JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.. Wiley-Blackwell andSociety for Research in Child Development are collaborating with JSTOR to digitize, preserve and extend access to Monographs of the Society for Research in Child Development.

We describe the development of a system for classifying attachment organization at age 6 on the basis of study of children's responses to unstructured reunions with parents in a laboratory setting. In a study of 33 families in Berkeley, California, sixth-year attachment classifications to mother were found to be highly predictable from infancy attachment classifications to mother (84%), with children secure in infancy identified as secure on reunion at age 6 (Group B); children insecure-avoidant in infancy identified as insecure-avoidant (Group A); and children who were insecure-disorganized/ disoriented identified as controlling of the parent (Group D). Lower predictability (61 %) was found for attachment to father. An insecure-ambivalent (Group C) sixth-year classification was developed following the Berkeley study. In a second study of child-mother dyads conducted in CharlottesviUe, Virginia (N = 50), 62% of children were stable in (A, B, C, and D) classification across a 1-month interval. When D children were reassigned to their best-fitting alternative A, B, or C categories, stability was high both for major classifications (A4-B+C = 86%) and for seven subclassincations (Al, A2, B1/B2, B3, B4, Cl, C2 = 76%). Avoidance of the mother was stable across both the 5-year and the 1-month periods.The description of individual differences in attachment organization has traditionally focused on infancy and, specifically, on the infant's response to the Ainsworth Strange Situation (Ainsworth, Blehar, Waters, & Wall, 1978), a laboratory observation that involves the infant in two brief separations from and reunions with the parent. The principal assessment of attachment in infancy is the attachment classification, which is based chiefly on the infant's response to the two 3-min reunions with the parent (Ainsworth, Bell, & Stayton, 1971;Ainsworth et al., 1978). Infants who actively attempt to regain proximity, contact, or interaction with the parent are classified as secure in their attachment to that parent (Group B); infants who actively avoid and ignore the parent on reunion are termed insecureavoidant (Group A); and infants who show anger or an inability to be settled by the parent (or both) on reunion are termed insecure-ambivalent (Group C). Insecure infants who could not be classified within this system were originally termed unclassifiable (Main & Weston, 1981).

Emotion regulation and quality of attachment are closely linked. It has been proposed here that one influence on individual differences in emotion regulation may be a child's attachment history. Individuals characterized by the flexible ability to accept and integrate both positive and negative emotions are generally securely attached; on the other hand, individuals characterized by either limited or heightened negative affect are more likely to be insecurely attached. While acknowledging the role of infant temperament, I have focused on the role of social factors in examining the link between emotion regulation and attachment. The approach to emotion regulation taken here--that emotion regulation is adaptive in helping a child attain her goals--is esentially a functionalist approach (Bretherton et al., 1986; Campos et al., 1983), consistent with earlier views of emotions as important regulators of interpersonal relationships (Charlesworth, 1982; Izard, 1977). It has been proposed that patterns of emotion regulation serve an important function for the infant: the function of maintaining the relationship with the attachment figure. Emotion regulation has been described as serving this function in two ways. First, the function of maintaining the relationship is thought to be served when infant emotion regulation contributes to the infant's more generalized regulation of the attachment system in response to experiences with the caregiver. Infants who have experienced rejection (insecure/avoidant infants) are thought to minimize negative affect in order to avoid the risk of further rejection. Infants whose mothers have been relatively unavailable or inconsistently available (insecure/ambivalent infants) are thought to maximize negative affect in order to increase the likelihood of gaining the attention of a frequently unavailable caregiver. Both these patterns of emotion regulation help ensure that the child will remain close to the parent and thereby be protected. Second, the function of maintaining the attachment relationship is thought to be served when the infant signals to the parent that she will cooperate in helping maintain the parent's own state of mind in relation to attachment. The minimizing of negative affect of the avoidant infant signals that the infant will not seek caregiving that would interfere with the parent's dismissal of attachment. The heightened negative emotionality of the ambivalent infant signals to the parent that the infant needs her and thus helps maintain a state of mind in which attachment is emphasized. The approach to emotion regulation presented here is congruent with much work examining the socialization of emotions (Lewis & Saarni, 1985; Thompson, 1990).

Researchers have used J. Bowlby's (1969/1982, 1973, 1980, 1988) attachment theory frequently as a basis for examining whether experiences in close personal relationships relate to the processing of social information across childhood, adolescence, and adulthood. We present an integrative life-span-encompassing theoretical model to explain the patterns of results that have emerged from these studies. The central proposition is that individuals who possess secure experience-based internal working models of attachment will process--in a relatively open manner--a broad range of positive and negative attachment-relevant social information. Moreover, secure individuals will draw on their positive attachment-related knowledge to process this information in a positively biased schematic way. In contrast, individuals who possess insecure internal working models of attachment will process attachment-relevant social information in one of two ways, depending on whether the information could cause the individual psychological pain. If processing the information is likely to lead to psychological pain, insecure individuals will defensively exclude this information from further processing. If, however, the information is unlikely to lead to psychological pain, then insecure individuals will process this information in a negatively biased schematic fashion that is congruent with their negative attachment-related experiences. In a comprehensive literature review, we describe studies that illustrate these patterns of attachment-related information processing from childhood to adulthood. This review focuses on studies that have examined specific components (e.g., attention and memory) and broader aspects (e.g., attributions) of social information processing. We also provide general conclusions and suggestions for future research.

No abstract