced by increased capacity to form Securely attached children also showed greater close, mutual friendships (e.g., Bigelow, 1977; social competence by adhering to peer-group Bigelow and LaGaipa, 1975; Furman & Bierman, norms. In middle-childhood children are expected 1984; Selman, 1980; Selman & Schultz, 1989). to favor interaction with same-gender peers Such friendships are distinctive in three ways: they (Maccoby, 1990). In the Minnesota sample, only are more qualitatively unique than earlier friend- 10% of observed interactions between 10-11 year ships; they are strongly differentiated in terms of olds were with peers of the opposite gender; and gender and in terms of depth of friendship; and they these exceptions almost always involved multiple often are emotionally deep. By middle childhood boys and multiple girls rather than a solitary boy or the expectations of earlier periods have become girl with a member of the other gender. Moreover, more elaborated by normative changes in such many of the interactions were accompanied by dis- competencies as enhanced role-taking, more exten- avowal (what Barrie Thorne [1986] calls sive communicative skills, and more sophisticated "borderwork") or by a "cover" which legitimized understanding of providing nurturance and reassur- the contact (e.g., boys frequently hurling insults ance (Barnett, King, Howard, & Dino, 1980; rather than expressions of interest at the same Hartup, 1984; Waters, Kondo-Ikemura, Posada, & group of girls; or interacting with a child of another Richters, 1991). gender because an adult has directed it). Children who violated gender boundary rules also generally These expected patterns are clearly apparent in showed lower social competence and less likeli- 10-11 year olds in the Minnesota longitudinal sam- hood than other children to have one or more ple. Children who had been secure in their attach- friends in the group. Longitudinally, gender bound- ments at 12 to 18 months of age were more likely to ary violation was associated with a history of anx- form a friendship than those who had been inse- ious attachment and with earlier observed interac- curely attached. Moreover, children who had been tions with parents in which parent or child or both securely attached in infancy tended to form friend- had shown peer-like behavior toward the other. ships with children who also had secure histories. Thus, paradoxically, maintaining separateness from Although this might be attributable to a natural at- the other gender in one developmental period may traction among competent children, more intensive well be a forerunner of adaptive relating in later examination showed that the qualities of these periods. For example, normative social separation friendships also were consistent with attachment of genders in middle childhood probably facilitates history (Shulman et al., 1994). Secure-secure "practice for intimacy" with peers that is relatively friendships clearly were apparent when the children

5 Capacity for Intimate Relationships Egeland & free of overtones of sexuality (e.g., Sullivan, 1953). intention. Likewise, independent ratings of the chil- These self-imposed norms parallel arrangements in dren's "positive expectations regarding peers," some cultures in which females and males are sepa- based on responses to TAT-like card and sentence rated by social structures and taboos against casual completions at age 11, significantly discriminated contact. attachment groups. Those with secure histories more frequently told stories in which peers cooper- Links to component skills for intimacy. Attach- ated and conflict was resolved and conlcuded sen- ment history also may contribute to capacities for tence items (e.g., "most kids...") with positive re- intimacy by supporting the development of specific sponses ("...like to play with me"). In short, key component or constituent skills. Theoretically, se- components of the internal working models de- curely attached infants, having experienced and scribed by Bowlby (1973) are linked to attachment therefore internalized a responsive relationship, history and are evinced in connection with peer later should be more empathic. Experiencing secure interactions in later periods. relationships should lead children to expect empathy from others, to believe that this is characteristic Relations among relationships in longitudinal of relationships, and thus to a tendency to be em- perspective. Middle-childhood peer competence is pathically responsive to others when cognitive ad- linked to the salient relationships of both infancy vances allow sufficient perspective taking (Elicker and early childhood. In the Minnesota data, secu- et al., 1992). In the Minnesota sample, both pre- rity of attachment in infancy strongly predicted school teachers' judgments of empathy and video- preschool characteristics of self-reliance, effective taped records from the classroom revealed more peer relationships (including empathy and affective empathic behavior by children with secure histories engagement), and positive relationships with teach- than children with anxious histories (Kestenbaum, ers (e.g., Sroufe, 1983: Sroufe et al., 1984). In turn, Farber, & Sroufe, 1989; Sroufe, 1983). Moreover, both quality of infant attachment and quality of children with avoidant histories, who are presumed preschool functioning subsequently predicted mid- to experience chronic rebuffs to their expressed dle childhood relationships, including capacity for needs, were significantly more likely than both forming close friendships, conflict resolution skill, other groups to show "anti-empathy" (behavior that and general effectiveness in the same-gender peer would make another person's distress worse; e.g., group (Elicker et al., 1992; Shulman et al., 1994). taunting a crying child); whereas those with resis- These findings carry two implications for thinking tant histories behaved as though the distress were further about the developmental precursors of their own, blurring the boundary between self and emerging romantic relationships in adolescence. other. One is that children actively contribute to the character of their relationships based upon their history One likely component of these differences is of experiences in earlier relationships. The second interpersonal cognitions or representations relevant is that these later relationships contribute signifi- to peer relationships. Comparing children who had cantly beyond the impact of early experiences to been securely or anxiously attached as infants, the subsequent development of capacities for form- Rosenberg (1984) found that those in the anxious ing and maintaining relationships. group were less likely to incorporate people into fantasy play and that their fantasized resolutions for Alternative Views of "Carry-Forward" Mechanisms misfortunes or interpersonal conflicts were less The impressive continuities in relationships out- likely to be positive. Such children may be less in- lined above may occur for a variety of reasons. terested in interpersonal relationships, may value Some explanations emphasize persistence of initial them less highly, or may expect negative outcomes cause, arguing either that an endogenous child trait in relationships, whereas securely attached children continues to be manifest (Kagan, Reznick, & Gib- may regard relationships more positively. Attach- bons, 1989) or that environmental influences re- ment history also appears to be related to distinctive main constant and account for the appearance of cognitive biases with respect to peers. Suess (1987; stability in child behavior (Lamb, 1984). Others Suess, Grossmann, & Sroufe, 1993) found that se- argue that continuity results from both prior adap- curely attached children usually made realistic attri- tation and current environment (e.g., Bowlby, butions or displayed a bias toward attributing be- 1973; Lewis, 1989). This latter view implies an nevolent intentions, whereas children with anx- interactional model, in which environmental ious-avoidant attachment manifested more unrealis- changes may have differential influences depend- tic or hostile/negative biases in their attributions of

6 Capacity for Intimate Relationships Egeland & ing on previous adaptation. 1981). Adolescents more often perceive reciprocity with and acceptance by parents than children do Our own view of continuity (Collins, 1995; (for reviews, see Collins, 1995, 1996; Collins & Sroufe, 1979; Sroufe & Fleeson, 1988; Sroufe, Ege- Repinski, 1994; Collins & Russell, 1991; Laursen land, & Kreutzer, 1990) embodies a transactional & Collins, 1994; Holmbeck, 1996; Holmbeck, Pai- process, whereby children with particular patterns koff, & Brooks-Gunn, 1995; Montemayor, 1983; of adaptation and expectations both assimilate and Steinberg, 1990). Transitory disruptions and accommodate to new circumstances. This transac- changes in the relative balance of positive and tional view is supported by two key points from the negative emotional expressions in parent-adoles- Minnesota longitudinal findings. First, while qual- cent relationships help to realign expectations in ity of care is indeed stable (Pianta, Sroufe, & Ege- response to developmental changes in adolescents, land, 1989), early adaptation or experience predicts while preserving affectional bonds (Collins, 1995, later behavior even after accounting for contempo- 1996). rary environmental influence both in childhood and in adolescence (Sroufe et al., 1990, Sroufe, 1995). Normative patterns in peer relationships also Second, later environmental influences are not in- support continuity in the development of intimate dependent of prior adaptation. For example, treat- functioning. Although specific friendships are not ment of children by both preschool teachers and typically stable throughout adolescence, relation- classmates is predicted by patterns of adaptation in ships with a "best" friend are more often stable than infancy and is in accord with earlier family experi- unstable over the course of a school year (Berndt, ences (Sroufe, 1983; Sroufe & Fleeson, 1988, Troy Hawkins, & Hoyle, 1986; Berndt & Hoyle, 1985; & Sroufe, 1987). Thus, in our view, environment is Berndt & Keefe, 1995; for a review, see Savin-Wil- not simply an independence force in interacting liams & Berndt, 1990). As networks expand and with child characteristics. Children, in part, create diversify, additional opportunities for expressing their own environments through differential en- and experiencing intimacy become available. Re- gagement and reaction based upon history of ex- lationships with friends, romantic partners, and perience. This partially created environment feed family members serve overlapping, but distinctive, back on adaptation in an on-going process. Early functions, and typical exchanges within each of relationships continue to account for unique vari- these types of dyads differ accordingly (Furman & ance in later relationships and, by implication, are Buhrmester, 1985, 1992; Furman & Wehner, 1994; likely to do so with respect to romantic relation- for reviews, see Collins, 1996; Collins & Repinski, ships in adolescence. 1994; Laursen & Collins, 1994; Savin-Williams & Berndt, 1990). Relationships with parents are re- Peer Relationships During Adolescence ported to be primary sources of support for children Adolescents experience a wider and more di- at fourth grade, but parents are viewed as less im- verse network of social relationships than children portant than same-sex friends as sources of social do. Romantic relationships thus emerge as part of a support in early and middle adolescence and less complex balancing of loyal friendships, intimate important than same-sex friends and romantic part- pair bonds, same-gender group affiliations, and ners at college age (Furman & Buhrmester, 1992). mixed-gender group associations (Sroufe, Egeland, By late adolescence, the functions of different & Carlson, in press). types of relationships are well differentiated. Per- Despite this complexity, the normative patterns ceptions of intimacy in cross-gender relationships and social structures of adolescence are conducive increase with age during early and middle adoles- to continuity between closeness in the relationships cence, with reported intimacy between close fe- of childhood and later romantic relationships male-male pairs at age 16 matching the level of (Collins, 1996; Collins & Repinski, 1994). In par- intimacy perceived in female-female friendships ent-adolescent relationships, popular stereotypes (but not that reported for male-male friendships) notwithstanding, surveys in European and North (Sharabany, Gershoni, & Hoffman, 1981). Peer American samples consistently reveal that parents relationships may serve functions of socialization and adolescents alike perceive their relationships for relations among equals and also satisfaction of with one another as warm and pleasant. Of the 20% affiliative needs, but romantic relationships may be or so of families that encounter serious difficulties equally or more important for mutual sharing and in this period, most have had a history of earlier emotional gratification, especially in late adoles- problems (Offer, 1969; Offer, Ostrov, & Howard, cence and early adulthood (Collins, 1996).

7 Capacity for Intimate Relationships Egeland &

Within the normative adaptations, individuals Quality of attachment in infancy and quality of show strikingly stable tendencies in the degree and caregiver-child interaction at age 13 (described in a nature of closeness. Individual variations in social later section) also were related to adolescent peer acceptance and in capabilities for forming and relationship measures. Infant and early adolescent maintaining friendships are among the best docu- measures together were especially strongly related mented of all developmental continuities (for re- to reunion assessments, with correlations in views, see Hartup, 1993, 1996; Parker & Asher, the .50s (Sroufe, et al., in press). The strongest rela- 1987; Parker & Gottman, 1989). Such continuities tion between attachment history and adolescence have repeatedly been linked to early and concurrent was for the "capacity for vulnerability" scale, relationship experiences (e.g., Dishion, Patterson, which clearly is in accord with attachment theory. & Griesler, 1994; Rubin, LeMare, & Lollis, 1990), Moreover, each of the eight participants involved implying that social competences transcend particu- in a couple relationships during the reunion had lar relationships (e.g., Dishion et al., 1994; Parker histories of secure attachment, a highly significant & Asher, 1987; Sroufe & Fleeson, 1986). finding, given that only half of the 41 participants had been secure. Those with secure histories were Data from our longitudinal study support the also rated higher on "leadership" and overall level developmental underpinnings of adolescent peer of competence in a revealed differences group dis- competence in general and, in a preliminary way, cussion situation; they also were significantly more the capacity for intimacy in particular. Both peer frequently elected as spokesperson for their group experiences and family experiences are strongly (Englund, Hyson, & Levy, in preparation). These predictive of individual differences in adolescence. correlations thus link initial attachment assess- We have found significant continuity in peer com- ments to indicators of relationships taken 14 years petence based on teacher ratings from early elemen- later. tary school (and, for a subsample of participants, from pre-school) through age 16 (Sroufe, et al., in Attachment history was also related to cogni- press). Even stronger evidence has come from the tive measures of adolescent experience. When re- sub-sample of individuals whom we have studied union participants were interviewed concerning intensively in nursery school and summer camps. their knowledge of the peer-group relationship When those children who had been assessed as structure, those having secure histories demon- competent at earlier ages were observed at a week- strated superior knowledge and perceptiveness. In end re-union, camp counselors rated them as more interviews conducted with all of our participants at competent and observed that they participated more age 16, a friendship intimacy index, based on actively in the group. The findings were most strik- "closeness" and "coherence of discourse" scales, ing for assessments keyed to age salient issues. For was significantly related to history of secure attach- example, we rated behavior at the reunion accord- ment. (Interview information concerning dating ing to a scale of "capacity for vulnerability." This will be discussed below.) scale was created especially to tap teens' abilities to Implications of Relationship History for Adole- participate in the range of reunion activities, includ- scent Intimacy and Romantic Relationships ing those in which ego-salient feelings would arise (such as engaging members of the other gender at Despite the stereotype of adolescence as "the the evening party). Scores on this scale were related age of sexual attraction and emergent love" (Zani, to competence indices at earlier ages. Correlations 1993), research on the transition to romantic rela- were especially strong with an observational- tionships during adolescence is in its infancy. Little ly-based intensity of same-gender friendship score evidence exists regarding specific romantic experi- in middle childhood. Scores also were correlated ences during adolescence. Consequently, there is with the gender boundary maintenance rating from little basis for specifying how experiences may be that period. Finally, for girls, both preschool and linked to relationship history. Some speculations middle childhood assessments, and especially the are possible, however, and may serve as an impetus friendship score (r=.64), were related to an inter- toward additional research. view based measure of "friendship intimacy" at age 16 (see Ostoja, 1996). (This interview format may In this section, we consider three phenomena not have been adequate for our male participants at that are commonly considered part of the develop- this age.) ment of romantic relationships: dating; being involved in a relationship that is perceived as

8 Capacity for Intimate Relationships Egeland &

"romantic" (i.e., going steady; believing that one is Several contrasts among the peer relations of in a relationship with at least some long-term po- children with different attachment histories proba- tential); and becoming sexually active. For each of bly forecast their dating experiences. Children with these topics, we address three key questions: (1) secure attachment histories have consistently been What is known about normative patterns for this found to be highest in popularity with peers, mas- form of romantic or proto-romantic involvement? tery of social skills, and positive engagement in (2) To what degree and in what ways might varia- peer-group activities. Attachment theory implies tions in capacity for intimacy be manifested within that these characteristics likely reflect relationships these normative patterns? (3) How, in turn, might with caregivers that foster positive expectancies variations in relationship history be associated with about interactions with others and a sense of variations in the phenomena of adolescent relation- self-worth and efficacy. ships? These characteristics of individuals with secure Dating histories carry several implications for dating. One is that early adolescents with such histories are Dating usually stems from involvement with relatively likely to be affiliated with crowds and social crowds, and thus is frequently no more than thus have ready support and social "cover" for dat- transitory and/or opportunistic affiliation, with no ing (Brown, Eicher, & Petrie, 1986; Dunphy, anticipation of the longer-term involvement or sex- 1963). As extensive involvement with crowds di- ual activity that mark romantic relationships. Dat- minishes normatively in middle and later adoles- ing typically begins in junior high school (for re- cence, the stage is set for securely attached indi- views, see Savin-Williams & Berndt, 1990; Zani, viduals to move smoothly toward increasingly inti- 1993). Although often attributed to hormonal mate relationships with smaller groups of friends changes at puberty, dating in early adolescence ac- and romantic partners. A second implication is that tually appears to be governed largely by age-graded securely attached individuals enter adolescence social expectations (e.g., Dornbusch, Carlsmith, with relatively high self-esteem, which is generally Gross, Martin, Jennings, Rosenberg, & Duke, correlated with involvement in dating (Long, 1983, 1981). Roscoe, Diana, and Brooks (1987) reported 1989; Samet & Kelly, 1987). High self-esteem that early and middle adolescents (i.e., 6th to 11th may support appropriate assertiveness and graders) in the midwestern U.S. say that they date self-confidence with potential dating partners and as a form of recreation, to establish a special rela- also may protect against negative emotional effects tionship with another person, and to gain status of such common experiences as rejection and com- with their peers. In contrast, college students gave petition (Mathes, Adams, & Davis, 1985). Third, greater emphasis to intimacy, companionship, and securely attached children's generally higher levels socialization to relationships as reasons for dating. of general social skills, popularity, and perceived Although little is known about dating among gay, social competence probably portend with smooth, lesbian, and bisexual adolescents, social benefits on-time transitions from primarily same-gender to such as enhanced peer status are probably less pow- more extensively cross-gender social relationships erful inducements for them, whereas a desire for during adolescence (e.g., Coleman & Hendry, intimacy may be a relatively greater motivation 1990; Miller, 1990). Finally, those with secure at- than in heterosexual dating couples. tachment histories will be more oriented toward the Historically, dating was highly ritualized and emotional depth that comes from ongoing, more governed by extensively prescribed social expecta- durable relationships. tions, and that is still true in some cultures. Among Preliminary data from our study support some U.S. adolescents, however, dating increasingly is an of these theoretical links. In our camp reunions, informal activity that often is carried out in connec- crowd-like phenomena occurred; that is, a defined tion with group activities and is marked by rela- group emerged that included couples and other tively superficial interactions between the partici- children of both genders who consistently inter- pants (Douvan & Adelson, 1966; Miller & Gordon, acted across the days of the reunion. This was 1986). Whether or not an adolescent dates and documented in independent sociograms made by when dating begins are probably more highly re- each counselor, which showed remarkable concor- lated to general social skills and acceptance by dance regarding membership in the crowd. In the peers than to a capacity for intimacy. one camp where this phenomenon was most clear,

9 Capacity for Intimate Relationships Egeland & all eight members of the crowd had secure histories can lead to increased isolation from the group and (vs. none of six with anxious attachment histories; difficulty in balancing the demands of couplehood Sroufe, Carlson & Shulman, 1993). with those of group activities and relationships (Surra, 1990; Zani, 1993). Relationship history Dating interviews with the total sample at age likely sets the stage for coping with these inevitable 16 revealed that participants with histories of anx- tensions. Even in middle childhood, anxiously at- ious-resistant attachment were significantly less tached pairs had notably greater difficulty in main- likely to have dated. This is consistent with the his- taining this balance than securely attached pairs. tory of social immaturity of these children docu- Anxious-avoidant pairs, for example, often isolated mented above. Those with secure and avoidant his- themselves from the group and rarely interacted tories at this age were similarly active in dating. with other individuals; and resistant-resistant pairs Those with secure histories, however, were signifi- tended to fragment, with one or both partners being cantly more likely to have consistently dated the absorbed separately by the group (Shulman et al., same person for three months, in accord with their 1994). theoretically predicted orientation toward depth and intimacy. Attachment history thus is probably linked to dating experiences in ways that are similar to its Relationship history also may provide clues association with other aspects of peer affiliation. about the likely selection of dating partners. Like During adolescence, availability of same-gender the social context in which dating is embedded, se- associates decreases as more and more couples are lection stems in predictable ways from the salient formed, and this increases pressure on the norms social motives of different phases of adolescent de- for selecting dating partners. The processes govern- velopment. Early adolescents, consistent with their ing these selections for one adolescent compared to emphasis on the social activity and status benefits another, however, probably changes little from the of dating, place relatively greater emphasis than processes that govern friendship selection during older adolescents to superficial features of potential childhood and preadolescence. partners (e.g., fashionable clothes) and approval by others (e.g., well liked by peers). In contrast, late Romantic Relationships adolescents give more weight to personality charac- Romantic relationships, in contrast to dating teristics (Roscoe et al., 1987). relationships, are marked by a higher level of com- These normatively preferred attributes (e.g., mitment by both partners (Diamond et al., this vol- Zani, 1993) may be especially influential in early ume). According to contemporary norms in the U. adolescence, when social expectations govern much S., romantic relationships may involve sexual inter- data behavior. Findings from our studies of friend- course or, if not, sexual relations are likely either to ship pairings in childhood lead us to believe, how- be anticipated in the future or to be actively de- ever, that individual differences in preferences are layed by explicit agreement between the partners likely, perhaps increasingly so in middle and late until some later stage of the relationship adolescence. Our observation of peer relationships (engagement or marriage) (Katchadourian, 1990). in preschool and middle childhood showed that se- Adolescents' self-reported experiences of love curely attached partners interact smoothly with both and related emotions are remarkably similar to the secure and anxious-resistant relationship histories. reported experiences of adults, according to recent With anxious-avoidant partners, however, even se- findings from an ethnically and socioeconomically cure children have difficulties. Pairs of securely diverse sample of high-school students (Levesque, attached children easily create a balance between 1993). Just as with adults, adolescents' satisfaction connectedness and autonomy for each child, with relationships was positively correlated with whereas pairs of either anxious-resistant or anx- passion, giving and getting communication, com- ious-avoidant children differ in the salience of con- mitment, emotional support and togetherness. In nectedness vs. autonomy. Relationship history may Levesque's sample, adolescent relationships were thus affect what characteristics likely make a poten- also characterized by measures of extremity of tial dating partner more or less salient among other positive emotion that have not typically been in- eligible partners. cluded in studies of adult relationships: feelings of Finally, the relation of adolescent dating to so- exhilaration, growth, appreciation, and specialness. cial crowds is both an advantage and a source of These adolescents' satisfactions were less related, controversy and tension in relationships. Pairing off however, to negative affect, perceived trouble, or

10 Capacity for Intimate Relationships Egeland & conflict -- variables that are inversely related in measures of adult "attachment styles" are corre- adults' reports of their romantic relationships. These lated with concurrent self-reported differences in findings await replication, but the initial patterns characteristics of romantic relationships, including suggest that entry into romantic relationships may orientations to intimacy (Feeney & Noller, 1990). conform to the popular stereotype of "young love" Secure subjects obtained low scores for avoidance as rosily optimistic, as compared to love between of intimacy, whereas avoidant adults scored high adults. on this measure. Investigators using other attachment-style instruments also report correlations be- For our purposes, the most important normative tween attachment representation and aspects of ro- distinction between dating and romantic relation- mantic relationships (N. L. Collins & Read, 1990; ships is the relatively greater degree of intimacy Cohn, Silver, Cowan, Cowan, & Pearson, 1992; between romantic partners (Reis & Shaver, 1988). Feeney & Noller, 1990; Hazan & Shaver, 1987; Relationships can, by definition, be close (i.e., ex- Pearson, Cohn, Cowan, & Cowan, 1994; Simpson, perience high levels of contact and causal interde- Rholes, & Nelligan, 1992). The measures of attach- pendence) without being intimate (i.e., mutually ment style used in these studies cannot be equated perceiving understanding, validation, and caring either to early measures of attachment security from each other) (Reis & Patrick, 1966). Research such as the Strange Situation or to the Adult At- findings with adults indicate that intimacy, in this tachment Interview. Nevertheless, the results are sense, differentiates well-functioning romantic rela- consistent with what might be expected if appropri- tionships from less well functioning ones (for a re- ate measures were used in a longitudinal design. view, see Berscheid & Reis, in press). Later, in the final section of this chapter, we will Similarly, adolescent couples, like adult roman- speculate that even stronger relations across time tic partners, vary in the degree of intimacy attained. might be expected with measures of closeness in The data outlined above, while concluding in childhood and adolescent peer relationships as ad- mid-adolescence and not focused on romantic rela- ditional predictors. tionships, lead us to predict that intimacy in adult Becoming Sexually Active relationships will be based on the foundation provided by earlier family and peer experiences. Those Conceptually, sexual activity is distinguishable with secure attachment histories had deeper rela- from both dating and romantic relationships; but tionships with peers in pre-school, more intense empirically, the intercorrelations of each are sig- friendships in middle childhood (and commonly nificant. Most researchers agree, based on some- with other children similarly oriented), and more what dated evidence, that first heterosexual inter- capacity for emotional vulnerability and sustained course occurs on the average at about 16 and that it dating relationships in adolescence than those with usually takes place in the context of a steady rela- histories of anxious attachment. Moreover, the tionship (for a review, see Katchadourian, 1990). quality of their peer relationships, when directly This generalization clearly does not capture the assessed in middle childhood, revealed a capacity experiences of adolescents in all ethnic and cultural for simultaneous autonomy and connectedness groups (Moore & Erickson, 1985). Comparable which we see as a prerequisite for intimacy. information is lacking for acknowledged gay, lesbian, and bisexual youth. Forecasting from these findings, future research likely will reveal that intimacy with roman- Many adolescents, nevertheless, experience tic partners is a joint function of contrasting care- sexual activity as part of dating relationship or as a giving histories and their functional sequelae in transitory encounter with little connection to peer relationships in childhood and adolescence. on-going social relationships. Such experiences One recent study, using the well-validated Adult may involve relatively low levels of emotional inti- Attachment Interview (AAI; Main & Kaplan, 1996) macy and commitment (Diamond et al., this vol- as a distillation of the individual's attachment his- ume). Relationships in which sexual activity is the tory, is suggestive (Owens, Crowell, Pan, Treboux, primary aspect nevertheless may arise for psycho- O'Connor, & Waters, in press). These authors found logical reasons. Adolescents may hope that such that differences on security on the AAI were related relationships will cause them to feel, or to be per- to both descriptions of romantic relationships and ceived as, more mature, enhance their social pres- behavior with romantic partners. Other cross-sec- tige, or compensate for a lack of intimacy in their tional research with adults shows that self-report lives (Martin, 1982; Tripp, 1975). Some may use

11 Capacity for Intimate Relationships Egeland & sexual relationships as a way of exploring or testing tence or pathology, or they may be seen as indica- their sexual identity (Savin-Williams, 1994). Dia- tions of functional coherence in the relationship mond et al. (this volume) suggest that such primar- patterns of individuals across time. Such coherence ily sexual liaisons may be especially likely among is consistent with individual histories in which a gay, lesbian, and bisexual teenagers, who often fear capacity for intimacy is limited because of experi- the more public nature of dating and romantic rela- ences in early and later relationships. tionships. We will soon have data on dating, romantic re- Relationship history, incorporating both care- lationships and, sexuality from our participants at giver and later peer relationships, may be associ- age 19. Our strongest prediction with regard to at- ated both with the likelihood of a primarily sexual tachment history is that those who were securely relationship and with the timing of beginning sex- attached will more often put a premium on sexual- ual intercourse. Involvement in one or multiple sex- ity in the context of intimacy. They will see inti- ual relationships, as opposed to romantic involve- macy as the foundation for sexuality and they will ments, at any age is likely to be correlated with see sexuality as having a role in deepening inti- consistent patterns of insecure attachment. In the macy. Although they certainly may explore their case of individuals with anxious-avoidant histories, sexuality in less intimate relationships, they are such a pattern would likely reflect avoidance of in- unlikely to be promiscuous or casual regarding timacy or inadequate capacity for intimacy, even if sexuality. desired. Such individuals demonstrate a tendency These speculative links between relationship toward suspicion, jealousy, and unavailability that history, the capacity for intimacy, and likely varia- would likely impede true intimacy. In the case of tions among aspects of romantic relationships anxious-resistant attachments, tendencies toward could constitute a daunting research agenda. Yet excessive dependency and anxiety, along with low they lie at the heart of widely held and compelling self-confidence and poor regulation of affect, would beliefs about the links between love in childhood interfere with the mutuality that supports intimate and in adulthood. In the next section, we address relationships. some of the criteria for research that will finally Early transitions to sexual activity (i.e., becom- examine these fundamental questions. ing sexually active at age 15 or younger) are associ- Some Criteria for Further Evidence on Pre- ated with broad-band assessments of personality, cursors of Intimacy in Romantic Relationships similar to those that characterize differences among secure and insecure individuals. Jessor and his col- Research on the development of romantic rela- leagues (Jessor, Costa, Jessor, and Donovan, 1983; tionships depends fundamentally on valid assess- Jessor, Donovan, & Costa, 1991) found that ear- ments both of attachment in infancy and intimate ly-active adolescents, compared to those who began relationships with a romantic partner in adoles- sexual intercourse at older ages, placed higher value cence and, later, in young adulthood. Our further on non-interference by adults, professed less con- criteria grow out of the conviction that links be- ventional values, and also made early transitions to tween earlier and later relationships reflect a trans- other behaviors tolerated in adults, but less so in active process. In this view expectations concern- early adolescents (e.g., alcohol use, smoking). ing self and relationships and patterns of arousal Some adolescents who manifest these values un- modulation characteristic of early relationships doubtedly have experienced secure, responsive rela- lead to particular forms of engagement with per- tionships with socially unconventional parents or sons and objects, and other persons commonly re- have been reared in a community that is skeptical act in a complementary way, thus perpetuating the toward conventional values; conversely, many teen- pattern, albeit in new forms and in new contexts, agers who show highly conventional behaviors may across developmental periods. have experienced insecure relationships in their earlier lives. The pattern described by Jessor et al. is Consequently, developmental research on ro- characteristic, however, of individuals identified in mantic relationships must include multiple, longitu- longitudinal research as having had insecure rela- dinal assessments of parent-child relationships, tionships with caregivers in infancy and corre- peer relationships, and relationship representations. spondingly marked relationships with peers in pre- With such data one can determine whether early school and middle childhood. These links may be attachment experiences predict adult relationship attributed to general patterns of social incompe- qualities beyond predictions from later family ex-

12 Capacity for Intimate Relationships Egeland & periences, how predictions from attachment meas- ships, similar to the Grotevant and Cooper (1985, ures fare in comparison to peer data, and whether 1986) and Allen et al. (1994) balance between indi- both family and peer data make independent contri- viduality and connectedness. Finally, we coded in- butions. Comprehensive information eventually formation about the balance between the relation- also will permit us to address process issues such as ship and the external world. Conceptually related to whether attachment experiences are mediated Reiss's (1984) work on "closure," this refers to the through peer relationships and whether representa- degree to which a relationship system can maintain tions are indeed the carriers of relationship experi- integrity while negotiating external demands (e.g., ences across phases of development. the strictures of an experimenter's instructions; or the stress of the parent's workplace or the child's Assessment of both parent-adolescent and peer school). This extensive coding provides a relatively interactions is integral to this approach. Measures full picture of key components of intimacy in par- of relationships with parents should tap both char- ent-adolescent relationships that may affect both acteristics of connectedness and autonomy. Grote- the desire to engage in romantic relationships and vant and Cooper (1985, 1986) have provided success in doing so. These measures were found to widely emulated models of coding procedures to be related not only to competence with peers in get at these qualities in laboratory-based observa- adolescence (above) but to Adult Attachment Inter- tions of adolescents with their mothers and fathers. view assessments at age 19 (Weinfield, Sroufe, & Using a similar conceptual framework, Allen, Egeland, submitted). Those children in relation- Hauser, Bell, & O'Connor (1994) developed codes ships rated higher on a composite of the three bal- for behaviors that encourage both autonomy and ance scales at age 13 were significantly more likely relatedness. For example, a low score on their to be secure ("autonomous") on the AAI. In addi- measure reflects: tion, those judged to be "dismissing" (the adult ... behaviors that make it more difficult category equivalent of avoidant) on the AAI had for family members to discuss their own earlier been in relationships more likely to be rated reasons for their position, including as low on engagement and conflict resolution and over-personalizing a disagreement, recant- high on negative affectivity. In profile, this is ex- ing a position without appearing to have actly the pattern one would expect in relationships ben persuaded the position is wrong..., undergirded by avoidance. and pressuring another person to agree; Assessing relationships with peers during ado- expressions of hostility or rudely inter- lescence requires methods that are both different rupting or ignoring the other's expressions from those appropriate for parents and adolescents of opinion. (p. 183) and from those that are valuable in research with Both Grotevant and Cooper and Allen et al. peers during childhood. Hartup's (1996) admoni- have demonstrated lawful links between these relation that multiple aspects of friendship are needed tional patterns and adolescent ego development and to comprehend their developmental significance skills that support a capacity for intimacy (e.g., implies that subjective, as well as observable quali- role-taking skills, identity development). Having ties of these relationships should be considered. such measures of family relationships, in addition Interviews with children and adolescents about to early attachment assessments, should enhance peer relationships are particularly valuable sources predictions of intimacy. of information about representations. One can learn much about the salience of a particular relationship In our longitudinal research, we have extended to a young person, the effectiveness of the proc- the concept of balance to other aspects of relation- esses that comprise the interactions between the ships between parent and adolescent that may be two peers, the coherence of representations of the relevant to eventual functioning in romantic rela- relationship, and the degree to which the relation- tionships. After observing parents and 13 year olds ship contributes positively to individual growth. complete tasks together, we coded the interactions in three ways (J. Sroufe, 1991). We first attended to Although more difficult, especially among older balance between individuals, with particular em- children and adolescents, direct observations of phasis on whether each person appeared to feel safe peer interaction may yield information about the in taking a position and maintaining an opinion nature and course of later romantic relationships. even in the face of disagreements. We next focused Cooper and Ayers-Lopez (1985) demonstrated that on the balance between individuals and relation- patterns of control and autonomy similar to those

13 Capacity for Intimate Relationships Egeland & observed in parent-child interaction can be detected ogy, 16, 243-244. in preadolescents' interactions with peers. Observa- Baumeister, R. F., & Leary, M. R. (1995). The tional studies also have revealed meaningful differ- need to belong: Desire for interpersonal at- ences between pairs of children and preadolescents tachments as a fundamental human motive. previously identified as friends or acquaintances. Psychological Bulletin, 117(3), 497-529. These differences appear to be related partly to contrasting degrees of intimacy and felt security be- Berndt, T. J., Hawkins, J. A., & Hoyle, S. G. tween partners (e.g., Daiute, Hartup, Sholl, & Za- (1986). Changes in friendship during a school jac, 1993; Nelson & Aboud, 1985; Newcomb, year: Effects on children's and adolescents' Brady, & Hartup, 1979). Finally, observations of impressions of friendship and sharing with adolescents in larger groups, although difficult to friends. Child Development, 57, 1284-1297. achieve, may provide valuable information about aspects of functioning in the social crowds in which Berndt, T. J., & Hoyle, S. G. (1985). Stability and dating relationships are embedded (e.g., Englund, change in childhood and adolescent friend- Hyson, & Levy, under review). ships. Developmental Psychology, 21, 1007-1015. Finally, representations of relationships should also be assessed Berndt, T. J., & Keefe, K. (1995). Friends' influence on adolescents' adjustment to school. This framework for collecting data affords both Child Development, 66(5), 1312-1329. the most comprehensive basis and the most promis- ing prospect for establishing a link between early Berscheid, E., & Reis, H. T. (in press). Attraction and intermediate close relationships and the emer- and close relationships. In S. Fiske (Ed.), gence of romantic relationships during adolescence. Handbook of social psychology, 4th ed. New Without intervening measures of experiences and York: Addison-Wesley. representation, questions of how much of the rela- Bigelow, B. J. (1977). Children's friendship expec- tion between infant and adult measures is direct or tations: A cognitive-developmental study. how much is mediated cannot be addressed, nor can Child Development, 48, 246-253. questions concerning change in internal working models or the relation between changing models Bigelow, B. J., & LaGaipa, J. J. (1975). Children's and changing relationship experiences. These alter- written descriptions of friendship: A multidi- native explanations must be examined if we are to mensional analysis. Developmental Psychol- understand how learning to love permeates the ogy, 11, 857-858. course of human development. Bowlby, J. (1973). Separation. New York: Basic References Books. Bretherton, I. (1990). Pouring new wine into old Ainsworth, M. D. S. (1989). Attachments beyond bottles: The social self as internal working infancy. American Psychologist, 44, 709-716. model. In M. Gunnar & L. A. Sroufe, (Eds.), Ainsworth, M. D. S., Blehar, M., Waters, E., Wall, Minnesota Symposia on Child Psychology. S. (1978). Patterns of attachment. Hillsdale, Hillsdale, NJ: Erlbaum. NJ: Erlbaum. Brown, B. B., Eicher, S. A., & Petrie, S. (1986). Allen, J. P., Hauser, S. T., Bell, K. L., & O'Connor, The importance of peer group affiliation in T. G. (1994). Longitudinal assessment of adolescence. Journal of Adolescence, 9, 73-96. autonomy and relatedness in adolescent-family Cohn, D. A., Silver, D. H., Cowan, C. P., Cowan, interactions as predictors of adolescent ego de- P. A., & Pearson, J. (1992). Working models velopment and self-esteem. Child Develop- of childhood attachment and couple relation- ment, 65, 179-194. ships. Journal of Family Issues, 13, 432-449. Baldwin, J. M. (1911). The individual and society. Coleman, J. C., & Hendry, L. (1990). the nature of Boston: Goreham. adolescence (2nd ed.). London: Routledge. Barnett, M., King, L., Howard, J., & Dino, G. Collins, N. L., & Read, S. J. (1990). Adult attach- (1980). Empathy in young children: Relation ment, working models, and relationship quality to parents' empathy, affection, and emphasis on in dating couples. Journal of Personality and the feelings of others. Developmental Psychol-

14 Capacity for Intimate Relationships Egeland &

Social Psychology, 58, 644-663. dating: A comparison of biological and social influences upon one set of behaviors. Child Collins, W. A. (1995). Relationships and develop- Development, 52, 179-185. ment: Family adaptation to individual change. In S. Shulman (Ed.), Close relationships and Douvan, E., & Adelson, J. (1966). The adolescent socioemotional development (pp. 128-154). experience. New York: Wiley. New York: Ablex. Dunphy, D. (1963). The social structure of urban Collins, W. A. (1996). Relationship and develop- adolescent peer groups. Sociometry, 26, ment during adoelscence: Interpersonal adap- 230-246. tation to individual change. Personal Relation- Egeland, B., & Brunnquell, D. (1979). ships, 3(4). Elicker, J., Englund, M., & Sroufe, L. A. (1992). Collins, W. A., & Repinski, D. J. (1994). Relation- Predicting peer competence and peer relationships during adolescence: Continuity and ships in childhood from early parent-child re- change in interpersonal perspective. In R. lationships. In R. Parke & G. Ladd (Eds.), Montemayor, G. Adams, & T. Gullotta (Eds.), Family-peer relationships: Modes of linkage Advances in adolescent development: Vol. 5. (pp. 77-106). Hillsdale, NJ: Erlbaum. Personal relationships during adolescence (pp. 7-36). Thousand Oaks, CA: Sage. Englund, M., Hyson, D., & Levy, A. (in preparation). Collins, W. A., & Russell, G. (1991). Mother-child and father-child relationships in middle child- Feeney, J. A., & Noller, P. (1990). Attachment hood and adolescence: A developmental analy- style as a predictor of adult romantic relation- sis. Developmental Review, 11, 99-136. ships. Journal of Personality and Social Psy- chology, 58, 281-291. Cooper, C., & Ayers-Lopez, S. (1985). Family and peer systems in early adolescence: New models Fonagy, P., Steele, M., & Steele, H. (1991). Mater- of the role of relationships in development. nal representations of attachment during preg- Journal of Early Adolescence, 5, 9-22. nancy predict the organization of infant-mother attachment at one year of age. Dishion, T. J., Patterson, G. R., & Griesler, P. C. Child Development, 62, 880-893. (1994). Peer adaptations in the development of antisocial behavior: A confluence model. In L. Furman, W., & Bierman, K. (1984). Children's R. Huesmann (Ed.), Current perspectives on conceptions of friendship; A multimethod aggressive behavior (pp. 61-95). New York: study of developmental changes. Developmen- Plenum. tal Psychology, 20, 925-931. Crowell, J., & Feldman, S. (1991). Mothers' work- Furman, W., & Buhrmester, D. (1992). Age and ing models of attachment relationships and sex differences in perceptions of networks of mother and child behavior during separation personal relationships. Child Development, 63, and reunion. Developmental Psychology, 27(4), 103-115. 597-605. Furman, W., & Wehner, E. (1994). Romantic Daiute, C., Hartup, W. W., Sholl, W., & Zajac, R. views: Toward a theory of adolescent roman- (1993, March). Peer collaboration and written tic relationships. In R. Montemayor, G. R. Ad- language development: A study of friends and ams, & T. P. Gullotta (Eds.), Advances in ado- acquaintances. Paper presented at the Society lescent development: Volume 6, Personal re- for Research in Child Development, New Or- lationships during adolescence (pp. 168-195). leans, LA. Thousand Oaks, CA: Sage. Dodge, K., & Somberg, D. (1987). Hostile attribu- Fury, G., & Carlson, E. (under review). Represen- tional biases among aggressive boys are exac- tational models of attachment in children's erbated under conditions of threats to the self. family drawings. Child Development, 58, 213-224. Grossmann, K., Fremmer-Bombik, E., Rudolph, J., Dornbusch, S. M., Carlsmith, J. M., Gross, R. T., & Grossmann, K. (1988). Maternal attachment Martin, J. A., Jennings, D., Rosenberg, A., & representations as related to patterns of in- Duke, P. (1981). Sexual development, age, and fant-mother attachment and maternal care dur-

15 Capacity for Intimate Relationships Egeland &

ing the first year. In R. Hinde & J. Steven- (167-200). Mahwah, NJ: Erlbaum. son-Hinde (Eds), Relations within families: Mutual influences (pp. 241-262). Oxford: Clar- Holmbeck, G. N., Paikoff, R. L., & Brooks-Gunn, endon Press. J. (1995). Parenting adolescents. In M. Born- stein (Ed.), Handbook of parenting: Vol. 1. Grotevant, H., & Cooper, C. (1985). Patterns of Child and parenting (pp. 91-118). Mahwah, interaction in family relationships and the de- NJ: Erlbaum. velopment of identity exploration in adolescence. Child Development, 56, 415-428. Jessor, R., Costa, F., Jessor, L., & Donovan, J. E. (1983). The time of first intercourse: A pro- Grotevant, H., & Cooper, C. (1986). Individuation spective study. Journal of Personality and So- in family relationships. Human Development, cial Psychology, 44, 608-626. 29, 82-100. Jessor, R., Donovan, J., & Costa, F. (1991). Beyond Hartup, W. W. (1984). The peer context in middle adolescence: Problem behavior and young childhood. In W. A. Collins (Ed.), Develop- adult development. New York: Cambridge ment during middle childhood: The years from University Press. six to twelve (pp. 240-282). Washington, D. C.: National Academy Press. Kagan, J., Reznick, J., & Gibbons, J. (1989). Inhib- ited and uninhibited types of children. Child Hartup, W. W. (1992). Friendships and their devel- Development, 60, 838-845. opmental significance. In H. McGurk (Ed.), Contemporary issues in childhood social de- Katchadourian, H. (1990). Sexuality. In S. S. velopment (pp. 175-205). London: Routledge. Feldman & G. R. Elliott (Eds.), At the threshold: The developing adolescent (pp. 330-351). Hartup, W. W. (1993). Adolescents and their Cambridge, MA: Harvard University Press. friends. In B. Laursen (Ed.), Close friendships in adolescence: New directions for child devel- Kestenbaum, R., Farber, E., & Sroufe, L. A. opment, No. 60 (pp. 3-22). San Francisco: (1989). Individual differences in empathy Jossey-Bass. among preschoolers: Relation to attachment history. In N. Eisenberg (Eds.), Empathy and Hartup, W. W. (1996). The company they keep: related emotional responses. San Francisco: Friendships and their developmental signifi- Jossey-Bass. cance. Child Development, 67, 1-13. Lamb, M. (1984). Fathers, mothers and childcare in Hauser, S. T., Powers, S. I., & Noam, G. G. (1991). the 1980s: Family influences on child develop- Adolescents and their families. New York: The ment. In K. Borman, D. Quarm, & S. Gide- Free Press. onese (Eds.), Women in the workplace (pp. 61-88). Hauser, S., Powers, S., Noam, G., Jacobson, A., Weiss, B., & Follansbee, D. (1984). Familial Larson, R., & Richards, M. H. (1991). Daily com- contexts of adolescent ego development. Child panionship in late childhood and early adoles- Development, 55, 195-213. cence: Changing developmental contexts. Child Development, 62, 284-300. Hazan, C., & Shaver, P. (1987). Romantic love conceptualized as an attachment process. Jour- Laursen, B., & Collins, W. A. (1994). Interpersonal nal of Personality and Social Psychology, 52, conflict during adolescence. Psychological 511-524. Bulletin, 115(2), 197-209. Hinde, R. A., & Stevenson-Hinde, J. (1987). Inter- Levesque, R. J. R. (1993). The romantic experience personal relationships and child development. of adolescents in satisfying love relationships. Developmental Review, 7, 1-21. Journal of Youth and Adolescence, 22, 219-251. Holmbeck, G. N. (1996). A model of family relational transformations during the transition to Lewis, M. (1989). Commentary. Human Develop- adolescence: Parent-adolescent conflict and ment, 32, 216-222. adaptation. In J. A. Graber, J. Brooks-Gunn, & A. C. Petersen (Eds.), Transitions through ado- Liberman, A. F. (1977). Preschoolers' competence lescence: Interpersonal domains and contexts with a peer: Relations with attachment and peer experience. Child Development, 48,

16 Capacity for Intimate Relationships Egeland &

1277-1287. Ostoja, E. (1994). Representation of friendships in adolescence: Antecedents and contemporary Long, B. H. (1983). A steady boyfriend: A step correlates. Unpublished manuscript. toward resolution of the intimacy crisis for American college women. Journal of Psychol- Ostoja, E., McCrone, E., Lehn, L., Reed, T., & ogy, 115, 275-280. Sroufe, L. A. (1995, March). Representations of close relationships in adolescence: Longitu- Long, B. H. (1989). Heterosexual involvement of dinal antecedents from infancy through child- unmarried undergraduate females in relation to hood. self-evaluations. Journal of Youth and Adoles- cence, 18, 489-500. Owens, G., Crowell, J., Pan, H., Trebonx, D., O'Connor, E., & Waters, E. (under review). Maccoby, E. E. (1990). Gender and relationships. The prototype hypothesis and the origins of American Psychologist, 45, 513-520. attachment working models: Parent-child rela- MacDonald, K. (1992). Warmth as a developmen- tionships and adult romantic relationships. tal construct. Child Development, 63, 753-773. Unpublished manuscript. State University of New York at Stony Brook. Martin, A. D. (1982). Learning to hide: The socialization of the gay adolescent. Adolescent Pancake, V. R. (1985, April). Continuity between Psychiatry, 10, 52-65. mother-infant attachment and ongoing dyadic peer relationships in preschool. Paper pre- Mead, G. H. (1934). Mind, self, and society. Chi- sented at the Biennial Meeting of the Society cago: University of Chicago Press. for Research in Child Development, Toronto. Miller, B., & Moore, K. (1990). Adolescent sexual Parker, J. G., & Asher, S. R. (1987). Peer relations behavior, pregnancy, and parenting: Research and later personal adjustment: Are low-ac- through the 1980s. Journal of Marriage and cepted children at risk? Psychological Bulletin, the Family, 52, 1025-1044. 102, 357-389. Miller, K. E. (1990). Adolescents' same-sex and Parker, J. G., & Gottman, J. M. (1989). Social and opposite-sex peer relations: Sex differences in emotional development in a relational context: popularity, perceived social competence and Friendship interaction from early childhood to social cognitive skills. Journal of Adolescent adolescence. In T. J. Berndt & G. W. Ladd Research, 5, 222-241. (Eds.), Peer relationships in child development Miller, R., & Gordon, K. (1986). The decline in (pp. 95-132). New York: Wiley. formal dating: A study in six Connecticut high Pearson, J. L., Cohn, D. A., Cowan, P. A., & schools. Marriage and Family Review, 10, Cowan, C. P. (1994). Earned- and continu- 139-156. ous-security in adult attachment: Relation to Montemayor, R. (1983). Parents and adolescents in depressive symptomatology and parenting conflict: All families some of the time and style. Development and Psychopathology, 6, some families most of the time. Journal of 359-373. Early Adolescence, 3, 83-103. Pianta, R., Sroufe, L. A., & Egeland, B. (1989). Moore, D. S., & Erickson, P. I. (1985). Age, gen- Continuity and discontinuity in maternal sensi- der, and ethnic differences in sexual and con- tivity at 6, 24, and 42 months in a high risk traceptive knowledge, attitudes, and behavior. sample. Child Development, 60(2), 481-487. Family and Community Health, 8, 38-51. Reis, H. T., & Patrick, B. C. (in press). Attach- ment and intimacy: Component processes. In Nelson, J., & Aboud, F. E. (1985). The resolution A. Kruglanski & E. T. Higgins (Eds.), Social of social conflict between friends. Child Devel- psychology: Handbook of basic principles. opment, 56, 1009-1017. New York: Guilford. Offer, D. (1969). The psychological world of the Reis, H. T., & Shaver, P. (1988). Intimacy as an teenager. New York: Basic Books. interpersonal process. In S. W. Duck (Ed.), Offer, D., Ostrov, E., & Howard, K. (1981). The Handbook of personal relationships (pp. adolescent: A Psychological self-portrait. New 367-389). NY: John Wiley. York: Basic Books.

17 Capacity for Intimate Relationships Egeland &

Reiss, D. (1984). The family's construction of real- Wiley. ity. Cambridge: Harvard University Press. Sharabany, R., Gershoni, R., & Hofman, J. (1981). Ricks, M. H. (1985). The social transmission of Girlfriend, boyfriend: Age and sex differences parental behavior: Attachment across genera- in intimate friendship. Developmental Psychol- tions. In I. Bretherton & E. Waters (Eds.), ogy, 27, 800-808. Growing points of attachment theory and re- Shulman, S., Elicker, J., & Sroufe, L. A. (1994). search. Monographs of the Society for Re- Stages of friendship growth in preadolescence search in Child Development, 50 (Whole No. as related to attachment history. Journal of 209), 211-227. Social and Personal Relationships. Roscoe, B., Diana, M. S., & Brooks, R. H. (1987). Simpson, J. A., Rholes, W. S., & Nelligan, J. S. Early, middle and late adolescents' views on (1992). Support seeking and support giving dating and factors influencing partner selec- within couples in an anxiety-provoking situation. Adolescence, 22, 59-68. tion: The role of attachment styles. Journal of Rosenberg, D. M. (1984). The quality and content Personality and Social Psychology, 62, of preschool fantasy play: Correlates in con- 434-446. current social-personality function and early Sroufe, J. (1991). Assessment of parent-adolescent mother-child attachment relationships. Unpub- relationships: Implications for adolescent de- lished doctoral dissertation, University of Min- velopment. Journal of Family Psychology, 5 nesota. (1), 21-45. Rubin, K. H., LeMare, L. J., & Lollis, S. (1990). Sroufe, L. A. (1979). The coherence of individual Social withdrawal in childhood: Developmen- development. American Psychologist, 34(10), tal pathways to peer rejection. In S. R. Asher & 834-841. J. D. Coie (Eds.), Peer rejection in childhood (pp. 217-249). Cambridge, MA: Cambridge Sroufe, L. A. (1983) Infant-caregiver attachment & University Press. patterns of adaptation in preschool: The roots of maladaptation & competence. In M. Samet, N., & Kelly, E. W. (1987). The relationship Perlmutter (Ed.), Minnesota Symposium in of steady dating to self-esteem and sex role Child Psychology (Vol. 16, pp. 41-83). Hills- identity among adolescents. Adolescence, 22, dale: Erlbaum. 231-245. Sroufe, L. A. (1989). Pathways to adaptation and Savin-Williams, R. C. (1994). Verbal and physi- maladaptation: Psychopathology as develop- cal abuse as stressors in the lives of lesbian, mental deviation. In D. Cicchetti (Ed.), Roch- gay male, and bisexual youths: Associations ester Symposia on Developmental Psychopa- with school problems, running away, substance thology (Vol. 1, pp. 13-14). Hillsdale, NJ: Erl- abuse, prostitution, and suicide. Journal of baum. Consulting and Clinical Psychology, 62, 261-269. Sroufe, L. A. (1994, February). Continuity of individual patterns of adaptation across changing Savin-Williams, R. C., & Berndt, T. J. (1990). developmental issues. Paper presented at the Friendship and peer relations. In S. S. Feldman biennial meeting of the Society for Research & G. R. Elliott (Eds.), At the threshold: The on Adolescence, San Diego. developing adolescent (pp. 277-307). Cam- bridge, MA: Harvard University Press. Sroufe, L. A. (1995, March). Evaluating the role of early experience in adult love relationships. Selman, R. L. (1980). The growth of interpersonal Discussant paper presented at the biennial understanding. New York: Academic Press. meeting of the Society for Research in Child Selman, R. L., & Schultz, L. H. (1989). Children's Development, Indianapolis. strategies for interpersonal negotiation with Sroufe, L. A., Bennett, C., Englund, M., Urban, J., peers: An interpretive/empirical approach to & Shulman, S. (1993) The significance of gen- the study of social development. In T. J. Berndt der boundaries in preadolescence: Contempo- & G. W. Ladd (Eds.), Peer relationships in rary correlates and antecedents of boundary child development (pp. 371-406). New York: violation and maintenance. Child Develop-

18 Capacity for Intimate Relationships Egeland &

ment, 64(2), 455-466. ment experiences on competence in preschool]. Unpublished doctoral dissertation, Sroufe, L. A., Carlson, E., & Shulman, S. (1993). Universitat Regensburg, Regensburg, Ger- The development of individuals in relation- many. ships: From infancy through adolescence. In D. C. Funder, R. Parke, C. Tomlinson-Keasey, & Suess, G. J., Grossmann, K. E., & Sroufe, L. A. K. Widaman (Eds.), Studying lives through (1993). Effects of infant attachment to mother time: Approaches to personality and develop- and father on quality of adaptation in pre- ment (pp. 315-342). Washington DC: Ameri- school: From dyadic to individual organization can Psychological Association. of self. International Journal of Behavioral Development, 15(1), 43-66. Sroufe, L. A., & Egeland, B., (1991). Illustrations of person and environment interaction from a Sullivan, H. S. (1953). the interpersonal theory of longitudinal study. In T. Wachs & R. Plomin psychiatry. New York: Norton & Co. (Eds.), Conceptualization and measurement of Surra, C. A. (1990). Research and theory on mate organism-environment interaction (pp. 68-84). selection and premarital relationships in the Washington, DC: American Psychological As- 1980s. Journal of Marriage and the Family, sociation. 52, 844-865. Sroufe, L. A., Egeland, B., & Carlson, E. (in press). Thorne, B. (1986). Girls and boys together . . . but One social world. In W. A. Collins & B. mostly apart: Gender arrangements in elemen- Laursen (Eds.), Relationships as developmental tary schools. In W. W. Hartup & Z. Rubin contexts: The Minnesota Symposia on Child (Eds.), Relationships and development (pp. Psycholoyg (Vol. 30). Mahtah, NJ: Erlbaum. 167-184). Hillsdale, NJ: Erlbaum. Sroufe, L. A., Egeland, B., & Kreutzer, T. (1990). Tripp, C. A. (1975). The homosexual matrix. New The fate of early experience following devel- York: McGraw-Hill. opmental change: Longitudinal approaches to individual adaptation in childhood. Child De- Troy, M., & Sroufe, L. A. (1987). Victimization velopment, 61, 1363-137 among preschoolers: The role of attachment relationship history. Journal of the American Sroufe, L. A., & Fleeson, J. (1986). Attachment and Academy of Child and Adolescent Psychiatry, the construction of relationships. In W. Hartup 26(2), 166-172. & Z. Rubin (Eds.), Relationships and development (pp. 51-71), Hillsdale, NJ: Erlbaum. Waters, E., Kondo-Ikemura, K., Posada, G., & Richters, J. E. (1991). Learning to love: Sroufe, L. A., & Fleeson, J. (1988). The coherence Mechanisms and milestones. In M. Gunnar & of family relationships. In R. A. Hinde & J. L. A. Sroufe (Eds.), Minnesota symposia on Stevenson-Hinde (Eds.), Relationships within child psychology: Vol. 23. Self processes and families: Mutual influences (pp. 27-47). Ox- development (pp. 217-255). Hillsdale, NJ: Erl- ford: Oxford University Press. baum. Sroufe, L. A., Schork, E., Motti, E., Lawroski, N., Waters, E., & Sroufe, L. A. (1983). Social compe- & LaFreniere, P. (1984). The role of affect in tence as a developmental construct. Develop- social competence. In C. Izard, J, Kagan, & R. mental Review, 3, 79-97. Zajonc (Eds.), Emotional cognition and behavior, New York: Plenum. Weinfield, N., Sroufe, L. A., & Egeland, B (under review). Early attachment as a pathway to Steinberg, L. (1990). Interdependency in the fam- adolescent peer competence. ily: Autonomy, conflict, and harmony in the parent-adolescent relationship. In S. Feldman Zani, B. (1993). Dating and interpersonal relation- & G. Elliot (Eds.), At the threshold: The devel- ship in adolescence. In S. Jackson & H. Rodrigu- oping adolescent (pp. 225-276). Cambridge, ez-Tome (Eds.), MA: Harvard University Press. Suess, G. J. (1987). Auswirkungen frukindlicher Bindungserfabrungen auf die Kompetenz im Kindergarten [Consequences of early attach-