Attachment Theory

Social, Developmental, and Clinical Perspectives

Susan Goldberg, Roy Muir, John Kerr, Susan Goldberg, Roy Muir, John Kerr

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781135890599
• Edition: 1
• Topic: Psychology
• Subtopic: Applied Psychology
• Index: Psychology

At a historic conference in Toronto in October 1993, developmental researchers and clinicians came together for the first time to explore the implications of current knowledge of attachment. This volume is the outcome of their labors. It offers innovative approaches to the understanding of such diverse clinical topics as child abuse, borderline personality disorder, dissociation, adolescent suicide, treatment responsiveness, false memory, narrative competence, and the intergenerational transmission of trauma.

1

Introduction

SUSAN GOLDBERG

John Bowlby, described as one of the “three or four most important psychiatrists of the twentieth century” (Storr, 1992), died in 1990. In October 1993 a conference was held in Toronto to honor Bowlby and to take stock of his theory of attachment, its origins, influences, and implications. Several previous conferences and volumes had focused on Bowlby’s attachment theory, but the 1993 conference was unique in scope. Whereas prior conferences and volumes had concentrated on specific themes within attachment theory, research, and practice (e.g., attachment across the lifespan), this volume, which grew out of the 1993 conference, is designed to place attachment theory in its cultural and historical context (Holmes Bretherton, Grossmann, Eagle), evaluate its relation to clinical (Adam, Fonagy, Cichetti, Liotti) and to research endeavors (Belsky, Hofer, Suomi), and identify new directions in attachment work (Crittenden, Main).

This breadth of interest will appeal to clinicians and researchers alike. Indeed, the most remarkable aspect of the conference itself was the intense excitement generated among clinicians previously unaware of the extensive developmental research on attachment and its relevance to their work. Their enthusiasm was matched by corresponding enthusiasm from researchers whose view of attachment had been primarily academic without concern for clinical applications.

The chasm between researchers and clinicians, particularly clinicians from the psychoanalytic tradition, is not unusual. Each group has its own language, produces and reads its own journals, and generally communicates with professionals who share the same orientation. John Bowlby was a rare figure in the psychoanalytic community for his insistence on the importance of empirical and extraclinical validation of theory. To that end, he formed alliances with researchers in many disciplines and brought research-trained psychologists to the Tavistock clinic as collaborators.

The most important of these was Mary Ainsworth. It was Ainsworth’s thoughtful observations, conceptual elaborations, and development of a standardized laboratory procedure (known as the Strange Situation) for assessing individual differences in patterns of attachment that opened the door to programmatic empirical research. Ainsworth’s own contributions initiated new directions in attachment theory and research. But her influence is multiplied many times because she nurtured a second generation of gifted researchers, who, in turn, trained major contributors.

Among Ainsworth’s earliest intellectual descendants, Mary Main is the one most responsible for bridging the researcher-clinician gap. Main is one of several in her generation to explore attachment and its measurement beyond infancy. The Adult Attachment Interview, which she and her colleagues developed, proved to be an important touchstone for clinicians who recognized its content as familiar clinical material and the meticulous discourse analysis used to code it as an objective translation of important clinical intuitions. Indeed, it was in an adult attachment workshop led by Mary Main and Erik Hesse that Andrew Brink, then head of the Humanities and Psychoanalytic Thought Program at Trinity College, University of Toronto, conceived the idea for the 1993 Toronto conference as an opportunity to unite clinicians and researchers in a common endeavor.

Thus, this volume marks a new era in attachment theory and research, one in which attachment theory is being rediscovered by clinicians, especially analysts (who discarded it on the basis of Bowlby’s early writings many years ago), and in which developmental researchers previously insulated from the front lines of clinical research are discovering the practical applications of attachment theory and research and thereby enriching theories of development.

It is our expectation that this volume will appeal to the same broad audience that enjoyed the conference. Like any specialized group, attachment researchers have developed a language and shorthand of their own to describe important behavioral patterns, a language that may not be familiar to this broader audience. Therefore, we have provided a guide to the labels for individual differences in attachment (attachment classifications) that are routinely used throughout the book. The appendices to this chapter summarize this information and may be used for quick reference. Readers who are already familiar with this material may want to skip this section and go directly to the next section.

ATTACHMENT AND ITS MEASUREMENT

This overview emphasizes observation and interview measures that are most germane to understanding the chapters in this volume. Other methods, such as the Water and Deane Q-sort (Waters and Deane, 1985) and projective measures currently being developed, for example, the separation anxiety test and doll play stories (Bretherton, Ridgeway, and Cassidy, 1990) are discussed in individual chapters. Since the infancy scheme was first to develop and other schemes have been largely derivative, it is described in detail. Several chapters in this volume rely on the Adult Attachment Interview (AAI) and its associated scoring scheme, and it too is discussed in detail. Schemes for preschoolers and for five-to seven-year-olds are discussed more briefly, and the reader is referred to descriptions in relevant appendices. In addition, Appendix 1 provides an overview of the relations between different schemes.

The original classification scheme for infant-caregiver attachment was based on observation of behavior in the Strange Situation (a standardized laboratory procedure involving two separations and reunions between the caregiver and the infant) but was validated by showing that infant behavior in the Strange Situation was a good marker of the kind of caregiving infants had experienced at home over the first year (Ainsworth et al., 1978).

Ainsworth described three major patterns, each with two or more subtypes. The most common and most optimal major pattern was labeled secure (or Type B, with four subtypes). These were infants who clearly used their mothers as a secure base for exploration. That is, they explored freely in her presence, checked on her whereabouts and reactions periodically, and restricted exploration in her absence. They showed varying levels of distress in her absence (one feature that distinguishes subtypes), ranging from simple inhibition of play/exploration to extreme distress, but all greeted her positively on her return. Those who were more distressed sought physical contact, were comforted by it, and were soon able to return to exploration.

The next most common pattern was labeled avoidant (Type A, with two subtypes). These infants appeared to explore without interest in their mothers’ whereabouts, were minimally distressed by her departure, and appeared to “snub” or ignore her when she returned.

The third pattern was labeled ambivalent or resistant (Type C, with two subtypes). These infants had difficulty separating from their mothers to explore, and their play was often impoverished. They were extremely distressed by their mothers’ departure, and, although they sought contact with her on her return, they did not readily settle down or return to exploration.

A notable feature of this classification scheme is that the amount of distress shown during separation (previously considered to reflect the strength of infant-mother relationships) does not distinguish between secure and insecure relationships. Within the secure group, some infants (those in the B1 and B2 subgroups) show little distress and seek minimal contact with the mother at reunion. In this respect they are like avoidant infants, but their generally positive approach to their mothers, particularly at reunions, identifies them as secure. Other secure infants (those in the B3 and B4 subgroups) may be very distressed during separations, and some (those in the B4 subgroup), like resistant infants, may also be slow to settle down at reunions. Nevertheless, it is their competence in expressing their needs directly, and their unambivalent acceptance of maternal ministrations, that places them in the secure group. Thus, one of the most important lessons from Ainsworth’s Baltimore study is that reunion behaviors are more telling than separation behaviors in revealing quality of attachment.

For many years, these three primary categories were adequate to capture individual differences in infant behavior in the Strange Situation. Research during this period documented the antecedents of these patterns in maternal and infant behavior in the home as well as the sequelae in the later social development of the child. The most notable example of the latter is a large longitudinal study by Sroufe and Egeland and their colleagues at University of Minnesota, which has followed infants for whom attachment was assessed at 12 and 18 months into late adolescence (e.g., Erickson, Sroufe, and Egeland, 1985; Sroufe et al., 1993). Two other major longitudinal studies, one at Berkeley led by Main and her colleagues (Main, Kaplan, and Cassidy, 1985) and one in Germany conducted by Klaus and Karin Grossmann and their colleagues (Wartner et al., 1994) have also made major contributions to this literature. In addition, numerous studies of normative and high-risk populations included attachment measures among other measures at discrete points in the life cycle. These studies occupied increasing space in leading developmental journals.

The original attachment classification scheme served these studies well, although there was always a small group of cases that were difficult to classify or that could not be readily assigned to one of the existing categories. As attachment studies expanded to include more clinical samples, such as maltreated children and infants of psychiatrically diagnosed mothers, it became evident that there were patterns of attachment behavior that could not be integrated into Ainsworth’s original three-category scheme. A “mixed” avoidant/resistant (A/C) pattern was documented by Crittenden (1985) and by Radke-Yarrow and her colleagues (1985).

Subsequently, Main and Solomon (1986, 1990) reviewed a large number of “unclassifiable” tapes and developed criteria for identifying disorganized/disoriented (Type D) attachment. Unlike children in the three primary categories, which represented organized strategies for managing arousal in the context of the attachment relationship, infants classified as disorganized/disoriented did not appear to have an organized strategy. Furthermore, they engaged in odd behaviors (e.g., they were unable to approach the caregiver directly even when very distressed), which made sense only if they were assumed to be either frightened or confused by the caregiver.

Because the D classification did not provide information about attachment strategies, infants in this group were also assigned to a “forced” classification that represented the best fit of the ABC alternatives. Thus, the most widely used classification scheme for infants now includes four classifications: avoidant (A), secure (B), resistant (C), and disorganized (D). Appendix 2 summarizes this scheme. While the A/C patterns described by Crittenden (1985) and Radke-Yarrow et al. (1985) have sometimes been subsumed under the D category, Crittenden (1992, this volume) has argued convincingly that the A/C pattern can be an organized strategy and should be considered distinct from the disorganized category.

For many years after Ainsworth’s development of the Strange Situation and the associated classification scheme, research focused on infants. Bowlby, however, clearly considered attachment to be a life-span construct, and it soon became evident that measures of attachment beyond infancy were needed.

One of the first steps in this direction was Main’s development of the Adult Attachment Interview (AAI). For the purpose of instrument development, Main assumed that parents of infants exhibiting different attachment patterns would differ in their “states of mind” regarding intimate relationships. By grouping and using a “guess and uncover” method to examine transcripts of interviews from parents whose infants’ attachment classification was known, Main and her colleagues established three primary adult attachment categories that corresponded to the infant scheme: dismissing (Ds, avoidant); autonomous or free to evaluate (F, secure); and preoccupied or enmeshed (E, resistant). It was also discovered that the parents of disorganized infants were likely to report salient loss or trauma with respect to attachment figures in childhood that had not been resolved. These parents were classified unresolved (U) and, as with the D classification of infancy, “forced” into one of the other categories that best captured their strategy for thinking about and discussing attachment experiences.

The method of classifying AAI transcripts is based on discourse analysis designed to assess one’s ability to provide a coherent and relevant narrative regarding at...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Dedication
• Contents
• Acknowledgments
• Contributors
• 1 Introduction
• Part 1 Origins and Context of Attachment Theory
• Part 2 Contemporary Research
• Part 3 Clinical Significance and Applications of Attachments
• Part 4 New Directions in Attachment Theory
• Index