Show Summary Details
Page of

(p. 427) Phillip Shaver, Mario Mikulincer, and the Experiences in Close Relationships Scale 

(p. 427) Phillip Shaver, Mario Mikulincer, and the Experiences in Close Relationships Scale
Chapter:
(p. 427) Phillip Shaver, Mario Mikulincer, and the Experiences in Close Relationships Scale
Author(s):

Robbie Duschinsky

DOI:
10.1093/med-psych/9780198842064.003.0005
Page of

PRINTED FROM OXFORD CLINICAL PSYCHOLOGY ONLINE (www.oxfordclinicalpsych.com). © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Clinical Psychology Online for personal use (for details see Privacy Policy and Legal Notice).

date: 29 September 2020

Biographical sketch

Shaver was a social psychologist, interested in combining the experimental approach of behaviourism with the concern for relationships and emotional life that characterised psychoanalysis. Hazan, a graduate student with Shaver, proposed in a seminar that aspects of adult romantic relationships resembled the qualities of an attachment relationship. In the mid-1980s, Shaver and Hazan developed a single-item self-report measure of romantic love modelled on the Ainsworth classifications of infant attachment. This initiated a decade in which a profusion of self-report measures appeared, a period mostly brought to an end by the synthesis of these measures into the Experiences of Close Relationships (ECR) scale in 1998 by Shaver’s group. This scale assessed beliefs about close relationships, and was based on two dimensions: avoidance of closeness and anxiety about closeness. Since the 1990s, a thriving tradition of attachment research has become established within social psychology. A central figure has been Mario Mikulincer, an Israeli experimental psychologist with a background in the study of learned helplessness and war trauma. From the 2000s, Shaver and Mikulincer became close collaborators. At times, the relationship between the social psychology tradition and the developmental tradition of attachment research has been rocky. Against expectations in the 1990s, research found that there is little relationship between the ECR (or other self-report methods) and either the Ainsworth Strange Situation or the Adult Attachment Interview (AAI). Nonetheless, in an apparent paradox, central hypotheses drawn from attachment theory are supported by research using the ECR. Shaver and Mikulincer have drawn on elegant experimental methodologies to further elaborate attachment theory. They have also supported the development of a younger generation of leaders of the social psychological tradition of attachment research in the USA and Israel. Over the decades, conflict between the social psychology and developmental approaches appears to have ebbed. An important contribution to this has been the three editions of the Handbook of Attachment, edited by Shaver and Cassidy. However, there remain difficulties in translating between the terminology and methodology of the two traditions of attachment research.

Introduction

In the early 1980s, the empirical research of Ainsworth, Sroufe, Egeland, and others had established the standing of attachment research as a credible paradigm for the study of early childhood within American developmental psychology (Chapters 2 and 4). Yet Bowlby’s theory claimed that attachment, as a behavioural system, continued to be relevant and influential after early childhood (Chapter 1). Writing in 1984, Bowlby reflected in correspondence that ‘I think we are still rather ignorant about the shift from a child–parent relationship (p. 428) to a spouse relationship’.1 Ainsworth and her immediate colleagues were engaged in extensive discussions in the early and mid-1980s about how to take the measurement of attachment beyond infancy, and felt sure that this would be possible. The first step taken in extending attachment methodology beyond infancy was Main’s ‘move to the level of representation’, and the development of the AAI (Chapter 3). A different route beyond infancy is the use of self-report to examine an individual’s perception of their attachment relationships. This route was adopted by some former students of Ainsworth, such as Mark Greenberg.2 However, in general, developmental psychology has preferred observational to self-report measures, and this was especially the case for attachment research. By contrast, social psychology as a subdiscipline was more favourable to the use of self-report measures. A sustained programme of research using self-report methods for assessing adult attachment styles was initiated by Shaver and colleagues.3

Shaver described his academic path to attachment theory as beginning as an undergraduate at Wesleyan University in the early 1960s. At the time, psychoanalysis was under heavy criticism within American academic psychology, especially from behaviourist approaches and their emphasis on experimental research and learnt responses (Chapter 2). Psychoanalysis had failed to develop much of a tradition of valid measurement techniques, and its account of motivation seemed speculative and untestable. The growing availability of federal research grants and the pressures to publish rapidly contributed to the attractiveness of highly focused mini-theories, rather than overarching models.4 However, to Shaver in the 1960s at Wesleyan, as later for Mukulincer at Bar-Ilan University in the 1970s, ‘the issues raised by psychoanalysts, beginning with Freud, are extremely important: sexual attraction and desire; romantic love; the development of personality beginning in infant–caregiver relationships; painful, corrosive emotions such as anger, fear, death anxiety, jealousy, hatred, guilt, and shame; intrapsychic conflicts, defenses, and psychopathology; individual and intergroup hostility; the brutality of war’. Compared to this, Shaver found that reading ‘academic social and personality psychology’ for his courses ‘seemed disappointingly superficial compared with psychoanalysis’.5

Shaver completed his PhD at the University of Michigan in 1970. Affirming the pivotal significance of internal experience against behaviourist approaches, Shaver’s doctoral research demonstrated the role of mental imagery in learning and problem solving. Shaver spent the next decade in New York, joining Columbia University in 1971 where he was asked to teach the undergraduate course in personality psychology.6 At that time, the Columbia faculty (p. 429) were offered the possibility of undergoing psychoanalysis at the Columbia Psychoanalytic Institute at a discounted rate. Shaver entered psychoanalysis with two presenting issues. A first was workaholism, which was emptying his life of other activities and harming his health and mood.7 A second was that Shaver had been ‘beginning to notice a repetitive and destructive pattern in my romantic relationships—always having a woman “in the wings” to whom I could flee if my primary relationship fell apart’.8

Shaver was in psychoanalysis for four days a week for two years. During this time, Shaver’s younger brother was diagnosed with cancer. The brothers had been close, often one another’s only friends for periods as the family moved repeatedly for their father’s work.9 On hearing the news, Shaver’s therapist said:

I’d like to step out of role and tell you, with great sadness, he will not live for more than 8–10 months. There is, unfortunately, no good treatment for testicular carcinoma. I think we should postpone talking about what we’ve been discussing lately and help you think through how you want to deal with this important situation. If you continue to focus on work and your personal problems, you may let this tragedy pass and then forever regret your lack of involvement.10

Shaver took leave from work, and moved for the summer to Minneapolis to be with his brother. The loss of his brother would prove critical to Shaver’s concern with attachment theory: ‘Eventually, he died, and I was present to receive his blessing. The intense grief that followed was unlike anything I had ever experienced, and it caused me to devote several therapy sessions to grieving. I also began reading books and articles about loss and grief, including some early papers by John Bowlby.’11

Throughout the 1970s, a central debate in American academic psychology was between behaviourist theories and new cognitive theories of the mind. Shaver acknowledged the strengths in experimental design of behaviourist research. And he acknowledged that any account of the mind had to include attention to cognition and communication. However, he criticised both paradigms, arguing that neither offered an adequate account of the role of emotion within human life: ‘Behaviorism is assaulted as if it were an oppressive ruler. But the powerful methods developed during the reign of strict behaviorism have now been combined with theoretical constructs borrowed from the sciences of communication and artificial intelligence to yield a robust cognitive psychology. Perhaps we can now begin extending it to include the notoriously uncomputerlike emotions.’12 By the end of the decade, he had (p. 430) come to conclude that his larger aim as a scholar was to ‘contribute to a theory of emotion which fits within an evolutionary and developmental framework’.13 More specifically, the loss of his brother led Shaver to a reorientation of his work around the themes of loneliness and depression.14 Shaver looked around him and saw a world in which these were the great themes of the age: ‘If eras can be characterised by salient emotions, ours must be the age of depression and loneliness.’15 To make sense of this, behaviourism and cognitive theories alone would not be adequate. He prophesised that ‘a revolution in ‘emotion science’ is about to follow the path blazed by the cognitive sciences. The emotion revolution is a logical outcome of the “age of depression and loneliness”.’16

The year 1979, however, would prove a turning point for Shaver, transfiguring his concern with depression and loneliness. One of the important factors contributing to this shift was meeting Robert Weiss, a Harvard-based sociologist, at a conference in 1979. Weiss had attended Bowlby’s seminars at the Tavistock during a sabbatical in 1971.17 Weiss accepted that the attachment behavioural system was less readily activated after infancy, and a different stock of attachment behaviours would be seen in adulthood. However, he highlighted that adults still provide one another with a secure base and safe haven, and can experience anxiety during unanticipated separations.18 Shaver reported that Weiss ‘encouraged me to conduct research more specifically on attachment and loss in adulthood’.19 Though familiar with Bowlby’s writings before 1979, Weiss’s urging gave Shaver impetus to explore beyond Bowlby’s account of grief, to consider attachment as a paradigm for conceptualising adult relationships. A second component of Shaver’s shift away from research on loneliness towards attachment was meeting and starting to date Gail Goodman at the University of Denver, where Goodman was a postdoctoral fellow with an interest in the psychology of children in court contexts. Shaver joined the University of Denver in 1980 and commenced a programme of work concerned with adult romantic relationships. Looking back in a recent paper, Shaver identified four great attachments in his life, which have offered him a secure base from which to explore and a safe haven providing regulation and comfort. The first two were his mother and then his analyst; 1979 saw the initiation of the two adult attachment relationships that would dominate his subsequent life: his marriage and attachment research.20

(p. 431) Specifically, in addressing his work to adult romantic relationships, Shaver’s interest was less in the concrete behaviour of a romantic dyad and more in the motivations and experiences of individual adults facing the predicament of intimacy or the rupture of relationships.21 Despite clear analogues, the concrete behaviour of an infant–caregiver dyad differed markedly from an adult romantic dyad. However, Shaver was excited by the possibility that at the level of motivation and experience, there were significant lines of continuity between infant and adult processes of intimacy and loss. In 1980, Bowlby published Loss, in which just such an argument was made. Similar claims were being made by Ainsworth and Sroufe.22 Themes of love, attraction, and commitment in romantic relationships had also, in the early 1980s, become established as areas for research in social psychology.23 New societies were established such as the Society for the Study of Personal Relationships, and the first issue of the Journal of Social and Personal Relationships appeared in 1984. In this context, one of Shaver’s early graduate students at Denver, Cindy Hazan, proposed in a graduate seminar that there might be individual differences in adult feelings about romantic relationships that mirrored the Ainsworth infant classifications.24 From these discussions emerged a co-authored 1987 article in the Journal of Personality and Social Psychology25 and a chapter in The Psychology of Love, edited by Sternberg and Barnes.26

These works presented the argument that adult romantic relationships exhibit the key features of the attachment system in infancy, above all the role of the partner as a secure base and safe haven.27 They also have other elements in common: reciprocal interaction and physical intimacy contribute to happiness; discoveries and experiences are shared with the other; and there is a mutual desire for the approval of the other. Hazan and Shaver pointed out that, for both infants and adults, the relationship can sometimes feel intrusive, contributing to anxiety and a wish for independence; and the availability of the other can sometimes feel in (p. 432) doubt, contributing to anxiety and hypervigilance about availability. More generally, Hazan and Shaver proposed that infant and adult behaviour were rooted in a common behavioural system, activated and terminated by the same kinds of conditions. In their writing, the authors drew on their own biographical experiences, including experiences of bereavement and of falling in love, to argue that the phenomenology and behaviours of adult attachment closely resemble the functioning of the attachment system in childhood. For instance, they observed that ‘as anyone knows who has been through such an experience, yearning for the lost person can continue for months or years and can be mingled with anger at the person for leaving’.28 On the basis of the analogies between infant attachment and adult romantic love, they developed a brief measure to assess individual differences in attachment in adulthood.

Part of the importance of this initial attempt at the development of a self-report measure of attachment was that it attracted Mario Mikulincer to attachment research. Mikulincer had finished graduate study in experimental psychology at Bar-Ilan University in 1985, and in 1986 joined the faculty at the University, where he would remain for over two decades. He also had a role in the Mental Health Department of the Israeli army, and worked with Zahava Solomon in conducting longitudinal research on the mental health and family life of combat veterans from the 1973 Yom Kippur War and the 1982 Lebanon War. These experiences would, decades later, be of great importance for his research in studying the effects of trauma on attachment in adulthood. Embarking on a research career following his doctorate, Mikulincer’s output was oriented by concerns reflecting his academic and military research responsibilities. From his doctoral research, he pursued inquiry into the origins of adult experiences of helplessness. The concept of ‘learned helplessness’ had been introduced by Seligman in the early 1970s to describe the way that repeated failures could contribute to a sense that it would not be worth continuing to try.29 Mikulincer was interested in the mental and emotional processes underpinning this effect and ways that it might be supressed.30 He drew on the work of Lazarus and Folkman in conceptualising helplessness as an avoidant response to a stressful stimulus, a response that could be circumvented through a variety of coping strategies.31 However, Mikulincer was curious also about the potential for failures to elicit rumination. In a sense this seemed like a coping strategy, since it focused attention on how the problem might be solved. However, rumination could become ‘the intrusion of autonomic failure-related thoughts into consciousness’. Rather than constructively addressing how a problem might be solved, these thoughts may ‘run ballistically on to completion, are irrelevant to one’s intensions, and are hard to suppress. They fill the cognitive system.’32

The ballistic metaphor signals a cross-over with Mikulincer’s research for the army, where his focus was on the role of intrusive images and emotions observed in combat veterans suffering from trauma and from loneliness. Mikulincer was interested in the role of the family in shaping the extent to which distress and fear intrude into the ordinary functioning of (p. 433) individuals. For instance, he was involved in research, published in 1990, which showed that family support could account for a quarter of variance in combat-related PTSD, and that this effect was fully mediated by feelings of loneliness.33 In the context of these concerns, Hazan and Shaver’s 1987 article caught his eye. Mikulincer recalls having ‘noticed similarities between (a) certain forms of helplessness in adulthood and the effects of parental unavailability in infancy; (b) intrusive images and emotions in the case of post-traumatic stress disorder and the anxious attachment pattern described by Ainsworth and her colleagues in 1978 and adapted by Hazan and Shaver; and (c) avoidant strategies for coping with stress and the avoidant attachment pattern described by these same authors’.34

According to a jointly written retrospective piece, a firm friendship and collaboration was begun in the 1990s when Mikulincer offered Shaver the chance to work together on a study of security-enhancement as a method of reducing intergroup hostility.35 However, already Shaver’s first publications in attachment theory reveal many facets of shared orientation between the two researchers, which included an ‘intense achievement motivation combined with distaste for egotism, a deep and perhaps neurotic split between the senses of ambition and sloth, and extended experiences with psychotherapy’.36 In Hazan and Shaver’s early writings on attachment, intense emotions of loss and of romantic love are palpable and explicitly personal, but also available to resource-thinking and theory-building. This explicit consideration of personal experiences as one source of information among others in the development of psychological theory became a hallmark of Shaver’s approach to research.37 At the same time, it is possible in Shaver’s first attachment publications to sense the relentless drive forcing the creation of a bridge between social psychological methods and psychodynamic concerns. In doing so, Shaver and Hazan attempted to grapple with the whole of human romantic life, and further at times, human close relationships in general.

The creation of this bridge was timely. The 1990s saw many other attempts to square psychoanalytic concerns with the tools of social and experimental psychology. Yet Shaver and Hazan’s work was unusual in reaching for the depths of the human experiences of intimacy and seeking to contain this in an assessment delivered to hundreds of people. The use of a self-report methodology was, however, especially attractive for researchers who had come to attachment theory without direct contact with Ainsworth. From the 1930s to the 1950s, Ainsworth had attempted to validate a self-report assessment of security, but had found that participants could report that they were secure as a defence against feelings of insecurity (p. 434) (Chapter 2). A person may be ‘so handicapped in his communication with others and in insight into his own needs and feelings that pencil-and-paper tests cannot reflect the nature and extent of his maladjustment’.38 Shaver appears not to have known about this work, and never cites it. In a paper written in 1984, Ainsworth urged future researchers: ‘do not take at its face value a person’s self reports of security, high self-esteem, high sense of competence or freedom from stress and anxiety, even though more credence may be given to self-reports of insecurity’.39

One attempt by developmental researchers to measure insecurity through self-report was Main, Hesse, and van IJzendoorn’s work on the Berkeley–Leiden Adult Attachment Questionnaire in the 1980s. This was a 200-item questionnaire, aiming to identify the four ‘states of mind regarding attachment’ from the AAI. This included two subscales for unresolved/disorganised states. A first was Unresolved States of Mind. This comprised items drawn directly from the coding manual for U/d on the AAI, with items designed to represent both lapses in reasoning (e.g. feelings of responsibility for a death in which the participant had no causal role) and lapses in discourse (e.g. feelings of lack of mental control when thinking about specific events). Additionally, ‘because unresolved trauma in certain cases may have a relation to post-traumatic disorders and to the dissociative disorders (Spiegel, 1989), some items indicating symptoms of these disorders were included’. A second subscale was items suggesting anomalous beliefs about time/space relations and causality, for instance beliefs in possession, mental telepathy, or precognition. The scales had strong reliability and stability over 12 months. The Unresolved States of Mind subscale had an association with U/d on the AAI of .63, accounting for 38% of variance. The Unusual Beliefs subscale also made a smaller but significant contribution, with the items for beliefs in possession performing especially well. Over 80% of participants could have been successfully classified on the AAI on the basis of their scores on the two scales. False positives were very rare; where misclassifications occurred, it was on the basis of false negatives. However, Main, van IJzendoorn, and Hesse were ultimate disappointed by the overall Berkeley–Leiden Adult Attachment Questionnaire measure, which failed to replace early successes. This was in contrast to the subscales for unresolved/disorganised states, which had proven quite successful. Nonetheless, the authors decided to leave the self-report scales for unresolved/disorganised states unpublished, and likewise their empirical studies validating these scales.40 As discussed in Chapter 3, Main has had difficulties bringing valuable work to publication where she remains somewhat dissatisfied with it.

In a private workshop in 1988 with close collaborators, including Main, Ainsworth stated that she regarded her own self-report measures as a failure: ‘I felt dissatisfied with the validity (p. 435) of my scales because of their inadequate coping with the whole matter of defensive maneuvers.’41 In her final publication in 1992, Ainsworth implied that self-report measures assess semantic memory (Chapter 1) and generalised propositions about the world, rather than episodic memory about concrete experiences.42 Any tendency for divergence between these two memory systems would be expected to produce distortions in the validity of self-report measures, in contrast to observational assessments which can capture both memory systems and their interaction.43 Following up on Ainsworth’s proposal in subsequent years, Pianta and colleagues, and Howard and colleagues, found that individuals classified as dismissing on the AAI self-report fewer adverse experiences and less distress on questionnaires than the assessment of independent observers, and preoccupied speakers on the AAI self-report more such experiences and distress.44 Furthermore, Bailey and colleagues reported that discrepancy between positive self-reported experiences in relationships and observer-reported assessment of insensitivity was associated with higher scores for avoidance and lower scores for proximity-seeking when the dyad were seen in the Strange Situation.45 Bailey and colleagues suggested that caregivers who hold positive accounts of relationships as a result of obliviousness or defences may miss or dismiss information relevant to emotions in relationships. This, in turn, would make them less likely to notice and respond to infant signals, or repair interactions when such signals are missed or misinterpreted.

Self-reporting attachment

A first attempt

From the start, then, the self-report methodology adopted by Shaver and Hazan was one that set them at odds with the community of researchers around Ainsworth. Nonetheless, Shaver received encouragement from Bowlby, who wrote a personal letter in 1985 to say (p. 436) that ‘I find your study of romantic love as related to childhood patterns most interesting’.46 In A Secure Base, Bowlby stated that Hazan and Shaver’s work suggested that there were, in adulthood, ‘features of personality characteristic of each pattern during the early years’ that could be measured using self-report. Bowlby affirmed his expectation that both the Hazan and Shaver adult classifications and the AAI classifications would both be predicted by the Strange Situation: ‘All our clinical experience strongly supports that view.’47 Nonetheless, he continued to worry, influenced by Ainsworth on this point, ‘that with the self-report paper-pencil method of appraisal it is well-nigh impossible to assess accurately how much defensive maneuvers have inflated security scores’.48

Shaver was certainly aware of the potential limitations of self-report measures. In the early 1980s, Rubenstein and Shaver acknowledged that ‘social psychology is, unfortunately, remarkable for its ability to reduce profound and fascinating human issues to rather superficial and uninteresting generalisations’.49 Part of the problem, Shaver felt, was that the operationalisation of variables in social psychology often abstracts away from the lived experiences of people in their complexity. Another part of the problem, however, was that self-report measures designed without sufficient care and creativity can end up reporting no more than the obvious. Throughout his career Shaver was fully willing to acknowledge the validity of Ainsworth’s concern about the potential for bias in self-report measures of security: ‘we agree that it is possible that some people defensively report that they do not worry about rejection and separation when actually they do worry about these issues, and some may have no conscious access to such worries, even though they exist’.50 Through the 1980s and 1990s, Shaver’s statements showed a certain hesitancy regarding what self-report assessments of attachment precisely measured.51 However, by the late 1990s he became confident that the participants’ report of their feelings, beliefs, expectations, and behaviours was not itself the object of self-report assessments of attachment, but rather markers of underlying psychological processes related to attachment that may not be fully accessible to the participants themselves.52 The challenge was to find a method of self-report that identified ‘noticeable feelings, beliefs, expectations, and behaviors that are related to, or arise from, underlying (p. 437) mental models and psychodynamic processes’.53 This approach to self-report measures can perhaps be placed in the context of a growing assumption within academic and clinical psychology, which became consensus by the 1990s, that responses to self-report questions about behaviours and attitudes would reflect underlying cognitive schemas.54

The Hazan and Shaver ‘love quiz’ was a single-item measure, asking participants to identify their relationships as corresponding to one of three adult attachment ‘styles’, modelled on the Ainsworth infant classifications. The concept of ‘lovestyle’ had been put forward by Hendrick and Hendrick to refer to an adult’s orientation to intimate relationships, on the analogy of ‘lifestyle’ as a person’s characteristic way of going about living.55 Hazan and Shaver used the concept of attachment styles in an analogous way, as a characterisation of an adult’s orientation to romantic relationships, conceptualised as an attachment relationship.56 Hazan and Shaver assumed that ‘conscious beliefs about love … are coloured by underlying, and perhaps not fully conscious, mental models’.57 While Hazan and Shaver headlined the applicability of attachment to romantic relationships, other adult relationships (for instance for a sibling) remained encompassed in what was, if at first implicitly, conceptualised as a comprehensive account of close relationships in general.58

Unlike Bowlby, Hazan and Shaver did not anticipate that there would be a longitudinal association between Ainsworth’s Strange Situation and adult mental models about relationships. And unlike Main, they did not have an underlying theory of common processes that might be in play in both the Strange Situation and adult cognition regarding attachment (Chapter 3). Instead, they situated their extension of Ainsworth’s categories to self-report adult attachment styles as purely pragmatic: ‘It would be naive to think that a style adopted in infancy remains unchanged or unelaborated all through life. Still, the search for connections between attachment in childhood and attachment in adulthood must begin somewhere.’59 The Ainsworth categories were knowingly torn up by their roots in the soil of developmental psychology and transplanted into social psychology: Hazan and Shaver were interested to see what might grow—though, as they would later acknowledge, Hazan and Shaver’s impression of what the Ainsworth categories entailed was partial and, at points, potentially erroneous.

The ‘love quiz’ appeared in the July 26th issue of the Rocky Mountain News, on the first and second pages of the Lifestyle section. Participants were asked ‘Which of the following best describes your feelings?’, and 599 participants replied. A secure attachment style was represented by the statement: ‘I find it relatively easy to get close to others and am comfortable depending on them and having them depend on me. I don’t often worry about being (p. 438) abandoned or about someone getting too close to me.’60 Close consideration of the statement is warranted as it contains Hazan and Shaver’s implicit theory of adult security, with implications for Shaver’s later work. As can be seen, security was not represented especially in terms of secure base or safe haven beliefs or expectations. Instead, the focus was on the absence of two feelings: (i) discomfort regarding closeness and (ii) worry about abandonment. In this, the item appeared to be in part working on the analogy of the behaviour of the infant from a secure dyad in the Strange Situation. The infant in a secure dyad approaches the caregiver on reunion to achieve physical closeness; by analogy, secure adults might be comfortable with emotional closeness. Lack of worry about abandonment appears to be a new element, however. It has a less clear analogue since infants in secure dyads do prototypically display distress on separation, besides the B1 subclassification. It might be that lack of worry about abandonment was conceptualised as an adult version of the capacity to be fully comforted on reunion. In the survey, 319 participants identified with the statement representing a secure attachment style.

An avoidant attachment style was represented by the statement ‘I am somewhat uncomfortable being close to others; I find it difficult to trust them completely, difficult to allow myself to depend on them. I am nervous when anyone gets too close, and often, love partners want me to be more intimate than I feel comfortable being’ (though there are conflicting sources regarding the exact wording used in the original study).61 In the statement as set out in the 1987 article, the focus seems to be on: (i) discomfort regarding closeness, (ii) nervousness regarding closeness, and (iii) distrust. Again the analogy with the Strange Situation is curiously partial. The infant in an avoidant dyad does not seek physical proximity, and when picked up may evidence discomfort. However, Ainsworth’s protocols certainly do not suggest that evidence of nervousness regarding closeness should be expected, as Shaver later acknowledged.62 In fact, 145 of Hazan and Shaver’s participants endorsed the statement representing an avoidant attachment style.

(p. 439) Finally, an anxious/ambivalent attachment style was represented by the statement ‘I find that others are reluctant to get as close as I would like. I often worry that my partner doesn’t really love me or won’t want to stay with me. I want to merge completely with another person, and this desire sometimes scares people away.’63 Here the focus seems to be on two feelings: (i) a wish for more closeness; and (ii) worry about abandonment. Once more, a partial analogy with the Strange Situation is apparent. Whereas the ambivalent/resistant infant classification in infancy includes inconsolability (understood as anxiety about caregiver availability), anger, or passivity, the Hazan and Shaver item includes no mention of anger or passivity. Instead, the analogy seems to be with the low threshold for activation of the infant attachment system and high threshold for its termination. It is not clear what, except the analogy with the infant classification, is ambivalent about the anxious/ambivalent statement in the ‘love quiz’. In the survey, 110 participants endorsed the statement representing an anxious/ambivalent attachment style. A further 25 participants checked more than one answer, despite having been instructed to select which statement best expressed their feelings. These participants were excluded from the analysis, a strategy that set the precedent for later research using the measure.

Hazan and Shaver stated that participants self-reported expectable correlates, though actually their findings offer a somewhat more complex story regarding the avoidant attachment style. Participants who endorsed the statement representing security reported that their romantic relationships were friendly, happy, and trusting. Hazan and Shaver stated that participants who endorsed the statement representing avoidance reported ‘fear of closeness’.64 They were not given options to report discomfort regarding closeness, and they did not, in fact, report more distrust than participants endorsing the anxious/ambivalent statement. According to the paper’s results, both avoidant and anxious/ambivalent participants had more fear of closeness than secure participants, though avoidant participants scored marginally higher.65 As such, the correlates of the avoidant attachment style were not exactly in line with Hazan and Shaver’s expectations. Anxious/ambivalent participants reported relationships marked by jealousy and emotional highs and lows. These findings were then replicated in a study with 108 undergraduate participants.

It was also quickly found that the ‘love quiz’ was capable of predicting behaviour beyond romantic relationships. A doctoral project completed by Michelle Hutt at Cornell, co-supervised by Hazan, used the ‘love quiz’ with 229 small business owners in the dairy and agricultural industry, and asked participants also about their patterns of exploration (e.g. innovation and curiosity in their work) and reliance on others (e.g. for support, help, or delegation), and asked them to respond to some vignettes with solutions to technical and personnel problems in the business context.66 Follow-up interviews were conducted with 15 participants. Hutt found that small business owners with a secure or avoidant attachment style ran more profitable businesses than participants with an anxious/ambivalent attachment style. This was despite the fact that the business of participants with different attachment styles (p. 440) did not differ in terms of size, number of employees, number of production units, location, or type of facility. Participants with an avoidant attachment style reported that they were less likely to seek help when they had a problem with the business, and were more likely to view their employees as ‘cogs in a machine’. Anxious/ambivalent participants stated that they were more likely to ask employees for emotional support when they confronted a technical problem. Hutt found that they were also more likely to have a chaotic management structure. Participants with avoidant and anxious/ambivalent attachment styles had higher turnover of staff than participants with a secure attachment style. They were also less likely to believe that their colleagues understood the goals and values of the business. The Hutt dissertation offered an early and elegant demonstration of the relevance of the ‘love quiz’ in predicting social behaviour, even in the work context, in theoretically expectable ways.

In an exploratory study, reported in 1990, Hazan and Shaver found that anxious/ambivalent subjects said that they struggled to finish tasks and slacked off after praise at work; avoidant subjects said that they were less likely to take vacations from work; and secure participants reported fewer psychosomatic symptoms and in general fewer illnesses.67 In another study from the 1990s, Cooper, Shaver, and Collins found that attachment style accounted for 5% of variance in mental health symptoms in a large student sample, after taking into account demographic covariates. The key findings from the study were that secure participants reported the lowest levels of depression and of problems managing anger.68 These findings were important for supporting the validity of the ‘love quiz’ categories, given the centrality of mental health to the conceptualisation of attachment since Bowlby.

In the early 1990s, Mikulincer published a flurry of dazzlingly inventive studies using a reformulation of the ‘love quiz’ measure. The Hazan and Shaver categories were reworked as Likert-style scales along three dimensions: secure, avoidant, and anxious-ambivalent.69 These papers established his reputation as among the most ingenious, nimble, and prolific experimental social psychologists of a talented generation. Mikulincer was promoted to full professor at Bar-Ilan in 1992 on the basis of these accomplishments, making him the youngest professor at any Israeli university at the time.70 Mikulincer’s particular skill was in articulating theoretical assumptions and then swiftly operationalising these in the laboratory. For instance, one key question faced by the emerging social psychological tradition of attachment research was whether self-report actually had reference to objective interaction and the shared, social world. Responding to this question, Banai, Weller, and Mikulincer had 72 student participants self-report their attachment style and receive a judgement as to their attachment style by other individuals, including by a friend and by a stranger asked to talk to them for five minutes. There were marked correlations between self-reported attachment style and the judgement made by friends and acquaintances. Friends agreed with the participant in 60% of cases, and acquaintances only slightly less in 56% of cases. This was essentially twice the rate of agreement expected by chance. These findings provided important evidence (p. 441) that self-reported attachment style was a construct representing a process sufficiently manifest in the individual’s behaviour that friends and acquaintances would come to the same conclusion.71

In the 1990s, Mikulincer and colleagues reported several findings suggesting differences in emotion regulation strategies between participants with different attachment styles. Mikulincer and Orbach found that an avoidant attachment style was associated with lower accessibility of negative memories. By contrast, participants who reported an anxious/ambivalent attachment style also had difficulty supressing negative emotions and memories. These participants experienced that one negative emotion or memory tended to spread into and elicit others.72 In another study, Mikulincer and Florian asked participants to complete a stressful task: handling a snake. They found that participants with an avoidant attachment style reported feeling more reassured by instrumental social support from a confederate, but not emotional support. Participants with a secure attachment style found reassurance in either form of interaction.73 By contrast, if they had an anxious/ambivalent attachment style, participants felt no different in response to emotional support, but had an increase in distress if they received instrumental social support. This was the first of many findings by Mikulincer and colleagues suggesting that an anxious/ambivalent attachment style is associated with a tendency to appraise experiences that others find reassuring or happy—such as gratitude—as a source of inner conflict, distress, self-doubt, or bitterness.74 Such associations seem particular to the anxious/ambivalent attachment style, and have been found to hold even after controlling for measures of self-esteem and social trust.75

In another study, Mikulincer asked participants to keep a diary over three weeks. He found that secure participants experienced more trust in their partners and adopted more constructive coping strategies in response to perceived violations of trust. Avoidant participants regarded the feeling of control as one of the functions of trust in a relationship. And anxious/ambivalent attached special importance to perceived violations of trust, and responded to these with rumination and worry.76 The theme of coping strategies was also explored by Mikulincer, Florian, and Weller, who studied response to missile attacks on Israel during the Gulf War. The researchers found that avoidant attachment style was associated with reports of using coping strategies that distanced the participant from other people, but (p. 442) only among participants living in more dangerous areas. Among participants living in less dangerous areas, there was no difference in the coping strategies used between participants with different attachment styles.77 This finding seemed to align well with Ainsworth’s observation that avoidant attachment behaviour was only seen in the Strange Situation; the same children did not display this conditional strategy in the lower-stress context of the home environment (Chapter 2). By contrast, illness, like long-term separations, was regarded as prompting a potentially chronic activation of the attachment system. Mikulincer and Florian found that men suffering from lower back pain only had worse mental health than matched controls when they reported an avoidant or anxious/ambivalent attachment style. Secure participants appraised their back pain in less-threatening terms, thought of themselves as more able to cope with pain, and used more problem-focused coping strategies.78

Together, the studies by Mikulicer and colleagues from the 1990s using the ‘love quiz’ suggested four conclusions. First, the studies offered evidence of the predictive power of the self-report categories, including the capacity to predict concrete observed behaviours. Second, the studies suggested that the attachment styles responded to contextual stimuli, in the laboratory and in the real world, in ways that were aligned with proposals by Bowby and Ainsworth. Third, the work by Mikulincer and colleagues showed that the difference between the avoidant and anxious/ambivalent attachment styles was evident across a range of domains, including recall of memories, interaction with their romantic partner in day-to-day life, and coping with emergencies and illness. This included intriguing findings such as that instrumental support provided to participants endorsing an anxious/ambivalent style actually made them feel less reassured. Fourth, the findings of Mikulincer and colleagues indicated that the secure attachment style was not reducible to the absence of avoidant or anxious/ambivalent attachment, but had its own correlates, a point that will be returned to in the section ‘The opposite of insecurity’.

Beliefs about self and others

In Separation, Bowlby had argued that internal working models contain at least two components: expectations regarding ‘(a) whether or not the attachment figure is judged to be the sort of person who in general responds to calls for support and protection; (b) whether or not the self is judged to be the sort of person towards whom anyone, and the attachment figure in particular, is likely to respond in a helpful way. Logically these variables are independent.’79 Bowlby’s term ‘independent’ inadvertently suggested to many readers the technical meaning that models of self and other would be unrelated and orthogonal; in fact, (p. 443) Gillath and colleagues are undoubtedly right that Bowlby meant only that the constructs should be regarded as conceptually distinguishable.80 Furthermore, it was not a distinction that he upheld elsewhere in his writings. This statement in Separation is an outlier: anchored by Hinde’s focus on interactions as the basis of relationships, Bowlby tended to treat the internal working model as rules and expectations regarding interaction between self and others, rather than as representations of the individual interactional partners themselves (Chapter 1).

Nonetheless, the idea of representations of the self and of others was a useful one for the social psychological tradition, and aligned with a wider interest in social psychology in individuals’ self-concept and their attitudes towards others. In line with this distinction, Shaver and colleagues interpreted attachment styles and their correlates as reflecting individual differences in an underlying psychological process, namely ‘beliefs about self and relationship partners’.81 In order to help situate their work as a kind of attachment research and to make the link with Bowlby, Shaver and colleagues felt obliged to say that the ‘love quiz’ was tapping individual differences in participants’ internal working models. However, in fact, Reis and Shaver had already expressed dissatisfaction with the concept of the internal working model, which they felt had been ‘defined so vaguely that different researchers employ rather different measures of it’.82 Reis and Shaver criticised Main for contributing to the further circulation of this unhelpful concept, which made it hard to understand her exact claims—a point that Main later fully accepted following feedback from Hinde (Chapter 3).83

In the 1990s, Hazan and Shaver sought to further articulate the relationship between attachment and romantic love. Following Ainsworth, they argued that adult romantic love includes elements of the attachment, caregiving, and sexual behavioural systems.84 The ‘love quiz’ might be influenced by the caregiving and sexual aspects of love, but they regarded the measure as primarily an assessment of attachment-related beliefs about the self and romantic partners.85 They agreed with Ainsworth (Chapter 2) that what distinguished the attachment system was that it related to the use of the partner as a secure base and safe haven, and to anxiety in response to unanticipated separations. On the basis of this disaggregation of elements of adult romantic love, Hazan and Shaver concluded that the ‘love quiz’ should be regarded as just a first attempt to distinguish the attachment components of adult relationships. Future research, they argued, should be built upon a theory about the beliefs about self and others at stake rather than treating their initial work as the model.86

(p. 444) Through the 1990s, following Hazan and Shaver’s lead, there was a boom in self-reported measures of attachment. Looking back on the period, Conradi and colleagues observe that ‘a bewildering variety of adult attachment typologies, adult attachment-related constructs, and measurement instruments’ were developed.87 Among the most consequential was the Bartholomew and Horowitz ‘Relationship Questionnaire’.88 This questionnaire was based on the idea that internal working models of self and other could be either positive or negative. In a paper from 1990, Bartholomew proposed that a secure attachment style was underpinned by positive models of the self and relationship partners, with the self considered worthy of care and others considered as generally available and willing to offer care. Avoidant attachment was underpinned by a positive model of the self and a negative model of relationship partners, resulting in self-reliance and distrust of others. Ambivalent/resistant attachment was underpinned by a negative model of the self and a positive model of others, resulting in emotional volatility and neediness. Bartholomew also introduced a new attachment classification, ‘fearful attachment’, which she conceptualised as underpinned by negative models of both self and other. For Bartholomew, the secure attachment style was associated with fluency and confidence in operating with both independence and intimacy. Fearful attachment, the inverse, was associated with problems in achieving either: ‘They desire social contact and intimacy, but experience pervasive interpersonal distrust and fear of rejection. The result is subjective distress and disturbed social relations.’89 Bartholomew felt that the Hazan and Shaver ‘love quiz’ had inappropriately collapsed the distinction between discomfort with intimacy and fear of rejection of attempts to achieve intimacy, which she believed represented substantially different profiles and predicaments.

The Bartholomew four-category model and questionnaire was a landmark in the history of self-report measures of attachment, since for the first time in this literature, categories used to assess individual differences in attachment were conceptualised as the effect of two underlying dimensions. Subsequently, attachment researchers in the social psychology tradition tended to conceptualise and analyse individual differences as dimensionally distributed.90 The addition of the fearful classification appeared to be of special value in characterising the attachment styles of participants experiencing adversity. This was shown by a series of studies by Shaver and colleagues using Bartholomew’s measure. The researchers (p. 445) found that the adult children of alcoholics predominantly fell into Bartholomew’s fearful category. They also found that a fearful attachment style was associated, as Bowlby had anticipated (Chapter 1), with reports of offering caregiving to attachment figures in order to get some attachment needs met.91 And in a study with a large student cohort, Brennan and Shaver found that a quarter of participants classified as fearful met criteria for personality disorder according to a self-report assessment compared to 6% of participants with other attachment styles.92

Shaver and colleagues concluded that Bartholomew’s fearful adult attachment style was the parallel, and likely a developmental correlate, of Main and Solomon’s disorganised attachment classification.93 They even used disorganised attachment and fearful attachment as synonyms through the 1990s.94 However, examination of Barthlomew’s 1989 unpublished doctoral thesis indicates an important qualification: Bartholomew regarded her fearful classification not as a general analogue for the disorganised attachment classification, but as an analogue for approach-avoidance conflict specifically.95 The fearful attachment style was based on the idea that the self is regarded poorly, causing distress, and other people are considered as a potential source of rejection. Jacobvitz insightfully observed at the time that this predicament might result in an approach–avoidance conflict—but not necessarily disoriented or direct apprehension regarding the caregiver.96 It would seem that the term ‘fearful attachment’ was complicit in the over-hasty comparison, suggesting a pervasive quality of fearfulness. In fact, in Bartholomew’s theory, the category was underpinned by fear of rejection rather than fear of the caregiver.97 Main and Stadtman had argued that whilst rejection by a caregiver would usually result in avoidant attachment, when the tone or context of the rejection was generally alarming, this would be a pathway to conflict behaviour that would be judged as disorganised attachment in the Strange Situation (Chapter 3). Rejection was not central to Main’s conceptualisation of the causes of disorganised attachment: it is not especially provided for by Main and Hesse among the indices of frightening/frightened caregiver behaviour, or anticipated in theory to result in infant disorientation or direct (p. 446) apprehension regarding the caregiver in the Strange Situation.98 There was also negligible association between a fearful attachment style and lack of resolution on the AAI.99

For a time in the 1990s, it seemed as if Bartholomew’s theoretical model and, to a lesser extent, her questionnaire, would become the dominant approach to adult attachment in the social psychology tradition. The approach found authorisation in a passage from Bowlby’s writings. It permitted the transition to dimensional approaches to measurement, which was becoming the norm in social psychology by the 1990s: the disadvantages of categories for statistical analyses were becoming increasingly widely known among attachment researchers, such as reductions in statistical power, decreases in scale reliability, and difficulties identifying non-linear relationships with other variables.100 And the division between positive/negative and self/other produced a neat model from which hypotheses regarding adult attachment styles could be generated and results interpreted. For instance, findings that avoidant attachment style was associated with stronger agreement that condoms protect against sexually transmitted diseases could be interpreted, in light of Bartholomew’s approach, as fuelled by negative associations regarding intimacy with romantic others.101 Furthermore, Bartholomew’s model seemed able to encompass infant and adult attachment together within the same frame, a serious asset given that the social psychology tradition was hoping to achieve acceptance from attachment researchers in developmental psychology. Finally, the addition of the fearful category was perceived as an advantage. This addition helped address an ambiguity in the operationalisation of avoidance in Hazan and Shaver’s original ‘love quiz’. It offered a parallel to the four-category model that was appearing in developmental psychology in the wake of Main’s innovations. And the fearful category seemed to offer a way for the self-report system to gain applicability for research with clinical samples.

Though it is certainly still used today, from the 2000s Bartholomew’s model increasingly lost ground to the two-dimension model of avoidance and anxiety underpinning the Experiences in Close Relationships (ECR) scale. The introduction of this scale is discussed in the section ‘Creating the ECR’. However, as well as the availability of the ECR, Batholomew’s model faced three salient limitations, which were pointed out by Shaver and his collaborators. First, the account of avoidant and anxious/ambivalent attachment was generally not well supported. An avoidant attachment style was not associated with a positive model of the self, but only with a desire to avoid acknowledging vulnerability. And this desire was relatively brittle, resulting in associations between an avoidant attachment style and self-blame when things go wrong.102 Furthermore, an anxious/ambivalent attachment style was not associated with a positive model of the other as ‘available and supportive’ as expected by Bartholomew, but with representations of others that were shaped by anger, jealousy, and a (p. 447) sense of partners as insensitive to their needs.103 This was hardly surprising, since the anxious/ambivalent attachment style had been operationalised specifically on the basis of espoused beliefs that romantic others are not adequately available.

Second, some of the intuitive pull and defensibility of Bartholomew’s approach came from the terms ‘positive’ and ‘negative’, which could have a variety of meanings. However, this was also a disadvantage in terms of usefulness for prediction.104 For instance, Mikulincer found that a secure attachment style did not appear to be associated merely with a positive working model of the self, but with one that was well balanced between acknowledgment of positive and negative aspects.105 It was not clear whether or not this constituted a falsification of the Bartholomew model precisely because it was difficult to pin down what was meant by a ‘positive’ working model.

Third, the Bartholomew questionnaire appeared to have weak cross-cultural validity, with the four categories not operating as expected along the two dimensions. In western samples, there was little overlap between participants endorsing items representing security and representing fearful attachment. By contrast, in African and Asian samples, there was a marked positive correlation between security and fearfulness, which ran counter to Bartholomew’s theory of positive and negative working models. Furthermore, whereas Bartholomew conceptualised avoidance and anxious/ambivalence as opposites, in fact they were negatively correlated only in a minority of countries around the world.106

Shaver and colleagues acknowledged that ‘it seems unlikely that Ainsworth’s research would have had such a profound influence on the field if she had eschewed types in favor of dimensions’ since ‘in the absence of a model of underlying processes, dimensions generally do not inspire theoretical advances’.107 Batholomew had circumvented the problem since she had offered a model of underlying processes in formulating her dimensions. Yet Fraley and Shaver argued that the dimensions Bartholomew identified were ultimately not the right ones.108 They hoped that further explicit discussion and debate about the dimensions underlying differences in adult attachment styles, and further psychometric inquiry, would help shed light on the relevant underlying processes.109

(p. 448) Avoidance and anxiety

Creating the ECR

Though Bartholomew and colleagues were able to source the idea of working models of the self and other in Bowlby’s writings, other accounts of security and insecurity were available too. At times, Ainsworth used the term ‘anxious attachment’ as a synonym for insecure attachment.110 This may have been motivated by a desire to remind readers that both infants in avoidant and ambivalent/resistant dyads showed anxiety about their caregiver’s availability in the home environment, and that it was only in the Strange Situation that the anxiety of the former group was masked. The use of ‘anxious attachment’ as a synonym for ‘insecure attachment’ was generally picked up by her immediate students and colleagues in the 1980s (with the exception of Main, who preferred the latter term). When discussing the Strange Situation, Bowlby sometimes used the term ‘anxious attachment’ to refer to the avoidant and ambivalent/resistant dyads, following Ainsworth.111 However, in the 1970s, Bowlby developed a different use of the term. This is most clearly seen in Separation where Chapter 15 is dedicated to an elaboration of the concept of anxious attachment. Already by the 1950s he had distinguished two classes of response to major separations: one that was manifest as clinging and protest, and another seen as detachment and avoidance of the caregiver (Chapter 1). Now, in Separation, he offered the term ‘anxious attachment’ as a label for the first of these two classes, with a particular emphasis on the vigilance regarding the caregiver’s availability: ‘Some children subjected to an unpredictable régime seem to despair. Instead of developing anxious attachment, they become more or less detached, apparently neither trusting nor caring for others.’112 This use of the term continued throughout the rest of his career. For instance, in Loss he argued that ‘one of the commonest forms of disturbance is the overready elicitation of attachment behaviour, resulting in anxious attachment. Another, to which special attention is given in this volume, is a partial or complete deactivation of attachment behaviour.’113

As Mikulincer and colleagues observed in 1990, the term ‘anxiety’ is potentially a very ambiguous one. It can imply various elements including worry, anger, and emotionality.114 Furthermore, as Ainsworth’s label of ambivalent/resistant attachment suggested, there may be a potential link with ambivalence (though Shaver felt that avoidant attachment was just as much a state of ambivalence regarding the attachment figure).115 However, despite such (p. 449) problems and ambiguities, the term ‘anxiety’ had been used by Bowlby, and this contributed to its appeal to Shaver and colleagues. The term would also have had an accessibility and intuitiveness, given its growing use and presence in public discourse from the 1940s to the early 1990s.116 Ultimately, the two classes identified by Bowlby were conceptualised by Shaver and colleagues from 1998 onwards as reflecting two dimensions in individual differences regarding close relationships: ‘The first dimension, Avoidance, captures variability in the tendency to feel uncomfortable with closeness or dependence. The second dimension, Anxiety, reflects a fear of abandonment’.117 Over the subsequent decade, the opposition came to be formulated rather differently: ‘The first dimension, typically called attachment avoidance, reflects the extent to which a person distrusts relationship partners’ goodwill and strives to maintain autonomy and emotional distance from partners. The second dimension, typically called attachment anxiety, reflects the degree to which a person worries that a partner will not be available in times of need.’118 However, contained in both early and late definitions is a technical sense of ‘anxiety’ to mean vigilance regarding the availability of attachment figures. In these definitions, ‘anxiety’ may be associated with anger and ambivalence, or indeed with general experiences or symptoms of anxiety, but these are not elements that formed the essence of the technical concept for Shaver and colleagues.119

In 1998, Brennan, Clark, and Shaver reflected that Shaver and Hazan ‘naively took for granted that Ainsworth et al. (1978) were correct in thinking of attachment patterns as categories or types. In retrospect, it is evident that Hazan and Shaver should have paid attention to Ainsworth et al.’s (1978) Figure 10 (p. 102), which summarized the results of a discriminant analysis predicting infant attachment type (secure, anxious, or avoidant) from the continuous rating scales.’120 Brennan, Clark, and Shaver were right that a dimensional approach to attachment had been napping peacefully on the backseat of Patterns of Attachment, even while the categories had control of the vehicle. However, close examination reveals that Ainsworth’s data did not completely support the claims of Shaver and colleagues to have identified a forerunner for their account of attachment in terms of two dimensions. True, in the discriminant function analysis in Patterns of Attachment, a first function had comprised essentially scores for avoidance. However, the second function was mainly constituted by scores on the resistance scale on first and second reunion, and with crying through the two reunion episodes. The function was as much or more constituted by anger or frustration as by anxious distress. Furthermore, whilst the two-function model reported in Patterns of (p. 450) Attachment was superb at predicting avoidant and secure dyads, it misclassified 30% of ambivalent/resistant dyads.

The conceptualisation of anxiety by Shaver and colleagues as vigilance regarding the availability of attachment figures was in part a theoretically motivated position, in which fear of abandonment was understood to be primary and to generate the other behavioural displays Ainsworth had identified.121 However, this conceptualisation of anxiety coincides with and seems to have been primarily influenced by a landmark piece of empirical research conducted with Kelley Brennan and Catherine Clark. The researchers created a pool of 323 questionnaire items from the different self-report measures of attachment available, which aimed to assess 60 named attachment-related constructs. These 323 items were then completed by 1,086 undergraduates enrolled at the University of Texas. When the items were factored, the result was 12 specific attachment-related dimensions, each with at least 10 items that could produce viable scales. The 12 scales ‘which can be viewed as “facets” (Costa & McCrae, 1992) of Avoidance and Anxiety, are as follows: (1) Partner is a Good Attachment Figure; (2) Separation Anxiety; (3) Self-Reliance; (4) Discomfort with Closeness; (5) Attachment-Related Anger at Partners; (6) Uncertainty About Feelings for Partners; (7) Discomfort with Dependence; (8) Trust; (9) Lovability/Relational Self-Esteem; (10) Desire to Merge with Partners; (11) Tough-Minded Independence; and (12) Fear of Abandonment.’122 A higher-order factor analysis of the 12 scales then revealed two underlying dimensions. The two-factor model accounted for 63% of variance.

At an item level, items reflecting the constructs ‘avoidance of intimacy’, ‘discomfort with closeness’, and ‘self-reliance’ were most strongly representative of the first factor, which was termed ‘Avoidance’. Items reflecting ‘preoccupation’, ‘fear of abandonment’, and ‘fear of rejection’ were most strongly representative of the second factor, which was termed ‘Anxiety’. Rather than part of an avoidant attachment style as assumed by Hazan and Shaver or an independent category as anticipated by Bartholomew, fear of rejection was in fact identified by the factor analysis as one of the three most important elements of the new ‘Anxiety’ dimension. Brennan, Clark, and Shaver found that the correlation between the ‘Avoidance’ and ‘Anxiety’ factors was only .12, ‘suggesting that the dimensions underlying attachment styles are essentially orthogonal’.123 The correlation appeared to be accounted for by a few constructs that loaded on both factors. These included perception of the unavailability of partners, perception of the self as unworthy of care, and a negative relationship with attachment security.

On the basis of the factor analysis, Brennan, Clark, and Shaver extracted 36 questionnaire items from the pool to craft the ECR scale. The items chosen were those that loaded most strongly on one or the other dimension, with the selection of 36—as opposed to 20 or 40—decided abitrarily.124 The stated aim of the ECR was to halt the raucous proliferation of new self-report measures of attachment, finding a neutral and common ground for the social (p. 451) psychology tradition of attachment research through ‘discovering and describing the essence’ of the self-report measures created to date.125 The ECR was also framed by the authors, in part as a result of its neutrality, as a supersession of Bartholomew’s theoretical model in presenting evidence for ‘anxiety’ and ‘avoidance’ as the two latent dimensions across all the other self-report measures. The measure was successful on both counts. Though other self-report measures continued to be used, the ECR soon became the ‘industry standard’ self-report assessment of attachment, and its two latent dimensions became the dominant explanatory model in the field.126

In 2003, a place was also reintroduced for Bartholomew’s ‘fearful’ attachment category by Mikulincer and Shaver. In Bartholomew’s model, fearful attachment occurred when a respondent was afraid of rejection by attachment figures and also had a negative model of their own worthiness. On the ECR, fearful attachment was re-envisioned as the co-occurrence of anxiety and avoidance. Mikuliner and Shaver’s stated reason for excluding fearful attachment as an independent category is quite curious in retrospect: ‘The fearfully avoidant self-description on the Bartholomew and Horowitz (1991) categorical measure … provides too low a hurdle for classification into the least secure category, and this creates too large a discrepancy between the clinical literature on disorganized attachment and the social-psychological literature.’127 Mikulincer and Shaver’s conceptualisation of fearful attachment as the coincidence of anxiety and avoidance, then, was an attempt to retain the bridge to clinical concerns and to the developmental psychology tradition, but with a narrower (and therefore hopefully more clinically meaningful) cut-off. This stance was also a development of Mikulincer’s longstanding interest in the intrusion of distress and fear into ordinary functioning. Already in his early work on learned helplessness in the early 1990s, Mikulincer had documented that anxiety can undermine an avoidant coping strategy, producing a state he called ‘disorganisation’.128

An anxious attachment style and an avoidant attachment style both, as Main had argued (Chapter 3), can be conceptualised as strategies, with the expectable result of offering some conditional form of the benefits of attachment relationships. However, the anxious strategy will be blocked if, when attachment or threat cues are present, a person fails to pursue the availability of their attachment figure. And the avoidant strategy will be undermined if it is intruded upon by anxiety or other overwhelming demands. Crawford, Shaver, and Goldsmith discovered that an avoidant attachment style was only associated with lower scores for attachment anxiety for participants with low or moderate levels of trait neuroticism (p. 452) (susceptibility to becoming distressed and worried). For those participants who became distressed and worried easily, an avoidant attachment style had the opposite effect, in that it was actually associated with increased scores for attachment anxiety.129 The authors interpreted this finding as suggesting that the avoidant attachment style contributes to attachment anxiety, rather than providing an alternative or defence against it, for individuals prone to becoming distressed and worried. Likewise, Mikulincer and colleagues demonstrated that an avoidant attachment style, which under ordinary circumstances is associated with low accessibility of attachment-related worries, in fact is associated with heightened accessibility of these worries when a cognitive load is added.130 The avoidant attachment style depends upon an infrastructure of attentional and regulatory control, even if this infrastructure is not the avoidant style itself.

In general, Shaver and Mikulincer conceptualised fearful attachment as a breakdown of attachment style, rather than having content in its own right.They repeatedly interpreted Main and Solomon’s disorganised attachment classification as ‘random fluctuations’ of anxious and avoidant behaviour.131 This, as Chapter 3 documented, is a misunderstanding, likely based on a lack of knowledge about how the construct is coded in practice.132 The status of ‘fearful attachment’ in the thinking of Mikulincer and Shaver has been neglected, perhaps because it too has been regarded as a state of meaninglessness. There has been little concern to inquire regarding the nature, form, and implications of fearful attachment. Indeed, it has not generally been regarded as necessary to test for fearful attachment or report results, leading in turn to little impetus for the design of research to investigate the phenomenon. For example, as Ein-Dor, Mikulincer, and Shaver put matters in a paper from 2011 about responses to threatening stimuli: ‘We did not include fearful avoidance in our analysis (i.e., a person above the median on both anxiety and avoidance) because attachment theory views this pattern as a breakdown of attachment strategies rather than a coherent strategy in its own right.’133

(p. 453) Minimising and maximising

From the early 2000s, one of the central aspects of the emergent collaboration between Shaver and Mikulincer was an attempt to develop a fleshed-out theory of the attachment system in adulthood, and the processes underlying individual differences in avoidance and anxiety. In this, they draw inspiration from developments in theory within the developmental tradition of attachment research. Main had conceptualised avoidance and ambivalent/resistance as conditional strategies, part of the human evolutionary repertoire for responding to environments that do not permit the direct satisfaction of the attachment system (Chapter 3). Cassidy and Kobak took Main’s concept of the infant attachment patterns as conditional strategies and argued that they represented two approaches to emotion-regulation: avoidant attachment was a minimising of the attachment system and of associated emotions which had the expectable outcome of avoiding rebuff and retaining regulation; ambivalent/resistant attachment was a maximising of the attachment system and associated emotions in the service of maintaining the attention of the caregiver.134

Shaver had worked on the 1999 edition of the Handbook of Attachment with Cassidy, and had become impressed by her ideas. In the early 2000s, Shaver and Mikulincer took her image of minimising and maximising of the attachment system as a central building block of their theory. They conceptualised the attachment system in adulthood as containing three modules.135 Together these form a cadence. The first and most basic module was the one described by Bowlby, and is based on the potential for perceiving threats. The first module is structured by the question: ‘Are there signs of threat?’ There can, naturally, be various sources of threat. Shaver and Mikulincer observed that these might include external dangers, such as the unexpected unavailability of attachment figures, the identification of a form of personal vulnerability. They might include internal states such as illness or worry about mortality. Shaver and Mikulincer proposed that symbolic threats to status, identity, or freedom would also be relevant to the first module of the attachment system.136 Where a sign of threat appears, the attachment system is activated, leading to ‘seeking proximity to external or internalised attachment figure’.137

Though the term ‘proximity’ was used by Shaver and Mikulincer, they did not mean physical proximity in the manner of Bowlby, but actually the sense of being close to a ‘source of felt security’, which might be physical or symbolic.138 Activation of the attachment system triggers a second module, which builds on the work of Ainsworth and Sroufe. The second module is structured by the question: ‘Is the attachment figure available, attentive, responsive, etc.?’ The etcetera signals Shaver and Mikulincer’s belief that the parameters for experiencing felt security in adulthood may be calibrated by factors including context and (p. 454) culture, as well as the nature of the perceived threat. It is not only physical proximity.139 Where the set-goal is achieved, Shaver and Mikulincer proposed that the result will be feelings of ‘relief, and positive affect’,140 grounded in the sense of being understood, validated, and cared about.141 These feelings in turn will contribute to confidence and competence, feeding back into the calibration of the first module, most notably in the appraisal of threats. Following Fredrickson, Shaver and Mikulincer referred to this as the ‘broaden-and-build cycle of attachment security’, since felt security, relief, and positive affect in turn come to shape the interpretation of threats, allowing for a more expansive, less cautious appraisal of the environment.142 Positive affects may also, as Sroufe documented, serve as social currency (Chapter 4). Recently, Mikulincer and Shaver qualified that Bowlby and Ainsworth’s concern with infancy has meant that too much has subsequently been made of an image of security as a state of being with attachment figures; in adulthood, broaden-and-build may also entail increased capacity to be comfortable with solitude, entering and exiting it without difficulties or shame, and benefiting from time alone in creative and flexible ways.143 There may be experiences of security, at least in adulthood, that are made for one.

The third module is initiated if the attachment figure is not ‘available, attentive, responsive, etc.’. This module is structured by the question: ‘Is proximity-seeking a viable option?’ Again, literal proximity is a potential meaning here. However, Shaver and Mikulincer appear to intend more the question of whether available, attentive, responsive (etc.) care can be expected to be achieved through directly seeking it. If yes, then the attachment system is ‘hyperactived’, in the form of ‘hypervigilance regarding threat- and attachment-related cues’. As for Main, there need be no conscious ‘decision’ regarding the deployment of a conditional strategy; in fact, it is a selection that generally occurs without conscious reflection. Hyperactivation entails a lowering of the threshold for activation of the first and second modules and a raising of the threshold for termination. This may entail, most directly, vigilance regarding internal or external information suggesting reasons for concern or alarm. Feelings such as jealousy, helplessness, and vulnerability will help maintain this vigilance.144 (p. 455) So will maintenance of pessimistic beliefs about whether coping alone is possible, and the attribution of threatening events to uncontrollable causes or inherent personal vulnerabilities.145 The difference between making a life and facing problems becomes collapsed.

Self-defeating actions may also keep the attachment system at ready alert: for instance, Mikulincer’s group documented associations on several occasions between the anxious attachment style and an odd escalation of investment and commitment as an activity becomes socially and/or economically unprofitable.146 As a result of various mechanisms that increase vigilance regarding threat- and attachment-related cues, the availability, attentiveness, responsiveness (etc.) of attachment figures is sought more readily in response to potential information about threat, and is sought in a manner more insistent, distressed, and frantic. This is likely to cause a tendency towards a cascade of thinking and worry about threats, vulnerabilities, and/or attachment, as one perception or thought will cue others linked to related feelings.147 The combination of vigilance and such a flow of associations may in turn contribute to the further identification of potential threats and reasons to wonder about the availability of attachment figures, burnishing the hyperactivating attachment strategy.148

On the other hand, Shaver and Mikulincer argued that if seeking the attachment figure’s availability is not appraised as viable, then the attachment system is ‘deactivated’, in the form of ‘distancing of threat- and attachment-related cues’.149 In Shaver and Mikulincer’s model, deactivation entails a raising of the threshold for activation of the first and second modules and a lower threshold for termination. In the formal presentation of their model of the attachment system, the selection of a conditional strategy is based on whether seeking the attachment figure’s availability is viable. However, there is significant ambiguity here. In other works, Shaver and Mikulincer acknowledged that selection of a strategy also depends on the perceived viability of the other strategy.150 As such, hyperactivation of the attachment system may in part reflect appraisals regarding the viability of seeking availability, and in (p. 456) part the extent to which deactivation is possible. More speculatively, deactivation of the attachment system may in part reflect appraisals regarding the viability of seeking availability, and in part the extent that hyperactivation is possible.151

In Shaver and Mikulincer’s model, when the attachment system is subject to deactivation, information incongruent with the deactivation of the attachment system is inhibited or appears less salient. For instance, the information itself may be altered, obstructed, or suppressed; or response to the information may be dampened, redirected, or postponed.152 This not only includes emotions associated with threat, vulnerability, or separation, which might reactivate the attachment system. Similiarly inhibited or reduced in salience are happy emotions like joy and comfort, which might otherwise promote interpersonal closeness or investment in close relationships.153 Someone may be radiating love, and all that is felt is mild warmth. Deactivation also means that the availability, attentiveness, responsiveness (etc.) of attachment figures is sought less readily, less insistently, and to a lesser extent, with a tendency towards taking cognitive, emotional, and physical distance from relationship partners. In turn, the individual feels less emotionally assailable. Self-control is held tightly, perhaps even with a white-knuckle grip, as both the condition and experience of the defensive exclusion of the attachment system. As a consequence, self-control can itself be felt as a form of comfort, as a substitute for attachment set-goals.

The deactivation of the attachment system is likely to cause difficulties moving between information about threats, vulnerabilities, and/or attachment, as one perception or thought is less likely to cue others linked to feelings.154 This reduces opportunities for identification of threats and for feedback indicating that attachment figures are available, attentive, responsive (etc.), contributing to a stabilisation of the deactivating strategy. The strategy was understood by Mikulincer and Shaver to be implicated across the whole range of attachment-based experiences, even those that generally remain private. For instance, Mikulincer, Shaver, and colleagues conducted a study in which they asked participants to report their dreams each morning for a month, as well as other information about how easy or difficult they found their day. The higher a participant’s score for avoidant attachment, the less likely they were to report dreams that included any active support-seeking. The higher a participant’s score for anxious attachment, the more likely it was that participants would report dreams in which (p. 457) they had wanted to be close to another person, and that this was especially likely following more stressful days.155

It has evidently been a source of frustration for Shaver and Mikulincer that characterisations of their work by attachment researchers from the developmental tradition have often assumed that a self-report measure could reflect merely conscious attitudes towards relationships. As well as research on dreams, another study they pursued was an examination of the relationship between the ECR and the Rorschach inkblot test used extensively by Ainsworth in her clinical work and early research. The Rorschach is a projective test assessing unconscious psychodynamic processes. Participants’ associations on the Rorschach measure were blind coded. It was found that Rorschach responses suggesting difficulties in controlling emotion and of feeling unworthy were associated with self-reported attachment anxiety. And Rorschach responses suggesting difficulties in acknowledging personal needs were associated with self-reported attachment avoidance. Associations were moderate, in the region of r = .32 to .34. Berant, Mikulincer, Shaver, and Segal interpreted the findings as suggesting that non-conscious projections in interpreting an ambiguous but evocative stimulus in part express self-report attachment styles. As such, they argued, the self-reports are not merely reflecting conscious attitudes, but tap non-conscious processing as well.156

Shaver, Mikulincer, and colleagues argued that each module of the attachment system is informed not by a static mental representation but by a hierarchical associative network, in which specific episodic memories become placed as exemplars of more generalised schemas about how relationships work.157 This associative network for each module is continually updated in light of social encounters, and also in light of information from the other modules. So, for instance, the appraisal of whether the attachment figure is available, attentive, responsive (etc.) will be informed by a variety of factors. Three, above all, are given precedence: how often and in what contexts threats have been experienced; feelings of felt security, relief, and positive affect from times when attachment needs were met; and tendencies to hyperactive or deactive the attachment system from occasions when these conditional strategies were employed and proved sufficiently successful. Shaver and Mikulincer propose that, given the existence of a hierarchical associative network, ‘with respect to a particular relationship and across different relationships, everyone has models of security-attainment, hyperactivation, and deactivation and so can sometimes think about relationships in secure (p. 458) terms and at other times think about them in more hyperactivating or deactivating terms’.158 Nonetheless, the combination of (i) the predominant form of actual social experiences and (ii) self-reinforcing cycles resulting from the adoption of particular strategies from childhood onwards will contribute to associative networks that are especially or chronically available. This produces a person’s characteristic attachment style. In characterising individual differences in attachment style, the ECR was conceptualised as a kind of snorkelling: though wholly submerged in manifest, espoused information about behaviours, beliefs, and feelings, it nonetheless is believed to breathe latent information from a hierarchical associative network offering generalised schemas about how relationships work.

In line with this model, Overall and colleagues asked participants to complete self-reports of attachment style for various different family relationships, friendships, and romantic relationships. They found that the attachment orientations for specific relationships were somewhat independent but nested within overarching tendencies towards a particular attachment style for each domain. Furthermore, each domain was somewhat independent but nested within an overall attachment style.159 Interpreting these findings, Shaver and Mikulincer described overall attachment styles as relatively stable, but responsive to changes in relevant social or psychological context.160 Social experiences are often stable over time as a result of structural factors in people’s lives. These are then reinforced by cycles of the broaden-and-build, minimising or maximising strategies. However, some relationships or components of relationships may feature experiences that run counter to the dominant attachment style. And over time, repeated social experiences or forms of appraisal that reinforce this style can become pervasive across the hierarchical associative network. This is illustrated well by an interesting unpublished doctoral study conducted in Mikulincer’s group. Lavi pursued a longitudinal study of 73 dating couples over eight months. She used an observational assessment at the start of the period, examining each partner’s sensitivity and responsiveness to the other. Relationship-specific and general attachment style was assessed at three time points. Lavi found that observed sensitivity and responsiveness predicted declines in attachment insecurity. First this occurred in relation to the specific relationship partner. Over time, however, there was a change in the participant’s general attachment style.161

Mikulincer argued that this elaboration of attachment theory gains its usefulness and particular value in ‘integrating different, perhaps even contradictory, views of human nature and maintaining a dialectical tension between opposites of four kinds’:162

(p. 459)

  1. 1) The shaping and constraining influences of past experiences versus the influence of current contexts and experiences;

  2. 2) the intrapsychic nature of behavioural systems and working models versus the relational, interdependent nature of feelings, experiences, and social behaviours;

  3. 3) the goal-oriented, promotive, expansive, self-regulatory function of behavioural system versus their defensive, protective, distress-regulating functions;

  4. 4) the centrality of basic fears, conflicts, and prevention-focused motivational mechanisms, as well as promotion-focused motives.163

Shaver and Mikulincer’s model of the attachment system can be usefully considered in these terms. The appraisal of the potential effectiveness of seeking the availability of the attachment figure is prompted in module one. The question that must be answered in module two is whether the goal of this system—an attachment figure who is available, attentive, responsive (etc.)—should be sought directly. This represents the hinge between ethology and personality psychology.164 If the goal of the system can be met directly, then this contributes to a broaden-and-build cycle that can support self-regulation. The availability of a secure base and safe haven are expected to contribute to increased richness and stability in both an individual’s inner world and social relations. If not, then module three enacts a defensive, protective, and distress-regulating strategy, which may also have promotive and self-regulatory potential in some circumstances.

The maximising strategy, on the one hand, promotes threat-sensitivity and the promotion of self-regulation through attempts to achieve the availability of attachment figures. The minimising strategy, on the other hand, promotes composure under low and moderate stress and the promotion of self-regulation under conditions in which attachment figures are appraised as not available. The selection of one strategy or another is shaped in part by learning from previous social encounters and the strategies used to respond to them, as well as available resources for emotion regulation. Yet the selection of a strategy is also fed in part by the actual nature of the present social interaction, cultural norms, and relational cues regarding the potential effectiveness of particular strategies in the particular moment. Individuals do not have, in Shaver and Mikulincer’s work, a single representational model of relationships. Instead, hierarchically organised associational networks make various attachment styles available. As such, they depend upon: the qualities of a particular relationship; previous experiences of and expectations about this kind of relationship; and the particularities of present-day interactions. That these sit nested within and interpreted through an overall attachment style shaped by the weight of personal history does not exclude this internal diversity and dynamic potential.

Shaver and Mikulincer claimed their elaboration of attachment theory as an integration of components of developmental, social, and personality psychology. This was in line with Bowlby, who always described attachment theory as an account of the relational development of personality.165 However, it was in contrast to the developmental tradition of (p. 460) attachment research. Among developmental researchers, appeals to ‘personality’ have been relatively rare and mostly appear in summaries of Bowlby.166 Yet in claiming attachment styles as part of personality psychology, the question was inevitably raised regarding how avoidant and anxious attachment styles relate to established personality constructs. And more pointedly, whether they added anything. An attempt to answer this concern was undertaken by Noftle and Shaver in a paper published in 2006. They examined the relationship between the ECR and a self-report measure of the ‘big five’ constructs in personality psychology: openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism. There were few major associations between the avoidant attachment style and the personality constructs. By contrast, in two studies Noftle and Shaver found associations between attachment anxiety and neuroticism of r = .42 and r = .52. The extent of this association suggested to the researchers that there is ‘some conceptual overlap between the constructs’.167 The strongest associations were for items for neuroticism that pertained to the susceptibility and frequency of negative affect (items for: being depressed, likely to be moody and nervous, excessively worrying, and not being emotionally stable). This suggests that the readiness of negative affect plays a key role in the overlap between the two psychological constructs. The readiness of negative affect is certainly shaped by culture and experiences past and present. However, it is also widely regarded by psychologists as partly a heritable trait.168

Following up this expectation, Crawford, Shaver, and colleagues reported in 2007 from a study of 239 twin pairs. Biometric models identified no contribution of heritable factors to the likelihood of an avoidant attachment style. However, 40% of variance in anxious attachment could be attributed to heritable factors. Furthermore, 63% of the association between anxious attachment and a self-report measure of personality disorder could be accounted for by common genetic effects. Such findings suggest that anxious attachment, and its association with mental health, may in part be an effect of heritable differences in the susceptibility and frequency of negative affect.169 However, Crawford, Shaver, and colleagues also argued that there are important differences between anxious attachment style and neuroticism. Firstly, anxious attachment is especially responsive to attachment-relevant cues, not merely generally anxiety-provoking stimuli. Secondly, there is still a substantial amount of variance not explained by heritable factors. Thirdly, even when neuroticism is statistically controlled, (p. 461) anxious attachment has been found to make a modest contribution to mental health outcomes such as depression and complex responses to bereavement.170 At a fundamental level, the experience of anxiety is solely an expression of neither individual personality nor attachment style. Both seem to play a role. However, this highlights the question of what convergence there is between the anxious attachment style and preoccupied states of mind regarding attachment, if both are conceptualised as maximising strategies. To address this question it is necessary to examine the ECR and the AAI in the context of social psychology and developmental psychology as two different subdisciplines concerned with attachment.

Two traditions of attachment research

As we have seen, Bowlby anticipated that both the AAI and the Hazan and Shaver ‘love quiz’ would both be predicted by the Strange Situation. This was also Shaver’s explicit position by the late 1990s: ‘some of their components, especially ability to depend on attachment figures, should be related if both stem from a person’s attachment history’.171 The first major attempt to assess whether the AAI and self-report measures would converge was published by Crowell, Treboux, and Waters in 1999. In their study, 81% of participants classified as secure-autonomous on the AAI identified themselves as having a secure attachment style on Bartholomew’s Relationship Questionnaire. However, only 42% of participants who were classified as preoccupied, dismissing, or unresolved on the AAI reported themselves as insecure on the Relationship Questionnaire.172 Crowell and colleagues therefore cautioned against equation of the AAI and self-report measures. They argued that it was a recipe for confusion that both systems used a version of Ainsworth’s terminology, such as ‘security’, when in fact the referents were quite different. Whereas Main was assessing ‘states of mind regarding attachment’, the self-report tradition was assessing a person’s sense of ‘comfort in close relationships’.173 There might be expected to be a little empirical overlap between these phenomena, but Crowell and colleagues argued that they were completely different constructs.174

It was not certain what the two traditions could mean for another. Nonetheless, the expectation that both the ECR and AAI were measures of ‘attachment’ implied that it should be possible to translate between the two measures, and that the two traditions should be to a (p. 462) large extent permeable to one another.175 In a 1998 book chapter, Bartholomew and Shaver offered a characterisation of the two traditions of attachment research. They described the developmental tradition as mostly led by direct students of Ainsworth. In Bartholomew and Shaver’s characterisation, ‘researchers in this group tend to think psychodynamically, be interested in clinical problems, prefer interview measures and behavioural observations over questionnaires, study relatively small groups of subjects, and focus their attention on parent–child relationships’. It is possible to quibble with some particulars—Kobak was conducting research on parent–adolescent relationships; Waters had little psychodynamic basis for his thinking except as mediated through Ainsworth and Bowlby—but in general this characterisation seems fair.176 By contrast, the social psychology tradition was described as shaped essentially by the characteristics of social and personality psychology of the period. Members of this community ‘tend to think in terms of personality traits and social interactions, be interested in normal subject populations, prefer simple questionnaire measures, study relatively large samples, and focus on adult social relationships’.177

The division between the two traditions was in part fed, it should be acknowledged, by the longer-standing professional tensions within the academic discipline between developmental and social/personality psychology. As Lapsley and Quintana observed, writing in the period in which the two traditions of attachment research were emerging, ‘it is hard to imagine an academic division of labor that is more formidable, and at the same time more artificial, than the division between social and developmental psychology’.178 Commenting on the division in the late 1990s, Bartholomew and Shaver noted that ‘not surprisingly, the members of these two research subcultures tend to speak past each other’.179 Shaver and Mikulincer added, in 2002, that another key difference between the traditions was that the developmental tradition only rarely tested causal propositions experimentally, whereas the use of experimental designs was a central part of the social psychology tradition, thanks in part to the work of Mikulincer. Furthermore, the developmental tradition tended to rely on rather cluttered coding systems supported by predictive validity but lacking rigorous psychometric analysis of categories and scales. By contrast, the social psychological tradition had a better track-record of cleaning and refining their measures to ensure psychometric rigour. Shaver and Mikulincer also argued that whereas their work shed light on the actual functioning of ‘the attachment behavioural system’ itself, by contrast the AAI was a measure of ‘individual differences in “state of mind with respect to attachment” ’.180 They claimed, in (p. 463) an unabashed polemic, that the AAI was likely not an assessment of attachment at all. Rather, given that it was validated against the Strange Situation, it might be better conceived as an assessment of cognitive aspects of the caregiving system. This is a claim they would later retract in 2004, only to swing to the other pole in asserting that both the AAI and self-report measures taped the same processes: ‘attachment-style scales differ from the AAI in content and method but not in the core attachment-related processes they index’.181

Yet in addition to specific methodological differences, Shaver and Mikulincer claimed that the most important difficulty was ‘professional ingroup–outgroup tensions between researchers in the two traditions’.182 In fact, looking back on the period, it seems as if professional and methodological tensions worked neatly to reinforce one another. The developmental tradition tended to regard the social psychologists as conducting quick-fix, superficial research on conscious attitudes towards intimacy, lacking fidelity to the ethological basis of attachment theory. The social psychology tradition tended to regard the developmentalists as lashed to Main’s four categories, and unable or unwilling to think critically or creatively about them. They also expressed scepticism whether a research agenda based on laborious observational and interview-based methodologies was practical in the context of contemporary academic psychology.183 This latter point certainly has purchase. One of the contributing factors to the growth of the social psychological tradition of attachment research has been that the entry requirements are much lower. The ECR is freely available and easy to administer on student populations and through online platforms, putting it within the reach of researchers even without a research grant or with only a side-interest in attachment. This cannot be said of the Strange Situation and AAI, which require substantial training and knowledge to be able to code effectively, or sometimes even interpret published results. However, recent developments like Fearon and colleagues’ cut-down AQS (Chapter 2), Dozier and colleagues’ cut-down AAI (Chapter 6), Madigan and colleagues’s cut-down assessment of frightening and disrupted caregiving (Chapter 6), and Ensink and colleagues’ short assessment of parent reflective functioning184 may reduce the resource burden of the developmentalists’ measures, if their validation proves successful.

Through the 1990s, the developmental and social psychologists blamed one other for failing to conduct comparative research. The developmentalists regarded it as rather unforgivable, and as an expression of a quick-fix mentality, that the social psychologists did not undertake the training to conduct observational research even for a measure as foundational as the Ainsworth Strange Situation. The social psychologists found it rather unforgivable that, even though it would have been quick and easy to do, the developmentalists did not include self-report measures of attachment when assessing parents in the Strange Situation, when conducting the AAI, or in the adolescent and adult phases of their longitudinal (p. 464) studies.185 At times these tensions operated in the background. So, for instance, research on parenting using self-report measures of attachment has remained quietly blunted, with leaders of the social psychological tradition giving the topic little attention.186 In part this was an effect of the division between social and developmental psychology in general, in which parenting is generally assigned to the latter. However, another factor may have been that, at best, self-report measures would draw with the AAI in predicting parenting, and there was the possibility that they would lose.187

Yet, at times, the tensions between the social and developmental traditions of attachment research moved from mutual inattention to a public disagreement. In the mid-1990s, Crowell and Waters described Shaver’s work as ‘thoughtful and provocative’ and as ‘exciting and potentially very useful’, but stated their view that in fact there is ‘little common ground’ with the tradition initiated by Ainsworth.188 In responding to the hostile reception from the developmental tradition, one tactic used by Shaver and colleagues to find mandate for their work was to offer anachronistic depictions of the field’s founders to make Bowlby and Ainsworth foreshadow their own work. For instance, Ainsworth has been repeatedly described by Shaver as using the concept of ‘a child’s attachment style’189, and Shaver and Mikulincer explicitly asserted that ‘Bowlby claimed that memorable interactions with others throughout life can alter a person’s working models and move him or her from one region of the two-dimensional attachment-style “space” to another’.190 These anachronisms reflected a claim to the right to inherit attachment theory from the first generation without using the observational methods that had come at least in part to constitute the basis for inheritance. They show the desire for legitimacy and heritage among the social psychological researchers, in the context of a general lack of acceptance—or even threats of symbolic dispossession—by the developmental tradition.

Despite trouble with the developmentalists, use of self-report measures of attachment was booming by the 1990s. Some researchers from the developmental tradition, such as Belsky and Cassidy, expressed concern that the social psychology tradition ‘might actually overshadow the work in infancy and childhood that gave rise to it’, and proposed that this threat gave added impetus to the need to achieve a ‘linkage of these two schools of attachment inquiry’.191 In this context, Seymour Weingarten, editor-in-chief of Guilford Press, approached (p. 465) Cassidy and Shaver at the 1995 meeting of the Society for Research in Child Development in Indianapolis with the idea of a Handbook of Attachment. The Society’s biennial conference is the central conference for the field of development psychology. Shaver had attended the 1993 conference and had presented a defence of the self-report approach to attachment as relevant for understanding intergenerational and developmental continuity of individual differences in attachment over time.192 Weingarten’s idea was for a handbook encompassing both traditions of research. Cassidy and Shaver ‘were barely acquainted when the project began’, and did not meet face-to-face after the initial conversation in Indianapolis during the four years of work on the first Handbook, which would be published in 1999. Nonetheless, in their preface to the Handbook, Cassidy and Shaver stated that ‘we have grown to know and respect each other professionally in ways that people rarely do’.193 In the late 1990s, Cassidy informed and nourished Shaver’s thinking about individual differences in attachment as minimising and maximising strategies.

Perhaps in an attempt at rapprochement, Crowell (a developmentalist) and Fraley and Shaver (social psychologists) co-authored a chapter in the first edition of The Handbook of Attachment. They acknowledged that there existed ‘considerable tension between the AAI and self-report traditions within the field of contemporary attachment research’.194 They observed that the two traditions had been primarily concerned with parenting and romantic love respectively, and pursued very different ways of eliciting and analysing their data. Nonetheless, both traditions appeared to be testing and elaborating aspects of attachment theory. This apparent common theoretical basis lent impetus for the need for comparative research. In pursuing this path, Shaver and Brennan collaborated with Jay Belsky, who had long been calling for integration (and, beyond this, was always something of a free spirit among researchers in the developmental tradition). In a paper from 2000, the researchers reported from their study with 138 Caucasian women from working- and middle-class families in central Pennsylvania. Strangely, the researchers did not report the relationship between self-reported anxiety and avoidance and the AAI categories. Given that the overall priority of the paper was to emphasise the compatibility of the two measures, the reader is perhaps left with the suspicion that this analysis was run, but the results were not regarded as desirable. Instead, the authors reported a few moderate associations on various subscales. Self-reported avoidance was associated with lack of memory for childhood events, anger at the speaker’s parents, and lower overall coherence.195 However, Shaver and colleagues acknowledged that ‘the degree of association was relatively modest, and due only to certain aspects of each measure’.196

Roisman, Fraley, and colleagues embarked on further work comparing the AAI and attachment style measures. A meta-analysis of nearly a thousand participants who had completed (p. 466) both kinds of assessment by 2007 revealed a trivial association (r = .09).197 One of the few specific findings was that unresolved trauma, but not unresolved loss, had a small association with self-report attachment anxiety (r = .20).198 In Attachment & Human Development, Roisman and Fortuna reported the results of a study with 160 college students. Participants completed the AAI, a self-report measure of attachment, a self-report measure of life stress, and a self-report measure of mental health symptoms. The association between the AAI and self-reported attachment style were, as expected, negligible. They both, however, predicted mental health symptoms. Self-reported attachment avoidance and attachment anxiety both had substantial associations with self-reported mental health symptoms (.43 and .48 respectively). These associations were unaffected by the participants’ self-reported life stress. The strength of the associations can in part be accounted for by the fact that both the assessments of attachment style and mental health were self-reported. However, it is nonetheless curious that an avoidant attachment style had an association with these symptoms of almost equivalent strength. By contrast, the AAI had a marked association (.34) with self-reported mental health under high experienced life stress, but only a weak association (.19) when life stress was low.199

Pursuing the question of the interrelation of measures from the two traditions, Bernier and Matte-Gagné reported a study of 59 Canadian families. They conducted the AAI, the ECR, and the Strange Situation, a measure of caregiver sensitivity conducted in the home, and self-report measures of the marital satisfaction of both partners. Rather than using the AAI categories, the researchers were convinced by Fraley and Roisman (Chapters 2 and 4) that attachment phenomena are better regarded as dimensions, and so they therefore reported only the AAI coherence scale. The researchers found no association between the AAI and the ECR. They also found, as anticipated, that coherence on the AAI was associated with greater caregiver sensitivity and increased incidence of secure infant–caregiver attachment in the Strange Situation. There was no statistical association between the ECR and either caregiver sensitivity or the Strange Situation.200 Marital satisfaction was not associated with (p. 467) coherence on the AAI. However, ECR anxiety of one partner was associated with dissatisfaction with the relationship by the other, controlling for coherence on the AAI. Bernier and Matte-Gagné concluded that ‘only romantic attachment was found to relate to marital satisfaction, while only the AAI was found to relate to caregiving’.201 Such findings suggest that, even if they claim a common theory, the two traditions are ultimately studying different phenomena.

In 2006, Shaver stated that he and his collaborators had essentially given up reading the outputs of the developmental psychology tradition of attachment research.202 Shaver’s observations have an ‘anger of despair’ quality to them. They mark trends in the mid-2000s that suggested that the two traditions could go their separate ways, segregated from one another and out of functional communication. Yet three important factors have contributed to better relations and productive interaction between the traditions. One has been sustained efforts to develop institutional and personal connections since the late 1990s. In 1999, the journal Attachment & Human Development was founded, with Howard Steele serving as Editor-in-Chief. Though the first issues were tilted more towards the developmental tradition, the journal sought to represent both traditions, and Shaver was an Associate Editor from the start. In its fourth year, the journal hosted a major discussion of the relationship between the traditions, with a lead article by Shaver and Mikulincer, eleven replies from attachment researchers from both camps, and a response from Shaver and Mikulincer.203 And over subsequent years, the journal has assiduously made space for contributions from both traditions, contributing to a sense of a common set of problems and theories.204 Likewise, Cassidy and Shaver worked together on a second edition of the Handbook of Attachment published in 2008. This mammoth project was evidently an experience that brought the researchers closer as colleagues, and convinced Cassidy of the benefits of the self-report methodology: in subsequent years, Cassidy and Shaver collaborated on theoretical papers and empirical research using self-report measures. Cassidy also used them in her own work, independently of Shaver.205 In 2010, Cassidy and Shaver also worked together in co-editing a special issue of Attachment & Human Development presenting an attachment perspective on incarcerated (p. 468) parents and their children, with contributions primarily using the AAI and Strange Situation as their methodologies.206

A second important development was the emergence of a new ‘third’ generation of attachment researchers, with tenured position and independent grants to pursue research on their own terms. As mentioned in Chapter 3, Sroufe’s former doctoral student Glenn Roisman has been one of a number of ‘third-generation’ attachment researchers who use ideas and methods from both the developmental and social psychology traditions, on the basis of personal networks of collaborators extending in both directions. Another has been Roisman’s colleague at Minnesota, Jeff Simpson. Simpson has drawn theoretically on the organisational perspective of Sroufe, but primarily adopts a social psychological approach to conceptualising and measuring attachment. Simpson and his group have documented that self-report measures of attachment and the AAI make non-redundant contributions in predicting the behaviour of dating couples in completing a tricky laboratory-based task. Secure-autonomous attachment on the AAI predicted offering their partner a degree of support that was sensitive to their signals, in a situationally contingent manner. A self-report of an avoidant attachment style predicted offering the partner less support no matter the degree to which the partner requested help.207 These results remained the same when controlling for personality traits (including neuroticism), perceptions of the quality of their relationships, and the attachment style of the partner. Another ‘third-generation’ attachment researcher—to be discussed further in the section ‘Attachment to God’—is Pehr Granqvist, who has collaborated directly with both Main and Hesse,208 and with Shaver and Mikulincer.209

A third development has been that both traditions have continued, quite successfully, to generate research findings that appear to test or elaborate much the same theory. This has meant that, though it is a bit awkward, both traditions have been able to find support in the work of the other for efforts to think about research or convince audiences of the merits of attachment theory. Attachment researchers in the social psychology tradition have long drawn on findings from the developmental tradition in the literature reviews of their papers, in interpreting their results, and in review articles. For instance, in their magisterial work Attachment in Adulthood offering the capstone to their careers of research in the area, Mikulincer and Shaver simply fold in findings from both self-report and AAI studies in their summary tables of results, rarely remarking on differences in findings.210

It is clear that, contrary to Shaver’s pessimistic remarks about his reading of works by developmentalist colleagues in 2006, researchers in the social psychological tradition have kept up to date with research using the AAI. Perhaps marking some aspects of a parallel (p. 469) transition, a curious shift has occurred in the developmental tradition of attachment research. Previously, with exceptions such as Cassidy and Belsky, researchers in this tradition excluded self-report measures from the literature reviews of their papers, discussions of results, and review articles. The grounds for this have been that the self-report measures have little association with the Strange Situation and the AAI, situated as the ‘gold standard’ measures of attachment. However, in the last few years there has been an increased trend for leaders in the developmental tradition to both acknowledge the contribution of self-report studies, and acknowledge it as attachment research. So, for instance, in a 2016 paper, Heckendorf, Huffmeijer, Bakermans-Kranenburg, and van IJzendoorn described Mikulincer’s work as having ‘convincingly shown … experimentally how feelings of more secure attachment facilitate supporting partners in distress’.211 And no less a defender of the developmental tradition than Sroufe has praised the social psychology approach in the most recent edition of the Handbook of Attachment. Sroufe stated that in the work of Mikulincer and Shaver, as well as Shaver’s Minnesota colleague Simpson, the ‘power’ of the approach for studying adult romantic relationships ‘has become manifest’.212

Attachment narrow and broad

Pivoting concepts

In a 2015 review paper, Jones, Cassidy, and Shaver expressed confusion over the fact that the ECR and AAI, unrelated measures, nonetheless ostensibly both assess attachment: ‘though the two measures are largely unrelated to each other, they are similarly related to a variety of attachment-relevant constructs’. They have termed this one of the ‘burning questions’ facing the field of attachment research.213 Shaver and Mikulincer have elsewhere presented this question as a challenge to researchers in the developmental tradition. If developmentalists want to claim that self-report measures are not true assessments of attachment, they need somehow to account for the fact that the results of studies using these measures correspond specifically to the predictions of attachment theory:

Self-reports of attachment anxiety validly predicted automatic preoccupation with attachment-related worries, whereas self-reports of attachment avoidance validly predicted defensive suppression of these worries. This suppression could be overcome by adding a cognitive load that interfered with defensive suppression. These effects are so closely related to the theoretical conception of anxiety and avoidance in attachment theory that they would be difficult to explain, as a whole, by any other theory.214

(p. 470) Shaver and Mikulincer legitimately described their work as having elaborated and tested concepts drawn from attachment theory. However, Roisman helpfully qualified that from this it should not be assumed that (i) ‘attachment’ itself is a unitary entity in adulthood, (ii) or that attachment theory is a unitary entity.215 The difficulties of translation between the developmental and social psychological traditions of attachment do not only arise from their differences in methodology. Perhaps of greater importance has been a confusion of tongues in the appeals of the two traditions to the work of Bowlby and Ainsworth, and in their use of the term ‘attachment’ itself.

Shaver and Fraley stated that ‘the term attachment is a metaphor. Its denotation is unclear.’216 Its meaning is not pre-set, but shaped by the surrounding network of theoretical concepts. As such, Shaver and colleagues readily acknowledged the differences between developmental and social psychological researchers in interpretations of the term ‘attachment’ and in the surrounding network of concepts. They argued that ‘just as a reader of Freud can focus mainly on his psychosexual theory of personality development or his theory of intrapsychic dynamics, a researcher interested in attachment theory can focus on one or more of its central components without being required to focus on all of them. This does not mean the researcher is not doing attachment research.’217 They sharply criticised Waters (Chapter 2) for his emphasis on the secure base and safe haven as the central denotation of the term ‘attachment’ and as the fundamental elements of attachment theory. To them, this seemed unnecessarily restrictive and overliteral. Waters’ position also seemed to Shaver and Mikulincer to be a power-conserving tactic, aiming to snatch the legacy of Bowlby and Ainsworth away from researchers who recognise that the ‘theory contained other central concepts’ and who wish to acknowledge these with a wider sense of the term ‘attachment’.218 From this perspective, the social psychological tradition of attachment research was not an unlicenced reapplication of the established attachment concepts; it was another way of generating concepts, methods, and research priorities on the basis of the ideas of the first generation of attachment researchers.

Shaver and Mikulincer were aware, of course, that varied use of the same terms put at risk the viability of effective communication within the research community, and between the research community and its audiences.219 However, this appeared a risk they were willing to take. One reason Shaver and Mikulincer may have felt warranted in adopting the broader notion of ‘attachment’ was that, like Waters and colleagues, they could easily point to passages in Bowlby’s work to support their interpretation of attachment and (p. 471) of attachment theory. This is because, as Chapter 1 showed, Bowlby’s integration of ethology and psychoanalysis brought together within the single term ‘attachment’ two quite different conceptualisations: one narrow account based on Hinde’s notion of following and proximity-seeking as activated by threat, and a broader account based on the psychoanalytic concern with all intimate relationships. A large swathe of the confusion among researchers lies in the fact that the major additions to Bowlby’s theory by Ainsworth and Main could apply to both the narrow and the broad concept of attachment. True, in a certain sense the narrow concept had primacy. Ainsworth’s identification of the secure base phenomenon, in the first instance, entered attachment theory as an addition to the ethological concern with proximity-seeking, as a cross-species response. Hinde’s idea that offspring will seek to retain proximity with their parent, especially under conditions of threat, was transformed by Ainsworth into the division between the secure base and safe haven. Individual differences in infant’s use of the caregiver as a secure base for exploration and for proximity-seeking as a safe haven were observed by Ainsworth at home and in the Strange Situation.

However, just as Bowlby’s concept of ‘attachment’ hinged on stringent and sprawling meanings, so did Ainsworth’s concepts of ‘secure base’ and ‘safe haven’. They could apply to the physical behaviour of an infant of any species. But they could also apply more generally to the symbolic aspects of all human intimate relationships. For instance, one friend can offer another regulation, comfort, and reprieve, without physical contact or any offer of protection. As Chapter 2 discussed, in a late publication Ainsworth criticised the work of her students and followers: ‘By focusing so closely on intimacies some attachment researchers have come to conceive of them as the only source of security—which is a pity.’220 For instance, without being an intimate relationship, even a short-term therapist may serve as a secure base and/or safe haven for a patient. Ainsworth’s sophisticated position in her 1985 article ‘Attachment across the lifespan’ was that the experience of another person as a secure base, a safe haven, and as a source of separation anxiety were separable in adulthood (Chapter 2).221 They did not necessarily form a unity. Ainsworth argued that it was futile to argue back and forth over whether a relationship, such as with a therapist or with a religious pastor, was or was not an attachment relationship. There was no binary. Instead, she argued that it was better to identify the extent to which a secure base, a safe haven, and separation anxiety were present, and to acknowledge that these phenomena could occur without other characteristics of an attachment relationship.

However, there remained a hinge in Ainsworth’s use of the concepts that hindered this specification. Ainsworth’s primary use of the terms ‘secure base’ and ‘safe haven’ was to indicated the physical movements of the infant away from and back towards the caregiver. However, she also extrapolated the concepts in remarks about adulthood to encompass the provision of other forms of support where these lead to an experience of security. The experience of feeling confident in a relationship as a basis of felt security may occur in relationships that are short term, relatively impersonal, and/or with only the barest or most symbolic link to a parent–child relationship. Probation officers, for example, describe secure base phenomena in their work with offenders as ‘screamingly obvious’; and similar observations have (p. 472) been made for short-term therapeutic involvement with clients by helping professionals.222 The secure base phenomenon is not limited to what are usually treated as attachment relationships when the term ‘attachment’ is construed narrowly.223

Likewise Main’s concept of conditional strategies was made to pivot in both directions, towards Bowlby’s narrower and broader concepts of attachment. What is curious is that this pivot was not intended by Main. Main’s concept of conditional strategies was grounded in the idea that ‘there exist species-wide abilities that are not part of the attachment system itself, but can, within limits, manipulate (either inhibit or increase) attachment behavior in response to differing environments’.224 Main came to regard the avoidant and ambivalent/resistant infant attachment classifications as based on the use of attentional processes to manipulate the activation and expression of the attachment system. Narrowly, in Main’s usage a conditional strategy was an ethological repertoire for achieving this effect: either directing attention towards potential threats and vigilance regarding caregiver availability and away from cues that might terminate the attachment system, or directing attention away from cues that might activate the attachment system, and towards potential sources of distraction in the environment. Main identified a similar process in adult autobiographical discourse, as attention seemed to be directed away from or towards attachment-relevant information. However, across both her published and unpublished writings, this never led her to refer to dismissing and preoccupied states of mind regarding attachment as ‘conditional strategies’.

The extension from infancy to adulthood for Main was based on analogous processes at the level of attention. Main argued that the caregiving system could likewise be deactivated or made especially vigilant, since the caregiving system also runs on attention allocated to the one being cared for. The adult sexual system was assumed to have a parallel set of potential conditional strategies. The extension from the Ainsworth Strange Situation to adult discourse and to the caregiving and sexual systems gave Main’s theory a wide scope of application. Yet Main’s use of the terms ‘attachment’ and ‘strategy’ always remained narrow and ethological. It was Jude Cassidy, especially, who made the concept of conditional strategies pivotal to the broader concept of attachment, in conceptualising them as the ‘maximising’ and ‘minimising’ of not attention but of attachment itself as a system for the regulation of intimacy and emotion (Table 5.1 and Chapter 3). And it was from Cassidy, not from Main directly, that Shaver and Mikulincer picked up the concept of ‘attachment strategies’.

Table 5.1 Table of pivoting concepts

Broad meaning

Narrow meaning

Used by social psychologists + a few developmentalists like Cassidy and Fonagy

Used by Main, Waters, and the large majority of other developmental psychologists

Safe haven function of attachment

The orientation to seek physical or symbolic comfort from close relationships under conditions of perceived threat

The capacity to trust in the physical and attentional availability of discriminated familiar individuals under conditions of perceived threat

Secure base function of attachment

The capacity of close relationships to provide felt security

The capacity to trust that attention can be turned to exploration, given the expectation of the availability of discriminated familiar individuals

Conditional strategies

The hyperactivation or deactivation of the orientation to seek physical or symbolic comfort from close relationships

A species-wide repertoire, made available by evolutionary processes, for manipulating the activation of the attachment behavioural system through the direction of attention vigilantly towards or away from cues about the availability of familiar caregivers or potential threats. This repertoire evolved because it has the predictable outcome of increasing the availability and support provided by attachment figures who may otherwise be unavailable.

Other ways of manipulating or overriding the output of the attachment behavioural system exist, and become increasingly available with developmental maturation. They can also be described as strategies. However, they are not conditional strategies in the technical sense unless—like the redirection of attention—they can be considered to express a species-wide repertoire, made available by evolutionary processes

Internal working models

The elaborated symbolic and affective representations made by humans about attachment figures and their availability, and the value of the self to these attachment figures

Variously:

  1. 1. Expectations about the availability of attachment figures

  2. 2. Elaborated symbolic meanings and images held by humans about attachment figures and their availability

  3. 3. A synonym for attachment representations, as used by Main in the 1980s (but subsequently abandoned)

In play across the attachment paradigm, then, can be seen narrow and broad concepts of attachment, narrow and broad concepts of secure base/safe haven, and narrow and broad concepts of maximising/minimising. Whereas the narrower concepts are based on (p. 473) a cross-species ethological account, the broader concepts encompass the felt experience characteristic of humans on the basis of symbolic capabilities. With both narrow and broad concepts in circulation, and with little or nothing to mark the distinctions between them, terms that researchers and their readers assume to have common meanings turn out to be quite treacherous. They have hidden compartments, allowing assumptions to be smuggled unnoticed and unscrutinised. Adding yet further to the confusion has been the concept of the ‘internal working model’, which can figure within both the broad and the narrow (p. 474) conceptualisations, albeit with different meanings (Chapter 1). The ‘internal working model’ concept has an unfortunate hypermobility, coming out of joint the moment weight is rested on it.

Given Waters’ criticisms of Shaver and Mikulincer, it might be thought that the narrow concepts are more characteristic of the developmental attachment tradition and the broader concepts more characteristic of the social psychological tradition. That is certainly the general impression. However, matters are more complicated. True, Waters has been unusually consistent in using the narrow concept of attachment, and argued against looser uses of the term.225 However, like Ainsworth, he has at times used a wider notion of secure base. And he does not appear to have ever accepted the concept of conditional strategies; the term is absent from his writings except in summaries of Main’s position. In Waters’ view, the distinction between avoidant and ambivalent/resistant attachment has seen insufficient predictive payoff, and insufficient validation against observations in naturalistic settings, to warrant its reification into discussions of minimising and maximising strategies (Chapter 2).

Among researchers in the developmental tradition, van IJzendoorn and Bakermans-Kranenburg are also unusually consistent in their use of the term ‘attachment’. In the 1980s, van IJzendoorn pulled no punches in describing Bowlby’s own tendency towards a broader usage of the term ‘attachment’ as ‘somewhat ridiculous’, since it implied identity between the process that brought a scared child to seek their caregiver and much more general feelings such as of identification with a political party.226 Van IJzendoorn had no doubt that more expansive and symbolic meanings of attachment could be meaningful. For instance, he suspected that the relationship children have with a soft toy or blanket might recruit certain facets of the attachment behavioural system, drawing on the emerging human capacity for symbolisation. But, for him, this needed to be defined in precise terms on the basis of use of the object as a secure base, as a safe haven and/or source of separation anxiety, and the prediction of later socioemotional functioning.227 Schuengel and van IJzendoorn discussed the narrower and broader uses of the terms ‘attachment’ and ‘secure base’ as contributing to a severe ‘confusion about concepts’, obscuring matters under discussion, for instance the precise relevance of attachment in therapeutic relationships.228 And van IJzendoorn and Bakermans-Kranenburg have alleged that the concept of attachment has been stretched and disfigured by broader uses until it has, by now, simply snapped, with researchers talking right past one another whilst using the same terms.229 For van IJzendoorn (p. 475) and Bakermans-Kranenburg what Shaver and Mikulincer have been measuring is simply a different behavioural system, one concerned with the adult pair relationship.230

However, it should not be thought that developmentalists are neatly aligned with the narrower use. Some developmental researchers slip and slide between the narrow and broad uses. Those who write for a clinical audience as well as for researchers and deliver commercial trainings in therapeutic approaches—such as Peter Fonagy, Patricia Crittenden, Allan Schore, Dan Hughes, and Susan Johnson—are generally consistent throughout their writings in using the broader concepts.231 The broad concepts of attachment, secure base/safe haven, and maximising/minisimising have significant advantages in this context. The broad concepts are closer to the connotations of ordinary language, and so are more accessible and evocative in writing aimed at both non-researchers and researchers. They offer greater attention to processes of symbolisation and meaning-making than the technical idea of an ethological behavioural system focused on proximity-seeking. And they also allow easier extrapolation across the life-cycle. For instance, Crittenden has proposed that what Ainsworth identified as patterns of attachment in infancy is in fact the local case of a broader phenomenon, linking human mental health and ill health. This broader phenomenon is the capacity of humans to exclude certain kinds of information from influencing experience or behaviour, in a manner responsive to that individual’s history and social context. This is therefore how Crittenden has used the concept of ‘attachment’. The term ‘strategies’ has been deployed to highlight the environmental responsiveness of individual behaviour. And the term ‘security’ has been used to mean the absence of information-processing exclusion or distortions.232

A few developmentalists, however, adopt some narrow and some broad uses. Cassidy, for instance, has generally been consistent in a comparatively narrow use of the concept of safe haven, but a broad use of the concepts of minimising and maximising.233 Researchers in the social psychology tradition, despite tending towards the broader usage, regularly shuttle between the narrow and broad formulations and only rarely flag these transitions to the reader. We have already seen an illustration of slippage in the work of Shaver and Mikulincer. Module 1 of their model of the attachment system is concerned with whether ‘proximity’ is achievable (attachment narrow), but module 2 begins with an evaluation, not of proximity, but of whether the attachment figure is ‘available, attentive, responsive, etc.’ (attachment broad). And again, module 3 is initiated by a concern with whether the attachment figure is ‘available, attentive, responsive, etc.’ (attachment broad). But this concern is framed as the question ‘Is proximity seeking a viable option?’ (attachment narrow). This oscillation (p. 476) in Shaver and Mikulincer’s account of the attachment system likely reflected a desire to assert that what they take to be attachment encompasses but is not reducible to concern with proximity with attachment figures. It allowed them to include the narrow notion of attachment within their broader concept, whilst also acknowledging that the set-goal of an adult attachment system is calibrated by context and culture.

Once the hinge in the concept of ‘attachment’ is acknowledged, Shaver and Mikulincer’s position might be regarded as an offshoot of Bowlby’s original insight—obscured by terminological confusion—that attachment in a broad sense has many components in common with, and may in some ways grow out of, attachment in a narrow sense. Shaver has had a tendency to overclaim links between adult espoused attachment styles and the attachment system in infancy. Stevenson-Hinde was surely right to admonish that ‘one is going far beyond the data to presume, as Hazan and Shaver do, that “the neural foundation of the attachment system remains largely unchanged” across the life span’.234 Nonetheless, already from the early works with Hazan, Shaver at times acknowledged that the construction of an adult bond as an attachment relationship was a piecemeal endeavour, forming a dappled surface that could, at most, only become consistent over time.235 In a conference presentation, Hazan reported that by late adolescence 75% of 17-year-olds reported displaying proximity-seeking and separation protest preferentially to peers rather than parents, but that parents remained favoured for providing a secure base.236 She theorised that at the start of a new adolescent or adult romantic relationship, proximity-seeking (in part motivated by the sexual system) contributes to finding a safe haven in the partner. This enthusiastic and clumsy process then leads to the establishment of pockets of comfort and shelter within the relationship and the incremental stabilisation of the partner as a secure base. The relationship itself comes to be felt as familiar, ordering, banal, and necessary. Hazan suggested that separation anxiety following unexpected separations follows on from experiencing the relationship as a secure base.

The speed and tone of the development of the relationship will be shaped by each individual’s previous experiences, which will have informed their sexuality, their trust in a safe haven, their capacity to use a secure base, and the meanings they give to separations. Hazan and Shaver theorised that experiences in childhood and adolescence that suggest that a safe haven will not be available will predispose an avoidant attachment style, which in turn will slow the development of the components of an attachment relationship. By contrast, experiences in childhood and adolescence that suggest that a secure base will not be available will predispose an anxious attachment style, which in turn will speed up the use of the caregiver as a safe haven. Hazan and Shaver speculated that a secure attachment style would make (p. 477) for a smoother transition from sexuality to caregiving as the primary ally of the attachment system in adult relationships. It might also facilitate the development of the full complement of components of an adult attachment relationship, with the mutual provision of a secure base, safe haven, and the experience of anxiety at unexpected separations.

Secure base/safe haven dynamics

That Shaver and colleagues would adopt a loose definition of ‘attachment’ was predisposed by their focus on adulthood. The wish for an adult partner to be ‘available, attentive, responsive, etc.’ will be shaped by culture and context. Indeed, culture and context can influence the relative priority of concern for availability, attentiveness, and responsiveness. Though their primary target was romantic relationships, Shaver’s initial impulse towards attachment theory had been as a resource for making sense of a close sibling relationship and its loss. Hazan and Shaver specified that they intended the introduction of attachment theory to social psychology as offering ‘a comprehensive theory of close relationships’.237 It was an individual’s ‘style’ with respect to intimacy that was the object of analysis. And Shaver and colleagues have frequently used the term ‘intimacy’, or sometimes ‘closeness’, as a functional synonym for attachment.238

What made attachment theory an exceptionally potent resource for social psychology was that, not just infant–caregiver relationships, but adult intimate relationships too seemed to have the potential for secure base/safe haven dynamics. This was already identified clearly by Bowlby in his writings on the therapeutic role of the secure base in clinical contexts. Whether or not this makes such adult interactions ‘attachment’ relationships is a matter of definition, and both sides of this debate can claim Bowlby’s authority and supporting passages from his writings. Furthermore, various behavioural systems can be subject to minimising and maximising: in Main’s original proposal, the parameters of both the infant attachment system and the adult caregiving and sexual systems could be altered through the manipulation of attention away from or towards the activating or deactivating conditions of the system. Whether or not they are therefore defined as ‘attachment’ relationships, the desire for intimacy in adult relations can be regarded as another site in which minimisation and maximisation may be deployed to respond to circumstances where the availability of a partner is in question. As such, the key ingredients of attachment theory—the attachment system (module 1), the secure base/safe haven dynamic (module 2), and maximising and minimising (module 3)—are all available for research on adult intimate relationships.

In a 2002 paper, ‘Activation of the attachment system in adulthood’, Mikulincer, Gillath, and Shaver reported a series of studies of the effects of subliminal threat on the activation of (p. 478) representations of attachment figures. The authors regarded the results as effective evidence that their work was studying the processes outlined by attachment theory. They challenged their readers to offer another theory that could account for their findings. In their view, ‘there is no alternative theory that predicts the results we obtained’, and also ‘no alternative theory that would have generated either these particular kinds of measures or our particular experiments’.239 The series of studies began by using the WHOTO measure, initially developed by Hazan and then adapted by Fraley and Davis.240 In the WHOTO, participants are asked to give a named person in response to questions about proximity-seeking/separation anxiety (Who is the person you most like to spend time with? Who is the person you don’t like to be away from?), safe haven (Who is the person you want to be with when you are feeling upset or down? Who is the person you would count on for advice?), and secure base (Who is the person you would want to tell first if you achieved something good? Who is the person you can always count on?). Mikulincer and colleagues were keen to emphasise that these were not merely conventional or obvious aspects of any close relationship, but were specifically the functions privileged by Ainsworth in defining an attachment figure.241

In study 1, Mikulincer, Gillath, and Shaver asked participants to complete a computerised lexical decision task, in which they read a string of letters and tried to identify whether it was a word or a non-word. Reaction times were used as a measure of the accessibility of thoughts related to the target word.242 The accessibility of the names of people identified in the WHOTO was assessed under two conditions. In the first condition, the name followed a neutral word (‘hat’). In the second condition, the name followed the word ‘failure’, which was chosen as a representation of threat but not necessarily one closely linked to attachment. Other names were also used: the names of close friends not selected by the WHOTO, the names of acquaintances, and names of strangers. Participants also completed the ECR, and a measure of neuroticism. The results of study 1 were that participants had a faster reaction time when the name was from their WHOTO list following the ‘failure’ prime, compared to reaction times for the names of other close people, acquaintances, strangers, and non-words. An anxious attachment style reduced reaction times for the names of attachment figures only, and this occurred in both the neutral and the ‘failure’ condition. As in other studies, there was a very substantial association between attachment anxiety and neuroticism (r = .54). However, neuroticism was unrelated to response times, and the effects for attachment anxiety remained the same even controlling for neuroticism.

(p. 479) Study 2 was a replication of study 1, except that instead of the threat word ‘failure’, the researchers substituted the threat word ‘separation’. Again, participants showed faster reaction time following the prime when the name was from their WHOTO list than other names. As in study 1, an anxious attachment style made a contribution to reaction times for attachment figures only, and in both the ‘separation’ and the neutral condition. However, a finding specific to study 2 was that an avoidant attachment style increased reaction times for the names of attachment figures following priming of the word ‘separation’ but not following the neutral word. This implied that the names of attachment figures were less available for participants in response to the theme of ‘separation’ to the degree that they were high in attachment avoidance. Again, the association between an anxious attachment style and neuroticism was large (r = .62), and again neuroticism had no association with reaction times or interaction with the effect of attachment style on reaction times.

Study 3 sought to replicate studies 1 and 2 using a different procedure, the Stroop task. This is a well-established procedure based on the observation that viewing words with salient personal meanings slows reaction times in identification of the colour of that word. That is to say, cognitive accessibility on this assessment was represented by longer reaction times, in contrast to studies 1 and 2. Participants were divided into three groups: one group received neutral primes, another group were primed by the word ‘failure’, and a third group were primed by the word ‘separation’. In the Stroop task, as expected, colour-naming was slower following the two threat primes in response to the name of attachment figures only, as opposed to other names, words, or non-words. An anxious attachment style increased reaction times in the neutral condition and following both threat primes in response to the names of attachment figures. An avoidant attachment style decreased reaction times in response to the names of attachment figures following only the separation prime, not the failure or neutral primes. Again, attachment anxiety overlapped substantially with neuroticism, but only the former was associated with any reaction time outcomes. Overall, the findings of the three studies were taken by Mikulincer and colleagues to show that threat contexts—whether general or attachment-specific—increase the cognitive accessibility of the names of attachment figures, in contrast to other close relationships and acquaintances (module 1). An anxious attachment style increased this cognitive accessibility even in neutral conditions, illustrating the feedback from attachment strategy (module 3) to the appraisal of threats and attachment figures (modules 1 and 2). An avoidant attachment style decreased cognitive accessibility only in response to the names of attachment figures and only in response to a separation prime rather than a general threat prime. The researchers demonstrated these effects over two different procedures and in three samples. The findings, the researchers argued, ‘increase our confidence in the psychological reality of the attachment system’ in adulthood.243

Mikulincer and colleagues regarded this study, one of their most cited by social psychology attachment researchers, as a gauntlet thrown down to the developmental tradition. Could they explain the findings except in terms of the operation of the attachment system in adulthood? The answer would appear to be no. Three research groups in the developmental (p. 480) tradition—those of Jude Cassidy,244 Anne Bernier,245 and Peter Fonagy246—have treated the study as evidence that the ECR measures attachment just as much as the assessments devised by Mary Ainsworth and Mary Main. Other developmental researchers have ignored the study, at least in print.247 The paper demonstrated that threat primes increase the salience of semantic information relevant to figures identified as serving as a secure base, safe haven, and potential source of separation anxiety. Whilst the study did not provide evidence that these figures are differentially sought as a secure base/safe haven in concrete behaviour, the evidence of increased cognitive accessibility in the laboratory is certainly relevant and intriguing, and points in the expected direction.248 The study also demonstrated that these effects are moderated by scores on the ECR in ways aligned with theory: an anxious attachment style increased vigilance to threats and to information about relationships with attachment components; and an avoidant attachment style decreased the cognitive availability of relationships with attachment components following a prime for separation.

The study suggested that secure base/safe haven processes operate in adulthood, and that these are moderated by maximising and minimising strategies. This was one important empirical source of legitimacy for the social psychology tradition to call itself a form of attachment research. However, it is important to note that the distinction between ‘attachment relationships’ and ‘close relationships’ in the study does not rule out other possibilities associated with the wider notion of attachment used by Shaver and colleagues. Shaver, Hazan, and Bradshaw claimed that ‘the attachment figure need not be physically present for such an “interaction” to take place, which makes it more understandable that people have imaginary but very convincing and affective interactions with rock singers, dead philosophers, and religious figures of all kinds. Attachment, separation distress, and grieving are primarily psychological processes; they require psychological, not physical, interaction partners.’249 Furthermore, some physical interactions with safe bases and secure havens are not with living partners. For instance, Bowlby stated that a person’s home would likely have the characteristics of a secure base and safe haven (Chapter 1). If a person’s home address were used in a study like that of Mikulincer and colleagues, it could be anticipated that responses would be more similar to names of attachment figures than to names of other close relationships. (p. 481) And as we have seen, Shaver himself has referred to his experience of attachment theory as an attachment relationship: it is quite possible that other phenomena that serve as a safe base and safe haven might also receive differential cognitive availability in response to threat primes. That secure base/safe haven processes operate in adulthood does not imply that these processes are exclusive to romantic or even close relationships.

Evidence for the wider relevance of the secure base/safe haven components of attachment relationships, and of minimising and maximising strategies, is offered by later research conducted by Zilcha-Mano, Mikulincer, and Shaver. The researchers developed a scale for assessing anxiety and avoidance in an adult’s sense of their pet’s availability as a secure base and safe haven:

Pet attachment anxiety consists of intense and intrusive worries that something bad might happen to one’s pet and that one might find oneself alone, a strong need for proximity to the pet, reassurance seeking from the pet in order to maintain self-worth, intense frustration when the relationship with the pet is not as close as one would like, and even anger when the pet prefers the proximity of others. Pet avoidant attachment consists of feelings of discomfort with physical and emotional closeness to a pet, striving to maintain emotional distance from the pet, avoiding intimacy with it, preventing the pet from intruding into one’s personal space, and difficulties in depending on the pet and turning to it when distressed.250

Zilcha-Mano and colleagues found that an avoidant attachment style on the ECR had no association with pet avoidance, and instead was associated with attachment anxiety regarding the pet’s availability as a secure base/safe haven (r = .35). The researchers interpreted this finding as suggesting that ‘avoidant people, who are unlikely to express attachment-related worries and anxieties in close human relationships (Mikulincer & Shaver, 2007), tend to express these worries and anxieties in relationships with pets’.251 By contrast, there was a large positive association (r = .60) between attachment anxiety regarding the pet and an anxious attachment style on the ECR. Yet Zilcha-Mano and colleagues reported that anxiety regarding the availability of the pet had a substantial association with reduced reported wellbeing (r = .34) and with symptoms of anxiety and depression (r = .41) on the Mental Health Inventory, even after controlling for scores on the ECR and a measure of personality traits. The researchers therefore emphasised that psychologically consequential secure base/safe haven effects, and minimising/maximising individual differences, can be identified in adulthood beyond close human relationships.

The researchers did not report the association between ECR anxiety and the Mental Health Inventory. However, if other studies with the ECR and the Mental Health Inventory are brought in for comparison,252 it is rather interesting that the pet attachment anxiety measure in the Zilcha-Mano study is actually a better predictor of scores on the Mental Health Inventory than the ECR. Given that many species of pet—unlike human partners—have been bred precisely to offer loyalty and reassurance, it may suggest an especially sorry (p. 482) state of affairs for someone to feel consciously anxious about whether their pet cares about them. Neuroticism as a measure of trait anxiety had only a weak association with anxiety regarding the pet’s availability as a secure base/safe haven (r = .15), indicating that association does not lie merely in personality. It suggests that the secure base/safe haven components occur in the relationship with the pet and that these are reducible to neither adult attachment style nor personality. Furthermore, the findings suggest that meaningful individual differences exist in anxiety regarding the relationship with the pet. Zilcha-Mano and colleagues interpreted this as individual differences in the deployment of a maximising strategy in order to ensure the availability of the pet as a secure base/safe haven. However, further research is required to demonstrate that this is indeed specifically a maximising strategy, as opposed to anxiety arising through a psychological process of a different kind.

Attention and cognitive schemas

The ECR was considered to tap schemas offer information about the meaning and availability of a secure base and safe haven within intimate relationships, and the relevance of minimising or maximising strategies as a way of achieving the availability, attentiveness, responsiveness (etc.) of the other. This is quite distinct from the AAI, as an assessment of individual differences in the capacity to attend to and communicate about personal experiences with attachment figures in childhood. Rather than assessing generalised schemas, this capacity is assessed in the AAI through comparison of particular episodic and generalised semantic information. Johnson has claimed that attachment patterns, attachment styles, attachment strategies, and states of mind regarding attachment should all be regarded as ‘equivalent terms’, with the same referent except for differences in age and assessment methodology.253 This runs counter to what Shaver and colleagues, and Main and colleagues have said that their assessments measure. It also seems a claim poorly aligned with the available evidence. Offering a more sophisticated and developmental reconciliation between the traditions, Cassidy proposed that (i) the experience of a secure base and safe haven in early child–parent relationships should facilitate (ii) a coherence in attending to and communicating about childhood experiences with attachment figures and (iii) generalised schemas about intimate relationships characterised by less anxiety or expectation of rejection.254 However, it should clearly be acknowledged that these are three vastly different meanings of the term ‘secure attachment’, and the relationship between any two of them is unlikely to be simple or strong. Not least, Cassidy’s proposal assumes extensive continuities from infancy to adult measures; though scaffolded by some of Bowlby’s claims, this assumption was and remains contrary to available data (Chapter 2).

There is no reason to assume that the capacity to attend to and communicate about personal experiences with attachment figures in childhood should correlate with generalised schemas about intimate relationships in adulthood. The first is an attentional process oriented by a comparison of episodic and semantic memories of attachment narrowly construed. This (p. 483) is prompted by questions designed to offer uncomfortable freedom to memory, setting episodic and semantic information into motion and providing an opportunity to see the extent to which they outgallop the speaker’s cooperation with the interviewer. The second is a generalised—in all likelihood largely semantic—schema of beliefs, emotions, and behaviours relevant to the minimising or maximising of the wish for others to be available, attentive, responsive, etc. This will certainly be linked to attentional processes, not least in terms of cognitive biases in the perception of new information.255 However, individual differences in attentional processes are at most a correlate, rather than what the ECR is specifically measuring. And whilst it might seem that both the AAI and ECR assess individual differences in terms of minimising and maximising strategies, this is a misapprehension. The AAI assesses the allocation of attention away from episodic information at the expense of fully answering the interview questions (Dismissing), towards negative aspects of experiences at the expense of focus on the actual questions asked in the interview (Preoccupation), or disruptions in working memory expressed as lapses in reasoning or discourse (Unresolved/disorganised). By contrast, the ECR assesses a variety of behaviours, beliefs, and feelings that index latent schemas. The two latent dimensions of the ECR are presumed by Shaver and Mikulincer to reflect an intensification or deintensification of the wish for others to be available, attentive, responsive, etc.

There may be several reasons for the lack of convergence between the AAI and ECR. They adopt different forms of measurement. And even within self-report assessments, there is only weak convergence between reports of attachment style with parents and attachment style in current intimate relationships. For example, self-reported avoidance with mother and with father have little association with a participant’s attachment style in their current relationship (r = .24 and r = .24 for anxiety; r = .17 and r = .12 for avoidance, for mother and father respectively).256 In addition to these factors, a further contribution is perhaps made by the fact that they tap different psychological processes, at least according to their originators. To the degree that Main and Hesse are right that the AAI is in the first instance an assessment of the allocation of attention, and that Shaver and colleagues are right that the ECR is ultimately an assessment of generalised schemas about intimate relationships, this suggests some specific conclusions about how and when the measures will diverge—but also when they will converge.

There should be poor association between the measures when correlates relate to semantic generalisations about need achievement within adolescent and adult relationships. This is perhaps why the ECR relates quite well to reported marital satisfaction and reported feelings of closeness with friends, but not the AAI.257 There should also be poor association between the measures when correlates relate specifically to the allocation of attention in responding to another—one component of the caregiving behavioural system (regardless of whether this is broadly construed as nurturance or narrowly construed as safe haven provision). This is (p. 484) perhaps why coherence on the AAI has been repeatedly found to be reliably associated with observational measures of sensitivity towards children and partners, whereas the link between the ECR and sensitivity seems to be more dependent on circumstances.258 Sensitivity is predicated on noticing and recognising the other’s signals (Chapter 2), primarily an attentional task. By contrast, the ECR is reliably associated with general supportiveness when the other is evidently stressed. Simpson and colleagues demonstrated this in assessments of couple interactions during a challenging laboratory task, as described in the section ‘Two traditions of attachment research’.259 And other studies have documented links between the ECR and generally supportive parenting when a child is distressed.260

The two measures will also be less well aligned when clinical phenomena are in question. The reason for this is foundational: part of the very definition of many clinical symptoms is the divergence between attentional processes and latent generalised schemas. In addition to differences in measurement, this may well be part of the reason why there is no evidence of overlap between the constructs of unresolved/disorganised states of mind and self-report measures of fearful attachment. Yet these two uncorrelated measures, the AAI and ECR, can nonetheless be anticipated to have similar correlates when the phenomenon in question is associated with both the allocation of attention in relation to attachment phenomena and generalised schemas about close relationships. There will be many occasions in which this alignment may occur. For instance, sexual behaviour in adulthood is likely influenced by both processes. The shaping of attention and sharing in relation to past experiences of tenderness and care are surely relevant, as are schemas about the possibilities and threats associated with intimacy. Likewise, both the allocation of attention in relation to attachment phenomena and generalised schemas about close relationships can be anticipated to be associated with the content—if not the form—of accounts of care in the course of childhood. And indeed, this is what evidence to date suggests.261

Nonetheless, it may also be anticipated that the processes underpinning the AAI and ECR would be activated and moderated by different circumstances. For instance, given Main’s theory of the importance of attention to individual differences in both measures, the Strange Situation and the AAI may each be particularly impacted by factors that affect attention and memory processes, such as Ritalin and other ADHD medications.262 Indeed, given that studies suggest use of these medications occur in 7–9% of adolescent community samples and 4% of adult samples,263 the use of the AAI and assessments modelled on it without (p. 485) controlling for attention-altering medication use may have introduced an unrecognised confound into analyses. In community samples, medication use would be a small confound, but in higher-risk samples where medication use is more common, this may have distorted research findings to some degree. The ECR by contrast is regarded as an assessment of generalised schemas, and would be expected to be relatively unaffected by the same medications, though the intensity of their availability may be altered.264 Such distinctions would need to be tested, though. It is not enough to rest on the authors’ own interpretations of what their measures measure.

Sexuality and dominance behavioural systems

Behavioural systems theory

Shaver and Mikulincer have been among the few attachment researchers, besides Main and Hesse (Chapter 3), to give serious consideration to Bowlby’s concept of behavioural systems. This concept has been taken for granted, and then ignored, by most researchers in the developmental tradition: its relevance to the Strange Situation seemed too obvious for further comment, and its relevance to the AAI was not explicitly articulated in Main’s published writings and remained unclear. However, Shaver and Mikulincer’s were brought to re-examine the concept as part of their attempts to establish and understand the meaning of attachment in adulthood, and in particular to operationalise its activation in the laboratory. Unlike researchers in the developmental tradition, Shaver and Mikulincer were forced to test and discriminate, rather than assume, the operation of psychological processes resembling those described by Bowlby.265

In the course of their exploration of Bowlby’s ideas about attachment and other behavioural systems, Shaver and Mikulincer came to the view that attachment theory should actually have been called ‘behavioural systems theory’.266 The concept of ‘attachment’ was introduced as ‘a metaphor that would refocus psychoanalytic theory on relational issues rather than imagined instincts or drives’, and so has served its usefulness already.267 If the banner of attachment theory was to be preserved, for Shaver and Mikulincer this would essentially be a matter of tradition rather than because this is the best name for Bowlby’s ultimate achievement. As documented in Chapter 1, in late correspondence Bowlby himself was dissatisfied with ‘attachment’ as the name for his theory. Additionally, whereas the developmental tradition of attachment research had already claimed the concept of ‘attachment’, (p. 486) the concept of behavioural systems was an uncontested ground in which Shaver and Mikulincer could engage with and claim Bowlby’s legacy.

Early in their work together, with the attachment system in childhood and adulthood as their central model, Shaver and Mikulincer defined a behavioural system as having six components.268 These can be readily illustrated with the examples of the attachment and exploratory system. First, any behavioural system had to offer an expected species-level benefit for survival and/or reproduction. The attachment system, for instance, was assumed to offer both protection and a source of regulation. The exploratory system was assumed to offer support for learning about the environment and the self’s capabilities.269 Second, any behavioural system should have a set of activating parameters. The attachment system in childhood and adulthood was assumed by Shaver and Mikulincer to be activated by the perception of some kind of threat, whether an external source of danger or separation or an internal state of concern. Likewise, activating parameters can be identified for the exploratory system. Whereas Bowlby regarded the activating conditions of the exploratory system the encounter with complexity and/or novelty, in their conceptualisation Shaver and Mikulincer limited the activation of the exploratory system to the encounter with novelty, but emphasised that novelty is defined in part through contradiction or challenge to existing knowledge.270

For Shaver and Mikulincer, a third defining characteristic of a behavioural system was that it should have a set-goal which deactivates the system when a given change is identified in the relationship with the environment. Shaver and Mikulincer conceptualised the attachment system as deactivated by signals of the availability, attentiveness, responsiveness (etc.) of attachment figures, though the nature of these signals differ between childhood and adulthood. They also made the innovative proposal that a behavioural system will have ‘anti-goals’, which the system attempts to avoid. In the case of the attachment system, these are ‘rejection, separation, and attachment figure unavailability’.271 To take the example of the exploratory system, Bowlby conceptualised this system as deactivated by familiarity. Shaver and Mikulincer’s concept was similar, though not identical: for them, the exploratory system would be deactivated once relevant skills or knowledge have been acquired. Shaver and Mikulincer at no point discuss the anti-goals of the exploratory system, but these might include feelings of incompetence and boredom.272 To achieve familiarity without the (p. 487) experience of relevance and growth might risk achievement of both a set-goal and an anti-goal, resulting in a happiness that continually drains itself. Bowlby was not wrong that familiarity terminates the exploratory system, but in Shaver and Mikulincer’s view this did not mean that the behavioural system is therefore fully realised and satisfied.

Fourth, a behavioural system will have a repertoire of functionally equivalent behaviours that can be used to achieve the set-goal. In infancy and adulthood, Shaver and Mikulincer emphasised that behavioural systems continually learn, mature, and reorganise in identifying new and better ways of achieving the set-goal.273 However, within this diversity, they conceived of the Ainsworth attachment classifications as representing the basic repertoires for achieving the direct or conditional availability of attachment figures. Similarly, in the case of the exploratory system, Shaver and Mikulincer proposed that maximising and minimising strategies are available.274 Maximising the exploratory system entails attempts to master novelty even when the context does not call for it and enough information is already in hand. This may contribute to procrastination and indecision, doubts, and worries. Maximisation of the exploratory system responds to an environment in which familiarity is distrusted, much like anxious attachment style expresses a lack of confidence in the availability of attachment figures even when they are physically present. Minimisation of the exploratory system entails the inhibition of cognitive openness, exploration, or curiosity even when confronted by a novel or challenging situation. This strategy responds to an environment in which previous attempts at exploration were punished or tended to backfire.

Fifth, Shaver and Mikulincer highlighted that any behavioural system will include cognitive components. In the case of the attachment and exploratory systems, maximising and minimising strategies are activated depending on the availability of cognitive schemas from within a hierarchical network of associations formed through past experiences. Following Bowlby, Shaver and Mikulincer downplayed the autonomy of the affective components of the behavioural system, often describing these components as mere ‘reflections’ of the cognitive and behavioural aspects.275 Sixth and finally, Shaver and Mikulincer argued that any behavioural system will have specific activating or inhibitory links with other behavioural systems. For instance, it would be expected that the activation of the attachment system would inhibit the operation of the exploratory system.

In a paper with Josh Hart and Jamie Goldenberg, Shaver argued that the attachment behavioural system also has a wider context: the ‘security metasystem’.276 Ainsworth argued that a sense of security could be fed by multiple sources, not just intimate relationships (Chapter 2). Apparently without awareness that they were reclaiming Ainsworth’s position, Shaver and colleagues proposed that a sense of security can be achieved variously through experiencing the availability of attachment figures, through experiencing the self as worthy (p. 488) and capable, and through experiencing the world as orderly and meaningful. These three psychological processes, for all their differences, therefore have in common that they can serve both to modulate experiences of anxiety and contribute to broaden-and-build cycles. As well as a behavioural system in its own terms, attachment was therefore situated as part of a security metasystem. The species-level function of the system, it was proposed, is self-regulation. The metasystem is activated by potential feelings of anxiety, which can arise from relational sources such as close relationships but also from non-relational sources such as reminders of mortality. It is deactivated by felt security, regardless of the source of this feeling. The achievement of this goal can be pursued, as needed, through attachment-relationships, drawing upon self-esteem or from appeal to cultural worldviews that make the world seem orderly and meaningful. Though they often can work together, at times when one strategy is hard-pressed by circumstances, the others can be brought in to substitute. This is facilitated by the fact that they have cognitive components in common. Shaver and colleagues also argued that all three may facilitate exploration and caregiving.

Hart’s work has subsequently focused on this macro level in articulating the concept of the ‘security system’.277 By contrast, this level of analysis has generally not been pursued by Shaver and Mikulincer, though it has remained conceptual background. Instead, Hart has developed ideas regarding the contribution of other behavioural systems to security. Initially, both Shaver and Mikulincer began with the assumption that attachment style would reflect the history of relationships with attachment figures. Attachment styles were defined as ‘systematic patterns of expectations, emotions, and behavior in close relationships that are viewed as the residue of particular kinds of attachment histories’.278 So, for instance, in the 1990s, Mikulincer reported from a study with a student sample that secure and anxious attachment styles were associated with greater self-reported curiosity and positive attitudes towards curiosity than participants with an avoidant attachment style. Furthermore, he found that participants with a secure attachment style were more likely to trust new information in making a social judgement than participants with avoidant or anxious attachment styles.279 It could be that cognitive openness contributes to, or even partly constitutes, a secure attachment style. However, Mikulincer’s interpretation of these findings was that insecure attachment hinders the full development and successful expression of the exploratory behavioural system, reducing occasions for felt curiosity. At the time, the self-report approach to attachment research was still being established, and part of the definition of secure attachment, since Ainsworth, had been an inverse relationship with exploration. Attachment was therefore conceptualised as antecedent and causal for curiosity.

By contrast, in the mid-2000s, with the self-report tradition established and thriving, Shaver and Mikulincer were willing to acknowledge that the qualities of adult attachment relationships are fed by other behavioural systems. Attachment styles were regarded as shaped by the variety of factors that shape perceptions of the availability, attentiveness, responsiveness (etc.) of others. In a 2005 article, Banai, Mikulincer, and Shaver argued that ‘a person who is psychologically injured in the process of seeking one selfobject provision (p. 489) tends thereafter to avoid seeking satisfaction of other selfobject needs’.280 Feelings of rejection in relation to exploration or sexuality, for instance, might well contribute to a ‘defensively avoidant stance so pervasive that it generalizes across different kinds of interpersonal experiences and leads people to dismiss social ties in general’.281 In 2007 they argued clearly that ‘just as good attachment experiences have beneficial effects on other behavioral systems, such as exploration and caregiving, good experiences related to those behavioral systems are likely to feed back on the attachment system in ways that allow it to function in a less defensive, less distorted way’.282 Elsewhere, Shaver expressed regret that he and Hazan had earlier described adult romantic relationships as grounded in attachment, sexuality, and caregiving. He felt that exploration and affiliation were also behavioural systems integral to adult romantic relationships, and to the functioning of the relationship as a secure base and safe haven.283 The capacity to explore is integral to utilisation of the attachment figure as a secure base. When exploration is blocked by a minimisation of the exploratory system, the experience of having a secure base is thinned or reduced. Conversely, when familiarity is distrusted as part of the maximisation of the exploratory system, this may have implications for being able to use attachment figures as a safe haven.

Over time, then, Shaver and Mikulincer came increasingly to the view that adult attachment style would be fed by many sources, rather than primarily reflecting a child’s history of attachment experiences. This position was supported in 2013, when Fraley and colleagues found little prediction from infant attachment classifications to the ECR completed in adolescence in a study with 707 participants besides a very small negative association between infant proximity-seeking and an anxious attachment style (β‎ = –.08). Even if it was somewhat theoretically expected, the social psychology paradigm nonetheless had a history of claims to legitimacy on the assumption of continuity from infancy. As a result, there appears to have been ambivalence about this finding: the result was reported by Fraley and colleagues, but only in a web-based appendix to the published article.284 As a result, the finding was effectively buried. It has been little noticed by subsequent researchers.

Sexuality

In pursuing their thinking about behavioural systems, Shaver and Mikulincer found the concept of the sexual behavioural system significantly underdeveloped. Bowlby had long acknowledged sexuality as a behavioural system. It was an essential point of comparison in Attachment, Volume 1 as Bowlby set out the concept of attachment and described it as a (p. 490) behavioural system. However, in this, Bowlby’s aim was also to distinguish attachment from sexuality, contesting versions of psychoanalytic theory that described all motivation as, in a sense, sexuality. Bowlby evidently felt that sexuality needed to be placed in the background in order to make space for attachment (Chapter 1). Though his discussions of the sexual system are, as a result, somewhat cursory, it is nonetheless clear that Bowlby regarded the adult sexual system as sharing some components of the attachment system, such as the behaviours clinging and kissing. Furthermore, he drew on ethological evidence to support the (psychoanalytic) argument that the qualities of the parent–child relationship go on to influence sexual preferences when the child has grown to adulthood.285 Ainsworth followed Bowlby in generally downplaying sexuality, despite her personal beliefs regarding its importance (Chapter 2). More than Bowlby, she drew out that adult romantic relationships entail attachment, sexuality, and caregiving. However, her remarks on the sexual behaviour system essentially agree with those of Bowlby, with one exception. Whereas Bowlby described gay and lesbian relationships as deviations in Attachment, Volume 1, Ainsworth argued that there was no basis for this position, implicitly acknowledging Bowlby’s stance as prejudice.286

Overall, the concept of the sexual behavioural system was acknowledged but left sleeping by the founders of attachment theory. And the focus of the developmental tradition of attachment research on childhood directed attention away from sexuality, with the exception of Fonagy and Crittenden.287 By contrast, the sexual behavioural system was central for Shaver and colleagues since their focus was on adult relationships, and above all adult romantic relationships. This concern also occurred against the backdrop of growing attention to sexuality within quantitative social science in general over the 1980s, in part reflecting the salience of sexuality for identity politics in the period, but also more narrowly the availability of federal funding to study social factors relevant to the spread of HIV.288 The initial model of the sexual behavioural system offered by Shaver and Mikulincer in the 1990s was not well formulated. They tended to pathologise people outside of monogamous romantic relationships. Individuals who experience sexual attraction to more than one person were situated as deviant, since it was expected by Shaver that ‘just like infants, adults are primed to select one special figure’.289 And Mikulincer argued in 2006 that ‘the set goal of this system is to impregnate an opposite-sex partner’.290 As in Bowlby’s work, this characterisation pathologises gay and lesbian sexualities, and can be regarded as reflecting Shaver and Mikulincer’s general lack of concern for sexual diversity in forms of adult relationships. Their characterisation of (p. 491) the set-goal of the sexual system also is clearly phallocentric: the sexual set-goal of women tends not to be ‘to impregnate’ their partner.

These problems in Shaver and Mikulincer’s initial reflections on the sexual behavioural system were grounded in a broader confusion of the evolutionary and individual levels of analysis, distinguished by Tinbergen and the ethological tradition (Chapter 1). Stevenson-Hinde had identified this tendency in Shaver’s work already in the 1990s.291 A behavioural system was, for ethologists like Stevenson-Hinde, a sequence of behaviours that might be inferred at a species level to contribute to survival and/or reproduction. The sexual system therefore did not need to have even intercourse, let alone impregnation, as its set-goal so long as in evolutionary history the behavioural system had contributed to the probability of reproduction. This distinction has been increasingly acknowledged by Shaver and Mikulincer over the past decade, as they have reflected on ideas from evolutionary biology292 and pursued collaborative research on the sexual behavioural system together with colleagues such as Gurit Birnbaum and Omri Gillath. At times, this group of collaborators have argued that the sexual system has multiple set-goals involving the initiation and maintenance of relationships.293 However, their stable position in recent years appears to have been that the set-goal of the sexual system is sexual access to a partner.294

As Shaver and Mikulincer have further developed their concept of the sexual system, one important area of change has been their attention to gender. The initial model of the sexual behavioural system was implicitly phallocentric, in situating the set-goal of the system as impregnation of the partner. However, empirical studies exploring sexual experiences indicated meaningful differences between male and female participants. A study by Gillath, Mikulincer, Birnbaum, and Shaver published in 2007 examined the effects of a subliminal prime of a picture of an attractive nude person of the opposite sex.295 In one version of the procedure, student participants were then asked to view pictures of members of the opposite sex and report their relative feelings of arousal. In a second version of the procedure, after the prime the student participants were exposed to a varied series of pictures and were asked to decide whether each contained or did not contain sexual content, and to do so as quickly as possible. Reaction times for accurate identification of sexual content was taken as a measure of the accessibility of sex-related thoughts. Gillath, Mikulincer, and colleagues found that the naked picture prime led to higher accessibility of sex-related thoughts in both male and female participants. However, the subliminal prime led women to report lower levels of sexual arousal. Rather than reflecting intrinsic sex-differences, the researchers interpreted these findings with the proposal that ‘it would not be surprising if women had more reasons than men to be threatened by certain kinds of sexual situations’.296 Women may experience (p. 492) implicit or explicit pressure around sex, including diverse reputational threats associated with both having and not having sex. The women students participating in the study were understood by the researchers as interpreting the subliminal sexual prime adversely, in the context of their experiences of what male sexuality, or male nudity specifically, represented to them.

By 2007, Shaver and Mikulincer had repeatedly tested for gender effects in attachment styles, but only found these when conducting studies of couples in interaction, not when individuals had been studied alone.297 In examining sexuality, however, they found both individual-level and couple-level differences. Brassard, Shaver, and Lussier asked 273 heterosexual couples aged 18–35 to complete the ECR alongside measures of sexual experiences in their relationships.298 The researchers found that for women, but not for men, an avoidant attachment style was associated with engagement in less sexual fantasy and daydreaming. They also found that for men, but not for women, an anxious attachment style was associated with pushiness for sex and the use of sex for reassurance. Women, in turn, responded differently to this behaviour by their partner depending on their attachment anxiety: only women low in attachment anxiety appeared to resist this pushiness and engage in sexual interactions according to their own desires and pacing.299

Recently, Birnbaum, Mikulincer, and colleagues developed the concept of the sexual system into a self-report scale modelled on the ECR, the Sexual System Functioning Scale (SSFS).300 Hyperactivation items on the SSFS were designed to tap the co-presence of desire and worries about sex, ‘much like anxiety items in the ECR tap into desire and worries related to emotional closeness’.301 Deactivation items on the SSFS were designed to tap disinterest and discomfort with sex, ‘much like avoidance items in the ECR tap into disinterest and discomfort with emotional closeness’.302 Both hyperactivation and deactivation were conceptualised as strategies for achieving the set-goal of sexual access to the partner. Hyperactivation is a strategy that entails vigilance regarding access to sex, and ready and intense striving for sexual interactions when sexual access seems possible. Sexual access may be sought insistently and abruptly, with less concern for context or the wishes of others, since it can be prioritised over other potential goals including caregiving, exploration, and attachment. Hyperactivation of the sexual system therefore entails reduced sensitivity, in the sense of recognising, acknowledging, and responding in a timely way to the cues of a partner. (p. 493) Deactivation is a strategy that entails reduced striving for sexual access, in the service of avoiding rejections that might further reduce the fragile sexual availability of the partner. Other behavioural systems may then have a relatively greater role in shaping decision-making regarding partners, including appraisals regarding the need for or meaning of sex.

Where hyperactivation and deactivation are low, the researchers proposed that the ‘primary strategy’ of the sexual system will entail communication of sexual interest with sensitivity to partner signals and context. This suggests the potential for some bidirectional causal links between secure adult attachment and the primary strategy of the sexual system, though the links will depend on the cultural and social factors that shape perceptions of the relational meaning of sexual interest and activities.303 However, the relationship between the sexual and attachment behavioural systems can be expected to vary substantially by life-stage. In adolescence and early adulthood, characteristically, sexual interactions with partners may be sought without the expectation of the partner becoming a secure base. The sexual encounter may offer instead an opportunity for a cocktail of other social and psychological processes, including the exploration of identity, gender, personal values, power, and status. It may also provide a temporary safe haven.304 Furthermore, Birnbaum and colleagues have also argued that causal links between individual differences in the attachment and sexual behavioural systems may be moderated by the context. For example, the conditions of success or failure of the sexual system may contribute especially to schemas regarding the availability of partners during periods when certainty regarding the relationship is low: in the early months of a relationship or following serious conflicts or separations.305

Birnbaum, Mikulincer, and colleagues examined the association between the SSFS and the ECR in two samples of undergraduates. There was a robust association between ECR attachment anxiety and SSFS hyperactivation (r = .52, .57). Both reflect concerns about the availability of intimacy and about the partner as a safe haven who can offer regulation and satisfaction in the context of arousal; a substantial association was therefore anticipated. However, ECR attachment avoidance had a much weaker association with SSFS deactivation (r = .28, .27). It had at least as strong a correlation with SSFS hyperactivation (r = .25, .38).306 An avoidant attachment style may at times be linked to the hyperactivation of the sexual system. This may be because sexuality can be made to serve as an alternative to or a substitution for intimacy, a tactic of a minimising attachment strategy.

The distinction between the sexual and attachment systems in adulthood suggests that they may have distinct networks of associations with other behaviours, beliefs, and experiences. In one study investigating the correlates of sexual system functioning, Birnbaum, Mikulincer, and colleagues showed undergraduate participants films of various people of the opposite sex, presented in the manner of a dating site. In examining their data for sexual (p. 494) hyperactivation, the researchers found that these were moderated by gender. Among men, hyperactivation was associated with greater reported interest in relatively less-attractive women. Among women, hyperactivation was associated with more interest in relatively more-attractive men. Birnbaum and colleagues interpreted these findings as suggesting that hyperactivation and deactivation represent conditional strategies, part of the human evolutionary repertoire. Where sexual access was uncertain for women, in human evolutionary history reproductive success would have been increased by focusing attention on healthy mates to ensure successful offspring. By contrast, where sexual access was uncertain for men, the equivalent strategy to increase reproductive success would have been to reduce the standards used for assessing prospective mates. In a replication study, effects remained significant when controlling for scores on the ECR, supporting the autonomy of sexuality as a distinct behavioural system and the incremental validity of the SSFS hyperactivation scale.307

The distinctive predictive power of the SSFS deactivation scale was also demonstrated in a later study by Szepsenwol, Mizrahi, and Birnbaum:308 62 couples who had recently begun dating completed the ECR and SSFS at four-month interviews on three occasions, alongside self-report measures of relationship satisfaction. Birnbaum and colleagues found that an individual’s report of SSFS deactivation was associated with their report of lower relationship satisfaction early in the relationship: this effect continued throughout the first year among participants whose partner had an avoidant attachment style, but disappeared among participants whose partner did not have an avoidant attachment style. Birnbaum and colleagues interpreted this finding in light of the idea that, though sexuality remains powerful and relevant, in many couples relationship satisfaction comes over time to rest more on the attachment system.309 By the end of the first year, this permits relationship security to compensate for potential sexual system deactivation in providing happiness in the relationship.

Overall, Birnbaum, Mikulincer, and colleagues’ work on the sexual behavioural system has represented the most sustained attempt to flesh out Bowlby’s early, and unfinished, characterisation, and respond to Ainsworth’s call for further inquiry. There has been healthy interest in what is still a relatively new model and self-report scale from within the wider area of relationship psychology. Predictably perhaps, the SSFS and surrounding theory has generally been ignored by researchers in the developmental tradition of attachment research, even those with a central concern with sexuality such as Fonagy. Nonetheless, a comparison of Birnbaum, Mikulincer, and colleagues’ account of the sexual behavioural system and that of Main can offer an indication of the contribution of this work to attachment theory.

Main took the concept of conditional strategies from ethological discussions of mating behaviour. The diverse repertoires of animal behaviour that can contribute to reproductive success formed the model for Main’s theory of repertoires of infant behaviour that promote survival (Chapter 3). Avoidance was theorised to be a ‘cut-off’ behaviour in the manner of Chance, used to maintain regulation and circumvent an approach–avoidance conflict by minimising attention to the stimuli provoking this conflict. Ambivalent/resistant attachment was conceptualised through Trivers as an intensification of signals to the caregiver of (p. 495) the need to attend to the child, achieved through a perceptual vigilance regarding signals that might suggest the caregiver’s unavailability. Both conditional strategies were anticipated at species level to increase the odds of infant survival when confidence is not possible in the caregiver’s availability. Birnbaum, Mikulincer, and colleagues’ account of the sexual system likewise described three different pathways to sexual access to the partner. The role of attentional vigilance is also highlighted in their account, though it seems to be a contributor rather than the essence of the maximising strategy as for Main.

The exact relationship between sexual behavioural repertoires and evolutionary success seems unstable and somewhat muted in the work of Birnbaum, Mikulincer, and colleagues. The social psychologists have only relatively recently stabilised ‘sexual access’ as the set-goal of the sexual behavioural system. However, an additional contributing factor may have been the tendency in the work of Shaver and Mikulincer to lose track of the distinction between two of Tinbergen’s four questions (Chapter 1), regarding the expected proximal and the evolutionary outcome of a behaviour sequence. The researchers have moved from reproductive success to sexual access as the set-goal of the sexual behavioural system, but without fully articulating the relationship between sexuality and reproduction, or between individual motivation and species-level behavioural repertoires. The inattention to the affective components of behavioural systems inherited from Bowlby has proven a hindrance here, since this might have prompted further attention to both individual-level motivations and the evolutionary basis of behavioural repertoires. For instance, the contribution of a deactivation of the sexual system to the achievement of either sexual access or reproductive success is both undertheorised and empirically unexamined. At the same time, the qualities of sexuality as at times insistent, disquieting, and above all polyvalent, which make this aspect of human life both uncomfortably personal and uncomfortably social, have yet to be adequately examined.310

Nonetheless, the efforts of Birnbaum, Mikulincer, and colleagues to describe the sexual behavioural system represent a major advance. With the exception of sexual abuse as a kind of trauma, Main did not further discuss sexual and/or reproductive behaviour in her writings after applying the concept of conditional strategies to infant behaviour. By contrast, Birnbaum, Mikulincer, and colleagues have provided evidence that concern about the availability of sexual access and hyperactivation of the wish to achieve it is associated in anticipated ways with concern about the availability of a safe haven and hyperactivation of the wish to achieve it. They have also shown that, though correlated, the two phenomena are not reducible to one another and have distinct correlates. It is easy to imagine that Bowlby would have been pleased by such findings, as evidence against the tendency he disliked in the psychoanalytic theory of his day to treat sexual and relational motivations as interchangeable.

Dominance

Whereas the sexual behavioural system was already described, if briefly, by Bowlby, an innovation made by Shaver and Mikulincer has been the introduction of a ‘dominance’ behavioural system. Though not discussed as a behavioural system, the concept of a behavioural (p. 496) repertoire associated with dominance can be found already in Bowlby’s reports from ethology. For instance, in Attachment, Volume 1, Bowlby discussed ethological observations indicating that dominance could be recruited in the service of the caregiving system in non-human primates. Dominance behaviours were described as those that signal the availability of coercion, or violence if pushed, if deference is not provided. Bowlby described ‘the important observation that when a dominant male senses a predator or other danger he commonly threatens or even attacks a juvenile that unwarily approaches the danger spot. The dominant male’s behaviour, by frightening the juvenile, elicits the juvenile’s attachment behaviour. As a result the juvenile seeks the proximity of an adult animal, as often as not that of the very male that frightened it; and by so doing the juvenile also removes itself from the danger.’311 However, beyond characterising dominance as a behavioural repertoire with links to other systems such as caregiving and attachment, as well as aggression, Bowlby did not give attention to the topic. Though strongly interested in questions of power and domination in his early work before the development of attachment theory, from the 1950s onwards he tended to consider the world from the point of view of the infant attachment system.312 Dominance therefore became relevant only insofar as it impacted the caregiving system. As Smith and Connelly astutely observed, in Bowlby’s importation of ethology into child development, the concepts of dominance and territory were both left behind, albeit for different reasons.313

Shaver’s first sustained consideration of the topic of dominance appeared in his writings on the concept of ‘narcissism’, which were prompted by his longstanding interest in the relationship between self-concept and affectionate relationships with others, though also by the wider context of interest in personality disorders within social psychology. In ‘Shamed into self-love: dynamics, roots, and functions of narcissism’ from 2001, Robins, Tracy, and Shaver argued that there are two paths to getting ahead in life, ‘through exerting influence over others and through accomplishments’.314 Here the researchers referred to a ‘getting-ahead orientation’ to achieve influence and accomplishments, fuelled by ‘the power and achievement motives’.315 They speculated that these motives may operate quite widely. Usually they are moderated by or combine with affiliative motives, so that the wielding of power and the attainment of achievements are either in the service of some others or at least not at their expense. However, Robins, Tracy, and Shaver speculated that where the goals of the attachment system have been chronically unmet, one response may be the recruitment of the ‘getting-ahead orientation’ in the service of attachment. This, they proposed, may be the origins of the spectrum towards ‘narcissistic personality disorder’. Minimising and maximising strategies are premised on the expectation that at least a conditional availability of attachment figures can be achieved in this way. By contrast, a temperamental predisposition towards aggression or irritability may predispose self-assertion as a kind of ‘tertiary strategy’ to coercing the notice and recognition of others, and sometimes perhaps their availability: ‘According to our reasoning, the search for love and affection could not have worked under the childhood (p. 497) conditions that fostered narcissism, particularly if the developing narcissist was temperamentally aggressive and irritable. Given these constraints, narcissists were, in a sense, correct to have pursued a more assertive, self-promoting strategy.’316

In 2011, Shaver and Mikulincer returned to these considerations, introducing the concept of a ‘power behavioural system’. They argued that the set-goal of the power system is to ‘remove threats and obstacles that interfere with a person’s sense of power’.317 They proposed ‘felt power’ as the set-goal for the power system. Their proposal was explicitly modelled on ‘felt security’ as the set-goal of the attachment system according to Sroufe and Waters.318 At first glance, this might appear to be a set-goal solely at the individual level, rather than the achievement of a particular self–environment relationship. However, Shaver and Mikulincer defined ‘felt power’ as the sense that one can control the environment to have one’s needs met ‘without undue social interference’.319 In a sense, ‘felt power’ might be regarded as another contributory to the ‘security metasystem’ theorised by Shaver in his earlier work with Hart and Goldenberg. Like attachment, self-esteem, and cultural meaning, the sense that the environment can be controlled to meet one’s needs may form one contributory factor to a sense of felt security.320

Shaver and Milulincer proposed that the primary strategy of the power system is to see felt power in a manner commensurate with the aims of other behavioural systems. Indeed, they suggested that ‘power often facilitates the smooth operation of other behavioral systems, such as exploration, caregiving, and sex’.321 For instance, control over the environment may be sought in order to ensure that the needs of other people can be met, perhaps with the power system operating in the service of caregiving or with power prompting the feeling of responsibility for dependents. However, in the context of repeated failures to predictably achieve felt power, feelings of fear and helplessness may instead be evoked. Then a maximising or minimising strategy will be brought online. Shaver, Sagev, and Mikulincer proposed that hyperactivation of the power behavioural system will occur when an individual experiences concern about the availability of sufficient control over their environment to ensure that their needs are met. The hyperactivation of the power system leads to increased vigilance regarding threats or obstacles to control of the environment to ensure that one’s needs can be met. This is fuelled by both ‘an excessive urge to gain power and an extreme fear (p. 498) of failure’, a combination that ‘results in chronic activation of the power system, even when there is no imminent threat or actual damage to one’s power; an indiscriminate urge to assert power over others; frequent anger and hostility toward others (who are viewed as potential rivals); and a proclivity to attack others following minimal or ambiguous signs of competition or provocation’.322 Other concerns, including attachment and exploration, are therefore regularly suppressed.

Shaver, Sagev, and Mikulincer’s account of deactivation of the power system was quite different to their account of deactivation of the attachment or sexual behavioural systems. Whereas deactivation of the attachment or sexual systems entails reduced acknowledgment of stimuli that might activate the behavioural system, Shaver and colleagues argued that deactivation of the power system entails a heightened sensitivity to threats, much like hyperactivation of the system. However, even minimal threats are interpreted as prompting submissive, coy, or self-abasing behaviours. What are avoided are contexts that would call for an assertion of rights and opinions, including conflicts with others.323 Gentleness is pressed into service, perhaps even deformed at times into a mode of submission or conciliation, rather than functioning as an expression of tenderness or of care. Shaver and colleagues developed the Power Behavioural System Scale to assess maximising and minimising of the power system. As predicted, both hyperactivation and deactivation of the power system were negatively associated with other existing self-report measures of feelings of power.324 Hyperactivation of the power system had a moderate positive association with both attachment anxiety and avoidance; deactivation of the power system was moderately associated with attachment anxiety.

In an observational study of 100 dating couples, Shaver, Sagev, and Mikulincer asked participants to discuss a live problem in their relationship. Hyperactivation of the power system predicted greater displays of hostility and distress to the partner, whereas deactivation predicted greater displays of submissiveness and distress to the partner. Both dimensions were associated with difficulties for the couple in engaging in a constructive discussion of the problem. These associations held when controlling for attachment anxiety and avoidance, and for neuroticism.325 In another study, Mikulincer and Shaver found that priming for a feeling of powerfulness led to greater feelings of optimism when participants were low on attachment anxiety, suggesting that the attachment anxiety interfered with the capacity to use felt power as a basis for felt security. They also found that the same prime led to greater objectification of others when participants were high on attachment avoidance. They interpreted this finding as suggesting that the feeling of connectedness and security associated with low attachment avoidance could counteract the potential for felt power to lead to haughtiness about others.326 Such findings suggest that the power system scale has expectable correlates in behaviour and perception, offering support for the new scale.

(p. 499) Shaver, Sagev, and Mikulincer were also pleased to report that ‘as intended, the correlation between the hyperactivation and deactivation scores was not statistically significant, r(360) = .07’.327 Yet it is not clear why hyperactivation and deactivation of the power system were assumed to be unrelated, rather than negatively associated. In principle, submissive behaviour and hyperactivated power behaviour would seem to be opposites, rather than unrelated. The implication is that Shaver and Mikulincer believed that individuals who display submissive behaviour to others when they feel threatened are no more or less likely to display dominant behaviour when they feel threatened than the rest of the population, and vice versa.328 In fact, the idea of a two-dimensional space defined by orthogonal variables seems rather to have been pressed into service from the ECR as a criterion of the validity of new scales of behavioural systems for Shaver and Mikulincer, without the presentation of a theoretical justification for why all behavioural systems should have this structure. Gillath, Karantzas, and Fraley have suggested that the idea of orthogonality of attachment dimensions was never justified: it originates, they suggest, in part from a misreading of Bowlby’s statement in Separation about the ‘independence’ of concepts of self and other.329

The concept of the power behavioural system is unquestionably an advance on the awkwardly named ‘getting-ahead orientation’ from Shaver’s earlier work. However, the power behavioural system is on much less firmer footing than, for instance, their more recent work on the sexual behavioural system. One problem is that the relationship between dominance and aggression remained poorly soldered down in Shaver and Mikulincer’s account, and the two elements rattle about audiably against one another. Shaver and Mikulincer highlighted the importance of the ‘feeling of anger, which in our view is an emotional signature of power-system activation’.330 And they have often described anger in general as linked to the power system. However, it remains unclear whether all assertions of power require anger. For instance, considering Shaver’s earlier remarks, the use of achievements and successes to claim glory and influence do not seem overtly aggressive. Conversely, it is not clear that all assertions of anger are linked to the power system. The secure infant’s protest on the departure of the caregiver in the Strange Situation may be angry, but this is presumably before the power behavioural system has been assembled. The relationship between the power system and the fear behavioural system is also opaque. Exactly the submissive or coy behaviours that Shaver and Mikulincer suggest characterise the deactivation of the power system were regarded by Ainsworth and Bretherton as an effect of the simultaneous activation of the fear and affiliative systems.331 Other attachment theorists such as Crittenden and Hilburn-Cobb have (p. 500) conceptualised submissive behaviour as an independent behavioural response that can be brought into the service of attachment strategies, rather than as a minimisation of the dominance system.332 Of course, several pathways could be possible, but the rationale for considering submissive behaviour as characterising the ‘deactivation’ of the system remains both underdeveloped and rather out of keeping with Shaver and Mikulincer’s general characterisation of the minimising strategy for a behavioural system. It may be suspected that Shaver and Mikulincer may have conflated deactivation of the dominance system with the operation of two further distinct behavioural systems discussed by contemporary ethologists: the reconciliation system and the submission system.333

Secondly, it is unclear whether the power behavioural system is even a behavioural system by Shaver and Mikulincer’s own definition. Shaver and Mikulincer have generally drawn their concepts about behavioural systems through an expansion of behavioural repertoires already described by ethology, such as attachment, caregiving, and sexuality. However, in other cases some link to a behavioural repertoire has been retained, permitting Shaver and Mikulincer to toggle backwards and forwards between specific behaviours and more general observations about how adults feel, respond, and create symbolic meanings about one another. By contrast, in the case of the power behavioural system, the link between the narrow behavioural focus of ethology and the broad terms of social psychology appears to have unspooled. By Shaver and Mikulincer’s own definition, a behavioural system is a repertoire of functionally equivalent behaviours made available by evolution, with activating and terminating conditions, which would have the expectable outcome at a species-level of achieving a particular kind of change in the relationship between individual and environment. This definition retained the fundamental concern of ethology with repertoires of functionally equivalent behaviours made available by evolution. Yet it is difficult to describe assertion of ‘power’ as a distinct behavioural repertoire made available by evolution. Perhaps for this reason, at some points Shaver and Mikulincer have referred to the ‘dominance behavioural system’, which then offers a link back to ethology and a specific and observable behavioural repertoire.334

Overall, Shaver and Mikulincer’s account of the power behavioural system seems to be still a little sketched in and shaggy. It has been discussed mainly in speculative book chapters rather than peer-review articles, generating only limited commentary to date by other scholars.335 And though the scales for the caregiving and sexual systems appear in Milkulincer and Shaver’s Attachment in Adulthood, situated as the capstone of their professional labours, the scale for measuring hyperactivation and deactivation of the power system does not feature. There is in fact no explicit mention of the power behavioural system in (p. 501) the enormous book.336 Nonetheless, their work on the power behavioural system has been an interesting line of inquiry, continuing Bowlby’s interest in dominance behaviour within the ethological literature. Not least, the self-report measure of hyperactivation of the power system is thought-provoking as the only measure of a non-clinical phenomenon mentioned by Shaver and Mikulincer with a powerful positive association with both attachment anxiety and avoidance. Sexual system deactivation is likewise associated with both attachment anxiety and avoidance, but the association is weaker than for deactivation of the power behavioural system.337

Religion

Attachment to God

Submission and supplication, the deactivation of the power system, has also featured in another aspect of Shaver’s work. In a collaboration with Lee Kirkpatrick, a doctoral student from the University of Denver, Shaver developed a concern with the psychology of religion in the 1980s,338 a subdiscipline that had seen rapid growth in the late 1970s.339 In a paper from 1985, Shaver and Kirkpatrick emphasised the importance of submissive behaviour as a fundamental aspect of religious life. Through such behaviour ‘the influencer advertises his or her helplessness and dependence in order to solicit sympathy and assistance. It is easy to see how these ideas might be applied to the process of prayer, in which the human petitioner can often be found to heap praise on the deity and offer various concessions (ingratiation), and to reaffirm one’s inferiority and dependence while entreating God for help and guidance (supplication)’.340 Shaver and Kirkpatrick were interested that ingratiation and supplication appeared to be a lowering of the self in order to achieve the beneficence of a deity. However, practices of submission and supplication might have only a short-term effect on the feeling of a deity as beneficent. Longer-term beneficence may need to be based on a longer-term positive relationship. The researchers found that stable images of religious figures as benevolent (p. 502) were empirically associated with an individual’s report of a supportive family during their childhood.341 This suggested to the researchers a potential link between attachment and the psychology of religious belief and religious practices.

In the 1980s, there was a strong tendency in the academic psychology of religion to explicitly or implicitly model religious experience on Christian faith and practice. Shaver and Kirkpatrick did tend in this direction at times.342 However, right from the start Shaver’s interest in Buddhism provided a counterweight. His first discussion of Buddhism in print was in a chapter for Everywoman’s Emotional Wellbeing, a self-help guide for women published in 1986. Shaver and O’Connor cited the Tibetan Buddhist concept of maitri, ‘unconditional friendliness toward oneself’. An attitude of maitri, Shaver and O’Connor proposed, puts aside grudges, hostility, rumination, and a sense of entitlement in the world. Whereas in his later work with Mikulincer, Shaver tended to think of deactivation of the power system in terms of submissiveness, in his remarks on Buddhism it is possible to see a different form of deactivation. Rather than an alternative strategy for achieving power, maitri signified for Shaver and O’Connor a relinquishment of the fight for power, ‘along with the corresponding attitude of openness and trust toward others and toward nature’.343

Shaver and O’Connor commented that ‘a common goal, according to Western psychology, is enhancement of one’s feelings of control and self-esteem’.344 However, rather than an inevitable aspect of human psychology, Shaver and O’Connor argued that this is an effect of how the human subject has been shaped and constructed within western culture. Dominance and submissive behaviours may be part of the human behavioural repertoire as a result of our evolutionary heritage. But the activation of these behaviours depends upon the perception that control or coercion is needed in order to have our needs met. Shaver contrasted this perception to an alternative one, which he associated with Buddhism, in which ‘control of nature and emotion is devastating to life and ultimately impossible. Nature is the source and sustainer of life, and human feelings are one of its brightest creations.’345 In the terms drawn from Shaver’s later thinking with Hart and Goldenberg, it might be said that the security metasystem can be fed by different tributaries: among these, one is the assertion of self and a quest for control; another is a cultural/religious worldview in which striving for control is not ultimately beneficial to ensuring one’s needs are met.

In the 1990s, Kirkpatrick and Shaver documented a number of aspects of religious experience that could be influenced by attachment style, focusing on Christian religious practices and beliefs in American undergraduate samples. Using the Hazan and Shaver ‘love quiz’, they found that sudden conversion experiences were almost exclusive to participants with an avoidant attachment style.346 Indeed, nearly a third of participants with an avoidant attachment style in their sample reported having experienced a religious conversion. In interpreting this result it is worth keeping in mind, as mentioned in the section ‘A first attempt’, that (p. 503) the ‘love quiz’ collapsed avoidance and fear regarding caregivers into one category. Another finding by Kirkpatrick and Shaver was that participants who endorsed an anxious/ambivalent attachment style on the ‘love quiz’ were distinctive in reporting experiences of speaking in tongues.347 Participants were also asked to self-report their attachment style to God. Those that endorsed a secure attachment style also reported that they experience greater life satisfaction, less anxiety, loneliness, and depression, and less physical illness than other participants. These were also the correlates of a secure attachment style in romantic relationships. Kirkpatrick and Shaver found that the effects of the two attachment variables seemed to be additive: a secure attachment style in relation to both romantic partners and God had a stronger relationship with report of these positive experiences than a secure attachment only in one domain, and a secure attachment style in one domain was associated with more positive experiences than an insecure attachment style in both domains.

Ainsworth was apparently enthusiastic about Kirkpatrick and Shaver’s work on attachment and religion.348 However, critics have alleged that the idea of attachment to God was an overextension of the concept of ‘attachment’, with little resemblance to the prototype of infant behaviour. Granqvist, Shaver, and Mikulincer responded to this criticism.349 They reported experimental studies that suggested that divine figures in particular, and religious practice in general, could serve as a safe haven in the context of distress, and to an extent as a secure base. Attachment to God could qualify as an attachment relationship on these grounds.350 Additionally, they pointed to the role of proximity-seeking within (Christian) religious practice, such as going to church during life transitions; the role of language suggesting ‘approaching God’ within supplicatory prayer; as well as metaphors of ‘separation from God’ in imagining Hell or forms of punishment. This suggests symbolic forms of separation anxiety. Granqvist and colleagues also highlighted the conditions under which religious life may become especially salient: ‘people are most likely to turn to God or other supernatural figures when they face situations that Bowlby (1982) believed activate the attachment system, such as illness, injury, or fatigue; frightening or alarming events; and separation or threat of separation from loved ones’.351 Part of the problem faced by Granqvist, Shaver, and Mikulincer in defining the relationship between attachment and religion is that the former term has had both narrow and broad meanings since Bowlby (Chapter 1). In this sense, it would appear from their arguments that the conditions that activate the attachment system—narrowly construed—seem to elicit phenomena associated with the attachment system—broadly construed—within religious life, perhaps given the role of religious practice as a possible contributory to the security metasystem. Above all, it is perhaps the role of divine figures or the religious community as a kind of safe haven that provides the hinge between attachment broad and narrow.

(p. 504) Mindfulness

Though enthusiastic about the idea of attachment to God, Shaver retained his interest in forms of religious life without a personal God or supplicatory prayer. This interest was further fuelled by an invitation received by Shaver and Mikulincer in October 2004 to visit the Dali Lama, who had been interested by their work.352 Following this experience, at the same time as their work with Granqvist, Shaver and Mikulincer developed a strand of empirical research exploring mindfulness practices. Mindfulness-oriented meditation techniques originally emerged within Buddhist religious life, but had become reformulated in the context of their transplantation into western wellness technique.353 Shaver and colleagues situated themselves in terms of this secularising trend: ‘American psychologists have lifted mindfulness out of this rich context (perhaps while attempting to separate it from religious considerations) and applied it in a more individualistic, less socially connected, and more ethically neutral way. In our opinion, placing mindfulness in an attachment-theoretical framework would allow it to benefit not only from additional kinds of empirical tests but also from an assortment of ethical, social, and developmental, yet not necessarily religious, concepts.’354 The reception of mindfulness in terms of attachment theory would, Shaver and colleagues argued, contrast with the individualistic reception of mindfulness as a western wellness technique. Mindfulness would entail finding a safe haven in the representation of attachment figures, the Buddha, in the tradition’s teachings, and in other members of the religious community:

English-language books about Buddhist meditation make the process of mindful meditation seem rather solitary and asocial. During our discussions with the Dalai Lama in 2004, however, it was pointed out that one of the simplest and most frequently spoken Buddhist prayers is: “I take refuge in the Buddha, the Dharma, and the Sangha,” which means (in our terms) the mental representation of the Buddha as a loving, compassionate, and wise teacher; the Buddha’s teachings (dharma); and the community of fellow Buddhists (sangha). In other words, the key concept is “taking refuge,” which is similar to Bowlby and Ainsworth’s notion of using an attachment figure as a “safe haven” …

In our conversations with the Dalai Lama, he said that a common form of cultivating maitri is to imagine, during meditation, experiencing love from someone who has deeply loved you, “such as your mother.” We replied that “mother” might not always be a good choice if one’s relationship with her in childhood was not comfortable. He said in that case one could imagine being loved by the Buddha, as in the common prayer, “I take refuge in the Buddha, the Dharma (the Buddha’s teachings), and the Sangha (the community of fellow practitioners)”.355

(p. 505) Shaver and colleagues hypothesised that attachment anxiety and avoidance would be anticipated to disrupt the capacity for mindfulness. They drew upon a self-report scale for mindfulness developed by Baer and colleagues.356 On the basis of a factor analysis of 112 items from previous self-report mindfulness scales, Baer and colleagues had developed a scale with five main factors which they termed 1) Nonreactivity to Inner Experience, 2) Observing/Noticing/Attending to Sensations/Perceptions/Feelings, 3) Acting with Awareness, 4) Describing/Labelling with Words, and 5) Nonjudging of Experience. Shaver, Lavy, Saron, and Mikulincer found that both an anxious and an avoidant attachment style made strong unique contributions to scores on the mindfulness scale, and that the two attachment dimensions accounted for 42% of variance in total mindfulness.357 In more detailed analysis, Shaver and colleagues found that an anxious attachment style was negatively correlated with non-reactivity to inner experience (r = –.54), acting with awareness (r = –.46), and non-judging experiences (r = .43). These are factors that might be especially vulnerable to rumination. They found that the avoidant attachment style was negatively associated with all of the factors of mindfulness. That is, unlike anxious attachment, avoidant attachment was also associated with difficulties noticing sensation, and difficulties giving words to such sensations even when they are noticed. These factors might be specifically vulnerable to the suppression of thoughts with content or associated affect related to attachment, characteristic of the avoidant attachment style. Additionally, both attachment styles may hinder broaden-and-build cycles, which Shaver and Mikulincer have theorised contribute to greater social and personal resources that support regulation, and through this contribute to mindfulness.

The remarkably strong association between attachment style and the mindfulness scale suggests a surprising degree of congruity, or even potential overlap, between the two variables. Like in earlier work disentangling neuroticism and the anxious attachment style, the overlap between a secure attachment style and mindfulness will need to be worked out in terms of their respective relationships with other variables, including observed behaviour and subjective phenomena such as self-compassion. Priming studies may also prove important in articulating factors mediating the relationship between attachment style and mindfulness.358 However, the strong association between attachment style and mindfulness also presents, as Sahdra and Shaver acknowledged, a fundamental conceptual question: ‘We encounter the paradox that “attachment security” or “secure attachment” is considered ideal or optimal in a major stream of Western psychology, attachment theory, whereas the ideal or optimal state in Buddhist psychology is called “nonattachment”.’359 Yet the paradox is an effect of two different technical uses of the term ‘attachment’. There are many strands of Buddhism, which makes generalisation difficult. Nonetheless, Sahdra and Shaver argued that, by and large, the Buddhist concept of non-attachment suggests a lack of fixation on mental representations of things in the world. This is not only compatible with connectedness with others and intimacy in relationships, but probably facilitates it. By contrast, non-attachment in attachment theory means the lack of a discriminated and preferential intimate (p. 506) relationship with a person who might serve as safe base and secure haven. Lack of a secure base and safe haven may be anticipated to contribute to reduced flexibility in mental representations about relationships.

Adaptation

Shaver and Mikulincer’s characterisation of religious life tends to give the impression that an insecure attachment style is always bad. In their writings, experiences that seem characteristic to certain forms of insecurity, such as conversion experiences or speaking in tongues, tend to be treated respectfully but as indirect signals of mental pathology. The same is true of their writings on adult romantic relationships. Like the majority of the developmental tradition of attachment research,360 Shaver and Mikulincer have tended to regard insecurity as a developmental adaptation, with the potential for some short-term benefits in achieving the availability of attachment figures, but with long-term harms across all other domains. As Waters and colleagues have observed (Chapter 2), there has been a tendency among attachment researchers to assume that ‘all good things go together’, based on positive associations with the word ‘security’ rather than precise knowedge of the psychological process actually under scrutiny, which may have advantages in some domains, no effect in others, and drawbacks under certain circumstances.361 At times, Shaver and Mikulincer too have acknowledged the problem. For instance, in a 2006 epilogue in an edited volume celebrating his work, Shaver wrote that ‘to insist on a model of perfect security, rather than a model that acknowledges human complexity, depth and intrapsychic conflicts and tensions is bound to be misleading and perhaps even dangerous … In the attachment field we are so accustomed to glorifying secure attachment that we rarely stop to wonder why there aren’t more saints in the world.’362

Shaver’s observation is in part self-criticism. There are many occasions where decisions in research design or interpretation made by Shaver and collaborators seem to have been shaped by a ‘glorification’ of security. For instance, in a 1996 paper Mehr and Shaver described personal inconsistency across contexts as a positive quality.363 Certainly it can be. (p. 507) However, at least from the way that the paper is written, the interpretation seems at least in part shaped by the fact that this quality was associated with a secure attachment style. If it had been associated with insecure attachment style, it is easy to imagine that the opposite evaluation could have been made of the same quality, with speculations regarding how emotion dysregulation or the lack of a safe base can disrupt the consistency of identity across contexts. To take another example, in a 2000 paper Mikulincer and Sheffi found that participants endorsing a secure attachment style were more likely in a cognitive test to miscategorise poor exemplars of types as similar. This finding was interpreted positively as indicating the expansiveness and freedom permitted by the feeling of having a secure base. However, if the finding had been reversed, security could easily have been praised as contributing to greater discrimination and discernment.364

Part of the trouble in achieving effective consideration of the issue has been carelessness by attachment researchers in the use of the term ‘adaptation’, a problem already identified by Ainsworth in an unpublished conference paper at the International Conference on Infant Studies in 1984.365 Following Hinde and other ethological researchers (Chapter 1), Ainsworth expressed concern that the term ‘adaptation’ represented a twig-thicket of different meanings. She acknowledged that, in using the term, she and other attachment researchers had hindered effective discussion and even at times misled readers. On the one hand, ‘adaptation’ could refer to processes at a species level, in identifying a behavioural system or trait as contributing to survival or reproduction. On the other hand, ‘adaptation’ could refer to an individual level, identifying a behaviour or trait as responsive to the available rewards and punishments of the immediate environment.366 However, Ainsworth observed a third meaning of the concept: ‘In the developmental mental health sense the focus is on how individual differences in development, and on evaluation of how well or how poorly such development equips the individual to cope with the impact of the environment in which he lives.’367 What distinguished this third meaning of the concept from the second was that an evaluation was entailed. The second meaning was merely an acknowledgement that an individual may ‘adapt’ to their circumstances. Ainsworth’s third meaning was to identify ‘adaptation’ as the capacity to thrive in the long term in some way within those circumstances.

The subtlety and complexity of Ainsworth’s argument may have contributed to her decision not to attempt to publish the article, despite its popularity with students and collaborators.368 Ainsworth’s argument was that individual adaptation (long-term thriving) may (p. 508) result from adaptation (changing oneself in order to respond) to the environment. However, there are forms of adaptation (long-term thriving) where refusal to adapt (change oneself in order to respond) is optimal, for instance in depleting or punitive environments that can be changed or exited. Some forms of adaptation (thriving) may come at the expense of other forms of adaptation (thriving), as in the familiar case in which the demands of one area of life—family, work—come at the expense of others—diet, exercise, self-care. A further complexity lies in the fact that there are forms of adaptation (responding and/or thriving) that are based very directly on adaptation (species-level natural selection), such as the deployment of conditional strategies as evolutionary-based behavioural repertoires. However, there are forms of adaptation (responding and/or thriving) that are more based on social learning or other processes based more on human plasticity rather than responses directly grounded in adaptation (species-level natural selection).

Ainsworth stated that attachment researchers all agreed that avoidant attachment should be regarded as an adaptation (a response to the caregiving environment). And many were becoming persuaded by Main’s argument, endorsed by Ainsworth, that the avoidant attachment pattern was an adaptation (part of the human evolutionary behavioural repertoire). However, she identified that there were significant disagreements among her students and collaborators about whether, over the long-run, avoidant attachment contributed to adaptation (long-term thriving). Ainsworth offered her conviction that this was essentially an empirical matter: the question of whether ‘avoidant attachment may be adaptive according to ultimate criteria in the mental health sense is clearly a researchable proposition’.369 Her personal expectation was that insecure attachment would tend to work against or undermine long-term thriving. And this attitude unquestionably shaped her published descriptions of the avoidant and ambivalent/resistant attachment classifications. She remained open to the potential for insecure attachment to contribute to long-term benefits, for instance in some domains. But she was not holding her breath in anticipation of such results.

Shaver’s position seems to have essentially been the same. Until around 2007, the dominant theme of his writing was the benefits of secure attachment, across whatever domain was under discussion. There are even some passages in his work that read as hymns to the glory of secure attachment for its capacity to ‘create a kinder and more tolerant, harmonious, and peaceful society’.370 And, even after 2007, Lawler, Shaver, and Goodman could still put forward the extreme proposal that mental health professionals working with families should be screened for their attachment style, claiming that it is the right of every child to ‘receive services from health and mental health interveners who are trained in supporting relationship quality and are themselves secure with respect to attachment (as assessed with adult attachment measures)’.371 Yet a whispering, subterranean theme also appears across Shaver’s writings before 2007, offering qualifications and at times pointing in the contrary direction. In the 1986 chapter in Everywoman’s Emotional Wellbeing, Shaver and O’Connor noted a (p. 509) shift in psychological theory towards a recognition that many apparent symptoms or forms of pathology might also or sometimes better be regarded as effective responses to challenging and intractable situations. Shaver and O’Connor offered, as an example, that their colleagues in psychology appeared to be moving in their discussions ‘from the concept of “defense” to the more constructive concept of “coping” ’.372 Bowlby was a forerunner for this transition (Chapter 1).

Reflecting on the difference between defence and coping, Shaver and O’Connor considered how an individual may direct their attention away from problems, in avoidance or denial. This reduces perceptual information about the problem, and had classically been treated in psychology as a form of pathology. However, Shaver and O’Connor countered: ‘in cases where reality serves up a problem for which there is no solution, or at least no immediate solution, what’s so bad about losing touch with it?’373 They advised their women readers that if a problem is solvable, then avoidance or denial will be a costly and unproductive strategy. However, if a problem is truly not solvable, then avoidance or denial are optimal responses to the situation, and ‘this strategy frees their energies for other, more rewarding activities’.374 In Ainsworth’s terms, Shaver and O’Connor’s argument was that avoidance is not always an adaptation (a response to the situation): sometimes it can be motivated by fear or habit rather than a genuine acknowledgement of the nature of the situation. Where avoidance is deployed as a strategy in response to a situation that could otherwise be resolved, this is neither adaptive (a response to the situation) nor adaptive (a contribution to longer-term thriving). However, when an avoidant strategy is used in response to an unsolvable problem, then this is both adaptive (a response to the situation) and adaptive (a contribution to longer-term thriving).

Evidence for Shaver and O’Connor’s claim came from an early study by Mikulincer and Florian published in 1995. This would be the first in a slow accumulation of discrepant findings by Mikulincer and colleagues, which documented the benefits of an insecure attachment style under specific circumstances. Mikulincer and Florian asked 92 Israeli army recruits to complete the ‘love quiz’ at the start of their basic combat training. Their appraisal of the training, coping strategies, and peer evaluations of their behaviour were assessed four months later. Individuals endorsing an avoidant attachment style reported the use of more distance forms of coping than other participants, and were just as likely to be nominated for leadership positions as participants endorsing a secure attachment style. Interpreting their results, Mikulincer and Florian reflected that ‘although the tendency of avoidant persons to maintain social distance may be negatively evaluated in an emotionally laden interaction, it may be that in purposive instrumental interaction, like daily activity during combat training, avoidant persons may provide concrete assistance and relief from distress to others’.375

Basic combat training may represent the kind of intractable problem described by Shaver and O’Connor where physical, mental, and ethical difficulties may not be resolvable, especially through rumination. There could be personal and social benefits to freeing energies for other more rewarding activities. However, a limitation of Mikulincer and Florian’s 1995 paper was its focus on individual coping. In a larger study of multiple units during basic (p. 510) combat training, published in 2007, Davidovitz, Mikulincer, Shaver, Izsak, and Popper reported that the higher the officer’s score on avoidance on the ECR, the more the self-reported mental health of their unit deteriorated. In the first two months this deterioration was moderated by the soldiers’ own attachment style. However, over the four months there was a negative association between officer avoidance and their soldiers’ mental health regardless of the attachment styles of the latter.376 Yet Davidovitz and colleagues also found a positive association between officers’ attachment anxiety and followers’ mental health. This positive association, however, seemed to come at the expense of the unit’s performance in exercises.377

Another finding from Mikulincer’s research group offered further indication that insecure attachment strategies can have certain benefits depending on the circumstances. In 2001, Berant, Mikulincer, and Florian published a study of mothers’ responses to the diagnosis of congenital heart disease in their infant. The researchers examined the relationship between attachment style, coping strategies, wellbeing, and severity of the infant’s diagnosis. They found that mothers endorsing a secure attachment style tended to utilise a combination of support-seeking and avoidant-coping strategies if their babies had severe forms of congenital heart disease. The temporary use of avoidance or denial was positively associated with better reported wellbeing. Berant and colleagues interpreted this finding as suggesting that avoidant strategies can be helpful for a time in the face of an irresolvable problem if they are combined with the security to permit support-seeking as needed. When mothers had babies without severe forms of congenital heart disease, then, as expected, participants endorsing an avoidant attachment style relied especially on avoidant and distancing coping strategies, and reported moderate levels of wellbeing.

However, for mothers who had babies with severe medical problems and who had an avoidant attachment style, avoidant and distancing coping strategies were not used; these mothers reported the very lowest levels of wellbeing. Berant and colleagues proposed that in the face of an irresolvable problem that is also too appallingly distressing to maintain a distancing strategy, the avoidant strategy breaks down into a less-strategic state, with some of the features of the anxious attachment style but characterised by a whirlpool of distress, rumination, and felt insecurity.378 A potential explanation for this finding was put forward by Gillath, Giesbrecht, and Shaver. These researchers argued that an avoidant strategy can help keep at bay thoughts and feelings that might otherwise undermine coping. However, where these thoughts and feelings intrude, the outcome is worse than had the strategy not been attempted. They used a computerised task to assess the capacity of participants to resist distractions from the task. An avoidant attachment style was associated with greater success at the task, except when participants had been primed to think about a past occasion in which they were made to feel insecure.379

The potential for insecure attachment styles to confer local advantages has been increasingly recognised since 2007. Nonetheless, Mikulincer and Shaver’s 2016 book Attachment (p. 511) in Adulthood, in part for narrative reasons and in part because it is surveying the existing literature, at times approaches a list of goods associated with the secure attachment style and a list of bads associated with insecure attachment styles.380 Attention to the circumstances in which insecure attachment styles can confer benefits has instead been the priority of one of Mikulincer’s former students, Tsachi Ein-Dor. Ein-Dor took a course with Mikulincer on attachment theory as an undergraduate student. He was struck by an apparent contradiction. On the one hand, the ethological-evolutionary basis of attachment theory suggested that behavioural repetoires common across a species likely have some survival or reproductive value in an expectable environmental niche. On the other hand, Mikulincer’s lectures emphasised the disadvantages of insecurity.

Ein-Dor served as the operations manager for Mikulincer’s research group between 2002 and 2009, undertaking masters’ and doctoral study.381 When Mikulincer moved from Bar-Ilan to become founding Dean of the School of Psychology at the Interdisciplinary Center Herzlyia in 2007, Ein-Dor followed to help set up the new laboratory. During this time, Ein-Dor paid close and critical attention to the data being gathered by Mikulincer’s group. Whilst the overarching story told in the published papers was about the virtues of a secure attachment style, he noticed an accumulation of research findings suggesting the benefits of insecure attachment styles under specific circumstances. Some of these findings might be due to chance. However, Ein-Dor began to see a logic to the findings, and developed hypotheses regarding the conditions under which an insecure attachment style might be an asset. He came to regard an anxious attachment style as a potential asset in contexts in which vigilance would be rewarded, and an avoidant attachment style as a potential asset in contexts in which instrumentalism would be rewarded. Though the co-author on many earlier papers published by Mikulincer’s research group, Ein-Dor’s first paper as lead author was published in 2010 entitled ‘The attachment paradox: how can so many of us (the insecure ones) have no adaptive advantages?’, with Mikulincer and Shaver both as co-authors.382

In this paper, Ein-Dor and colleagues acknowledged that Mikulincer and Shaver had time and again reported results indicating the benefits of a secure attachment style. Yet they also pointed to theoretical work suggesting that insecure attachment can be adaptive, not only in the sense of responsive to the environment but also in the sense of conferring some advantages. This position, already under discussion by Ainsworth in the mid-1980s (Chapter 2), was developed in print by researchers in the developmental tradition such as Belsky and Crittenden.383 Belsky’s emphasis on the contribution of attachment strategies to reproductive fitness was especially influential for Ein-Dor.384 As well as this theoretical tradition, (p. 512) Ein-Dor and colleagues could also point to the accumulation of studies showing some benefits of the insecure attachment style. In addition to earlier findings from Mikulincer’s group, Ein-Dor and colleagues were able to report new findings from their longitudinal follow-up of ex-prisoners of war from the 1973 Yom Kippur war. Participants were followed up in 1991, 2003, and 2008. This research found that veterans’ avoidant attachment scores were inversely associated with the extent to which their wives showed symptoms of trauma, such as intrusion or hyperarousal.385 The avoidant attachment strategy of the ex-prisoners of war appeared to have kept their spouse safe from contamination by their symptoms of trauma. However, wives’ avoidant attachment style was positively associated with veterans’ PTSD symptoms, suggesting that, when it was not a matter of their own choosing, reduced opportunity to seek intimacy and share emotional experiences with their partner was harmful for the veterans.

Ein-Dor and colleagues criticised the wider field of attachment research generally on two grounds. First, following Belsky, they argued that researchers have generally designed studies based on an implicit model of mental health as behaviour that is adaptive in industrialised society, rather than developing hypotheses considering how apparently maladaptive behaviours might have their niche within human evolutionary history and/or particular contemporary contexts. Second, they criticised the focus of attachment researchers on individuals and, sometimes, dyads. Few attachment researchers besides the Minnesota group (Chapter 4) have examined the contribution of individual differences in attachment to group-level processes. Yet contemporary evolutionary biology, in contrast to the evolutionary theory of Bowlby’s day, has come to place greater emphasis on group processes within natural selection. For instance, the survival of the group as a whole may benefit if there are some members who are more wary and alert to threat, some more focused on instrumental concerns, and others more capable of coordination, negotiation, and compromise.

Ein-Dor, Mikulincer, and Shaver published a first empirical article based on this agenda in 2011:386 46 groups of participants were observed in the laboratory room as it gradually filled with smoke from an apparently malfunctioning computer. The results were in line with Ein-Dor’s theory of the potential benefits of insecure attachment styles under particular circumstances. Individuals high on attachment anxiety detected the smoke more quickly and alerted the group. This led the group as a whole to a faster response to the threat, since they had been notified of it earlier. Effects were quite marked: a 1-point increase in attachment anxiety was associated with an 11.5-second decrease in detection time. Individuals high on attachment avoidance were faster at getting out the door once the danger was detected, and contributed to a faster exit for the group as a whole. Furthermore, Ein-Dor and colleagues found a linear association between diversity of attachment scores in the group and the group’s effectiveness at evacuating the room. This implied that although security might not be associated with vigilance regarding threat or directness in avoiding the threat, it was associated with the holistic effectiveness of the group, perhaps by facilitating coordination. They found that the effects remained significant even with temperament measures of extraversion and neuroticism statistically controlled. As a conceptual replication and extension (p. 513) of the finding that groups benefited from heterogeneity of attachment styles, Ein-Dor and colleagues assessed the attachment styles of students at their university enrolled in courses including a team project. The researchers found that when reported team cohesion was high, heterogeneity in attachment anxiety and avoidance scores in the group was associated with better grades on the group project. However, only heterogeneity in anxiety was associated with better perceived group functioning. As such, heterogeneity of attachment avoidance was associated with better performance evaluations, but without individuals being aware of the increased effectiveness of their team.387

Ein-Dor and Orgad conducted another study. They led participants to believe that they had accidently activated a computer virus that wiped the experimenter’s computer. They were then asked to alert the department’s computer technicians. On the way to the technicians’ office, they were presented with four decision-points where they could either choose to delay the warning or continue on to their destination. Only the anxious attachment style, and not neuroticism, was associated with continuing on to deliver the warning rather than responding to the distractions.388 Ein-Dor and Orgad concluded that whereas anxious attachment and trait anxiety might both influence threat perception, the communication of concerns to someone who should help is a priority specific to attachment anxiety.

Such findings provided support for Ein-Dor’s claim that there can be beneficial correlates of heterogeneity of attachment styles for groups, since there are specific situations in which an insecure attachment style is a local advantage. The studies had been designed to offer a certain analogy to human evolutionary history and the question of group survival. However, Ein-Dor felt that he could prove his claims also on the ‘home terrain’ of attachment research in the study of individual differences, and even in the study of contemporary professional life. Ein-Dor, Reizer, Shaver, and Dotan proposed that individuals with avoidant attachment styles will profit in professional fields that reward self-reliance and the ability to work without social support.389 The researchers elegantly demonstrated this with a study of professional singles tennis players: 58 players completed the ECR, and their progress in the national tennis rankings was assessed over a 16-month period. Amount of training, feelings of self-efficacy, and avoidant attachment style all predicted change in ranking movement, accounting for 13.5% of variance.390 However, in a regression, only avoidant attachment style was significant, perhaps suggesting that amount of training was in part a function of avoidance. In another study, Ein-Dor and colleagues were able to demonstrate the potential benefits of an anxious attachment style for individuals. Researching a card game tournament, they found that the higher the player’s attachment anxiety score, the better their ability to cheat without being caught and to detect others cheating.391 Research findings from other groups have offered support for Ein-Dor’s claims regarding the potential advantages of insecure attachment styles under particular circumstances. Of special note are findings from (p. 514) a randomised control trial of the Circle of Security eight-session parenting intervention. Cassidy and colleagues reported that this intervention had no main effect on either infant attachment security or on child mental health. However, positive effects were seen when caregivers were high in attachment avoidance; by contrast, the intervention reduced infant–caregiver security for caregivers who began the intervention low in attachment avoidance.392

Evaluating Ein-Dor’s position, Gillath, Karantzas, and Fraley have argued for only qualified acceptance. They acknowledge that in specific circumstances, insecure attachment styles may have local advatanges. However, they emphasise that the vast majority of studies have found that secure attachment is associated with more positively regarded outcomes.393 Yet Shaver and Mikulincer have acknowledged Ein-Dor’s point that this has partly been a result of how studies have been designed and conceptualised. They have accepted that Ein-Dor’s work has provided a helpful corrective to the tendency in their writing to treat insecure attachment as, in itself, simply bad. Furthermore, Ein-Dor’s concern with the relationship between individual attachment style and group processes has been an important development for attachment research with adults. To date, there have not been sufficient studies of small group processes to build the critical mass for a research agenda, and research energies have been directed elsewhere.394 However, interest in small groups has definitely been growing in social psychology,395 and the developmental tradition has developed attachment-based parenting interventions that specifically make use of group dynamics such as GABI (Chapter 4). It may be that Ein-Dor’s ongoing research efforts will contribute to small-group research as an important new direction for attachment research.

Some remaining questions

The ECR items

For two decades, the social psychology tradition of attachment research has been underpinned both theoretically and methodologically by the ECR. As discussed in the section (p. 515) ‘Creating the ECR’, Brennan and colleagues developed the ECR through a two-level factor analysis of items from existing measures. The model of two orthogonal factors that came out of this analysis comprised items representing attachment anxiety and attachment avoidance. And the idea of orthogonal dimensions in a two-dimensional space, representing minimising and maximising, became the foundation for Shaver and Mikulincer’s approach to all other behavioural systems. Yet there has been astonishingly little discussion of the ECR items and the inner machinery of the measure, with researchers seeming to rest comfortably on the original factor analysis conducted by Brennan and colleagues. An exception is Allen and colleagues, who have called for a close study of the ECR items, but not conducted this study themselves.396 Likewise, Banai, Mikulincer, and Shaver, in a paper from 2005, urged that distinctions may be drawn between forms of attachment anxiety, and that ‘researchers should attempt to distinguish among these potentially different kinds of individuals who score high on the anxiety dimension’.397 However, these calls have not been followed up. Social psychology research on attachment seems to have been firmly gripped in the beak of the two dimensions the ECR is understood to embody, with the items themselves treated as of little consequence. Yet examination of the items makes this assumption all the more strange, since the items—drawn as they are from a variety of pre-existing scales—are quite a menagerie. This was acknowledged by Mikulincer and Shaver: ‘the items consequently range from ones concerned with relationships in general to ones concerned with a particular partner. Some deal with “comfort” and other feelings; some deal with desires and motives.’398 They expressed surprise at how well the items have performed, ‘given their relative crudeness’. In fact, ‘it is remarkable how systematic and cumulative our research findings have been’ on the basis of items that were ‘not designed component-by-component with a coherent theoretical model in mind’.399

The ECR is 36 items, which is unusually long for a self-report measure. This has led to attempts to produce a shortened version of the scale with the same properties. Work on translations and the ECR and attempts to produce a shortened version have represented the only sustained conversation in the published literature about the ECR items and the latent phenomena they measure, and as such the only fine-grained consideration of what participants may be endorsing. In a 2007 paper, Wei and colleagues presented the first attempt to develop a short version of the ECR, drawing 12 items from the 36.400 They cited a personal communication from Shaver (July 2004) that the avoidant and anxious attachment styles each have three ‘critical components’. According to Shaver, the ‘critical components’ of attachment avoidance are:

  1. i) concern about closeness;

  2. ii) reluctance to depend on others;

  3. iii) reluctance to self-disclose.

(p. 516) The critical components of attachment anxiety are:

  1. i) concern about abandonment;

  2. ii) an extensive desire for reassurance from others;401

  3. iii) distress about the unavailability of one’s partner.

Whilst these often occur together or cause one another, the three components of each of the forms of insecure attachment are articulated. It is possible for them to occur unalloyed. For instance, reluctance to depend on others may often come with a concern about closeness, but it need not do so. Attachment figures could be regarded, for instance, as generally untrustworthy but nonetheless attractive. Or again, distress about the availability of one’s partner may entail concern about abandonment, but it need not do so. Attachment figures could be regarded as distracted and inattentive, but still invested in the relationship and its continuation.

Wei and colleagues developed the ECR-short, drawing items from the ECR with reference to the theoretical rationale of the three ‘critical components’ per form of insecure attachment. However, in an independent attempt to make a short form of the ECR, Lo, Mikulincer, and colleagues argued that there was a need to reconsider the Brennan study itself. They observed that the Brennan et al. paper is usually described as creating the two dimensions of the ECR out of the items of a factor analysis. But in fact this was a two-stage process. The data presented by Brennan and colleagues indicated that when factored at the item level, a higher-order factor structure was present. Twelve first-order factors were initially extracted, which Brennan and colleagues termed ‘facets’ of the two latent dimensions found using the higher-order factor analysis.402 The 36 items did not reduce in one step to two latent dimensions; an intermediate layer was in operation. Lo, Mikulincer, and colleagues conducted an exploratory factor analysis of the 36 items and found that avoidance and anxiety each broke down into two factors. Within the avoidant attachment style, one of the factors was discomfort with closeness. This factor accounted for 16% of variance in endorsement of items on the ECR. The other factor was lack of willingness to rely on others or willingness to disclose to others (which clustered two of the ‘critical components’ separated by Wei and colleagues on the basis of Shaver’s personal communication). This factor accounted for 14% of variance. The distinction between these two factors has been supported in other studies.403

The anxious attachment style also broke down into two factors. The first factor represented items indicating frustration about the unavailability of the attachment figure. This (p. 517) is similar to the ‘critical component’ of distress about the unavailability of the partner from Wei and colleagues’ formulation, except for the focus on frustration. As we saw in the section ‘Creating the ECR’, Shaver and colleagues shifted the second dimension of attachment theory away from Ainsworth’s concern with inconsolability (understood as anxiety about caregiver availability), anger, and passivity towards a focus solely on anxiety about attachment relationships. However, the fact that one of the constitutive factors of the anxious attachment style constitutes frustration suggests that the ECR has continued to tap anger about the unavailability of the attachment figure, though with only distress/worry about partner availability featuring focally and theoretically in accounts of what is being measured or in the design of research studies. The ‘frustration’ factor accounted for 18% of variance, the largest share of any of the factors. The other factor constituting the anxious attachment style represented items indicating worry about the relationship. This seems to be similar to the concern about abandonment from Wei and colleagues’ formulation. This factor accounted for 9% of variance.404 There was no equivalent in Lo, Mikulincer, and colleagues’ work of the wish for reassurance, the final critical component for Wei and colleagues. However, there is only one item directly about the wish for reassurance in the ECR: ‘I need a lot of reassurance that close relationship partners really care about me’ (Item 18).405 This item is loaded with frustration in the factor analysis, though not strongly. Here again, there appears to be some holdover from Ainsworth’s original model, since the wish for reassurance without active efforts to achieve it was how Ainsworth defined the C2 ‘passive’ classification.

Overall, close consideration of the ECR suggests that even if the items can be considered to ultimately tap two latent dimensions, they do so through an intermediate layer of ‘facets’ in which relevant differences can be identified.406 Indeed, some studies do not find a two-factor model as the optional solution, and instead identify these facets as distinct factors in accounting for variance for endorsement of items on the ECR.407 Part of the issue is that the items that constitute the ECR were not designed for the purpose of tapping the two latent (p. 518) dimensions, and item-total correlations for the anxiety and avoidance scales are frequently low.408 Differences in the ‘facets’ that constitute the intermediate layer may contribute to discrepancies between theory and measurement, reducing the precision and coherence of the measure and the tradition of research built upon it. For instance, attachment anxiety is generally assumed by Shaver and Mikulincer to be correlated with, but distinct from, frustration in relationships. However, one of the two components of the anxiety scale is tapping frustration whereas the other is not. Is frustration a correlate or a component of an anxious attachment style? Wongpakaran and colleagues have argued that in fact frustration, though related, is ultimately ‘extraneous’ to what is truly meant by attachment anxiety, and its inclusion within the ECR and self-report measures based on it may be a cause of unintended noise and imprecision.409 They also expressed concern that some ECR items may simply reflect poor self-regard rather than an attachment-specific experience, even if this loads with attachment anxiety in factor analyses. And indeed, Esbjørn and colleagues found a five-factor solution was superior to a two-factor solution in their data, and that the ‘extra’ factor beyond the four identified by Lo, Mikulincer, and colleagues represented those items in the ECR that suggest poor global self-worth (e.g. Item 34: ‘When others disapprove of me, I feel really bad about myself’).410 Karantzas and colleagues have argued for the ‘conceptual and empirical importance of including both broad factors and specific facets of attachment style’, and alleged that the constructs of attachment anxiety and avoidance are decidedly ‘blunt instruments’. They contend that consideration of the facets will have much to offer to work in clinical and therapeutic contexts, for instance in understanding how exactly attachment insecurity may contribute to mental health symptoms.411

Another potential issue with the anxiety items is that a single item (Item 18) suggestive of Ainsworth’s concept of passivity appears to have smuggled aboard the ECR, but without a second item to stabilise measurement and without theoretical acknowledgement. Yet passivity may also be implicated elsewhere in the measure. Item 29 is the statement ‘I feel comfortable depending on others’; this item, reversed, was intended to represent avoidance. However, it may also represent the passive desire for care Ainsworth characterised as C2. In this light it is notable that several samples, including work by Shaver and colleagues themselves, have found Item 29 negatively associated with avoidant attachment and positively associated with anxious attachment, rather than unrelated to anxiety as anticipated by Brennan and colleagues.412

(p. 519) In relation to the avoidance dimension, there seem to be distinct facets representing relative comfort with closeness on the one hand, and trust on the other hand. Again, these differences in ‘facets’ in the intermediate layer may introduce discrepancies between theory and measurement, especially in contexts where one or the other facets may be more important. It is easy to imagine social contexts in which trust is more salient, contexts in which comfort with closeness is more important, as well as contexts in which both play a substantial role. In a higher-order factor analysis, Lo, Mikulincer, and colleagues found that whereas willingness to disclose and rely on others was entirely distinct from attachment anxiety, the items representing the facet of discomfort with closeness loaded partially on both attachment avoidance and attachment anxiety.413 This implies that discomfort with closeness may be fed by both attachment avoidance and attachment anxiety, albeit for different reasons.414 Consider, for instance, Item 26 on the ECR: ‘I find that my partners don’t want to get as close as I would like’. With what is ‘liked’ unspecified, endorsement of this item may be fed by both anxiety and avoidance, as a factor analysis conducted by Shaver and colleagues indeed showed.415

Lafontaine, Shaver, and colleagues conducted a study to further refine the items for a short version of the ECR. One analysis they pursued was the extent to which items could discriminate particular portions of the avoidance scale. At the high end of the avoidance scale, some items that were particularly effective were Item 29: ‘I feel comfortable depending on others’ (Reversed) and Item 9: ‘I don’t feel comfortable opening up to others’. However, ‘almost all avoidance items performed relatively poorly on the lower portion of the avoidance scale continuum’.416 The items were quite good at picking up the avoidant attachment style, but were poor at discriminating security. There was only one item that worked adequately, though even then not well, in picking out security (Item 35). Yet this item had the highest standard error and lowest discrimination of all the avoidance items. On the anxiety scale, again, the low end was not well discriminated, perhaps because there is only one reverse item in the anxiety scale.417 Whereas the ECR promises to capture security as the absence of anxiety (p. 520) and avoidance, such findings raise the prospect that there are aspects of secure attachment that are not well captured or discriminated by the existing scales, or by theory shaped in the image of these scales. Already in 2000, Fraley and colleagues had called on researchers to ‘write items that tap the low ends of the Anxiety and Avoidance dimensions with better precision’.418 However, two decades later this problem with the ECR remains little discussed. One of the issues that this neglects is whether security really is merely the inverse of anxiety and avoidance, or itself makes an independent contribution to individual differences in adult attachment style. Given trends in the developmental tradition towards Individual Participant Meta-analysis (Chapter 6) in order to address psychometric questions, it will be interesting to see whether the social psychological tradition will also adopt such methodologies for further articulating the relationship between the ECR and its latent dimensions.

The opposite of insecurity

In their landmark 1998 paper, Brennan and colleagues reported a two-factor solution as the fundamental structure of individual differences in adult attachment. In the years after the introduction of the ECR, the two-factor model quickly became accepted. This marvellously elegant characterisation of individual differences in adult attachment style subsequently became orthodoxy among social psychological attachment researchers, with theory and methodology looping around one another to direct research questions and interpretations of findings. Yet leading figures of the social psychological tradition of research acknowledged that the model is a pragmatic simplification, and that in reality security might represent more than the absence of avoidance and resistance. Judith Feeney, for instance, writing four years after the introduction of the ECR, stated:

I would like to take issue with the implicit suggestion that in defining the two attachment dimensions of avoidance and anxiety, researchers have settled basic questions concerning the structure of self-reported attachment. There now seems to be considerable consensus that avoidance and anxiety are the two primary dimensions underlying adult attachment, and that these dimensions generally provide moderately strong prediction of relationship outcomes (especially if both partners’ characteristics are taken into account). However, in reducing such a complex construct as romantic attachment to two dimensions, important information is inevitably lost.419

This had already been a concern of Brennan and Shaver themselves. In a 1995 paper, Brennan and Shaver first offered the proposal that ‘secure adults can be characterized as the opposite of all these insecure tendencies’.420 This was aligned with the ‘love quiz’ which had, to an extent, characterised security in terms of the absence of discomfort regarding closeness or the (p. 521) absence of worry about abandonment. Characterisations of security as no more than the opposite of insecurity would remain the dominant narrative in Shaver’s work in subsequent years. Yet already in the 1995 paper, Brennan and Shaver offered a qualification: that secure attachment is not solely the opposite of insecurity, but also has positive characteristics of its own such as ‘being able and willing to trust romantic partners and share ideas and feelings with them in a flexible, appropriate manner that is sensitive to their partners’ needs and concerns’.421 This idea was not abandoned after 1998 and the introduction of the ECR. For instance, in the Shaver and Mikulincer model of the attachment system developed in the early 2000s (discussed in the section ‘Minimising and maximising’), it would appear that broaden-and-build cycles are facilitated by specific and vital qualities of security, and not merely through the absence of their interruption by the insecure attachment styles. For instance, broaden-and-build cycles were anticipated by Shaver and Mikulincer to contribute to confidence and trust, which was not merely the absence of avoidance or attachment anxiety.422 They argued that feeling understood, feeling validated, and feeling cared about are distinct aspects of the broaden-and-build cycle characteristic of security.423

In 2004, Shaver and Mikulincer reported their impression that by this point ‘most recent adult attachment studies are based on a two-dimensional model’, theoretically and in terms of measurement.424 They acknowledged that some studies, including their own work using the ‘love quiz’,425 and research by Rainer Banse426 had produced evidence that security represented a distinct construct not reducible to the absence of anxiety and avoidance.427 However, they argued that this position was compatible with their theory, since a security-insecurity dimension represented a 45-degree turn within the two-dimensional space of anxiety and avoidance.428 So long as the ECR could tap the low end of anxiety and avoidance effectively, the contribution of security to individual differences could be captured by the measure. Colleagues such as Fraley had criticised the capacity of the ECR to capture security (p. 522) effectively. In 2004, Shaver and Mikulincer appear to have regarded the problem as meaningful, but not of great pragmatic or theoretical significance.

Yet later work from their laboratories has continued to trouble the two-dimensional model. In 2006, Al-Yagon and Mikulincer published a study of children’s experiences of loneliness. Since at that time no version of the ECR had been validated for use with children, they instead used an adaptation of the ‘love quiz’. The researchers found that secure attachment made a negative contribution to loneliness over and above the positive contribution to loneliness of the avoidant and anxious attachment styles.429 Further evidence for a three-factor model came from work by Omri Gillath. Gillath had been a graduate student with Mikulincer between 1998 and 2003, and then a postdoctoral fellow with Shaver from 2003 to 2006 at the University of California. Following his appointment as faculty at the University of Kansas, Gillath worked with Joshua Hart and colleagues to develop a ‘state adult attachment measure’ (SAAM) to assess a participant’s current feelings rather than enduring generalised schemas about relationships (see also the section ‘Conclusion’ for a discussion of security priming).430 The purpose of the SAAM was to allow researchers to explore the extent to which particular stimuli or circumstances, for example changes in partner interactions over a week, might contribute to attachment-relevant experience. The items for the SAAM were based on the ECR. However, Gillath, Hart, and colleagues ‘also wrote additional items to reflect aspects of attachment styles that are underrepresented on current measures, such as the low end of anxiety (e.g., “I feel relaxed knowing that close others are there for me right now”), which is represented by only a single item in the ECR’.431 They conducted seven studies with the SAAM, with a total of 2,327 participants. Across the seven studies, they repeatedly found that a three-factor model was the best fit for the data, with independent factors for security, avoidance, and anxiety. They concluded that security appears to be an autonomous dimension of attachment states.432

Gillath and colleagues emphasised that Shaver and Mikulincer’s theory gives a place to dynamics specific to security, even if these are not captured well by the ECR. However, Gillath, Karantzas, and Fraley have urged that more needs to be done within the social psychological tradition of attachment research to capture security-specific processes within theory and methodology. They are unconvinced that these processes are merely the opposite of anxiety and avoidance, and anticipate that they will have distinct correlates. They argue that greater (p. 523) understanding of the secure attachment style will contribute to insights into both clinical phenomena and experiences of human thriving. For example, it will help clarify whether traumatic experiences ‘either increase attachment insecurity or wear away at attachment security’.433 In a review of self-report measures in 2014, Shaver and Mikulincer acknowledged the problems with the ECR for capturing security, and praised the work of Gillath and colleagues. They argued that ‘the inclusion of a separate security subscale may suggest a way out of the problem identified by Fraley et al. (2000) … that most previous attachment insecurity scales discriminated poorly at their “secure” ends’.434 They noted, however, that when the low end of avoidance and anxiety are captured, a result is that the two forms of insecurity are no longer orthogonal, as they are both negatively associated with security: ‘expanding the scales at their secure ends in similar ways causes the two kinds of security items to correlate with each other, which in turn makes the scales as wholes correlate more with each other. Whether or not this leads to weaker detection of distinct effects of anxiety and avoidance remains unclear.’435

High anxiety/high avoidance

Yet even for the unmodified ECR, questions have been growing about its psychometric properties. Since Shaver and Mikulincer’s theory of adult attachment has been based on the two dimensions of the ECR, this is not simply a minor matter of methodological rigour but a concern stretching to the very basis for their scientific project. Whereas the original factor analysis conducted by Brennan and colleagues accounted for 63% of variance, subsequent studies have not accounted for such a high proportion of variance. Factor analytic exploration of the ECR after Brennan and colleagues has been rare among American researchers, who have generally taken its psychometric properties for granted. However, in the wider international literature, factor analytic studies of the ECR or its translation tend to report solutions that account for around 45% of variance.436

Furthermore, there has been growing evidence against the orthogonality of the two ECR dimensions. In a review of several of their studies in 2005, Mikulincer, Shaver, and colleagues reported a small association between the scales (r = .18).437 In a paper from the next year, Condradi and colleagues conducted a further analysis of the published literature, observing (p. 524) that ‘intercorrelations vary considerably from .04 to .30’.438 The developmental and relationship tasks of students as a population may well have some differences from populations in other countries and at different lifestages.439 Condradi and colleagues observed that these differences are evidently not sufficient to block the successful application of the ECR to diverse research populations, as years of research has demonstrated. However, the characteristics of the population used to develop the sample may have nonetheless introduced unrecognised assumptions into the measure, denting the reliability and validity of its scales. Specifically, Condradi and colleagues were concerned by Brennan and colleagues’ conclusion that attachment anxiety and attachment avoidance were, and should be, orthogonal dimensions. This seemed to be a holdover from Bartholomew’s model of orthogonal dimensions, rather than based on a cogent theoretical justification. They expressed concern that there was no basis in Bowlby’s theory for assuming that the two dimensions would be unrelated. Indeed, researchers in the developmental tradition have neither expected nor found orthogonality between avoidance and resistance.440

Condradi and colleagues sought to empirically examine whether orthogonality might be an effect of the student sample used by Brennan and colleagues. A two-factor model of orthogonal dimensions was indeed the best fit for the American and Dutch student samples.441 However, the result was less clear for the Dutch sample of adults. Condradi and colleagues argued that anxiety and avoidance may indeed be unrelated for participants without experience in lasting relationships. In more mature relationships they predicted that anxiety and avoidance would no longer be unrelated. Both dimensions would be negatively associated with the security that a long-established and healthy relationship can build. And both dimensions would be positively associated with the anxiety and avoidance that can become entwined within a long-established and unhealthy relationship.442

These proposals were supported by two meta-analyses conducted by Cameron and colleagues and by Graham and Unterschute.443 The meta-analyses found that orthogonality between anxiety and avoidance was the exception rather than the rule in the published literature. Furthermore, older samples had larger correlations between anxiety and avoidance than younger samples, long-term relationships had larger correlations than newly established ones, non-student samples had larger correlations than college samples, and non-American (p. 525) samples had larger correlations than American samples. Cameron and colleagues argued that ‘such findings call into question the largely implicit assumption that attachment dimensions are orthogonal. As such, our results may inspire researchers to revisit theoretical assumptions, measurement choices, and measure creation techniques. If researchers do not wish to seek other measurement options, they should at the very least adapt their statistical analyses to accommodate shared variance between dimensions … One method of accommodating shared variance is to include both dimensions as predictors in the same step in a regression, and thus control for shared variance.’444 The proposal to accommodate or explore shared variance has not, as yet, been pursued. As a result, it remains unknown whether the association between the anxiety and avoidance scales in the ECR lies on the basis of their joint negative association with security or on the basis of their joint positive association with a state of high anxiety/high avoidance.

In recent work with Birnbaum, Mikuliner and Shaver acknowledged both possibilities.445 They also offered a thought-provoking speculation that the tendency towards orthogonality of the ECR dimensions might actually be an effect of two processes, invisible in the data because they have suppressed one another. On the one hand, anxiety and avoidance may function as exact opposites: Shaver and colleagues have at times argued that each is the inverse of the other in terms of psychological processes.446 To the extent that they represent maximisation and minimisation of the attachment system, there can be anticipated to be a negative association between the two dimensions. However, this negative association may be counteracted by a positive association between the scales at their endpoints. They may have a positive association in their mutual opposition to security. Furthermore, ‘both hyperactivation and deactivation … represent problems in the system’s functioning, which may push their correlation in a positive direction’.447 The result of both positive and negative associations may be an apparent tendency towards orthogonality.

Whereas Bartholomew gave fearful attachment a major place in her system, the conjunction of high anxiety and high avoidance has not been a major concern of Shaver and Mikulincer over their careers. In part, this is likely due to the fact that there are few high anxiety/high avoidance participants in Shaver and Mikulincer’s samples. One essential reason for this, as they have acknowledged, has been their sampling: ‘The issue of “fearful avoidance” is, in any case, less likely to arise in normal samples of college students and community adults. Extremely high scores on both the anxiety and avoidance dimensions are more common in samples of abused or clinical samples … In most of our studies, the results can be (p. 526) adequately described in terms of either anxiety or avoidance … with the distinction between dismissing and fearful avoidance mattering only … when abuse or psychopathology are at issue.’448 For instance, they cited as evidence a published study by Fraley and Bonanno and unpublished data from Colin Murray Parkes showing that ‘combinations of attachment anxiety and avoidance produced the highest levels of anxiety, depression, grief, trauma-related symptoms, and alcohol consumption’.449

Shaver and Mikulincer’s overriding use of student samples has been a pragmatic one. They have readily admitted that it supplies a limitation to their work. For instance, they have noticed that it can be difficult to recruit student participants with an avoidant attachment style when adverts are frank with participants that the study is concerned with close relationships.450 And Makariev and Shaver have acknowledged that middle-class college populations were not ‘the kinds of people Bowlby had in mind when he began to develop attachment theory’.451 Nonetheless, Shaver and Mikulincer have defended their use of student samples on two grounds. In a first defence, they have argued that adversities and psychopathology is dimensionally distributed, and so there is no particular problem with using a student sample—so long as clinical variance is captured.452 This defence is relatively weak. It is not clear that clinical variance is generally well captured by the measures Shaver and Mikulincer tend to use; in this context, it is perhaps unsurprising that studies of adult attachment styles and PTSD have tended not to use the ECR.453 Furthermore, the idea that adversities and psychopathology can be treated as dimentional in terms of their implications for attachment style is simply asserted, rather than adequately empirically demonstrated in their work.

Shaver and Mikulincer have also offered a second defence of their approach in describing themselves as pragmatists, seeing to achieve cogent simplifications whilst remaining aware of what has been simplified:

In our work, we always try to keep in mind that there is, on the one hand, complex everyday reality as we experience it subjectively and encounter it in the behavior of other people. And, on the other hand, there is psychological theory, with its associated hypothetical constructs, and an ever-evolving toolbox of psychological measures. The trick is to discover and document something important and valid about real life, thereby nudging (p. 527) psychological science forward, without mistaking our tentative, overly simplified picture for everything that is actually there.454

However, the inattention to high anxiety/high avoidance attachment styles is more than a happenstance of sampling. Instead, the sampling strategy appears to have reflected a broader lack of concern with adverse and clinical experiences, and states where both anxiety and avoidance are in play.

Even before the introduction of the ECR, Shaver and Mikulincer demonstrated a marked lack of interest in what it meant when participants endorsed both an anxious and avoidant attachment style. In his work in the early 1990s, Mikulincer reformulated the ‘love quiz’ as three scales to produce dimensional ratings of attachment style.455 In a validation study of the scales with 127 undergraduates, nine participants scored differently on the scales compared to the original category-based ‘love quiz’. Mikulincer dropped these participants from the analysis. This approach set a precedent. Cooper, Shaver, and Collins used both the category-based and the dimension-based measures in their 1998 paper ‘Attachment styles, emotion regulation, and adjustment in adolescence’.456 In the study, 411 participants—20% of the sample—were inconsistent across the two measures. In a footnote, the authors acknowledged that this inconsistent group had some interesting properties. They were more likely to be non-White, more likely to endorse an insecure attachment style, and more likely to have been held back in school. Furthermore, ‘inconsistent responders reported higher levels of phobic anxiety, paranoia, and psychoticism and were less satisfied with their body image’.457 Though this ‘inconsistent’ group represented hundreds of participants, they were thrown out of the analysis.458

Considering this study alone, the decision makes methodological sense. However, in the broader context of Shaver’s work, it illustrates a trend to exclude the effects of adversity, stigma, pharmacology, and clinical complexity from the study of adult attachment styles. The most influential instance of this trend was in work with Brennan and Clark in developing the ECR. In their 1998 paper, the factor analysis was conducted on an undergraduate sample from the University of Texas. However, no report was made or analysis conducted for demographic or ethnic diversity within the sample. And no data were collected on clinical differences. Across their published studies, Shaver and Mikulincer have rarely discussed the implications of both high anxiety and avoidance, even in studies where they included measures of mental health.459 There seems to have been a self-perpetuating cycle between lack of interest and lack of scientific basis for interest in high anxiety/high avoidance phenomena in existing theory and research. This has shaped the tone and priorities of social psychological (p. 528) research on attachment, hindering dialogue with the developmental tradition where theorising and measuring the implications of trauma and fear for attachment have been a major focus (Chapter 3).

Despite the general trend, there have been a few occasions on which Shaver and Mikulincer have reported associations for high anxiety/high avoidance attachment.460 One was a study by Schachner and Shaver of sexual motives, which found that high anxiety/high avoidance—over and above the anxious attachment style alone—was associated with greater report of having sex due to insecurity.461 Hart, Shaver, and Goldenberg found that high anxiety/high avoidance was distinctively associated with a desire for closeness after receiving information that was anticipated to threaten the participants’ cultural worldview, but not in the control condition. This was in contrast to the anxious attachment style which was associated with a desire for closeness in both conditions.462 The researchers concluded that the avoidant strategy held high anxiety/high avoidance participants back from seeking support in the control condition, but that the strategy was overwhelmed by the combination of high attachment anxiety and a worldview threat, leading to behaviour shaped by the anxious strategy.463

The effects of trauma

A major qualification must be offered, however, to any allegation that inattention to attachment phenomena characterised by high anxiety and high avoidance is a limitation in the work of Shaver and Mikulincer.464 The qualification lies in Mikulincer’s two decades of research on the effects of war and military trauma. However, this research has not always been well linked-up with Mikulincer’s theoretical work on adult attachment. The relationship between PTSD and the insecure attachment styles in Mikulincer’s work has remained untamed, especially in terms of what this relationship means for the conceptualisation of high anxiety/high avoidance states and of adult attachment in general. This issue appears to be coming to a head in recent years, given the very sharp increase in studies of PTSD and attachment styles in the past decade. However, refinements in theory and methodology appear generally not to have caught up with this empirical concern.

(p. 529) In the late 1980s, Solomon, Mikulincer, and colleagues examined Israeli army veterans who had experienced symptoms of post-traumatic combat stress during the 1982 Lebanon war. On the one hand, the researchers observed that one of the symptoms associated with this condition was ‘psychic numbing’, which was associated with low emotional expressiveness in the veterans’ adult relationships with family members. The researchers accepted that ‘there is little way of knowing whether it is the family that did not permit the veteran to express himself or the veteran who did not make use of the avenues of expression open to him’.465 However, they expressed their suspicion that ‘many PTSD casualties who report a low level of expressiveness may themselves have avoided discussion’.466 On the other hand, Mikulincer, Solomon, and colleagues also found that, among veterans who had experienced symptoms of post-traumatic combat stress, ‘the anxiety feelings aroused during battle by the fear of death are crystalized one year later in anxiety that is not specifically related to death, but is generalized to every potential threat’.467 This would, presumably, include relational threats. These early observations suggest, though this point was not drawn out by the researchers, that general forms of avoidance and anxiety stemming from trauma may feed more specific forms of avoidance and anxiety within adult close relationships.

Mikulincer, Solomon, and Benbenishty also conducted clinical interviews with a sample of 104 veterans with histories of post-traumatic combat stress symptoms a year after the end of the Lebanon war. From these interviews, they identified 26 different manifestations of post-traumatic combat stress and nine battle events that seemed relevant to the emergence or maintenance of symptoms. Analysis of the manifestations resulted in six factors, accounting for 62% of variance. The six factors were: psychic numbing, anxiety reactions, guilt, loneliness, loss of bodily control (such as uncontrollable crying or vomiting), and disorientation. It is interesting, if perhaps not surprising, how well these agree with Bowlby’s own clinical observations of war veterans during his time as an army psychiatrist (Chapter 1). In the work of Mikulincer and colleagues, psychic numbing and anxiety reactions appeared to be the most important, between them accounting for 31% of variance in symptoms reported in the clinical interview. They explicitly argued that psychic numbing should be conceptualised as a kind of ‘avoidance’ and considered as quite an all-purpose response, since it was not associated with any particular kind of battle event. By contrast, anxiety reactions were predicted best by poor unit functioning. Mikulicer and colleagues argued that psychic numbing and anxiety reactions seemed to correspond well, respectively, to ‘avoidance and numbing’ and ‘hyperarousal’ as two of the main clusters of general PTSD symptoms. They therefore suggested that post-traumatic combat stress symptoms should not be regarded solely as localised responses to particular war experiences, but as reflecting the basic forms of the human trauma response.468

Pursuing these questions further, in the mid-1990s Mukulincer and colleagues asked 40 Israeli Jewish settlers living within Palestinian Authority territory in the Gaza Strip to (p. 530) fill out the ‘love quiz’ and measures of mental distress and PTSD. The settlers were physically cut-off from the rest of Israel. The settlement had been repeatedly attacked, including a suicide-bombing in 1994 which killed six Jewish members of the settlement and six Palestinians.469 The settlement was also at the time facing the threat of eviction by the Israeli government (indeed, this threat would later be realised in August 2005, when the government sent in the army to forcibly remove the settlers). Though the identities and lives of the settlers would be potentially quite different from the combat veterans he had studied, Mukulincer nonetheless hoped that he could treat this group as representing a ‘high threat’ condition. The settler participants were compared to 40 matched controls living in villages of equivalent size but within the borders of Israel. They also examined the correlates of different symptom clusters of PTSD: intrusive symptoms (e.g. flashbacks), avoidant symptoms (e.g. emotional numbing), and hyperarousal (e.g. sudden irritability or startle responses).470 Mukulincer and colleagues found that overall attachment styles were a much better predictor of mental health than whether participants were settlers or in the control group, accounting for five to six times more variance in general distress and mental health and PSTD-specific symptoms.471

However, the threat condition was a powerful moderator of attachment style. The anxious-ambivalent attachment style was associated with general mental health symptoms and avoidant, hyperarousal, and intrusive PTSD symptoms for both the settlers and the matched controls; this association was not stronger for the settlers. By contrast, the avoidant attachment style was associated with general distress and mental health symptoms only among the settler group. It was also associated with PTSD symptoms from the avoidance cluster. An association with PTSD symptoms from the hyperarousal/intrusion cluster was marked but fell just short of significance, which was interpreted as an effect of the small sample. Mikulincer and colleagues interpreted their findings as suggesting that the high threat condition was putting strain on the avoidant attachment strategy, and that this strain was most clearly seen in the form of PTSD symptoms that reflected this attachment strategy.472 However, there were indications that the high threat condition additionally made a contribution to distress and mental illness among participants with an avoidant attachment style.

(p. 531) The avoidant PTSD symptoms reported by the settlers could be interpreted as a consequence of an avoidant strategy under severe strain.473 It might also be that as the avoidant strategy became overwhelmed and unviable, individuals would experience other forms of distress and anxiety. Conversely, Shaver and Mikulincer considered that the anxious attachment strategy might become unviable in some situations and start to fail. The comparison of the settlers and matched controls did not reveal differences in the anxious attachment strategy or its correlates. However, Shaver and Mikulincer acknowledged that, in other conditions, it might be ‘possible to speak about the failure of hyperactivating strategies’.474 Indeed, Mikulincer and colleagues’ work with the combat veterans from the Lebanon war suggested that war trauma could perhaps contribute to forms of PTSD resembling or expressing both deactivating and hyperactivating strategies.475

In the early 2000s, Mukulincer and Shaver identified the need for further attention to the diverse conditions that might promote the activation of the avoidant attachment style in different forms.476 They proposed that these conditions might include (i) consistent rejection from attachment figures; (ii) threats of punishment by attachment figures for the display of attachment behaviours; (iii) traumatic or abusive experiences in the context of the desire for comfort; and (iv) contexts that encourage self-reliance. The conditions that would promote an anxious attachment style might include: (i) care unrelated to signals about need; (ii) care that punishes or prevents the development of self-regulation skills or autonomy; (iii) messages from the attachment figure that emphasise the individual’s helplessness; and (iv) traumatic or abusive experiences in the context of separation from attachment figures. So, for instance, trauma in both the context of the wish for comfort and the context of separations could be expected to promote both anxious and avoidant strategies. The most potent conditions shaping the selection of attachment strategy, Shaver and Mikulincer suggested, (p. 532) would be those that related directly to the use of the caregiver as secure base or safe haven.477 However, as discussed in the section ‘Minimising and maximising’, over time, Shaver and Mikulincer came to explicitly acknowledge that other behavioural systems could feed into and alter schemas about relationships. This would, presumably, include the fear behavioural system: threat experiences are anticipated to inform Module 1 in Shaver and Mikulincer’s model of the attachment system. The researchers argued that, overall, more needed to be done to understand the conditions that would allow the avoidant and anxious attachment strategies to mitigate, intensify, or express mental illness.478

Yet in the 2000s, the lines of causality remained tangled. It could be that traumatic experiences were relevant to the selection and intensification of maximising and minimising strategies, including in the domain of attachment. This was suggested by Mikulincer’s original work with combat veterans from the Lebanon war. However, it could be that attachment strategies from early life were predisposing and shaping traumatic experiences. This was suggested by Mikulincer’s study of the Israeli settlers, and would be congruent with Shaver and Mikulincer’s emphasis in the early 2000s on attachment styles as shaped primarily by childhood experiences and then predisposing later mental illness. Of course, it was also possible that both processes were taking place simultaneously and reciprocally. To pick apart the temporal relationship between attachment styles and trauma, Solomon, Dekel, and Mikulincer reported findings in 2008 from a longitudinal study of Israeli former prisoners of war from the 1973 Yom Kippur War.479 Participants were followed up in 1991 and 2003, and asked at both times to complete the scaled version of the ‘love quiz’ and a self-report measure of PTSD symptoms. The researchers found that both attachment anxiety and attachment avoidance increased over time among the former prisoners of war, whereas they remained at least stable among the matched control veterans. Both anxious and avoidant attachment styles were positively associated with PTSD symptoms for the former prisoners of war and the matched controls. PTSD avoidance was a potent predictor of later attachment avoidance (r = .68), but it was also predicted by PTSD hyperarousal (r = .59). PTSD avoidance and hyperarousal both also predicted later attachment anxiety (r = .58, .50).480

The longitudinal nature of the study allowed Solomon and colleagues to analyse the respective contribution of these factors to one another over time. Contrary to expectations, the results showed that early PTSD symptoms predicted later attachment styles much better than early attachment styles predicted later PTSD symptoms. The researchers accepted that ‘this finding cannot be easily explained by adult attachment theory (Mikulincer & Shaver). According to this theory, attachment insecurities are a risk factor for the emergence and increase of PTSD symptoms and not the reverse. As a result, this finding is a major novelty of the current prospective study. It seems that traumatic events and post-traumatic responses cause changes in people’s resources and resiliency and then deteriorate their sense (p. 533) of attachment security.’481 By the late 2000s, Shaver and Mikulincer had come to acknowledge that attachment style would reflect not only early experiences with attachment figures, but also information relevant to close relationships from other behavioural systems and from adolescent and adult experiences. Nonetheless, Mikulincer was still taken aback by the fact that PTSD symptoms seemed to be better predictors of attachment style in the sample than vice versa. It also remained unclear whether the additional attachment anxiety and avoidance predicted by the PTSD was the same kind of anxiety and avoidance as before the trauma. It could be that the attachment measure was actually picking up PTSD avoidance and hyperarousal as it figured within the attachment relationship.

Conclusion

Writing in the New York Times in 1974, Bowlby acknowledged that it might often be perplexing why psychological researchers fell into agreement or disagreement when ostensibly discussing the same phenomena, and even sometimes the same theory or data. Acknowledging that there could be a variety of factors involved, he underlined one as central. When psychological knowledge is in dissensus, ultimately ‘we differ, I believe, on how we picture the raw human nature than emerges into the world when a baby is born’.482 The difference between the developmental and social psychological traditions must be recognised as only partially a divergence in method. Differences in method have at times reflected and at times stabilised important differences in how the two traditions conceptualise attachment and its role within the human condition. As Chapter 3 described, Mary Main’s account of conditional strategies was not merely a lens on the Strange Situation, but a model of the three basic ways that humans can direct attention in responding to distress and the activation of the attachment system. It was this general model that led to the development of the AAI coding system, since infant behaviour and adult autobiographical discourse could be regarded as forms affected by the same processes. In turn, Main’s methodological innovations and their predictive value have helped stabilise the general theoretical idea of minimising and maximising strategies as the best way to capture individual differences in the Strange Situation and AAI. In a similar manner, the methodological decisions of the social psychological tradition have had a bidirectional relationship with their model of human nature.483

(p. 534) The developmental tradition has more often, though not exclusively, adopted a narrower definition of attachment, focused on the properties of a behavioural system shared with other primates and not requiring elaborate symbolic capacities. By contrast, the social psychological tradition has tended to adopt a much broader image of attachment as manifest in all human relational dynamics, including how humans ‘co-explore their social world, share their enjoyment in the pleasures of their world, and cope with stress in that world’.484 This has followed a thread in Bowlby’s writings, where sometimes he used the idea of attachment to refer to this broader sense, including the symbolisation of safe haven and secure base phenomena in relationships with ideas and institutions. However, more structurally, this broader notion of attachment reflected and continues to express the concerns of social psychology as a subdiscipline. In very general terms, and especially in its American incarnation, since the 1960s, developmental psychology has tended to focus on the study of the emergence of the capacities of individuals, especially as these occur in the context of their family relationships.485 The work of Ainsworth and second-generation attachment researchers contributed to this discipline formation and was shaped by it. From the 1980s, again especially in America, experimental social psychology has tended to examine the role of personality and social processes within the diversity of adult individual perceptions, attitudes, and practices, including but not limited to intimate relationships.486 (Exceptions can readily be found to these limited characterisations, of course, which are intended to convey general and predominant differences. Nonetheless, researchers who seem the prime exceptions can often be found expressing concern about the limitations of predominant disciplinary trends.)487

The subdisciplines can be regarded as aligned by a basic commitment to imagining humans as interdependent with others, and therefore to the study of interactions. And they find many points of cross-over, for instance in the study of the development of moral reasoning. However, in general terms, in developmental psychology humans are attended to as interdependent in their formation. Main’s theory is an account of repetoires of response to this interdependence as it is expressed by the demands of the attachment system. This is sometimes confused by the assumption that the AAI measures a ‘thing’ called ‘attachment’ and by the individualising language of ‘secure’ and ‘insecure’ infants. Nonetheless, following Ainsworth, the official line of the developmental tradition has been that attachment is, or at least starts as, a dyadic property. This is a central reason why developmental attachment (p. 535) researchers have been reluctant to support the ‘attachment disorder’ diagnosis of individual children, leaving this diagnosis an odd and poorly integrated appendage to the developmental research tradition (Chapter 1).

By contrast, in experimental social psychology the focus tends to be on the study of already largely formed adult humans as influenced by social and relational factors in their perceptions, attitudes, and practices. As Stainton Rogers and colleagues have argued, there is a liberal humanist quality to social psychology: humans are registered at the point that they can act independently, though the concern is then with how they ultimately live together.488 This has helped make use of college student samples a backbone of the subdiscipline. It has also helped make self-report methodology more acceptable than for developmental psychologists. Shaver and Mikulincer are more committed than most social psychologists to the idea that, at a fundamental level, relationships are primary, individuals are secondary: ‘Rather than conceptualizing human beings as separate entities whose interactions with each other need to be understood, it makes more sense to consider social relatedness and its mental correlates as the normal “baseline” condition’.489 Nonetheless, aligned with the focus on individuals and their interactions in wider American social psychology, their predominant methodological orientation has been towards the study of the schemas of individuals about close relationships as reflected in their self-reports. Using self-report measures of attachment styles, Shaver and Mikulincer’s work has explored how an individual’s schemas about close relationships shape a diverse range of their perceptions, attitudes, and practices, ranging from how new romantic relationships are established to how recklessly they drive a car.

One important by-product of the self-report methodology used by social psychological attachment researchers was that it lowered an important barrier to experimental research. This contributed to the identification of the effects of security priming by social psychologists, which is discussed in the next chapter. It also lowered a barrier to public engagement. As documented in previous chapters, Bowlby’s work had contributed to a widespread public for attachment theory and research. However, the developmental tradition of attachment research had faced obstacles in continuing this legacy. Concepts such as the attachment behavioural system were complex and subtle, and the observational methodology favoured by Ainsworth was complex and required extensive training to fully understand. Furthermore, (p. 536) Ainsworth and her students were focused on creating a differentiated space within which empirical attachment research could thrive, giving little time to communicating in popular forums. There may also have been some reticence to take on Bowlby’s mantle as a public intellectual. For the social psychological tradition, the barrier to public understanding was substantially lowered. The Hazan and Shaver ‘love quiz’ has its subtlties in the formulation of the statements, but the measure’s presumption that there are distinct adult attachment styles that can be reported by any individual about themselves contributed to a fundamental accessiblity.

An interesting illustration is the recommendation by Fonagy and colleagues of the use of the Hazan and Shaver ‘love-quiz’ at the start of brief therapeutic work with patients rather than the AAI or the ECR. Though they acknowledge that the ‘love-quiz’ is generally too crude to pick up any changes in attachment style associated with therapy, they argued that it offers an excellent initial basis for a therapeutic conversation in which both therapist and patient are treated as knowledgeable but capable of change in their perspectives.490 There is no need for the tallying of responses indexing latent constructs of anxiety and avoidance required for the ECR, let alone the labour-intensive process of having an AAI coded. Part of the appeal of the ‘love-quiz’ for Fonagy and colleagues in a therapeutic context was that, despite being a self-report measure, the ECR is a deceptively subtle tool, oriented towards mental schemas expressed in feelings, beliefs, expectations, and behaviours, but which may not be consciously known. Fonagy and colleagues at the Anna Freud Centre have also shown increasing interest in the use of self-report measures of attachment-related constructs in research contexts, in part because of the greater potential viability of these measures for use at scale.491

Over the past 20 years or more, the ECR has successfully served as the dominant methodological tool in the social psychological tradition, and additionally the basis for the model of anxiety and avoidance as the dominant theory. Further discussion was largely excluded by the fact that the ECR seemed to work well in predicting various expected correlates, and the factor analysis by Brennan and colleagues was taken as proof of the psychometric standing of the measure. Yet, for the third-generation researchers most versed in the social psychological tradition, questions have been raised about the theory and method that have organised the approach for the past 20 years. For example, Gillath, Fraley, and Birnbaum have wondered whether enough variance in individual difference is captured in theory and method by the constructs of anxiety and avoidance. Of course, they acknowledge, pragmatic compromises are always a part of scientific measurement.492 Yet it is also possible for these compromises to be renegotiated if these compromises are recognised as having incrementally caused enough hinderance over time to warrant the effort. On the one hand, psychometric inquiries have suggested that security may not be reducible to the absence of anxiety and avoidance, but may represent its own dimension. On the other hand, it is not clear that the ECR as a measure or existing theory in the tradition is adequate to capturing the effects of trauma on experiences in close relationships. These are questions that Shaver and Mikulincer have acknowledged as limitations to their work. Their former students and younger colleagues will need to appraise whether these limitations are significant enough to warrant methodological or theoretical change, or, if measures are left unchanged, how to respond to the ensuing limitations.

Notes:

1 Bowlby, J. (1984) Letter to Marco Bacciagaluppi, 3 January 1984. PP/Bow/B.3/40.

2 E.g. Greenberg, M.T., Siegel, J.M., & Leitch, C.J. (1983) The nature and importance of attachment relationships to parents and peers during adolescence. Journal of Youth and Adolescence, 12(5), 373–86.

3 In 2014, in tributes on the occasion of Shaver’s retirement as Distinguished Professor of Psychology at the University of California, Davis, Jude Cassidy observed that ‘no scholar has done more to expand attachment theory and research’. Pehr Granqvist claimed that Shaver’s work ‘surpasses nearly all other research programs in psychological science, both in terms of originality and sheer quantity. I’ll stand on Mary Ainsworth’s coffee table in my cowboy boots and say that!’ Such staunch statements of praise are especially salient given that, unlike Cassidy, or Granqvist’s mentor Mary Main, Shaver was not a student of Ainsworth’s. His route into attachment research was less direct, and the acceptance his work has gained has taken a lot longer. http://www.foundationpsp.org/shaver.php.

4 Shaver, P.R. & Mikulincer, M. (2005) Attachment theory and research: resurrection of the psychodynamic approach to personality. Journal of Research in Personality, 39, 22–45, pp.23–4.

5 Shaver, P.R. & Mikulincer, M. (2011) Analysis of a collaborative working relationship. Relationship Research Newsletter, 9(2), 7–9, p.7.

6 Shaver, P. (2017) Attachment to attachment theory. Voices: Journal of the American Academy of Psychotherapists, 53(2), 35–9, p.36.

7 In a chapter for a psychology textbook written during this period, Shaver’s tone shifts towards the personal: Shaver, P.R. (1975) Psychoanalytic theories of personality. In Psychology Today: An Introduction, 3rd edn (pp.402–425). New York: CRM/Random House: ‘Lest the superego seem to be a dry abstraction, you should realise that intense guilt and the wearying pursuit of perfection are two of the most common reasons people give for entering psychoanalysis’ (409). He also asks the reader, ‘have you ever gone on a long walk during cold, wet weather, suspecting that you might get sick but feeling vaguely that it would serve you right?’ (409).

8 Shaver, P. (2017) Attachment to attachment theory. Voices: Journal of the American Academy of Psychotherapists, 53(2), 35–9.

9 Goodman, G.S. (2006) Attachment to attachment theory: a personal perspective on an attachment researcher. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.3–22). New York: Guilford, p.8.

10 Shaver, P. (2017) Attachment to attachment theory. Voices: Journal of the American Academy of Psychotherapists, 53(2), 35–9.

11 Ibid. p.38.

12 Shaver, P.R. (1972) Review of cognition and affect, edited by J.S. Antrobus. American Journal of Psychology, 85, 297–9, p.297. See also Shaver, P.R. (1975) Emotional experience and expression. In Psychology Today: An Introduction, 3rd edn. New York: CRM/Random House: ‘Surely emotion, not variety, is the spice of life—though the two are obviously related … with fear, with indignation, jealousy, ambition, worship. If they are there, life changes’ (333).

13 Shaver, P.R. & Klinnert, M. (1982) Schachter’s theories of affiliation and emotion: implications of developmental research. In L. Wheeler (ed.) Review of Personality and Social Psychology, Vol. 3 (pp.37–72). Beverly Hills: Sage Publications, p.39.

14 Shaver, P. (2017) Attachment to attachment theory. Voices: Journal of the American Academy of Psychotherapists, 53(2), 35–9, p.39.

15 Shaver, P.R. & Brennan, K.B. (1991) Measures of depression and loneliness. In J.P. Robinson, P.R., Shaver, & L.S. Wrightsman (eds) Measures of Social Psychological Attitudes, Vol. 1 (pp.195–289). San Diego: Academic Press, p.195.

16 Ibid. p.281.

17 The experience is described by Weiss in Weiss, R.S. (1994) Foreword. In M.B. Sperling & W.H. Berman (eds) Attachment in Adults: Clinical and Developmental Perspectives (pp.iv–xvi). New York: Guilford, p.xiv.

18 Weiss, R. (1974) The provisions of social relationships. In Z. Rubin (ed.) Doing Unto Others (pp.17–26). Englewood Cliffs, NJ: Prentice-Hall; Weiss, R.S. (1982) Attachment in adult life. In C. Parkes & J. Stevenson-Hinde (eds) The Place of Attachment in Human Behavior (pp.171–84). New York: Basic Books.

19 Shaver, P. (2017) Attachment to attachment theory. Voices: Journal of the American Academy of Psychotherapists, 53(2), 35–9, p.39.

20 Ibid.: ‘Looking back over the decades, I see that I have been attached to my mother, my analyst, and my wife, and also—viewing attachment theory as a “safe haven” and “secure base”—to attachment theory itself’ (39).

21 Shaver, P.R. (2010) My appreciation of Caryl Rusbult. Personal Relationships, 17(2), 172–3: ‘My own work is inspired by psychoanalytic theory and tilts heavily toward personality rather than, or in addition to, situations … I wanted to dive deeply into the individual’s mind, viewing a mentally represented couple relationship as at least as important as the dyad’s actual behavioral interactions’ (172).

22 E.g. Ainsworth, M.D.S. (1985) Attachments across the life span. Bulletin of the New York Academy of Medicine, 61, 792–812.

23 Perlman, D. & Duck, S. (2006) The seven seas of the study of personal relationships: from ‘the thousand islands’ to the interconnected waterways. In D. Perlman & A.L. Vangelisti (eds) The Cambridge Handbook of Personal Relationships (pp.3–34). Cambridge: Cambridge University Press.

24 Goodman, G.S. (2006) Attachment to attachment theory: a personal perspective on an attachment researcher. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.3–22). New York: Guilford, p.14.

25 Hazan, C. & Shaver, P. (1987) Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–24.

26 Shaver, P.R., Hazan, C., & Bradshaw, D. (1988) Love as attachment: the integration of three behavioral systems. In R.J. Sternberg & M. Barnes (eds), The Psychology of Love (pp. 68–99). New Haven, CT: Yale University Press. Reflecting on the wider context of the publication of these works, Shaver has recalled that when they submitted the paper for publication, it was so far outside the mainstream of social psychology that he worried that it might derail his career. Instead, the paper immediately attracted interest. ‘A lot of young researchers, including the growing number of young women in the field, followed up the initial study in interesting and unanticipated ways,’ he said. ‘I feel fortunate to have come along when the number of women in the field was growing, which made close relationships a more acceptable topic, and when the divorce rate in the U.S. was of concern to government research funders and to the American population. At the same time, new research techniques were developed, making it possible to pursue personality and relationship phenomena that had not been studied empirically before.’ Holder, K. (2018) Interview with Phillip Shaver on winning the Society of Personality and Social Psychology Legacy Award. https://lettersandscience.ucdavis.edu/news/phil-shaver-research-legacy-award.

27 Shaver, P.R., Hazan, C., & Bradshaw, D. (1988) Love as attachment: the integration of three behavioral systems. In R.J. Sternberg & M. Barnes (eds) The Psychology of Love (pp.68–99). New Haven, CT: Yale University Press, p.77.

28 Ibid. p.92. The personal experiences that contributed to Cindy Hazan’s interest in attachment and loss do not appear in the public record.

29 Seligman, M.E.P. (1972) Learned helplessness. Annual Review of Medicine, 23(1), 407–412.

30 Mikulincer, M. & Caspy, T. (1986) The conceptualization of helplessness: I. A phenomenological-structural analysis. Motivation and Emotion, 10, 263–78; Mikulincer, M. (1986) Attributional processes in the learned helplessness paradigm: the behavioral effects of globality attributions. Journal of Personality and Social Psychology, 51, 1248–56.

31 Lazarus, R.S. & Folkman, S. (1984) Coping and adaptation. In W.D. Gentry (ed.) Handbook of Behavioral Medicine (pp.282–325). New York: Guildford.

32 Mikulincer, M. (1994) Human Learned Helplessness: A Coping Perspective. New York: Plenum Press, p.72.

33 Solomon, Z., Waysman, M., & Mikulincer, M. (1990) Family functioning, perceived societal support, and combat-related psychopathology: the moderating role of loneliness. Journal of Social and Clinical Psychology, 9(4), 456–72.

34 Shaver, P.R. & Mikulincer, M. (2011) Analysis of a collaborative working relationship. Relationship Research Newsletter, 9(2), 7–9, pp.7–8.

35 Shaver’s CV lists a visit to Bar Ilan in 1995: Shaver, P.R. (1995) Psychodynamic and representational aspects of adult attachment. Invited address, Bar Ilan University, Ramat Gan, Israel. There is also a visit in 2000: Shaver, P.R. (2000) An updated theory of romantic (pair-bond) attachment. Invited address at 3rd International Conference of the Peleg-Bilig Center for the Study of Family Wellbeing: ‘Towards a science of couple relationships.’ Bar-Ilan University, Ramat Gan, Israel. Their first co-authored work would begin appearing in 2001. https://adultattachment.faculty.ucdavis.edu/wp-content/uploads/sites/66/2014/06/Shaver.pdf.

36 Shaver, P.R. & Mikulincer, M. (2011) Analysis of a collaborative working relationship. Relationship Research Newsletter, 9(2), 7–9, p.8.

37 Shaver, P.R. (2006) Dynamics of romantic love: comments, questions, and future directions. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.423–56). New York: Guilford: ‘There is no theory of personality, emotions, social relationships or psychological development that holds much more than a flickering candle to actual experience … It behoves us as relationship researchers to keep attachment theory, alternative theories of love, and our own actual experiences of love in mind’ (426).

38 Ainsworth, M. & Ainsworth, L. (1958) Measuring Security in Personal Adjustment. Toronto: University of Toronto Press, p.17. Van Rosmalen, van IJzendoorn, and van der Veer reported that a slightly revised version of the Ainsworth self-report questionnaire ‘correlated significantly with the two ECR-RS scales for avoidance and anxiety to mother, –.61 and –.32, respectively (p < .001; n = 230). The same was true for the ECR-RS to father: –.54 and –.39, respectively (p < .001; n = 222)’ (93). Such findings indicate sufficient convergence between Ainsworth’s abandoned questionnaire and Shaver’s work to suggest that they represent highly aligned, if not necessarily identical, research strategies. Van Rosmalen, L. (2015) From security to attachment: Mary Ainsworth’s contribution to attachment theory. Unpublished doctoral dissertation, Leiden University, Chapter 4.

39 Ainsworth, M.D.S. (1985) Attachments across the life span. Bulletin of the New York Academy of Medicine, 61, 792–812, p.798.

40 Main, M., van IJzendoorn, M.H., & Hesse, E. (1993) Unresolved/Unclassifiable Responses to the Adult Attachment Interview: Predictable from Unresolved States and Anomalous Beliefs in the Berkeley–Leiden Adult Attachment Questionnaire. Unpublished manuscript. https://openaccess.leidenuniv.nl/dspace/bitstream/1887/1464/1/168_131.pdf.

41 Ainsworth, M.D.S. (1988) Security. Unpublished discussion paper prepared for the Foundations of Attachment Theory Workshop, convened for the New York Attachment Consortium by G. Cox-Steiner & E. Waters, Port Jefferson, NY. http://www.psychology.sunysb.edu/attachment/online/mda_security.pdf.

42 Ainsworth, M. (1992) A consideration of social referencing in the context of attachment theory and research. In S. Feinman (ed.) Social Referencing and the Construction of Reality in Infancy (pp.349–67). New York: Plenum Press: ‘Some find their models too conflicting for such integration to be achieved. It would appear that in such cases the conflicting models become separately consolidated, and often enough this is because what has been conveyed by the partner in expressive behaviour does not match the semantic content of what is verbally conveyed’ (365).

43 Over time, developmental attachment researchers have repeatedly found substantial evidence for this conclusion. For instance, in a recent study, Howard and colleagues found that participants classified as dismissing on the AAI disclosed fewer adverse childhood experiences on a self-report measure than independent scorers rated in their AAI, and participants classified as preoccupied on the AAI self-reported more adverse childhood experiences than independent coders. Howard, A.R.H., Razuri, E.B., Copeland, R., Call, C., Nunez, M., & Cross, D.R. (2017) The role of attachment classification on disclosure of self and rater-reported adverse childhood experiences in a sample of child welfare professionals. Children and Youth Services Review, 83, 131–6.

44 Pianta, R., Egeland, B., & Adam, E. (1996) Adult attachment classification and self-reported psychiatric symptomatology as assessed by the Minnesota Multiphasic Personality Inventory–2. Journal of Consulting and Clinical Psychology, 64(2), 273–81; Howard, A.R.H., Razuri, E.B., Copeland, R., Call, C., Nunez, M., & Cross, D.R. (2017) The role of attachment classification on disclosure of self and rater-reported adverse childhood experiences in a sample of child welfare professionals. Children and Youth Services Review, 83, 131–6. See also Borelli, J.L., Palmer, A., Vanwoerden, S., & Sharp, C. (2019) Convergence in reports of adolescents’ psychopathology: a focus on disorganized attachment and reflective functioning. Journal of Clinical Child & Adolescent Psychology, 48(4), 568–81.

45 Bailey, H.N., Redden, E., Pederson, D.R., & Moran, G. (2016) Parental disavowal of relationship difficulties fosters the development of insecure attachment. Revue Canadienne des Sciences du Comportement, 48(1), 49–59.

46 Bowlby, J. (1985) Letter to Phillip Shaver, 30 October 1986. PP/Bow/J.9/181.

47 Bowlby, J. (1988) The role of attachment in personality development. In A Secure Base (pp.134–54). London: Routledge, p.145.

48 Ainsworth, M. & Bowlby, J. (1991) 1989 APA award recipient addresses: an ethological approach to personality development. American Psychologist, 46(4), 333–41, p.334.

49 Rubenstein, C. & Shaver, P.R. (1982) The experience of loneliness. In L.A. Peplau & D. Perlman (eds) Loneliness: A Sourcebook of Current Theory, Research, and Therapy (pp.206–23). New York: Wiley, p.221.

50 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford, p.24.

51 E.g. Shaver, P.R., Collins, N.L., & Clark, C.L. (1996) Attachment styles and internal working models of self and relationship partners. In G.J.O. Fletcher & J. Fitness (eds) Knowledge Structures in Close Relationships: A Social Psychological Approach (pp.25–61). Mahwah, NJ: Lawrence Erlbaum: ‘Research on attachment styles—relatively coherent and stable patterns of emotion and behaviour exhibited in close relationships—is based on the assumption that relationship orientations are due to, or perhaps consist in, something called internal working models of self and others’ (25); Fraley, R.C. & Shaver, P.R. (1998) Airport separations: a naturalistic study of adult attachment dynamics in separating couples. Journal of Personality and Social Psychology, 75(5), 1198–212. ‘Technically, the term attachment style refers to the observable patterns of behavior exhibited by an individual, not the unobservable variables (such as working models) that shape these patterns. Nonetheless, as is customary in the literature on attachment, we use the terms attachment style and working models interchangeably’ (1199).

52 Shaver, P.R. & Mikulincer, M. (2002) Attachment-related psychodynamics. Attachment & Human Development, 4(2), 133–61: ‘We use self-report measures in somewhat the same way that physicians use simple indicators of health and illness—e.g. body temperature measured with a thermometer or verbal reports of insomnia. Although such indicators do not provide direct access to underlying disease processes, they are very helpful in assessing a person’s health’ (154).

53 Shaver, P.R., Belsky, J., & Brennan, K.A. (2000) The adult attachment interview and self-reports of romantic attachment: associations across domains and methods. Personal Relationships, 7(1), 25–43, p.41.

54 This approach to interpreting the meaning of self-report measures of mental health has diverse roots, but the role of cognitive-behavioural models may be considered an important contributory. See Beck, A.T. (1991) Cognitive therapy: a 30-year retrospective. American Psychologist, 46(4), 368–75.

55 Hendrick, C. & Hendrick, S. (1986) A theory and method of love. Journal of Personality and Social Psychology, 50, 39–42.

56 Levy, M.B. & Davis, K.E. (1988) Lovestyles and attachment styles compared: their relations to each other and to various relationship characteristics. Journal of Social and Personal Relationships, 5(4), 439–71.

57 Hazan, C. & Shaver, P. (1987) Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–24, p.513.

58 This claim would be made explicitly a few years later in Hazan, C. & Shaver, P.R. (1994) Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22, p.1.

59 Hazan, C. & Shaver, P. (1987) Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–24, p.521.

60 Ibid. p.515.

61 Ibid. p.151. This is the statement as presented by Hazan and Shaver in their 1987 paper. However, it appears that it was not actually the one circulated in the newspaper. The original study also included the additional phrase ‘I am comfortable having others depend on me’, though this was taken out shortly after. Shaver, P.R., Belsky, J., & Brennan, K.A. (2000) The adult attachment interview and self-reports of romantic attachment: associations across domains and methods. Personal Relationships, 7(1), 25–43: ‘Hazan & Shaver (1987), in the first self-report measure of romantic attachment style, used the statement “I am comfortable having others depend on me.” They did not realize that attachment theory does not include this comfort as a legitimate part of being attached; it is, instead, part of being an attachment figure, or caregiver, for others (Cassidy, 1999). In later work by Hazan & Shaver (1990) and Kunce and Shaver (1994), the potentially misleading statement was edited out of the self-report romantic attachment measure. For good or ill, however, it had already been incorporated into self-report measures constructed by Bartholomew and Horowitz (1991) and Collins and Read (1990)’ (39). However, it would not figure within the ECR scale.

62 Shaver, P.R. & Mikulincer, M. (2002) Dialogue on adult attachment: diversity and integration. Attachment & Human Development, 4(2), 243–57: ‘The original Hazan and Shaver (1987) measure of avoidance tapped fearful, rather than dismissing, avoidance’ (253). An additional disanalogy with Ainsworth’s categories can be identified: the avoidant attachment style was represented by Hazan and Shaver by two feelings and a belief/expectation. Distrust in close relationships and difficulty depending on them is perhaps an analogue for Ainsworth’s idea that avoidance is a response to experiences of less-sensitive care. However, Ainsworth did not discuss avoidance in terms of distrust. Instead, she characterised ambivalent/resistant pattern as representing distrust in the caregiver. Ainsworth, M.D.S., Bell, S.M., & Stayton, D.J. (1974) Infant–mother attachment and social development: ‘socialisation’ as a product of reciprocal responsiveness to signals. In J.M. Richards (ed.) The Integration of a Child into a Social World (pp.9–135). Cambridge: Cambridge University Press: ‘It may be viewed as advantageous for an infant whose mother seems to him to move unpredictably and inconsistently (and whom he has not been able to learn to trust) to monitor her movements with exceptional alertness and to evince disturbance whenever she moves off’ (125).

63 Hazan, C. & Shaver, P. (1987) Romantic love conceptualized as an attachment process. Journal of Personality and Social Psychology, 52(3), 511–24, p.515.

64 Ibid. p.518.

65 Ibid. Table 3 and Table 6.

66 Hutt, M. (1991) Influences of attachment in everyday problem solving. Unpublished doctoral dissertation, Cornell University. Shaver drew on this study in support of the predictive validity of self-report measures of assessment in Shaver, P.R. & Norman, A.J. (1995) Attachment theory and counseling psychology: a commentary. The Counseling Psychologist, 23(3), 491–500.

67 Hazan, C. & Shaver, P.R. (1990) Love and work: an attachment-theoretical perspective. Journal of Personality and social Psychology, 59(2), 270–80, Table 3.

68 Cooper, M.L., Shaver, P.R., & Collins, N.L. (1998) Attachment styles, emotion regulation, and adjustment in adolescence. Journal of Personality and Social Psychology, 74(5), 1380–97.

69 The astonishing volume of outputs from Mikulincer’s group in these years means that only a selection can be mentioned. The slight adaptations to the measure enacted by Mikulincer and colleagues are detailed in Mikulincer, M., Florian, V., & Tolmacz, R. (1990) Attachment styles and fear of personal death: a case study of affect regulation. Journal of Personality and Social Psychology, 58(2), 273–80.

71 Banai, E., Weller, A., & Mikulincer, M. (1998) Inter-judge agreement in evaluation of adult attachment style: the impact of acquaintanceship. British Journal of Social Psychology, 37(1), 95–109, p.102.

72 Mikulincer, M. & Orbach, I. (1995) Attachment styles and repressive defensiveness: the accessibility and architecture of affective memories. Journal of Personality and Social Psychology, 68(5), 917–25.

73 Mikulincer, M. & Florian, V. (1997) Are emotional and instrumental supportive interactions beneficial in times of stress? The impact of attachment style. Anxiety, Stress, and Coping, 10, 109–127.

74 Shaver, P.R. & Mikulincer, M. (2008) Adult attachment and cognitive and affective reactions to positive and negative events. Social and Personality Compass, 2, 1844–65, p.1857. For instance, in a study among newlyweds from Mikulincer’s laboratory, expressions of happiness in the relationship were usually associated with positive feelings, but for anxious participants were associated with envy. Sofer-Roth, S. (2008) Adult attachment and the nature of responses to a romantic partner’s expression of personal happiness. Unpublished doctoral dissertation, Bar-Ilan University, Ramat Gan, Israel.

75 Mikulincer, M., Shaver, P.R., & Slav, K. (2006) Attachment, mental representations of others, and gratitude and forgiveness in romantic relationships. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.190–215). New York: Guilford.

76 Mikulincer, M. (1998) Attachment working models and the sense of trust. Journal of Personality and Social Psychology, 74(5), 1209–224. These findings were fleshed out at the level of the dyad by a later study. Mikulincer and Florian found that couples in which both partners were secure reported more cohesion and adaptability in their relationship than couples in which one was insecure, who in turn reported more cohesion and adaptability than when neither endorsed a secure attachment style. Mikulincer, M. & Florian, V. (1999) The association between spouses’ self-reports of attachment styles and representations of family dynamics. Family Process, 38, 69–83.

77 Mikulincer, M., Florian, V., & Weller, A. (1993) Attachment styles, coping strategies, and posttraumatic psychological distress: the impact of the Gulf War in Israel. Journal of Personality and Social Psychology, 64(5), 817–26.

78 Mikulincer, M, & Florian, V. (1998) The relationship between adult attachment styles and emotional and cognitive reactions to stressful events. In J. Simpson & S. Rholes (eds) Attachment Theory and Close Relationships (pp.143–65). New York: Guilford.

79 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.238. In the manuscript version of the book, the formulation is a little different: Bowlby, J. (not dated) Notes towards Chapter 6 of Volume 2. PP/Bow/K.5./17: ‘in the working model of the world that anyone builds a key feature is his notion of who his attachment figures are and how accessible or inaccessible they may be. Similarly in the working model of the self that anyone builds a key feature is his notion of how acceptable or unacceptable he is in the eyes of his attachment figures. On the structure of these communications turns that person’s forecasts of how available and responsive his attachment figures will be should he turn to them for support.’

80 Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, p.120.

81 Reis, H.T. & Shaver, P.R. (1988) Intimacy as an interpersonal process. In S. Duck (ed.) Handbook of Research in Personal Relationships (pp.367–89). London: Wiley, p.372.

83 Reis and Shaver suspected that Main was ultimately assessing ‘defensiveness and information-processing distortions’. Ibid. p.372.

84 Ainsworth, M. (1984) Attachment. In N.S. Endler & J. McVicker Hunt (eds) Personality and the Behavioral Disorders (pp.559–602). New York: Wiley: ‘There seem to be three major behavioural systems involved in forming and maintaining heterosexual pair bonding: 1) the reproductive or mating system, which seems likely to be initially the most important in bond formation, regardless of whether the biological function of reproduction is fulfilled; 2) the caregiving behavioural system, which is involved in two ways—giving care to the partner and sharing with the partner caregiving to the young that may result from the union; and 3) the attachment system, which implies each partner seeks security—comfort and reassurance—through maintaining contact with the other’ (595).

85 Hazan, C. & Shaver, P.R. (1994) Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22.

86 Hazan, C. & Shaver, P.R. (1994) Deeper into attachment theory. Psychological Inquiry, 5(1), 68–79, p.73.

87 Conradi, H-J., Gerlsma, C., van Duijn, M., & de Jonge, P. (2006) Internal and external validity of the experiences in close relationships questionnaire in an American and two Dutch samples. European Journal of Psychiatry, 20(4), 258–69, p.259. The array of self-report attachment measures from this period include: Armsden, G.C. & Greenberg, M.T. (1987) The inventory of parent and peer attachment: individual differences and their relationship to psychological well-being in adolescence. Journal of Youth and Adolescence, 16(5), 427–54; West, M., Sheldon, A., & Reiffer, L. (1987) An approach to the delineation of adult attachment: scale development and reliability. Journal of Nervous and Mental Disease, 175(12), 738–41; Pottharst, K. & Kessler, R. (1990) The search for methods and measures. In K. Pottharst (ed.) Explorations in Adult Attachment (pp.9–37). New York: Peter Lang; Griffin, D.W. & Bartholomew, K. (1994) The metaphysics of measurement: the case of adult attachment. In K. Bartholomew & D. Perlman (eds) Advances in Personal Relationships: Vol. 5. Attachment Processes in Adulthood (pp.17–52). London: Jessica Kingsley.

88 Bartholomew, K. & Horowitz, L.M. (1991) Attachment styles among young adults: a test of a four-category model. Journal of Personality and Social Psychology, 61(2), 226–44.

89 Bartholomew, K. (1990) Avoidance of intimacy: an attachment perspective. Journal of Social and Personal Relationships, 7(2), 147–78, p.163.

90 Fraley, R.C. & Waller, N.G. (1998) Adult attachment patterns: a test of the typological model. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.77–114). New York: Guilford: ‘Bartholomew’s model, combined with an emerging consensus that two latent dimensions underlie Hazan and Shaver’s (1987) and Bartholowew’s (1990) attachment types (see Feeney & Noller 1996; Griffin & Bartholomew 1994a, 1994b; Hazan & Shaver 1994), has encouraged researchers to use continuous measures’ (82).

91 Brennan, K.A., Shaver, P.R., & Tobey, A.E. (1991) Attachment styles, gender, and parental problem drinking. Journal of Social and Personal Relationships, 8, 451–66; Kunce, L.J. & Shaver, P.R. (1994) An attachment-theoretical approach to caregiving in romantic relationships. In K. Bartholomew & D. Perlman (eds) Advances in Personal Relationships, Volume 5 (pp.205–237). London: Jessica Kingsley.

92 Brennan, K.A. & Shaver, P.R. (1998) Attachment styles and personality disorders: their connections to each other and to parental divorce, parental death, and perceptions of parental caregiving. Journal of Personality, 66(5), 835–78.

93 Brennan, K.A., Shaver, P.R., & Tobey, A.E. (1991) Attachment styles, gender, and parental problem drinking. Journal of Social and Personal Relationships, 8, 451–66.

94 Shaver, P.R., Collins, N.L., & Clark, C.L. (1996) Attachment styles and internal working models of self and relationship partners. In G.J.O. Fletcher & J. Fitness (eds) Knowledge Structures in Close Relationships: A Social Psychological Approach (pp.25–61). Mahwah, NJ: Lawrence Erlbaum: ‘Disorganised, or fearful, children lack self-confidence and have low self-worth’ (36).

95 Bartholomew, K. (1989) Attachment styles in young adults: Implications for self-concept and interpersonal functioning. Unpublished doctoral dissertation, Stanford University.

96 Jacobvitz, D., Curran, M., & Moller, N. (2002) Measurement of adult attachment: the place of self-report and interview methodologies. Attachment & Human Development, 4(2), 207–215: ‘The fearful attachment style seems to differ qualitatively from the disorganized infant pattern. A mixture of avoidance and ambivalence is only one of the many behavioral indices of attachment disorganization during infancy’ (209).

97 Bartholomew, K. (1989) Attachment styles in young adults: implications for self-concept and interpersonal functioning. Unpublished doctoral dissertation, Stanford University. This is in contrast to the new self-report ‘disorganised’ attachment style introduced by Paetzold and colleagues in 2015, which does seek to assess fear of attachment figures. Paetzold, R.L., Rholes, W.S., & Kohn, J.L. (2015) Disorganized attachment in adulthood: theory, measurement, and implications for romantic relationships. Review of General Psychology, 19(2), 146–65.

98 Duschinsky, R. (2018) Disorganization, fear and attachment: working towards clarification. Infant Mental Health Journal, 39(1), 17–29.

99 Roisman, G.I., Holland, A., Fortuna, K., Fraley, R. C., Clausell, E., & Clarke, A. (2007) The Adult Attachment Interview and self-reports of attachment style: an empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–97.

100 Simpson, J.A. (1990) Influence of attachment styles on romantic relationships. Journal of Personality and Social Psychology, 59(5), 971–80; Collins, N.L. & Read, S.J. (1990) Adult attachment, working models, and relationship quality in dating couples. Journal of Personality and Social Psychology, 58(4), 644–63; Fraley, R.C. & Waller, N.G. (1998) Adult attachment patterns: a test of the typological model. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.77–114). New York: Guilford.

101 Feeney, J.A., Peterson, C., Gallois, C., & Terry, D.J. (2000) Attachment style as a predictor of sexual attitudes and behavior in late adolescence. Psychology & Health, 14(6), 1105–122.

102 E.g. Gillath, O., Canterberry, M., & Atchley, P. (2017) Attachment as a predictor of driving performance. Transportation Research Part F, 45, 208–217.

103 Bartholomew, K. & Shaver, P.R. (1998) Methods of assessing adult attachment: do they converge? In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.25–45). New York: Guilford: ‘The “positivity of the other” model indicates the degree to which others are generally expected to be available and supportive.’ Yet an anxious attachment style was found empirically to be associated with negative representations of others. Simpson, J.A., Rholes, W.S., & Phillips, D. (1996) Conflict in close relationships: an attachment perspective. Journal of Personality and Social Psychology, 71(5), 899–914. Extensive criticism of Bartholomew’s assumption was presented in Fraley, R.C. & Shaver, P.R. (2000) Adult romantic attachment: theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4(2), 132–54.

104 Stein, H., Jacobs, N.J., Ferguson, K.S., Allen, J.G., & Fonagy, P. (1998) What do adult attachment scales measure? Bulletin of the Menninger Clinic, 62, 33–82.

105 Mikulincer, M. (1995) Attachment style and the mental representation of the self. Journal of Personality and Social Psychology, 69(6), 1203–215.

106 Schmitt, D. P., Alcalay, L., Allensworth, M., et al. (2004) Patterns and universals of adult romantic attachment across 62 cultural regions: are models of self and of other pancultural constructs? Journal of Cross-Cultural Psychology, 35(4), 367–402.

107 Brennan, K.A., Shaver, P.R., & Tobey, A.E. (1991) Attachment styles, gender, and parental problem drinking. Journal of Social and Personal Relationships, 8, 451–66, p.465.

108 Crowell, J.A., Fraley, R.C., & Shaver, P.R. (1999) Measures of individual differences in adolescent and adult attachment. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (pp.434–65). New York: Guilford; Fraley, R.C. & Shaver, P.R. (2000) Adult romantic attachment: theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4(2), 132–54.

109 E.g. Fraley, R.C. & Waller, N.G. (1998) Adult attachment patterns: a test of the typological model. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.77–114). New York: Guilford: ‘An explicit focus on the latent dimensions, however, may facilitate inquiry into the underlying operation of the attachment system rather than remaining at the level of manifest behaviour … We believe that the prototype approach leaves the ontological status of the attachment patterns unclear. Are the prototypes advocated by Griffin and Bartholomew supposed to represent “fuzzy” groups that exist in nature or “fuzzy” groups that exist in the minds of perceivers of nature?’ (107).

110 E.g. Ainsworth, M.D.S. (1983) Patterns of infant–mother attachment as related to maternal care: their early history and their contribution to continuity. In D. Magnusson & V.L. Allen (eds) Human Development: An Interactional Perspective (pp.35–55). New York: Academic Press, p.49.

111 E.g. Bowlby, J. (1969, 1982) Attachment. London: Penguin, p.365.

112 Bowlby, J. (1973) Separation: Anxiety and Anger. New York: Basic Books, p.273.

113 Bowlby, J. (1980) Loss. London: Pimlico, p.41.

114 Mikulincer, M., Paz, D., & Kedem, P. (1990) Anxiety and categorization—2. Hierarchical levels of mental categories. Personality and Individual Differences, 11(8), 815–21.

115 See e.g. Brennan, K.A. & Shaver, P.R. (1995) Dimensions of adult attachment, affect regulation, and romantic relationship functioning. Personality and Social Psychology Bulletin, 21(3), 267–83, p.280.

116 Wilkinson, I. (2001) Anxiety in a Risk Society. London: Routledge; Horwitz, A.V. (2013) Anxiety: A Short History. Baltimore: Johns Hopkins University Press.

117 Fraley, R.C. & Shaver, P.R. (1998) Airport separations: a naturalistic study of adult attachment dynamics in separating couples. Journal of Personality and Social Psychology, 75(5), 1198–212, p.1199.

118 Mikulincer, M., Shaver, P.R., & Slav, K. (2006) Attachment, mental representations of others, and gratitude and forgiveness in romantic relationships. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.190–215). New York: Guilford, p.192.

119 There is some lack of clarity on this point of theory, however. In general, attachment anxiety is understood to be the degree to which a person worries that a partner will not be available in times of need. However, at other times it is defined as anger and ambivalence, with worries fading into the background, e.g. Shaver, P.R. & Mikulincer, M. (2005) Attachment theory and research: resurrection of the psychodynamic approach to personality. Journal of Research in Personality, 39, 22–45: ‘Hyperactivating strategies reflect a compromise between conflicting, ambivalent tendencies toward attachment figures—overwhelming anger and hostility toward unavailable attachment figures together with an intense need for proximity to these frustrating figures’ (28).

120 Brennan, K.A., Clark, C.L., & Shaver, P.R. (1998) Self-report measurement of adult romantic attachment: an integrative overview. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.46–76). New York: Guilford, p.47; Ainsworth, M., Blehar, M., Waters, E., & Wall, S. (1978, 2012) Patterns of Attachment: A Psychological Study of the Strange Situation. Bristol: Psychology Press.

121 Ibid.: ‘Right from the start, Ainsworth’s three major attachment “types” could be conceptualized as regions in a two-dimensional space, the dimensions being Avoidance (discomfort with closeness and dependency) and Anxiety (crying, failing to explore confidently in the absence of mother, and angry protest directed at mother during reunions after what was probably experienced as abandonment)’ (49).

122 Ibid. pp.66–7.

123 Ibid. p.57.

124 This is implied but not detailed in the chapter. Confirmed by Kelly Brennan-Jones, personal communication, August 2019.

125 Frıas, M.T., Shaver, P.R., & Mikulincer, M. (2014) Measures of adult attachment and related constructs. In G.J. Boyle & D.H. Saklofske (eds) Measures of Personality and Social Psychological Constructs. Philadelphia: Elsevier, 417–47, p.437.

126 Pittman, J.F. (2012) Attachment orientations: a boon to family theory and research. Journal of Family Theory & Review, 4(4), 306–310, p.308.

127 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press, p.136.

128 Mikulincer, M. (1994) Human Learned Helplessness: A Coping Perspective. New York: Plenum Press: ‘People for whom emotion acts as an internal stimulus that recalls both the unresolved mismatch and their own weakness and vulnerability may perceive emotion as an uninvited intruder that counteracts their avoidance coping. They may also experience their emotions as overwhelming and disorganising forces that demand that they resolve the mismatch, which is precisely what they feel they cannot do. In this case, emotions do not facilitate adaptation and mastery but rather reflect failure of control and flooding, and they may disorganise the person’s avoidance activities’ (143).

129 Crawford, T.N, Shaver, P.R., & Goldsmith, H.H. (2007) How affect regulation moderates the association between anxious attachment and neuroticism. Attachment & Human Development, 9(2), 95–109.

130 Mikulincer, M., Birnbaum, G., Woddis, D., & Nachmias, O. (2000) Stress and accessibility of proximity-related thoughts. Journal of Personality and Social Psychology, 78(3), 509–523. See also Edelstein, R.S. & Gillath, O. (2008) Avoiding interference: adult attachment and emotional processing biases. Personality and Social Psychology Bulletin, 34(2), 171–81.

131 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood, 2nd edn. New York: Guilford, p.143. For another example see Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press: ‘Recently, a fourth category, “disorganized/disoriented,” has been added. It is characterized by odd, awkward behavior during separation and reunion episodes and random fluctuations between signs of anxiety and avoidance’ (66).

132 In discussions of the phenomenon, Mikulincer, Shaver, and colleagues have been prone to cite social psychologists in turn discussing developmental psychologists, rather than the relevant primary literature, e.g. Mikulincer, M., Solomon, Z., Shaver, P.R., & Ein-Dor, T. (2014) Attachment-related consequences of war captivity and trajectories of posttraumatic stress disorder: a 17-year longitudinal study. Journal of Social and Clinical Psychology, 33(3), 207–228: ‘In extreme cases, attachment insecurities can result in a disorganized attachment pattern—an incoherent blend of contradictory approach and avoidance behaviors or paralyzed inaction (Simpson & Rholes, 2002)’ (210). Simpson, J.A. & Rholes, W.S. (2002) Fearful-avoidance, disorganization, and multiple working models: some directions for future theory and research. Attachment & Human Development, 4(2), 223–9.

133 Ein-Dor, T., Mikulincer, M., & Shaver, P.R. (2011) Effective reaction to danger: attachment insecurities predict behavioral reactions to an experimentally induced threat above and beyond general personality traits. Social Psychological and Personality Science, 2(5), 467–73, p.473.

134 Cassidy, J. & Kobak, R. (1988) Avoidance and its relation to other defensive processes. In J. Belsky & T. Neworski (eds) Clinical Implications of Attachment (pp.300–323). Hillsdale, NJ: Lawrence Erlbaum; Kobak, R., Cole, H., Fleming, W., Ferenz-Gillies, R., & Gamble, W. (1993) Attachment and emotion regulation during mother–teen problem-solving: a control theory analysis. Child Development, 64, 231–45.

135 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood. New York: Guilford, Figure 2.1, p.29.

136 Shaver, P.R. & Mikulincer, M. (2012) An attachment perspective on coping with existential concerns. In P.R. Shaver & M. Mikulincer (eds) Meaning, Mortality, and Choice: The Social Psychology of Existential Concerns (pp.291–307). Washington, DC: American Psychological Association, p.296.

137 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood. New York: Guilford, p.29, Figure 2.1.

138 Mario Mikulincer, personal communication, July 2019.

139 Sketching a tightening of this position, Mikulincer has recently described the set-goal of the attachment system in adulthood as evidence that the partner is, specifically, ‘available, responsive and engaged’ (A.R.E.). This riffs on the same acronym with slightly different components—‘accessibility, responsiveness, and emotional engagement’—offered by Johnson, S.M. (2011) Hold Me Tight: Your Guide to the Most Successful Approach to Building Loving Relationships. London: Hachette, Chapter 4. Mikulincer, M. (2019) Advances in the study of the broaden and build cycle of attachment security. Presentation at the Adult Attachment Research Legacy: 32 Years Since Hazan and Shaver Symposium, Society of Personality and Social Psychology Conference, Portland, Oregon, 9 February 2019.

140 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood. New York: Guilford, p.29, Figure 2.1.

141 Shaver, P.R. & Mikulincer, M. (2008) Augmenting the sense of security in romantic, leader–follower, therapeutic, and group relations: a relational model of personality change. In J.P. Forgas & J. Fitness (eds) Social Relationships: Cognitive, Affective, and Motivational Processes (pp.55–73). New York: Psychology Press. Shaver and Mikulincer attributed the idea of ‘security’ as entailing feeling understood, validated, and cared about to Reis, H.T. & Shaver, P.R. (1988) Intimacy as an interpersonal process. In S. Duck (ed.) Handbook of Research in Personal Relationships (pp.367–89). London: Wiley. However, the argument in this text is that these are characteristics of ‘intimacy’. The textual history illustrates the potential slide between intimacy and security at times in Shaver’s work.

142 Fredrickson, B.L. (2001) The role of positive emotions in positive psychology. The broaden-and-build theory of positive emotions. American Psychologist, 56(3), 218–26.

143 Mikulincer, M. & Shaver, P. (2014) An attachment perspective on loneliness. In R.J. Coplan & J.C. Bowker The Handbook of Solitude: Psychological Perspectives on Social Isolation, Social Withdrawal, and Being Alone (pp.34–50). New York: Wiley, p.46.

144 Shaver, P.R., Mikulincer, M., Lavy, S., & Cassidy, J. (2009) Understanding and altering hurt feelings: an attachment-theoretical perspective on the generation and regulation of emotions. In A.L. Vangelisti (ed.) Feeling Hurt in Close Relationships. Cambridge: Cambridge University Press, pp.92–119: ‘Attachment anxiety intensifies the expression of emotions such as jealousy and anger and exaggerates the expression of vulnerability, helplessness, and need’ (109).

145 Shaver, P.R. & Mikulincer, M. (2007) Adult attachment strategies and the regulation of emotion. In J.J. Gross (ed.) Handbook of Emotion Regulation (pp.446–65). New York: Guilford, p.454.

146 E.g. Jayson, Y. (2004) An attachment perspective to escalation of commitment. Unpublished doctoral dissertation, Bar-Ilan University; Mikulincer, M. & Shaver, P. (2008) Contributions of attachment theory and research to motivation science. In J.Y. Shah & W.L. Gardner (eds) Handbook of Motivational Science (pp.201–216). New York: Guilford; Erez, A., Sleebos, E., Mikulincer, M., van IJzendoorn, M.H., Ellemers, N., & Kroonenberg, P.M. (2009) Attachment anxiety, intra-group (dis)respect, actual efforts, and group donation. European Journal of Social Psychology, 39(5), 734–46.

147 Shaver, P.R. & Mikulincer, M. (2003) The psychodynamics of social judgments: an attachment theory perspective. In J.P. Forgas, K.D. Williams, & W. von Hippel (eds) Social Judgments: Implicit and Explicit Processes (pp.85–114). Philadelphia: Psychology Press: ‘Hyperactivating strategies create a chaotic, undifferentiated mental architecture that is constantly pervaded by negative affect’ (104).

148 A further factor stabilising the anxious attachment style was proposed, speculatively, in 2009: Mikulincer, M., Shaver, P.R., Cassidy, J., & Berant, E. (2009) Attachment-related defensive processes. In J.H. Obegi & E. Berant (eds) Attachment Theory and Research in Clinical Work with Adults. New York: Guilford, pp.293–327: ‘For attachment-anxious people, histrionics may seem to have two beneficial effects. First, they sometimes elicit the desired attention, care, and love from others, which is, theoretically, the reason the anxious strategy was adopted in the first place. But there may be a second, less obvious, benefit: The hubbub and distraction generated by strident, impulsive expressions of pain, need, and anger may direct attention and energy away from a deeper problem—sensing oneself as not very substantial at all and not worthy of anyone’s care. Agitating and grabbing someone’s attention is at least likely to make something happen, and even if that something is unpleasant, it may feel better than nothing—that is, better than existential isolation and worthlessness’ (309). This proposal has not been much elaborated and remains as yet untested, except insofar as there exists correlational evidence linking the anxious attachment style to self-report of feelings of isolation and self-report of lower self-esteem.

149 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood. New York: Guilford, Figure 2.1, p.29.

150 See e.g. Crawford, T.N, Shaver, P.R., & Goldsmith, H.H. (2007) How affect regulation moderates the association between anxious attachment and neuroticism. Attachment & Human Development, 9(2), 95–109.

151 Mikulincer, M., Shaver, P.R., & Pereg, D. (2003) Attachment theory and affect regulation: the dynamics, development, and cognitive consequences of attachment-related strategies. Motivation and Emotion, 27(2), 77–102.

152 Shaver, P.R. & Mikulincer, M. (2007) Adult attachment strategies and the regulation of emotion. In J.J. Gross (ed.) Handbook of Emotion Regulation (pp.446–65). New York: Guilford, p.450. See also Edelstein, R.S. (2006) Attachment and emotional memory: investigating the source and extent of avoidant memory impairments. Emotion, 6(2), 340–45; Simpson, J.A., Rholes, W.S., & Winterheld, H.A. (2010) Attachment working models twist memories of relationship events. Psychological Science, 21(2), 252–9. Neuroimaging studies conducted by Gillath and colleagues have been interpreted as suggesting that there is no attachment-specific suppressive mechanism, but that in the case of the avoidant attachment style, general mental capacities for suppression are applied to attachment-related content. Gillath, O., Bunge, S.A., Shaver, P.R., Wendelken, C., & Mikulincer, M. (2005) Attachment-style differences in the ability to suppress negative thoughts: exploring the neural correlates. Neuroimage, 28(4), 835–47.

153 Schachner, D.A., Shaver, P.R., & Mikulincer, M. (2005) Patterns of nonverbal behavior and sensitivity in the context of attachment relations. Journal of Nonverbal Behavior, 29(3), 141–69, p.157.

154 Shaver, P.R. & Mikulincer, M. (2003) The psychodynamics of social judgments: an attachment theory perspective. In J.P. Forgas, K.D. Williams, & W. von Hippel (eds) Social Judgments: Implicit and Explicit Processes (pp.85–114). Philadelphia: Psychology Press, p.106.

155 Mikulincer, M., Shaver, P.R., Sapir-Lavid, Y., & Avihou-Kanza, N. (2009) What’s inside the minds of securely and insecurely attached people? The secure-base script and its associations with attachment-style dimensions. Journal of Personality and Social Psychology, 97(4), 615–33.

156 Berant, E., Mikulincer, M., Shaver, P.R., & Segal, Y. (2005) Rorschach correlates of self-reported attachment dimensions: dynamic manifestations of hyperactivating and deactivating strategies. Journal of Personality Assessment, 84(1), 70–81.

157 This proposal was in part an elaboration of Mikulincer’s earlier emphasis on the structure of the human mind as a set of strategic resources rather than constituted by static representations. Mikulincer, M., Paz, D., & Kedem, P. (1990) Anxiety and categorization—2. Hierarchical levels of mental categories. Personality and Individual Differences, 11(8), 815–21. The image of internal working models as a hierarchical associative network has been criticised by Fraley as, even if of heuristic value, ultimately an overrigid reification. Fraley, R.C. (2007) A connectionist approach to the organization and continuity of working models of attachment. Journal of Personality, 75(6), 1157–80. Fraley and colleagues have also demonstrated that people who have more differentiated attachment styles across their varying relationships are more likely to have an insecure overall attachment style. Fraley, R.C., Heffernan, M.E., Vicary, A.M., & Brumbaugh, C.C. (2011) The Experiences in Close Relationships—Relationship Structures Questionnaire: a method for assessing attachment orientations across relationships. Psychological Assessment, 23(3), 615–25.

158 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press, p.64.

159 Overall, N.C., Fletcher, G.J., & Friesen, M.D. (2003) Mapping the intimate relationship mind: comparisons between three models of attachment representations. Personality and Social Psychology Bulletin, 29(12), 1479–93.

160 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford, pp.52–3.

161 Lavi, N. (2007) Bolstering attachment security in romantic relationships: the long-term contribution of partner’s sensitivity, expressiveness, and supportiveness. Unpublished doctoral dissertation, Bar-Ilan University, Ramat Gan, Israel. For a recent review of the effect of partners on attachment style in adulthood see Arriaga, X.B. & Kumashiro, M. (2019) Walking a security tightrope: relationship-induced changes in attachment security. Current Opinion in Psychology, 25, 121–6.

162 Mikulincer, M. (2006) Attachment, caregiving and sex within romantic relationships. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.23–44). New York: Guilford, p.39.

164 Fraley, R.C. & Shaver, P.R. (2008) Attachment theory and its place in contemporary personality research. In O. John & R.W. Robins (eds) Handbook of Personality: Theory and Research, 3rd edn (pp.518–41). New York: Guilford: ‘The attachment behavioral system is an important concept in attachment theory because it provides the conceptual bridge between ethological models of human development (e.g., Hinde, 1966) and modern theories of emotion regulation and personality (e.g., John & Gross, 2007)’ (523).

165 E.g. Bowlby, J. (1988) The role of attachment in personality development. In A Secure Base (pp.134–54). London: Routledge.

166 Second-generation developmental attachment researchers have generally considered ‘personality’ a reified concept in the academic psychology of their day, better conceptualised as an effect of developmental processes rather than a unitary construct to be measured in itself. See e.g. Sroufe, L.A. & Fleeson, J. (1986) Attachment and the construction of relationships. In W. Hartup & Z. Rubin (eds) Relationships and Development (pp.51–71). Hillsdale, NJ: Lawrence Erlbaum, p.51.

167 Noftle, E.E. & Shaver, P.R. (2006) Attachment dimensions and the big five personality traits: associations and comparative ability to predict relationship quality. Journal of Research in Personality, 40(2), 179–208, p.187.

168 John, O.P., Neumann, L., & Soto, C.J. (2010) The Big Five trait taxonomy: history, measurement, and conceptual issues. In L.A. Pervin & O.P. John (eds) Handbook of Personality: Theory and Research, 3rd edn (pp.114–58). New York: Guilford; Vukasović, T. & Bratko, D. (2015) Heritability of personality: a meta-analysis of behavior genetic studies. Psychological Bulletin, 141(4), 769–85.

169 Crawford, T.N., John Livesley, W., Jang, K.L., Shaver, P.R., Cohen, P., & Ganiban, J. (2007) Insecure attachment and personality disorder: a twin study of adults. European Journal of Personality, 21(2), 191–208. See also Donnellan, M.B., Burt, S.A., Levendosky, A.A., & Klump, K.L. (2008) Genes, personality, and attachment in adults: a multivariate behavioral genetic analysis. Personality and Social Psychology Bulletin, 34(1), 3–16. In contrast to Crawford and colleagues, Donnellan and colleagues reported that genetic effects accounted for variability also in avoidant attachment style. Troisi and colleagues have reported potential gene × environment interactions in the origins of fearful attachment. Troisi, A., Frazzetto, G., Carola, V., et al. (2012) Variation in the μ‎-opioid receptor gene (OPRM1) moderates the influence of early maternal care on fearful attachment. Social Cognitive and Affective Neuroscience, 7(5), 542–7.

170 E.g. Roberts, J.E., Gotlib, I.H., & Kassel, J.D. (1996) Adult attachment security and symptoms of depression: the mediating roles of dysfunctional attitudes and low self-esteem. Journal of Personality and Social Psychology, 70(2), 310–20; Meier, A.M., Carr, D.R., Currier, J.M., & Neimeyer, R.A. (2013) Attachment anxiety and avoidance in coping with bereavement: two studies. Journal of Social and Clinical Psychology, 32(3), 315–34, Supplement 3; Wijngaards-de Meij, L., Stroebe, M., Schut, H., et al. (2007) Neuroticism and attachment insecurity as predictors of bereavement outcome. Journal of Research in Personality, 41(2), 498–505.

171 Shaver, P.R., Belsky, J., & Brennan, K.A. (2000) The adult attachment interview and self-reports of romantic attachment: associations across domains and methods. Personal Relationships, 7(1), 25–43, p.25.

172 Crowell, J.A., Treboux, D., & Waters, E. (1999) The Adult Attachment Interview and the Relationship Questionnaire: relations to reports of mothers and partners. Personal Relationships, 6(1), 1–18.

173 Ibid. p.16.

174 This was a consensus position among active attachment researchers from the developmental tradition by the end of the 1990s. See e.g. Stein, H., Jacobs, N.J., Ferguson, K.S., Allen, J.G., & Fonagy, P. (1998) What do adult attachment scales measure? Bulletin of the Menninger Clinic, 62, 33–82: ‘Both approaches have something important to offer, but are likely to be looking at different, though valid, constructs. The self-report measures are consistently measuring an aspect of intimate relationships that relates to individual perceptions of how relationships are managed’ (77).

175 Carnelley, K.B. & Brennan, K.A. (2002) Building bridges. Attachment & Human Development, 4(2), 189–92: ‘The biggest obstacle to the fledgling field of attachment theory appears to be a lack of co-operation among researchers wielding different measurement techniques. As it is, each researcher appears to be using a favorite measure and readers of their work are left to make the attempt to translate the results into their own measurement rubric before interpreting the meaning of the findings’ (191).

176 Bartholomew, K. & Shaver, P.R. (1998) Methods of assessing adult attachment: do they converge? In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.25–45). New York: Guilford, p.27.

178 Lapsley, D.K. & Quintana, S. (1985) Integrative themes in social and developmental theories of the self. In J. Pryor & J. Day (eds) The Development of Social Cognition (pp.153–78). New York: Springer, p.154. Also on the state of broader relations between developmental and social psychology in the period of the emergence of the two traditions of attachment research see Masters, J.C. & Yarkin-Levin, K. (1984) Boundary Areas in Social and Developmental Psychology. New York: Academic Press.

179 Bartholomew, K. & Shaver, P.R. (1998) Methods of assessing adult attachment: do they converge? In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.25–45). New York: Guilford, p.27.

180 Shaver, P.R. & Mikulincer, M. (2002) Attachment-related psychodynamics. Attachment & Human Development, 4(2), 133–61: ‘The AAI and CRI are focused entirely on individual differences in “state of mind with respect to attachment” and therefore do not reveal much about the normative workings of the attachment behavioral system’ (155).

181 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford, p.29.

182 Shaver, P.R. & Mikulincer, M. (2002) Attachment-related psychodynamics. Attachment & Human Development, 4(2), 133–61, p.134.

183 Brennan, K.A., Clark, C.L., & Shaver, P.R. (1998) Self-report measurement of adult romantic attachment: an integrative overview. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.46–76). New York: Guilford, p.46. See also Sassenberg, K. & Ditrich, L. (2019) Research in social psychology changed between 2011 and 2016: larger sample sizes, more self-report measures, and more online studies. Advances in Methods and Practices in Psychological Science, 2(2).

184 Ensink, K., Borelli, J.L., Roy, J., Normandin, L., Slade, A., & Fonagy, P. (2019) Costs of not getting to know you: lower levels of parental reflective functioning confer risk for maternal insensitivity and insecure infant attachment. Infancy, 24(2), 210–27.

185 Shaver, P.R & Mikulincer, M. (2002) Dialogue on adult attachment: diversity and integration. Attachment & Human Development, 4(2), 243–57, p.245.

186 Though see e.g. Rholes, W.S., Simpson, J.A., & Friedman, M. (2006) Avoidant attachment and the experience of parenting. Personality and Social Psychology Bulletin, 32(3), 275–85.

187 A meta-analysis by Lo and colleagues, however, found more studies using self-report measures (12) than using the AAI (5). The odds ratio for child maltreatment for self-report measures was 2.75; the odds ratio for the AAI was 5. Lo, C.K., Chan, K.L., & Ip, P. (2019) Insecure adult attachment and child maltreatment: a meta-analysis. Trauma, Violence, & Abuse, 20(5).

188 Crowell, J.A. & Waters, E. (1994) Bowlby’s theory grown up: the role of attachment in adult love relationships. Psychological Inquiry, 5(1), 31–4, pp.33–4. See also Kobak, R. (2009) Defining and measuring of attachment bonds: comment on Kurdek (2009). Journal of Family Psychology, 23(4), 447–9.

189 E.g. Shaver, P.R. & Mikulincer, M. (2010) Mind–behavior relations in attachment theory and research. In C.R. Agnew, D.E. Carlston, W.G. Graziano, & J.R. Kelly (eds) Then a Miracle Occurs: Focusing on Behavior in Social Psychological Theory and Research (pp.342–67). New York: Oxford University Press: ‘Ainsworth et al. (1978) provided persuasive evidence for the impact of parental behavior on the formation of an infant’s attachment style’ (358). See also Levy, K.N., Blatt, S.J., & Shaver, P.R. (1998) Attachment styles and parental representations. Journal of Personality and Social Psychology, 74(2), 407–419, p.408.

190 Shaver, P.R. & Mikulincer, M. (2012) An attachment perspective on coping with existential concerns. In P.R. Shaver & M. Mikulincer (eds) Meaning, Mortality, and Choice: The Social Psychology of Existential Concerns (pp.291–307). Washington, DC: American Psychological Association, p.293.

191 Belsky, J. & Cassidy, J. (1994) Attachment and close relationships: an individual-difference perspective. Psychological Inquiry, 5(1), 27–30, p.27.

192 Shaver, P.R. (1993) Where do adult romantic attachment styles come from? Paper presented at symposium entitled ‘Mental Representations of Relationships: Intergenerational and Temporal Continuity’, Meeting of the Society for Research in Child Development, New Orleans, 27 March 1993. https://adultattachment.faculty.ucdavis.edu/wp-content/uploads/sites/66/2014/06/Shaver.pdf.

193 Cassidy, J. & Shaver, P. (1999) Preface. In The Handbook of Attachment, 1st edn. New York: Guilford, p.xiv.

194 Crowell, J.A., Fraley, R.C., & Shaver, P.R. (1999) Measures of individual differences in adolescent and adult attachment. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications (pp.434–65). New York: Guilford, p.452. The passage would be repeated verbatim in the 2008 edition on p.618.

195 Shaver, P.R., Belsky, J., & Brennan, K.A. (2000) The adult attachment interview and self-reports of romantic attachment: associations across domains and methods. Personal Relationships, 7(1), 25–43.

196 Ibid. p.39.

197 Roisman, G.I., Holland, A., Fortuna, K., Fraley, R.C., Clausell, E., & Clarke, A. (2007) The Adult Attachment Interview and self-reports of attachment style: an empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–97. A point of comparison here are findings showing that indirect (implicit) and direct (explicit) measures of self-esteem are virtually unrelated. Like ‘attachment’, ‘self-esteem’ is likely an umbrella term within psychological discourse, and captures more than one autonomous process, even if the correlates of these processes may be similar—producing the superficial effect of a single phenomenon. Pietschnig, J., Gittler, G., Stieger, S., et al. (2018) Indirect (implicit) and direct (explicit) self-esteem measures are virtually unrelated: a meta-analysis of the initial preference task. PLoS One, 13(9), e0202873.

198 Roisman, G.I., Holland, A., Fortuna, K., Fraley, R.C., Clausell, E., & Clarke, A. (2007) The Adult Attachment Interview and self-reports of attachment style: an empirical rapprochement. Journal of Personality and Social Psychology, 92(4), 678–97, p.689. More recently, in an unpublished doctoral study by Watkins, 87 participants (43 with borderline personality disorder) completed the ECR and AAI. There was no association with attachment categories. But when a variable was used that combined scales for preoccupied and unresolved states of mind (Chapter 3), this was associated with ECR anxiety (r = .34). Watkins, C.D. (2016) Convergence versus divergence of social and developmental measures of adult attachment: testing Jay Belsky’s proposals. Unpublished doctoral thesis, University of Tennessee.

199 Fortuna, K. & Roisman, G.I. (2008) Insecurity, stress, and symptoms of psychopathology: contrasting results from self-reports versus interviews of adult attachment. Attachment & Human Development, 10(1), 11–28.

200 For other relevant studies see Mayseless, O., Sharabany, R., & Sagi, A. (1997) Attachment concerns of mothers as manifested in parental, spousal, and friendship relationships. Personal Relationships, 4(3), 255–69; Volling, B.L., Notaro, P.C., & Larsen, J.J. (1998) Adult attachment styles: relations with emotional well-being, marriage, and parenting. Family Relations, 47(4), 355–67; Laurent, H.K., Kim, H.K., & Capaldi, D.M. (2008) Prospective effects of interparental conflict on child attachment security and the moderating role of parents’ romantic attachment. Journal of Family Psychology, 22(3), 377–88; Howard, K.S. (2010) Paternal attachment, parenting beliefs and children’s attachment. Early Child Development and Care, 18(1–2), 157–71.

201 Bernier, A. & Matte-Gagné, C. (2011) More bridges: investigating the relevance of self-report and interview measures of adult attachment for marital and caregiving relationships. International Journal of Behavioral Development, 35(4), 307–316, p.313. This outcome was generally predicted by Mikulincer and Cowans a decade earlier, though they anticipated some association based on the expectation that marital satisfaction will influence caregiving: Mikulincer, M., Florian, V., Cowan, P.A., & Cowan, C.P. (2002) Attachment security in couple relationships: a systemic model and its implications for family dynamics. Family Process, 41(3), 405–434, p.424.

202 Shaver, P.R. (2006) Dynamics of romantic love: comments, questions, and future directions. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.423–56). New York: Guilford: ‘Most of us self-report attachment researchers tend to ignore the AAI literature (although Bartholomew, Furman and Simpson are important exceptions)’ (445).

203 Shaver, P.R. & Mikulincer, M. (2002) Dialogue on adult attachment: diversity and integration. Attachment & Human Development, 4, 243–57.

204 Part of the contribution made by Attachment & Human Development as a topic-specific journal was that it could overcome some of the division between developmental and social psychology as subdisciplines. Compare, for instance, the Journal of Social and Personal Relationships. Mikulincer served as Editor-in-Chief from 2010 to 2015. Though an important venue for attachment research in the social psychology tradition, Mikulincer’s premiership did not coincide with publications in the journal by researchers in the developmental tradition of attachment research. It may well have been that articles were not solicited, submitted, accepted—or all three.

205 E.g. Duggan, A., Berlin, L., Cassidy, J., Burrell, L., & Tandon, D. (2009) Examining maternal depression and attachment insecurity as moderators of the impacts of home visiting for at-risk mothers and infants. Journal of Consulting and Clinical Psychology, 77(4), 788–99.

206 Cassidy, J., Poehlmann, J., & Shaver, P.R. (2010) An attachment perspective on incarcerated parents and their children: introduction to the special issue. Attachment & Human Development, 12, 285–8.

207 Simpson, J., Rholes, W.S., Orina, M.M., & Grich, J. (2002) Working models of attachment, support giving, and support seeking in a stressful situation. Personality and Social Psychology Bulletin, 28(5), 598–608.

208 Granqvist, P., Hesse, E., Fransson, M., Main, M., Hagekull, B., & Bohlin, G. (2016) Prior participation in the strange situation and overstress jointly facilitate disorganized behaviours: implications for theory, research and practice. Attachment & Human Development, 18(3), 235–49.

209 Granqvist, P., Mikulincer, M., & Shaver, P.R. (2010) Religion as attachment: normative processes and individual differences. Personality and Social Psychology Review, 14(1), 49–59. A further, more limited example of transcendence of the opposition between the developmental and social psychological traditions is Pasco Fearon. Though one of the leaders of the new generation of developmental attachment researchers, Fearon nonetheless incorporated psychometric methodology from the social psychological tradition in developing a brief version of the Attachment Q-Sort (Chapter 2). Cadman, T., Belsky, J., & Fearon, R.P. (2018) The Brief Attachment Scale (BAS-16): a short measure of infant attachment. Child: Care, Health and Development, 44(5), 766–75.

210 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood, 2nd edn. New York: Guilford.

211 Heckendorf, E., Huffmeijer, R., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2016) Neural processing of familiar and unfamiliar children’s faces: effects of experienced love withdrawal, but no effects of neutral and threatening priming. Frontiers in Human Neuroscience, 10, 231.

212 Sroufe, L.A. (2016) The place of attachment in development. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research, and Clinical Applications, 3rd edn (pp.997–1011). New York: Guilford, p.999.

213 Jones, J.D., Cassidy, J., & Shaver, P.R. (2015) Parents’ self-reported attachment styles: a review of links with parenting behaviors, emotions, and cognitions. Personality and Social Psychology Review, 19(1), 44–76, p.69.

214 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford, p.26.

215 Roisman, G.I. (2009) Adult attachment: toward a rapprochement of methodological cultures. Current Directions in Psychological Science, 18(2), 122–6.

216 Shaver, P.R. & Fraley, R.C. (2000) Attachment theory and caregiving. Psychological Inquiry, 11, 109–114, p.111.

217 Shaver, P.R & Mikulincer, M. (2002) Dialogue on adult attachment: diversity and integration. Attachment & Human Development, 4(2), 243–57, p.246.

218 Ibid. See also Shaver, P.R. & Mikulincer, M. (2002) Attachment-related psychodynamics. Attachment & Human Development, 4(2), 133–61, p.155.

219 The potential for imprecise or diverging definitions of concepts to reduce clarity in communication is acknowledged in various places in Shaver and Mikulincer’s work, even from their early publications, e.g. Mikulincer, M. & Caspy, T. (1986) The conceptualization of helplessness: I. A phenomenological-structural analysis. Motivation and Emotion, 10, 263–78: ‘Many psychological concepts are not well defined, so their meanings vary somewhat from person to person. Variations in usage are not surprising because psychological phenomena are often abstract terms that summarize the verbal report of a lay person regarding his perceptions and behavior in reference to specific real-life situations. Because the terms are imprecise in meaning, they also generate low interjudge reliability’ (263–4).

220 Ainsworth, M. (1990, 2010) Security and attachment. In R. Volpe (ed.) The Secure Child: Timeless Lessons in Parenting and Childhood Education (pp.43–53). Charlotte, NC: Information Age Publishing, p.49.

221 Ainsworth, M.D.S. (1985) Attachments across the life span. Bulletin of the New York Academy of Medicine, 61, 792–812.

222 Ansbro, M. (2018) Integrating attachment theory into probation practice: a qualitative study. British Journal of Social Work, 48(8), 2235–52.

223 Ainsworth’s emphasis on a secure base, a safe haven, and separation anxiety was inherited by Waters—though often, confusingly, he has used the term ‘secure base’ to refer collectively to the three components, and especially the first two. Waters regarded the concept of ‘safe haven’ as partially superseded by the concept of ‘felt security’, with anxiety as the inverse of security (Chapter 2). To him, this made the feeling of security the overarching concept, linking infant behaviour to adult experience. It also contributed to his frustration with the focus of other attachment researchers on separations and reunions at the expense of attention to ordinary experiences that contribute to feelings of trust and security in the relationship. Crowell, J. & Waters, E. (1989) Separation anxiety. In M. Lewis & S. Miller (eds) Handbook of Developmental Psychopathology (pp.209–218). New York: Plenum Press.

224 Main, M., Hesse, E., & Kaplan, N. (2005) Predictability of attachment behavior and representational processes at 1, 6, and 19 years of age: the Berkeley Longitudinal Study. In K.E. Grossmann, K.Grossmann, & E. Waters (eds) Attachment from Infancy to Adulthood: The Major Longitudinal Studies (pp.245–304). New York: Guilford, p.256.

225 E.g. Crowell, J.A., Treboux, D., & Waters, E. (1999) The Adult Attachment Interview and the Relationship Questionnaire: relations to reports of mothers and partners. Personal Relationships, 6(1), 1–18. For an argument for the return of the concept of strategy to its narrower origins in ethology see Stevenson-Hinde, J. (1994) An ethological perspective. Psychological Inquiry, 5(1), 62–5.

226 Van IJzendoorn, M.H., Tavecchio, L.W.C., Goossens, F.A., & Vergeer, M.M. (1982) Opvoeden in Geborgenheid: Een Kritische Analyse van Bowlby’s Attachmenttheorie. Amsterdam: Van Loghum Slaterus, p.59.

227 Van IJzendoorn, M.H., Goossens, F.A., Tavecchio, L.W.C., Vergeer, M.M., & Hubbard, F.O.A. (1983) Attachment to soft objects: its relationship with attachment to the mother and with thumbsucking. Child Psychiatry and Human Development, 14(2), 97–105; van IJzendoorn, M.H., Sagi, A., & Lambermon, M.W.E. (1992) The multiple caregiver paradox. Some Dutch and Israeli data. New Directions for Child Development, 57, 5–25, p.9.

228 Schuengel, C. & van IJzendoorn, M.H. (2001) Attachment in mental health institutions: a critical review of assumptions, clinical implications, and research strategies. Attachment & Human Development, 3(3), 304–323, p.307. See also Harder, A.T., Knorth, E.J., & Kalverboer, M.E. (2013) A secure base? The adolescent–staff relationship in secure residential youth care. Child & Family Social Work, 18, 305–317.

229 Van IJzendoorn, M.H. & Bakermans-Kranenburg, M.J. (2010) Stretched until it snaps: attachment and close relationships. Child Development Perspectives, 4(2), 109–111.

230 Verhage, M.L., Schuengel, C., Fearon, R.P., et al. (2017) Failing the duck test: reply to Barbaro, Boutwell, Barnes, and Shackelford (2017). Psychological Bulletin, 143(1), 114–16, p.114. This is in contrast, for instance, to Bernier and Dozier who likewise worry that the broader use of the concept of ‘attachment’ jeopardises its meaning, but who still described self-report measures and the AAI as assessing individual differences in a single behavioural system, even if they tap different aspects. Bernier, A. & Dozier, M. (2002) Assessing adult attachment: empirical sophistication and conceptual bases. Attachment & Human Development, 4(2), 171–9.

231 E.g. Schore, A.N. (2001) Minds in the making: attachment, the self-organizing brain, and developmentally-oriented psychoanalytic psychotherapy. British Journal of Psychotherapy, 17(3), 299–328; Hughes, D. (2007) Attachment-Focused Family Therapy. New York: Norton; Sacco, F.C., Twemlow, S.W., & Fonagy, P. (2008) Secure attachment to family and community. Smith College Studies in Social Work, 77(4), 31–51; Johnson, S.M. (2019) Attachment Theory in Practice. New York: Guilford.

232 Crittenden, P.M., Dallos, R., Landini, A., & Kozlowska, K. (2014) Attachment and Family Therapy. London: McGraw-Hill; Crittenden, P.M. & Landini, A. (2015) Attachment relationships as semiotic scaffolding systems. Biosemiotics, 8(2), 257–73.

233 This consistency over time is well illustrated by Cassidy’s opening chapters to the three editions of the Handbook of Attachment, in 1999, 2008, and 2016.

234 Stevenson-Hinde, J. (1994) An ethological perspective. Psychological Inquiry, 5(1), 62–5, p.63, citing Hazan, C. & Shaver, P.R. (1994) Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22.

235 Hazan, C. & Shaver, P.R. (1994) Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22: ‘The process of attachment formation, at any age, is hypothesized to involve the same sequence: proximity seeking followed by safe-haven behavior followed by the establishment of a secure base’ (12).

236 Hazan, C., Hutt, M.J., Sturgeon, J., & Bricker, T. (1991) The process of relinquishing parents as attachment figures. Paper presented at the Biennial Meetings of the Society for Research in Child Development, Seattle, WA. These findings were replicated by later studies: Fraley, R.C. & Davis, K.E. (1997) Attachment formation and transfer in young adults’ close friendships and romantic relationships. Personal Relationships, 4, 131–44; Trinke, S.J. & Bartholomew, K. (1997) Hierarchies of attachment relationships in young adulthood. Journal of Social and Personal Relationships, 14, 603–625; Nickerson, A.B. & Nagle, R.J. (2005) Parent and peer attachment in late childhood and early adolescence. Journal of Early Adolescence, 25(2), 223–49.

237 Hazan, C. & Shaver, P.R. (1994) Attachment as an organizational framework for research on close relationships. Psychological Inquiry, 5(1), 1–22, p.1. This was, in fact, a limitation criticised by social psychologists in the 1990s, who felt that any comprehensive theory in social psychology should deal with relationships in general, not just intimate ones, e.g. Duck, S. (1994) Attaching meaning to attachment. Psychological Inquiry, 5(1), 34–8.

238 Hence statements such as ‘attachment-related avoidance is a tendency to withdraw from situations involving intimacy’. Brassard, A., Darveau, V., Péloquin, K., Lussier, Y., & Shaver, P.R. (2014) Childhood sexual abuse and intimate partner violence in a clinical sample of men: the mediating roles of adult attachment and anger management. Journal of Aggression, Maltreatment & Trauma, 23(7), 683–704, p.686. Other attachment researchers, however, have treated intimacy and attachment as distinct constructs. Karantzas, G.C., Feeney, J.A., Goncalves, C.V., & McCabe, M.P. (2014) Towards an integrative attachment-based model of relationship functioning. British Journal of Psychology, 105(3), 413–34.

239 Mikulincer, M., Gillath, O., & Shaver, P.R. (2002) Activation of the attachment system in adulthood: threat-related primes increase the accessibility of mental representations of attachment figures. Journal of Personality and Social Psychology, 83(4), 881–95, p.892.

240 Fraley, R.C. & Davis, K.E. (1997) Attachment formation and transfer in young adults’ close friendships and romantic relationships. Personal Relationships, 4, 131–44.

241 Trinke and Bartholomew found an association of .45 between the WHOTO and a measure of ‘supportive’ relationships, suggesting that the WHOTO is materially related to supportiveness, but not reducible to it. They interpreted this finding as indicating that supportiveness may contribute to the formation of attachment components in a relationship, but that many relationships with attachment components in adulthood form and endure without supportiveness. Trinke, S.J. & Bartholomew, K. (1997) Hierarchies of attachment relationships in young adulthood. Journal of Social and Personal Relationships, 14, 603–625. More recently, Gillath and colleagues have argued against the use of self-reported separation anxiety as an index of an attachment relationship, as they anticipate that participants have to speculate about what it would be like to lose the other person, whereas they can report more concretely on secure base and safe haven use. Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, p.54.

242 The methodology had been introduced into attachment research a decade earlier by Baldwin, M.W., Fehr, B., Keedian, E., & Seidel, M. (1993) An exploration of the relational schemata underlying attachment styles: self-report and lexical decision approaches. Personality and Social Psychology Bulletin, 19, 746–54.

243 Mikulincer, M., Gillath, O., & Shaver, P.R. (2002) Activation of the attachment system in adulthood: threat-related primes increase the accessibility of mental representations of attachment figures. Journal of Personality and Social Psychology, 83(4), 881–95, p.891. Aligned findings were later reported in Edelstein, R.S. & Gillath, O. (2008) Avoiding interference: adult attachment and emotional processing biases. Personality and Social Psychology Bulletin, 34(2), 171–81. The researchers found that an avoidant attachment style interfered with the availability of only attachment-relevant emotional words, and not emotional words not relevant to attachment.

244 Dykas, M.J. & Cassidy, J. (2011) Attachment and the processing of social information across the life span: theory and evidence. Psychological Bulletin, 137(1), 19–46.

245 Bernier, A., Larose, S., & Whipple, N. (2005) Leaving home for college: a potentially stressful event for adolescents with preoccupied attachment patterns. Attachment & Human Development, 7(2), 171–85.

246 Fonagy, P. & Luyten, P. (2009) A developmental, mentalization-based approach to the understanding and treatment of borderline personality disorder. Development & Psychopathology, 21(4), 1355–81.

247 Granqvist might be anticipated as an exception, given that his work spans both traditions. However, the only occasions on which he has cited the paper are in works actually co-authored with Mikulincer. Roisman is another figure straddling the two traditions, and cites the paper in a report co-authored with Chris Fraley: Haydon, K.C., Roisman, G.I., Marks, M.J., & Fraley, R.C. (2011) An empirically derived approach to the latent structure of the Adult Attachment Interview: additional convergent and discriminant validity evidence. Attachment & Human Development, 13(5), 503–524.

248 Fraley and Shaver demonstrated that adult attachment style predicts observable behaviour on separations, accounting for around 8% of variance. Fraley, R.C. & Shaver, P.R. (1998) Airport separations: a naturalistic study of adult attachment dynamics in separating couples. Journal of Personality and Social Psychology, 75(5), 1198–212. Mikulincer and colleagues would later supply evidence that anxious and avoidant attachment styles bias motor responses on a push-pull task in response to attachment primes. Mikulincer, M., Shaver, P.R., Bar-On, N., & Ein-Dor, T. (2010) The pushes and pulls of close relationships: attachment insecurities and relational ambivalence. Journal of Personality and Social Psychology, 98(3), 450–68.

249 Shaver, P.R., Hazan, C., & Bradshaw, D. (1988) Love as attachment: the integration of three behavioral systems. In R.J. Sternberg & M. Barnes (eds) The Psychology of Love (pp.68–99). New Haven, CT: Yale University Press, p.73.

250 Zilcha-Mano, S., Mikulincer, M., & Shaver, P.R. (2011) An attachment perspective on human–pet relationships: conceptualization and assessment of pet attachment orientations. Journal of Research in Personality, 45(4), 345–57, p.354.

252 Birnbaum, G.E., Orr, I., Mikulincer, M., & Florian, V. (1997) When marriage breaks up-does attachment style contribute to coping and mental health? Journal of Social and Personal Relationships, 14(5), 643–54.

253 Johnson, S.M. (2003) Introduction to attachment: a therapist’s guide to primary relationships and their renewal. In S.M. Johnson & V. Whiffen (eds) Attachment Processes in Couple & Family Therapy (pp.3–17). New York: Guilford, p.10.

254 E.g. Cassidy, J. (2001) Truth, lies, and intimacy: an attachment perspective. Attachment & Human Development, 3(2), 121–55.

255 Fraley, R.C., Davis, K.E., & Shaver, P.R. (1998) Dismissing avoidance and the defensive organization of emotion, cognition, and behavior. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.249–79). New York: Guilford.

256 Fraley, R.C., Heffernan, M.E., Vicary, A.M., & Brumbaugh, C.C. (2011) The Experiences in Close Relationships—Relationship Structures Questionnaire: a method for assessing attachment orientations across relationships. Psychological Assessment, 23(3), 615–25, Table 2.

257 On associations with affiliation with friends see e.g. Mikulincer, M. & Selinger, M. (2001) The interplay between attachment and affiliation systems in adolescents’ same-sex friendships: the role of attachment style. Journal of Social and Personal Relationships, 18(1), 81–106; Mayseless, O. & Scharf, M. (2007) Adolescents’ attachment representations and their capacity for intimacy in close relationships. Journal of Research on Adolescence, 17(1), 23–50.

258 E.g. Mills-Koonce, W.R., Appleyard, K., Barnett, M., Deng, M., Putallaz, M., & Cox, M. (2011) Adult attachment style and stress as risk factors for early maternal sensitivity and negativity. Infant Mental Health Journal, 32(3), 277–85.

259 Simpson, J., Rholes, W.S., Orina, M.M., & Grich, J. (2002) Working models of attachment, support giving, and support seeking in a stressful situation. Personality and Social Psychology Bulletin, 28(5), 598–608.

260 E.g. Edelstein, R.S., Alexander, K.W., Shaver, P.R., et al. (2004) Adult attachment style and parental responsiveness during a stressful event. Attachment & Human Development, 6(1), 31–52.

261 For instance, convergence between the ECR and scales for inferred childhood experience coded on the AAI has been substantial for avoidance (r = .30 and .41) and material for anxiety (r = .24 and .22). Haydon, K.C., Roisman, G. I., Marks, M.J., & Fraley, R.C. (2011) An empirically derived approach to the latent structure of the Adult Attachment Interview: additional convergent and discriminant validity evidence. Attachment & Human Development, 13(5), 503–524.

262 Cf. Storebø, O.J., Skoog, M., Rasmussen, P.D., et al.(2014) Attachment competences in children with ADHD during the Social-Skills Training and Attachment (SOSTRA) randomized clinical trial. Journal of Attention Disorders, 19, 865–71.

263 Bachmann, C.J., Wijlaars, L.P., Kalverdijk, L.J., et al. (2017) Trends in ADHD medication use in children and adolescents in five western countries, 2005–2012. European Neuropsychopharmacology, 27(5), 484–93; Anderson, K.N., Ailes, E.C., Danielson, M., et al. (2018) Attention-deficit/hyperactivity disorder medication prescription claims among privately insured women aged 15–44 years—United States, 2003–2015. Morbidity and Mortality Weekly Report, 67(2), 66.

264 Fraley, R.C. & Shaver, P.R. (2000) Adult romantic attachment: theoretical developments, emerging controversies, and unanswered questions. Review of General Psychology, 4(2), 132–54: ‘We hypothesize that anxiety-reducing drugs affect the intensity but not the avoidant-nonavoidant orientation of attachment behaviors’ (146).

265 Bernier, A. & Dozier, M. (2002) Assessing adult attachment: empirical sophistication and conceptual bases. Attachment & Human Development, 4(2), 171–9.

266 Shaver, P. & Mikulincer, M. (2004) Attachment in the later years: a commentary. Attachment & Human Development, 6(4), 451–64, p.454.

267 Mikulincer, M. & Shaver, P.R. (2007) Reflections on security dynamics: core constructs, psychological mechanisms, relational contexts, and the need for an integrative theory. Psychological Inquiry, 18, 197–209, p.208.

268 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press, p.57. Compare the definition of behavioural system offered in Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.72. In this later text the elements especially highlighted are the first, third, and fourth. There is no mention, for instance, of cognitive components as required as part of the definition of a behavioural system.

269 Mikulincer, M. & Shaver, P.R. (2012) Attachment theory expanded: a behavioral systems approach to personality. In K. Deaux & M. Snyder (eds) Oxford Handbook of Personality and Social Psychology (pp.467–92). Oxford: Oxford University Press, p.473.

270 Ibid.: ‘The primary strategy of the exploration system, seeking new information about oneself and the world, is activated whenever a person encounters novel situations, objects, or people or experiences novel internal states that contradict or challenge existing knowledge and working models’ (473).

271 Mikulincer, M. & Shaver, P. (2008) Contributions of attachment theory and research to motivation science. In J.Y. Shah & W.L. Gardner (eds) Handbook of Motivational Science (pp.201–216). New York: Guilford, p.204.

272 An interesting case discussed by Shaver in his early work is that of prank calls to the fire service. In a subtle and thoughtful paper, Shaver speculated that implicated in these prank calls was a pervasive combination of anger, boredom, and powerlessness, and a lack of other legitimate situations for expressing these feelings. In terms of his later thinking, this might be described as the interaction of the anti-goals of the exploratory systems and dominance systems. Shaver, P.R., Schurtman, R., & Blank, T. (1975) Conflict between ghetto-dwellers and firemen: environmental and attitudinal factors. Journal of Applied Social Psychology, 5, 240–61.

273 Mikulincer, M. & Shaver, P. (2008) Contributions of attachment theory and research to motivation science. In J.Y. Shah & W.L. Gardner (eds) Handbook of Motivational Science (pp.201–216). New York: Guilford, p.204.

274 Mikulincer, M. & Shaver, P.R. (2012) Attachment theory expanded: a behavioral systems approach to personality. In K. Deaux & M. Snyder (eds) Oxford Handbook of Personality and Social Psychology (pp.467–92). Oxford: Oxford University Press, pp.473–4.

275 E.g. Mikulincer, M. (2006) Attachment, caregiving and sex within romantic relationships. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.23–44). New York: Guilford: ‘The profound joy and affection, self-protective anxiety, numbing boredom, corrosive anger, lustful passion, uncontrollable jealousy, and intense sorrow experienced in romantic relationships are reflections on the central importance of these behavioural systems’ (23–4).

276 Hart, J.J., Shaver, P.R., & Goldenberg, J.L. (2005) Attachment, self-esteem, worldviews, and terror management: evidence for a tripartite security system. Journal of Personality and Social Psychology, 88, 999–1013.

277 Hart, J. (2014) Toward an integrative theory of psychological defense. Perspectives on Psychological Science, 9, 19–39.

278 Shaver, P.R. & Mikulincer, M. (2003) The psychodynamics of social judgments: an attachment theory perspective. In J.P. Forgas, K.D. Williams, & W. von Hippel (eds) Social Judgments: Implicit and Explicit Processes (pp.85–114). Philadelphia: Psychology Press, p.88.

279 Mikulincer, M. (1997) Adult attachment style and information processing: individual differences in curiosity and cognitive closure. Journal of Personality and Social Psychology, 72(5), 1217–30.

280 Banai, E., Mikulincer, M., & Shaver, P.R. (2005) ‘Selfobject’ needs in Kohut’s self psychology. Psychoanalytic Psychology, 22(2), 224–60, p.251.

281 Ibid. p.251.

282 Mikulincer, M. & Shaver, P.R. (2007) Boosting attachment security to promote mental health, prosocial values, and inter-group tolerance. Psychological Inquiry, 18, 139–56, p.152.

283 Shaver, P.R. (2006) Dynamics of romantic love: comments, questions, and future directions. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.423–56). New York: Guilford, p.432.

284 Fraley, R.C., Roisman, G.I., Booth-LaForce, C., Owen, M.T., & Holland, A.S. (2013) Interpersonal and genetic origins of adult attachment styles: a longitudinal study from infancy to early adulthood. Journal of Personality and Social Psychology, 104(5), 817–38, web-based supplement C. Additionally, a later study using the same sample demonstrated that the ECR has essentially no relationship with secure base scripts in adulthood. Steele, R.D., Waters, T.E., Bost, K.K., et al. (2014) Caregiving antecedents of secure base script knowledge: a comparative analysis of young adult attachment representations. Developmental Psychology, 50(11), 2526–38.

285 Bowlby, J. (1969, 1982) Attachment., London: Penguin, p.233.

286 Ainsworth, M. (1984) Attachment. In N.S. Endler & J. McVicker Hunt (eds) Personality and the Behavioral Disorders (pp.559–602). New York: Wiley, p.596.

287 Fonagy, P. (2008) A genuinely developmental theory of sexual enjoyment and its implications for psychoanalytic technique. Journal of the American Psychoanalytic Association, 56(1), 11–36; Crittenden, P.M. (2015) Raising Parents, 2nd edn. London: Routledge. The inattention to sexuality among attachment researchers after Bowlby matched a trend in post-war psychoanalytic theory to downplay the topic. Some of this was blamed on attachment theory. See Zamanian, K. (2011) Attachment theory as defense: what happened to infantile sexuality? Psychoanalytic Psychology, 28(1), 33–47.

288 Mottier, V. (1997) Sex and discourse. The politics of the Hite reports. In T. Carver & M. Hyvärinen (eds) Interpreting the Political: New Methodologies (pp.39–59). London: Routledge; Epstein, S. (2006) The new attack on sexuality research: morality and the politics of knowledge production. Sexuality Research and Social Policy Journal of NSRC, 3(1), 1.

289 Morgan, H.J. & Shaver, P.R. (1999) Attachment processes and commitment to romantic relationships. In J.M. Adams & W.H. Jones (eds) Handbook of Interpersonal Commitment and Relationship Stability (pp.109–24). New York: Plenum Press, p.111.

290 Mikulincer, M. (2006) Attachment, caregiving and sex within romantic relationships. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.23–44). New York: Guilford.

291 Stevenson-Hinde, J. (1994) An ethological perspective. Psychological Inquiry, 5(1), 62–5.

292 Mikulincer, M. & Shaver, P.R. (2012) Attachment theory expanded: a behavioral systems approach to personality. In K. Deaux & M. Snyder (eds) Oxford Handbook of Personality and Social Psychology (pp.467–92). Oxford: Oxford University Press: ‘As evolutionary psychologists have explained, however, the proximal motivation for an act (i.e., wishing to have sex with an attractive person) need not be the same as the evolutionary reason for the existence of the motives involved’ (475).

293 Gillath, O., Mikulincer, M., Birnbaum, G.E., & Shaver, P.R. (2008) When sex primes love: subliminal sexual priming motivates relationship goal pursuit. Personality and Social Psychology Bulletin, 34(8), 1057–69.

294 Szepsenwol, O., Mikulincer, M., & Birnbaum, G.E. (2013) Misguided attraction: the contribution of normative and individual-differences components of the sexual system to mating preferences. Journal of Research in Personality 47, 196–200, p.197.

295 Gillath, O., Mikulincer, M., Birnbaum, G.E., & Shaver, P.R. (2007) Does subliminal exposure to sexual stimuli have the same effects on men and women? Journal of Sex Research, 44(2), 111–21.

296 Ibid. p.119.

297 Mikulincer, M. (2007) Building personal relationships theory. Personal Relationships, 14(3), i–iv: ‘In general, I have noticed informally that few gender differences appear in attachment-related studies of individuals, yet they are common when both members of couples are studied’ (ii).

298 Brassard, A., Shaver, P.R., & Lussier, Y. (2007) Attachment, sexual experience, and sexual pressure in romantic relationships: a dyadic approach. Personal Relationships, 14(3), 475–93.

299 Little and colleagues found that an avoidant attachment style was unrelated to marital satisfaction among partners who had frequent sex, and that an anxious attachment style was unrelated to marital satisfaction among partners who frequently had pleasurable sex. Such findings suggest that sex can serve as an alternative source of assurance of partner availability, even for individuals who would find it otherwise difficult to seek (avoidant) or feel assured of (anxious) this availability in other aspects of their relationship. Little, K.C., McNulty, J.K., & Russell, V.M. (2010) Sex buffers intimates against the negative implications of attachment insecurity. Personality and Social Psychology Bulletin, 36(4), 484–98.

300 Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42.

301 Szepsenwol, O., Mikulincer, M., & Birnbaum, G.E. (2013) Misguided attraction: the contribution of normative and individual-differences components of the sexual system to mating preferences. Journal of Research in Personality 47, 196–200, p.197.

303 Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42, p.823; Mizrahi, M., Hirschberger, G., Mikulincer, M., Szepsenwol, O., & Birnbaum, G.E. (2016) Reassuring sex: can sexual desire and intimacy reduce relationship-specific attachment insecurities? European Journal of Social Psychology, 46(4), 467–80.

304 See also Collins, W.A. (2003) More than myth: the developmental significance of romantic relationships during adolescence. Journal of Research on Adolescence, 13(1), 1–24.

305 Birnbaum, G.E. (2018) The fragile spell of desire: a functional perspective on changes in sexual desire across relationship development. Personality and Social Psychology Review, 22(2), 101–127.

306 Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42.

307 Ibid. p.835.

308 Szepsenwol, O., Mizrahi, M., & Birnbaum, G.E. (2015) Fatal suppression: the detrimental effects of sexual and attachment deactivation within emerging romantic relationships. Social Psychological and Personality Science, 6(5), 504–512.

309 See also Birnbaum, G.E. & Finkel, E.J. (2015) The magnetism that holds us together: sexuality and relationship maintenance across relationship development. Current Opinion in Psychology, 1, 29–33.

310 Target, M. (2015) A developmental model of sexual excitement, desire and alienation. In A. Lemma & P.E. Lynch (eds) Sexualities (pp.43–62). London: Karnac.

311 Bowlby, J. (1969, 1982) Attachment. London: Penguin, p.227.

312 Bowlby, J. (1946) Psychology and democracy. The Political Quarterly, 17(1), 61–76; Mayhew, B. (2006) Between love and aggression: the politics of John Bowlby. History of the Human Sciences, 19(4), 19–35.

313 Smith, P.K. & Connolly, K. (1972) Patterns of play and social interaction in pre-school children. In N. Blurton Jones (ed.) Ethological Studies of Child Behaviour (pp.65–96). Cambridge: Cambridge University Press, p.68.

314 Robins, R.W., Tracy, J.L., & Shaver, P.R. (2001) Shamed into self-love: dynamics, roots, and functions of narcissism. Psychological Inquiry, 12(4), 230–36.

315 Ibid. p.233.

316 Ibid. p.234. The potential for tertiary strategies is not discussed explicitly by Shaver and colleagues. However, the existence of tertiary strategies or minor secondary strategies is suggested by Mikulincer and Shaver’s claim that anxiety and avoidance are the ‘major’, not only, secondary strategies, e.g. Mikulincer, M. & Shaver, P.R. (2012) Attachment theory expanded: a behavioral systems approach to personality. In K. Deaux & M. Snyder (eds) Oxford Handbook of Personality and Social Psychology (pp.467–92). Oxford: Oxford University Press. Mary Main and Erik Hesse, personal communication, August 2019, have agreed that controlling or dominant behaviour may be a ‘tertiary’ strategy if primary and conditional strategies fail. However, they would not term this a ‘conditional strategy’ since this is a technical term from ethology to refer to behavioural repetoires made available by human evolution for the purpose of survival. Dominance behaviour could be such a repertoire, but they are not sure and would regard a cross-species review as necessary for addressing the question.

317 Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.75.

318 Mikulincer, M. & Shaver, P.R. (2011) Attachment, anger, and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.241–57). Washington, DC: American Psychological Association, p.250.

320 Cf. Bauman, Z. (2001) Community: Seeking Safety in an Insecure World. Cambridge: Polity Press.

321 Mikulincer, M. & Shaver, P.R. (2012) Attachment theory expanded: a behavioral systems approach to personality. In K. Deaux & M. Snyder (eds) Oxford Handbook of Personality and Social Psychology (pp.467–92). Oxford: Oxford University Press, p.476.

322 Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.77.

323 Ibid. p.78.

324 Ibid. p.80–81.

325 Ibid. p.84.

326 Mikulincer, M. & Shaver, P.R. (2011) Attachment, anger, and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp. 241–57). Washington, DC: American Psychological Association, p.253.

327 Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.81.

328 It is possible that passive-aggressive behaviour is an example of the two dimensions coming together; though if so, then this raises the question of how anxiety, passivity, and aggression interrelate. Anxious attachment in Shaver and Mikulincer’s account emphasised only one of the three components of Ainsworth’s category, and left to the side aggression and/or passivity. Ainsworth explicitly characterised Group C infants as passive-aggressive in correspondence, e.g. Ainsworth, M. (1967) Letter to John Bowlby, 17 October 1967. PP/Bow/K.4/12. However, it could well be that this is not the kind of conjunction of passivity and aggression/coercion that Shaver and Mikulincer have in mind.

329 Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, p.120.

330 Shaver, P.R., Segev, M., & Mikulincer, M. (2011) A behavioral systems perspective on power and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.71–87). Washington, DC: American Psychological Association, p.76.

331 Bretherton, I. & Ainsworth, M.D.S. (1974) Responses of one-year-olds to a stranger in a strange situation. In M. Lewis & L. A. Rosenblum (eds) Origins of Fear (pp.131–64). New York: Wiley.

332 Crittenden, P.M. (1995) Attachment and psychopathology. In S. Goldberg, R. Muir, & J. Kerr (eds) John Bowlby’s Attachment Theory: Historical, Clinical and Social Significance (pp.367–406). New York: Analytical Press; Hilburn-Cobb, C. (2004) Adolescent psychopathology in terms of multiple behavioral systems. In L. Atkinson & S. Goldberg (eds) Attachment Issues in Psychopathology and Intervention (pp.95–135). Mahwah, NJ: Lawrence Erlbaum.

333 Verbeek, P. & de Waal, F.B. (2001) Peacemaking among preschool children. Peace and Conflict: Journal of Peace Psychology, 7(1), 5–28; Kutsukake, N. & Clutton-Brock, T.H. (2008) Do meerkats engage in conflict management following aggression? Reconciliation, submission and avoidance. Animal Behaviour, 75(4), 1441–53.

334 Mikulincer, M. & Shaver, P.R. (2011) Attachment, anger, and aggression. In P.R. Shaver & M. Mikulincer (eds) Human Aggression and Violence: Causes, Manifestations, and Consequences (pp.241–57). Washington, DC: American Psychological Association, p.250. See also Leedom, L.J. (2014) Human social behavioral systems: ethological framework for a unified theory. Human Ethology Bulletin, 29(1), 39–65.

335 An exception is Overall, N.C. (2019) Attachment insecurity and power regulation in intimate relationships. Current Opinion in Psychology, 25, 53–8.

336 Some relevant behaviours are mentioned on page 154 of the 2010 edition and page 150 of the 2016 edition: ‘avoidant people often entertain fantasies of perfection and power, exaggerate their achievements and talents, and avoid situations that challenge their defences and threaten their grandiosity’. However, no mention is made of the power behavioural system, or Mikulincer and Shaver’s own data discussed above, published by 2011, showing that hyperactivation of the power system is moderately associated with both avoidance and anxiety. Such findings would rather qualify the characterisation of ‘avoidant people’ in general.

337 This is discussed in Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42.

338 As well as working with Shaver, Kirkpatrick was also influenced by Bernard Spilka at Denver, a leading figure in the emergent specialism and president of Division 36 (1985–86). Spilka, B., Shaver, P., & Kirkpatrick, L.A. (1985) A general attribution theory for the psychology of religion. Journal for the Scientific Study of Religion, 24, 1–20.

339 The growth of psychology of religion in America reflected wider public interest in religious identities and movements, in the context of substantial sociological change. Marty, M.E. (1985) Transpositions: American religion in the 1980s. Annals of the American Academy of Political and Social Science, 480(1), 11–23. The growing basis for psychology of religion as a subdiscipline was expressed in and further spurred on in the specific academic context by the founding of the American Psychological Division 36 ‘Psychology of Religion’ in 1976. Paloutzian, R.F. (2017) Invitation to the Psychology of Religion, 3rd edn. New York: Guilford, Chapter 2.

340 Spilka, B., Shaver, P., & Kirkpatrick, L.A. (1985) A general attribution theory for the psychology of religion. Journal for the Scientific Study of Religion, 24(1), 1–20, p.16.

341 Ibid. p.11.

342 E.g. Kirkpatrick, L.A. & Shaver, P.R. (1990) Attachment theory and religion: childhood attachments, religious beliefs, and conversion. Journal for the Scientific Study of Religion, 29(3), 315–34: ‘The most striking (and perhaps obvious) point of contact, simply stated, is that the God of most Christian traditions seems to correspond very closely to the idea of a secure attachment figure’ (318).

343 Shaver, P.R. & O’Connor, C. (1986) Coping with stress: problems in perspective. In C. Tavris (ed.) Everywoman’s Emotional Wellbeing (pp.305–329). New York: Doubleday, p.326.

344 Ibid. p.336.

345 Ibid. p.327.

346 Kirkpatrick, L.A. & Shaver, P.R. (1990) Attachment theory and religion: childhood attachments, religious beliefs, and conversion. Journal for the Scientific Study of Religion, 29, 315–34, p.326.

347 Kirkpatrick, L.A. & Shaver, P.R. (1992) An attachment-theoretical approach to romantic love and religious belief. Personality and Social Psychology Bulletin, 18(3), 266–75, p.266. See also Kirkpatrick, L.A. (2005) Attachment, Evolution, and the Psychology of Religion. New York: Guilford.

348 Personal communication to Mary Main, cited in Granqvist, P. (2020) Attachment, Religion, and Spirituality: A Wider View. New York: Guilford.

349 On a mutual interest in religion as one contributing factor to the Shaver–Mikulincer collaboration see Shaver, P.R. & Mikulincer, M. (2011) Analysis of a collaborative working relationship. Relationship Research Newsletter, 9(2), 7–9.

350 Granqvist, P., Mikulincer, M., & Shaver, P.R. (2010) Religion as attachment: normative processes and individual differences. Personality and Social Psychology Review, 14(1), 49–59, p.50.

351 Ibid. p.52. See also Granqvist, P. (2020) Attachment, Religion, and Spirituality: A Wider View. New York: Guilford.

352 Goodman, G.S. (2006) Attachment to attachment theory: a personal perspective on an attachment researcher. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.3–22). New York: Guilford.

353 The extent and nature of this reformulation remains contested. See e.g. Shonin, E. (2016) This is not McMindfulness. The Psychologist, 29(2), 124–5.

354 Shaver, P.R., Lavy, S., Saron, C.D., & Mikulincer, M. (2007) Social foundations of the capacity for mindfulness: an attachment perspective. Psychological Inquiry, 18, 264–71.

355 Shaver, P.R., Mikulincer, M., Sahdra, B.K., & Gross, J.T. (2016) Attachment security as a foundation for kindness toward self and others. In K.W. Brown & M.R. Leary, (eds) The Oxford Handbook of Hypo-Egoic Phenomena, pp.223–42. Oxford: Oxford University Press, p.231.

356 Baer, R.A., Smith, G.T., Hopkins, J., Krietemeyer, J., & Toney, L. (2006) Using self-report assessment methods to explore facets of mindfulness. Assessment, 13(1), 27–45.

357 Shaver, P.R., Lavy, S., Saron, C.D., & Mikulincer, M. (2007) Social foundations of the capacity for mindfulness: an attachment perspective. Psychological Inquiry, 18, 264–71, pp.269–70.

358 E.g. Melen, S., Pepping, C.A., & O’Donovan, A. (2017) Social foundations of mindfulness: priming attachment anxiety reduces emotion regulation and mindful attention. Mindfulness, 8(1), 136–43.

359 Sahdra, B.K., & Shaver, P.R. (2013) Comparing attachment theory and Buddhist psychology. International Journal for the Psychology of Religion, 23(4), 282–93, p.287.

360 E.g. Juffer, F., Struis, E., Werner, C., & Bakermans-Kranenburg, M.J. (2017) Effective preventive interventions to support parents of young children: illustrations from the Video-feedback Intervention to promote Positive Parenting and Sensitive Discipline (VIPP-SD). Journal of Prevention & Intervention in the Community, 45(3), 202–214: ‘Secure attachment relationships are essential for children’s current and later development’ (202).

361 Waters, E., Corcoran, D., & Anafarta, M. (2005) Attachment, other relationships, and the theory that all good things go together. Human Development, 48(1–2), 80–84.

362 Shaver, P.R. (2006) Dynamics of romantic love: comments, questions, and future directions. In M. Mikulincer & G.S. Goodman (eds) Dynamics of Romantic Love: Attachment, Caregiving, and Sex (pp.423–56). New York: Guilford, pp.426–7. An illustrative case appears in Cooper, M.L., Shaver, P.R., & Collins, N.L. (1998) Attachment styles, emotion regulation, and adjustment in adolescence. Journal of Personality and Social Psychology, 74(5), 1380–97: ‘Results of our mediation analyses provide further support for the distinctiveness of the three types by raising the possibility that unique constellations of underlying processes account for the differential involvement of the three attachment groups in risky or problematic behaviors. Evidence supporting the distinctiveness of these profiles and, in particular, differences between the two insecure groups should help to mitigate concerns that attachment style differences can be summarized along a single good–bad, or secure–insecure dimension’ (1394). Here, the researchers distanced themselves from the idea that security–insecurity is the same as good–bad. However, their argument was that ‘bad’ divides into two distinct categories. They did not contest that security is equivalent to ‘good’.

363 Mehr, D.G. & Shaver, P.R. (1996) Goal structures in creative motivation. Journal of Creative Behavior, 30(2), 77–104, p.81.

364 Mikulincer, M. & Sheffi, E. (2000) Adult attachment style and cognitive reactions to positive affect: a test of mental categorization and creative problem solving. Motivation and Emotion, 24(3), 149–74.

365 Ainsworth, M. (1984) Attachment, adaptation and continuity. Paper presented at International Conference on Infant Studies, April 1984. PP/Bow/J.1/57.

366 Ibid.: ‘In the phylogenetic or evolutionary sense adaptation implies that in the course of natural selection those behaviours that yield survival advantage in the environment in which the evolutionary changes are taking place become part of the behavioural repertoire characteristic of a species … In the ontogenetic sense adaptation refers to the process through which an organism adjusts to its environment in the course of development.’

367 Ibid. An explicit scale for the extent to which an individual’s behaviour appears adaptive, in this sense, would later be developed by Steele, H., Steele, M., & Kriss, A. (2009) The Friends and Family Interview (FFI) Coding Guidelines. Unpublished manuscript: ‘This scale refers specifically to responses to the question asking what the respondent does when distressed or upset. An adaptive strategy may involve seeking comfort from others (e.g. parents, friends, or siblings), engaging in a favorite activity that relieves their unhappiness (e.g. listening to music, walking the dog), or simply thinking things through.’

368 E.g. Crittenden, P.M. (1992) Quality of attachment in the preschool years. Development & Psychopathology, 4(2), 209–241; Grossmann, K. (1995) Kontinuität und Konsequenzen der frühen Bindungsqualität während des Vorschulalters. In G. Spangler & P. Zimmermann (eds) Die Bindungstheorie—Grundlagen, Forschung und Anwendung (pp.191–202). Stuttgart: Klett Cotta.

369 Ainsworth, M. (1984) Attachment, adaptation and continuity. Paper presented at International Conference on Infant Studies, April 1984. PP/Bow/J.1/57.

370 E.g. Mikulincer, M. & Shaver, P.R. (2007) Boosting attachment security to promote mental health, prosocial values, and inter-group tolerance. Psychological Inquiry, 18, 139–56: ‘If human beings were helped by their families, communities, schools, religious institutions, and cultural media to become more secure, they would be better able to create a kinder and more tolerant, harmonious, and peaceful society’ (150); ‘without a sizeable proportion of secure, mindful, and self-efficacious citizens, political will alone is unlikely to accomplish desirable ethical goals’ (152).

371 Lawler, M.J., Shaver, P.R., & Goodman, G.S. (2011) Toward relationship-based child welfare services. Children and Youth Services Review, 33(3), 473–80, p.478.

372 Shaver, P.R. & O’Connor, C. (1986) Coping with stress: problems in perspective. In C. Tavris (ed.) Everywoman’s Emotional Wellbeing (pp.305–329). New York: Doubleday.

373 Ibid. p.323.

375 Mikulincer, M. & Florian, V. (1995) Appraisal of and coping with a real-life stressful situation: the contribution of attachment styles. Personality and Social Psychology Bulletin, 21(4), 406–414, p.413.

376 Davidovitz, R., Mikulincer, M., Shaver, P.R., Izsak, R., & Popper, M. (2007) Leaders as attachment figures: leaders’ attachment orientations predict leadership-related mental representations and followers’ performance and mental health. Journal of Personality and Social Psychology, 93(4), 632–50, p.645.

377 Ibid. p.647.

378 Berant, E., Mikulincer, M., & Florian, V. (2001) The association of mothers’ attachment style and their psychological reactions to the diagnosis of infant’s congenital heart disease. Journal of Social and Clinical Psychology, 20(2), 208–232, p.227.

379 Gillath, O., Giesbrecht, B., & Shaver, P.R. (2009) Attachment, attention, and cognitive control: attachment style and performance on general attention tasks. Journal of Experimental Social Psychology, 45(4), 647–54.

380 Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood, 2nd edn. New York: Guilford.

382 Ein-Dor, T., Mikulincer, M., Doron, G., & Shaver, P.R. (2010) The attachment paradox: how can so many of us (the insecure ones) have no adaptive advantages? Perspectives on Psychological Science, 5(2), 123–41.

383 Belsky, J., Steinberg, L., & Draper, P. (1991) Childhood experience, interpersonal development, and reproductive strategy: an evolutionary theory of socialization. Child Development, 62, 647–70; Crittenden, P.M. (1992) Quality of attachment in the preschool years. Development & Psychopathology, 4(2), 209–241. For reflection on Ein-Dor’s claims from critical psychology see Carr, S. & Batlle, I.C. (2015) Attachment theory, neoliberalism, and social conscience. Journal of Theoretical and Philosophical Psychology, 35(3), 160–76. These researchers interpret Ein-Dor’s perspective as essentially a positive framing of the fit between an avoidant attachment style and the depleting, dehumanising late capitalist labour market. They argue that, even if there are some advantages to insecure attachment styles, these advantages are vastly outweighed by their cost, to the point that claims about advantages are misleading.

384 See Ein-Dor, T. & Hirschberger, G. (2016) Rethinking attachment theory: from a theory of relationships to a theory of individual and group survival. Current Directions in Psychological Science, 25(4), 223–7, p.226.

385 Ein-Dor, T., Doron, G., Solomon, Z., Mikulincer, M., & Shaver, P.R. (2010) Together in pain: attachment-related dyadic processes and posttraumatic stress disorder. Journal of Counseling Psychology, 57(3), 317–27.

386 Ein-Dor, T., Mikulincer, M., & Shaver, P.R. (2011) Effective reaction to danger: attachment insecurities predict behavioral reactions to an experimentally induced threat above and beyond general personality traits. Social Psychological and Personality Science, 2(5), 467–73.

387 Lavy, S., Bareli, Y., & Ein-Dor, T. (2015) The effects of attachment heterogeneity and team cohesion on team functioning. Small Group Research, 46(1), 27–49.

388 Ein-Dor, T. & Orgad, T. (2012) Scared saviors: evidence that people high in attachment anxiety are more effective in alerting others to threat. European Journal of Social Psychology, 42(6), 667–71.

389 Ein-Dor, T., Reizer, A., Shaver, P.R., & Dotan, E. (2012) Standoffish perhaps, but successful as well: evidence that avoidant attachment can be beneficial in professional tennis and computer science. Journal of Personality, 80(3), 749–68.

390 Ibid. Table 2.

391 Ein-Dor, T., Perry-Paldi, A., Zohar-Cohen, K., Efrati, Y., & Hirschberger, G. (2017) It takes an insecure liar to catch a liar: the link between attachment insecurity, deception, and detection of deception. Personality and Individual Differences, 113, 81–7.

392 Cassidy, J., Brett, B.E., Gross, J.T., et al. (2017) Circle of security-parenting: a randomized controlled trial in Head Start. Development & Psychopathology, 29(2), 651–73. However, the assessment of outcomes was conducted as soon as feasible after completion of the intervention, when participants low in avoidant attachment style may have still been processing the intervention, requiring time to stabilise their caregiving. Longer before follow-up may have provided more opportunity for positive intervention effects. It should also be noted that the researchers appear not to have measured or controlled for the involvement of other services (e.g. child welfare involvement). This might be another relevant moderator.

393 Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, p.272.

394 On the potential for self-perpetuating neglect of certain research agendas within attachment research see Fearon, R.M.P., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2010) Jealousy and attachment: the case of twins. In S.L. Hart & M. Legerstee (eds) Handbook of Jealousy. Theory, Research, and Multidisciplinary Approaches (pp.362–86). New York: Wiley: ‘Without an accumulation of empirical data and novel findings it may be that we have not seen a sufficient number of new phenomena for researchers to get their teeth into, and so their energies have, to a large extent, been directed elsewhere’ (364).

395 Marmarosh, C. & Markin, R. (2007) Group and personal attachments. Group Dynamics: Theory, Research, and Practice, 11(3), 153–64; Yip, J., Ehrhardt, K., Black, H., & Walker, D.O. (2018) Attachment theory at work: a review and directions for future research. Journal of Organizational Behavior, 39(2), 185–98; DeMarco, T.C. & Newheiser, A.K. (2019) Attachment to groups: relationships with group esteem, self-esteem, and investment in ingroups. European Journal of Social Psychology, 49(1), 63–75.

396 Allen, J.G., Stein, H., Fonagy, P., Fultz, J., & Target, M. (2005) Rethinking adult attachment: a study of expert consensus. Bulletin of the Menninger Clinic, 69(1), 59–80: ‘We believe there is room to continue sharpening the conceptual boundaries of adult attachment by further examining the content of putative attachment scale items’ (60).

397 Banai, E., Mikulincer, M., & Shaver, P.R. (2005) ‘Selfobject’ needs in Kohut’s self psychology. Psychoanalytic Psychology, 22(2), 224–60, p.253.

398 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp. 53–152). New York: Academic Press, p.141.

400 Wei, M., Russell, D.W., Mallinckrodt, B., & Vogel, D.L. (2007) The Experiences in Close Relationship Scale (ECR)—short form: reliability, validity, and factor structure. Journal of Personality Assessment, 88(2), 187–204.

401 Wei and colleagues only had one item represent this component, Item 18 on the ECR. This item reads: ‘I need a lot of reassurance that close relationship partners really care about me’. Wei and colleagues characterised this item as ‘an excessive need for approval from others (Item 18)’. This is an improbable characterisation of the item! It would seem most likely that the word ‘assurance’ was intended, rather than ‘approval’. There are items in the ECR that represent need for approval, such as Item 34: ‘when other people disapprove of me, I feel really bad about myself’. However, it is Item 18, not these items about approval, that feature in the list of 12 items offered by Wei and colleagues as the short version of the ECR. It is perhaps worth noting that a desire for approval is explicitly one of the features specified by Shaver and Mikulincer as associated with hyperactivation of the power behavioural system, which may account for a portion of the link between the two measures.

402 Brennan, K.A., Clark, C.L., & Shaver, P.R. (1998) Self-report measurement of adult romantic attachment: an integrative overview. In J A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.46–76). New York: Guilford, p.66.

403 E.g. Olssøn, I. Sørebø, O., & Dahl, A.A. (2010) The Norwegian version of the Experiences in Close Relationships measure of adult attachment: psychometric properties and normative data. Nordic Journal of Psychiatry, 64(5), 340–49.

404 Lo, C., Walsh, A., Mikulincer, M., Gagliese, L., Zimmermann, C., & Rodin, G. (2009) Measuring attachment security in patients with advanced cancer: psychometric properties of a modified and brief Experiences in Close Relationships scale. Psycho-Oncology, 18(5), 490–99, p.495.

405 Reassurance is also mentioned by Item 35: ‘I turn to close relationship partners for many things, including comfort and reassurance’. However, the formulation of the item is astonishingly unspecific—‘many things’—so it is not clear that participants would answer based primarily on their experiences of wanting or receiving reassurance. More importantly, the item, reversed, is one that contributes to the scoring of attachment avoidance, not attachment anxiety. A study by Lafontaine and colleagues found that Item 35 had the highest standard error and worst discrimination of any of the avoidance items in the ECR. Lafontaine, M.F., Brassard, A., Lussier, Y., Valois, P., Shaver, P.R., & Johnson, S.M. (2016) Selecting the best items for a short-form of the Experiences in Close Relationships questionnaire. European Journal of Psychological Assessment, 32(2), 140–54, p.146.

406 Intermediate ‘facets’ have also been identified between the items and the latent dimensions of anxiety and avoidance in the Attachment Style Questionnaire: confidence; relationships as secondary; discomfort with closeness; preoccupation with relationships; and need for approval. Karantzas, G.C., Feeney, J.A., & Wilkinson, R. (2010) Is less more? Confirmatory factor analysis of the Attachment Style Questionnaires. Journal of Social and Personal Relationships, 27(6), 749–80.

407 E.g. Olssøn, I., Sørebø, O., & Dahl, A.A. (2010) The Norwegian version of the Experiences in Close Relationships measure of adult attachment: psychometric properties and normative data. Nordic Journal of Psychiatry, 64(5), 340–49: ‘The five-factor solution. Factors 1 and 4 consist of avoidance items only. The items of factor 1 describe avoidance of getting close or discomfort by coming close. In factor 4, the content of all items is reluctance to self-disclosure or dependence on others … Factors 2, 3 and 5 all consist of anxiety items. The items in factor 2 describe worrying about abandonment or being alone. Factor 3 content statements about one’s need for partner’s availability or reassurance, and factor 5 concerns worry that the individual wants more closeness than the other person does’ (344).

408 Questions have been raised about the item-total correlations of the ECR and measures based on it by several psychometricians, e.g. Hanak, N. & Dimitrijevic, A. (2013) A Serbian Version of Modified and Revised Experiences in Close Relationships Scale (SM–ECR–R). Journal of Personality Assessment, 95(5), 530–38.

409 Wongpakaran, T., Wongpakaran, N., & Wannarit, K. (2011) Validity and reliability of the Thai version of the Experiences of Close Relationships–Revised questionnaire. Singapore Medical Journal, 52(2), 100–106, p.103.

410 See also Esbjørn, B.H., Breinholst, S., Niclasen, J., Skovgaard, L.F., Lange, K., & Reinholdt-Dunne, M.L. (2015) Identifying the best-fitting factor structure of the Experience of Close Relations–Revised in a Scandinavian example. PLoS One, 10(9), e0137218. Though this study was of the ECR-R, not the ECR, there were items falling in the ‘extra’ factor from the original ECR, such as Items 6 and 10.

411 Karantzas, G.C., Feeney, J.A., & Wilkinson, R. (2010) Is less more? Confirmatory factor analysis of the Attachment Style Questionnaires. Journal of Social and Personal Relationships, 27(6), 749–80, p.774; Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, pp.120, 245. See, for instance, work by Tasca and colleagues, which found that the association between attachment anxiety and symptoms of depression and disordered eating was fully mediated by emotion regulation strategies. Tasca, G.A., Szadkowski, L., Illing, V., et al. (2009) Adult attachment, depression, and eating disorder symptoms: the mediating role of affect regulation strategies. Personality and Individual Differences, 47(6), 662–7.

412 See e.g. Alonso-arbiol, I., Balluerka, N., & Shaver, P.R. (2007) A Spanish version of the Experiences in Close Relationships (ECR) adult attachment questionnaire. Personal Relationships, 14(1), 45–63.

413 Lo, C., Walsh, A., Mikulincer, M., Gagliese, L., Zimmermann, C., & Rodin, G. (2009) Measuring attachment security in patients with advanced cancer: psychometric properties of a modified and brief Experiences in Close Relationships scale. Psycho-Oncology, 18(5), 490–99, p.495. The items representing discomfort with closeness are:

  • Item 7. ‘I get uncomfortable when other people want to be very close to me’

  • Item 13. ‘I am nervous when other people get too close to me’

  • Item 17. ‘I try to avoid getting too close to other people’

  • Item 23. ‘I prefer not to be too close to other people’

  • Item 9. ‘I don’t feel comfortable opening up to other people’.

414 Ibid.: ‘Conceptually, the double-loading of Discomfort on the higher-order attachment dimensions suggests that both attachment anxiety and avoidance tap some discomfort with the experience of intimacy and closeness’ (498).

415 E.g. Alonso-arbiol, I., Balluerka, N., & Shaver, P.R. (2007) A Spanish version of the Experiences in Close Relationships (ECR) adult attachment questionnaire. Personal Relationships, 14(1), 45–63. Matters are further complicated by findings by Mikulincer and colleagues that attachment anxiety on the ECR is associated with both preconscious approach and avoidance goals with respect to relational closeness, which affect their motor responses in a push–pull task. It may be that when individuals with an anxious attachment style feel overclose, as may occur in longstanding relationships, they will endorse items representing a desire for avoidance of closeness. Mikulincer, M., Shaver, P.R., Bar-On, N., & Ein-Dor, T. (2010) The pushes and pulls of close relationships: attachment insecurities and relational ambivalence. Journal of Personality and Social Psychology, 98(3), 450–68, p.463.

416 Lafontaine, M.F., Brassard, A., Lussier, Y., Valois, P., Shaver, P.R., & Johnson, S.M. (2016) Selecting the best items for a short-form of the Experiences in Close Relationships questionnaire. European Journal of Psychological Assessment, 32(2), 140–54.

417 Frıas, M.T., Shaver, P.R., & Mikulincer, M. (2014) Measures of adult attachment and related constructs. In G.J. Boyle & D.H. Saklofske (eds) Measures of Personality and Social Psychological Constructs. Philadelphia, PA: Elsevier, pp.417–47, p.446.

418 Fraley, R.C., Waller, N.G., & Brennan, K.A. (2000) An Item Response Theory analysis of self-report measures of adult attachment. Journal of Personality and Social Psychology, 78(2), 350–65: ‘Also notice that the two ECR-R scales, like the original ECR scales, are not adept at assessing individuals with trait levels less than –1.00 on Anxiety or Avoidance … An important next step for future research on scale development is to write items that tap the low ends of the Anxiety and Avoidance dimensions with better precision’ (361).

419 Feeney, J.A. (2002) Attachment-related dynamics: what can we learn from self-reports of avoidance and anxiety? Attachment & Human Development, 4(2), 193–200, p.198.

420 Brennan, K.A. & Shaver, P.R. (1995) Dimensions of adult attachment, affect regulation, and romantic relationship functioning. Personality and Social Psychology Bulletin, 21(3), 267–83, p.280.

422 Mikulincer, M. & Shaver, P.R. (2007) Boosting attachment security to promote mental health, prosocial values, and inter-group tolerance. Psychological Inquiry, 18, 139–56: ‘We conceptualize the sense of attachment security as an inner resource’ (139). On the contribution of distrust to both anxiety and avoidance, or as a reciprocal pathway between the two, see McWilliams, L.A. & Fried, E.I. (2019) Reconceptualizing adult attachment relationships: a network perspective. Personal Relationships, 26(1), 21–41.

423 Shaver, P.R. & Mikulincer, M. (2008) Augmenting the sense of security in romantic, leader–follower, therapeutic, and group relations: a relational model of personality change. In J.P. Forgas & J. Fitness (eds) Social Relationships: Cognitive, Affective, and Motivational Processes (pp.55–73). New York: Psychology Press.

424 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford, p.51.

425 Concurrent with the publication of the ECR was work by Mikulincer suggesting a three-factor solution, in which security played the smallest but nonetheless material role in predicting variance. A factor analysis conducted by Banai, Weller, and Mikulincer in 1998 found that 13.3% of variance in the classification of an individual’s attachment style by themselves, their friends, and acquaintances could be accounted for by endorsement of the secure attachment style on an adaptation of the ‘love quiz’, even taking into account avoidance and ambivalence-resistance. Banai, E., Weller, A., & Mikulincer, M. (1998) Inter-judge agreement in evaluation of adult attachment style: the impact of acquaintanceship. British Journal of Social Psychology, 37(1), 95–109, p.104.

426 E.g. Banse, R. (2004) Adult attachment and marital satisfaction: evidence for dyadic configuration effects. Journal of Social and Personal Relationships, 21(2), 273–82; Asendorpf, J.B., Banse, R., Wilpers, S., & Neyer, F.J. (1997) Relationship-specific attachment scales for adults and their validation with network and diary procedures. Diagnostica, 43(4), 289–313.

427 See also Holmes, J.G. & Murray, S.L. (2007) Felt security as a normative resource: evidence for an elemental risk regulation system? Psychological Inquiry, 18(3), 163–7.

428 Shaver, P.R. & Mikulincer, M. (2004) What do self-report attachment measures assess? In W.S. Rholes & J.A. Simpson (eds) Adult Attachment: Theory, Research, and Clinical Implications (pp.17–54). New York: Guilford: ‘This 45-degree rotation of the measurement axes fits well with the process model proposed by Shaver and Mikulincer (2002)’ (51).

429 Al-Yagon, M. & Mikulincer, M. (2006) Children’s appraisal of teacher as a secure base and their socio-emotional and academic adjustment in middle childhood. Research in Education, 75(1), 1–18.

430 Gillath, O., Hart, J., Noftle, E.E., & Stockdale, G.D. (2009) Development and validation of a state adult attachment measure (SAAM). Journal of Research in Personality, 43(3), 362–73.

431 The security items in the SAAM are: I feel loved; I feel like I have someone to rely on; I feel secure and close to other people; If something went wrong right now I feel like I could depend on someone; I feel like others care about me; I feel relaxed knowing that close others are there for me right now; I feel I can trust the people who are close to me.

432 Parallel findings have also been reported for the ECR-RC, the adaptation of the ECR for children and adolescents to report about their relationship with their parents. Here too security has emerged as an independent latent factor. This may be an effect of developmental stage. However, like well-established couples and unlike college students, children and adolescents are also structurally entangled in their attachment relationships, which may reduce the orthogonality of avoidance and resistance and contribute to the autonomy of broaden-and-build cycles. Such a conclusion would suggest that Brennan and colleagues’ two-factor solution with orthogonal dimensions may have been influenced by the disembedded social conditions of college students within American culture, where neither secure-base effects are fully online, nor anxiety and avoidance are able to become especially tangled. Lionetti, F., Mastrotheodoros, S., & Palladino, B.E. (2018) Experiences in Close Relationships-Revised Child version (ECR-RC): psychometric evidence in support of a security factor. European Journal of Developmental Psychology, 15(4), 452–63.

433 Gillath, O., Karantzas, G.C., & Fraley, R.C. (2016) Adult Attachment: A Concise Introduction to Theory and Research. London: Academic Press, p.247. See also Kanninen, K., Punamaki, R.L., & Qouta, S. (2003) Personality and trauma: adult attachment and posttraumatic distress among former political prisoners. Peace and Conflict, 9(2), 97–126.

434 Frıas, M.T., Shaver, P.R., & Mikulincer, M. (2014) Measures of adult attachment and related constructs. In G.J. Boyle & D.H. Saklofske (eds) Measures of Personality and Social Psychological Constructs (pp.417–47). Philadelphia, PA: Elsevier, p.443.

435 Ibid. p.446.

436 The ECR-R has fared even less well on this front than the ECR. See e.g. Rotaru, T.Ş. & Rusu, A. (2013) Psychometric properties of the Romanian version of Experiences in Close Relationships-Revised questionnaire (ECR-R). Procedia-Social and Behavioral Sciences, 78, 51–5; Busonera, A., Martini, P.S., Zavattini, G.C., & Santona, A. (2014) Psychometric properties of an Italian version of the Experiences in Close Relationships-Revised (ECR-R) Scale. Psychological Reports, 114(3), 785–801.

437 Mikulincer, M., Shaver, P.R., Gillath, O., & Nitzberg, R.A. (2005) Attachment, caregiving, and altruism: boosting attachment security increases compassion and helping. Journal of Personality and Social Psychology, 89(5), 817–39, p.821.

438 Conradi, H.-J., Gerlsma, C., van Duijn, M., & de Jonge, P. (2006) Internal and external validity of the Experiences in Close Relationships questionnaire in an American and two Dutch samples. European Journal of Psychiatry, 20(4), 258–69, p.268.

439 The principle is one that Shaver acknowledged in his early work, if less in his writings on attachment: Felton, B.F. & Shaver, P.R. (1984) Cohort variation in adults’ reported feelings. In C.Z. Malatesta & C.E. Izard (eds) Emotions in Adult Development (pp.103–123). Beverly Hills: Sage Publications: ‘Adults’ choices of coping strategies are linked to the nature of the problems they face’ (118–19).

440 See e.g. Kroonenberg, P.M., Dam, M.V., van IJzendoorn, M., & Mooijaart, A. (1997) Dynamics of behaviour in the strange situation: a structural equation approach. British Journal of Psychology, 88(2), 311–32.

441 Research findings from Shaver’s American and Mikulincer’s Israeli undergraduate samples have generally aligned well in terms of the correlates of attachment styles. Cultural differences have been reported at times, but not ones that interact with the ECR. See e.g. Mikulincer, M., Shaver, P.R., Gillath, O., & Nitzberg, R.A. (2005) Attachment, caregiving, and altruism: boosting attachment security increases compassion and helping. Journal of Personality and Social Psychology, 89(5), 817–39, p.836.

442 Conradi, H.-J., Gerlsma, C., van Duijn, M., & de Jonge, P. (2006) Internal and external validity of the experiences in close relationships questionnaire in an American and two Dutch samples. European Journal of Psychiatry, 20(4), 258–69.

443 Cameron, J.J., Finnegan, H., & Morry, M.M. (2012) Orthogonal dreams in an oblique world: a meta-analysis of the association between attachment anxiety and avoidance. Journal of Research in Personality, 46(5), 472–6; Graham, J.M. & Unterschute, M.S. (2015) A reliability generalization meta-analysis of self-report measures of adult attachment. Journal of Personality Assessment, 97(1), 31–41.

444 Cameron, J.J., Finnegan, H., & Morry, M.M. (2012) Orthogonal dreams in an oblique world: a meta-analysis of the association between attachment anxiety and avoidance. Journal of Research in Personality, 46(5), 472–6, p.475.

445 Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42, p.826.

446 E.g. Fraley, R.C., Davis, K.E., & Shaver, P.R. (1998) Dismissing avoidance and the defensive organization of emotion, cognition, and behavior. In J.A. Simpson & W.S. Rholes (eds) Attachment Theory and Close Relationships (pp.249–79). New York: Guilford: ‘It should be noted that the processes and mechanisms we discuss with respect to dismissing-avoidance apply to preoccupation inversely’ (275).

447 Birnbaum, G.E., Mikulincer, M., Szepsenwol, O., Shaver, P.R., & Mizrahi, M. (2014) When sex goes wrong: a behavioral systems perspective on individual differences in sexual attitudes, motives, feelings, and behaviors. Journal of Personality and Social Psychology, 106(5), 822–42, p.826. An example here may be the role of distrust in contributing to both avoidance and anxiety, or their reciprocal reinforcement. See McWilliams, L.A. & Fried, E.I. (2019) Reconceptualizing adult attachment relationships: a network perspective. Personal Relationships, 26(1), 21–41.

448 Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press, pp.70, 88.

449 Shaver, P.R. & Mikulincer, M. (2008) Adult attachment and cognitive and affective reactions to positive and negative events. Social and Personality Compass, 2, 1844–65, p.1853. Citing Fraley, R. C. & Bonanno, G. A. (2004) Attachment and loss: a test of three competing models on the association between attachment-related avoidance and adaptation to bereavement. Personality and Social Psychology Bulletin, 30(7), 878–90.

450 Gillath, O., Bunge, S.A., Shaver, P.R., Wendelken, C., & Mikulincer, M. (2005) Attachment-style differences in the ability to suppress negative thoughts: exploring the neural correlates. Neuroimage, 28(4), 835–47, p.945.

451 Makariev, D.W. & Shaver, P.R. (2010) Attachment, parental incarceration and possibilities for intervention: an overview, Attachment & Human Development, 12(4), 311–31, p.325.

452 Brennan, K.A. & Shaver, P.R. (1998) Attachment styles and personality disorders: their connections to each other and to parental divorce, parental death, and perceptions of parental caregiving. Journal of Personality, 66(5), 835–78: ‘One obvious limitation of our research is its use of a nonclinical sample. To some extent, this limitation turns on whether one accepts a purely categorical (vs. dimensional) understanding of personality disorders. If, as we do, one assumes that personality disorders can be arrayed on continua, then our results ought to generalize to clinical populations’ (870).

453 Woodhouse, S., Ayers, S., & Field, A.P. (2015) The relationship between adult attachment style and post-traumatic stress symptoms: a meta-analysis. Journal of Anxiety Disorders, 35, 103–117.

454 Shaver, P. & Mikulincer, M. (2004) Attachment in the later years: a commentary. Attachment & Human Development, 6(4), 451–64, p.462. Other laboratories have confirmed that high anxiety/high avoidance attachment is more prevalent in psychiatric samples, e.g. Alessandri, G., Fagnani, C., Di Gennaro, G., et al. (2014) Measurement invariance of the experiences in close relationships questionnaire across different populations. Spanish Journal of Psychology, 17(2), E22.

455 Mikulincer, M. & Nachshon, O. (1991) Attachment styles and patterns of self-disclosure. Journal of Personality and Social Psychology, 61(2), 321–31.

456 Cooper, M.L., Shaver, P.R., & Collins, N.L. (1998) Attachment styles, emotion regulation, and adjustment in adolescence. Journal of Personality and Social Psychology, 74(5), 1380–97.

457 Ibid. p.1385.

458 Ibid.: ‘We adopted the procedure used by Mikulincer and Nachshon and excluded inconsistent respondents from further analyses’ (1386).

459 E.g. Crawford, T.N., Livesley, W.J., Jang, K.L., Shaver, P.R., Cohen, P., & Ganiban, J. (2007) Insecure attachment and personality disorder: a twin study of adults. European Journal of Personality, 21(2), 191–208.

460 Shaver and Mikulincer have at times reported relevant findings from other research groups, though this has been exceptionally rare. E.g. Mikulincer, M. & Shaver, P.R. (2005) Attachment theory and emotion in close relationships: exploring the attachment-related dynamics of emotional reactions to relational events. Personal Relationships, 12, 149–68, p.160.

461 Schachner, D.A. & Shaver, P.R. (2004) Attachment dimensions and sexual motives. Personal Relationships, 11(2), 179–95, p.192.

462 Hart, J.J., Shaver, P.R., & Goldenberg, J.L. (2005) Attachment, self-esteem, worldviews, and terror management: evidence for a tripartite security system. Journal of Personality and Social Psychology, 88, 999–1013, p.1008.

463 For another example of specific correlates of high anxiety/high avoidance attachment see Gillath, O., Giesbrecht, B., & Shaver, P.R. (2009) Attachment, attention, and cognitive control: attachment style and performance on general attention tasks. Journal of Experimental Social Psychology, 45(4), 647–54, p.651.

464 It might be thought that an additional qualification could be offered on the basis of the major efforts by Shaver and Mikulincer to review empirical studies of psychopathology, e.g. Mikulincer, M. & Shaver, P. (2016) Attachment in Adulthood, 2nd edn. New York: Guilford, Chapter 13. However, these reviews are lacking precisely in consideration of high anxiety/high avoidance attachment styles except, on a handful of occasions, when reporting findings from Bartholomew’s measures. The question of whether studies using the ECR have found effects for high anxiety/high avoidance, and more generally the question of whether the high ends of the scales are individually or in interaction effective at capturing clinical phenomena, is not in view. The focus is instead on anxiety and avoidance as, separately, dimensionally associated with forms of mental pathology. Throughout the reviews, a linear and dimensional association between the ECR scales and psychopathology is assumed rather than demonstrated or explored.

465 Solomon, Z., Mikulincer, M., Freid, B., & Wosner, Y. (1987) Family characteristics and posttraumatic stress disorder: a follow-up of Israeli combat stress reaction casualties. Family Process, 26(3), 383–94, p.390.

467 Mikulincer, M., Solomon, Z., & Benbenishty, R. (1988) Battle events, acute combat stress reaction and long-term psychological sequelae of war. Journal of Anxiety Disorders, 2(2), 121–33, p.131.

468 Solomon, Z., Mikulincer, M., & Benbenishty, R. (1989) Combat stress reaction—clinical manifestations and correlates. Military Psychology, 1(1), 35–47, p.44. Later work by the researchers documented powerful longitudinal effects of these traumatic symptoms. See Solomon, Z. & Mikulincer, M. (2007) Post traumatic intrusion, avoidance, and social functioning: a 20-year longitudinal study. Journal of Consulting and Clinical Psychology, 75(2), 316–24.

469 See also Haberman, C. (1994) Palestinians arrest 100 Islamic militants after bicycle bombing. New York Times, 13 November 1994.

470 The distinction between these symptom clusters had been introduced in DSM-III-R. Brett, E., Spitzer, R., & Williams, J.B. (1988) DSM-III-R criteria for posttraumatic stress disorder. American Journal of Psychiatry, 145(10), 1232. The distinction was embedded in psychological tools such as the Post-traumatic Stress Disorder Inventory, developed by Solomon, Weisenberg, Schwarzwald, and Mikulincer in the late 1980s, and used in the 1999 study of the Jewish settlers.

471 Mikulincer, M., Horesh, N., Eilati, I., & Kotler, M. (1999) The association between adult attachment style and mental health in extreme life-endangering conditions. Personality and Individual Differences, 27(5), 831–42, p.837.

472 Ibid.: ‘Interestingly, this distress was mainly manifested in avoidance rather than intrusive responses, which seem to be avoidant persons’ habitual affect regulation strategy’ (839). The proposal that PTSD avoidance and hyperarousal symptoms parallel the Ainsworth avoidant and ambivalent/resistant classifications had first been made a few years earlier by Crittenden, P.M. (1997) Toward an integrative theory of trauma: a dynamic–maturational approach. In D. Cicchetti & S. Toth (eds) The Rochester Symposium on Developmental Psychopathology, Vol. 10. Risk, Trauma, and Mental Processes (pp.34–84). New York: University of Rochester Press. In a later study, Besser and colleagues used the ECR with civilians exposed to terrorist attacks in southern Israel. Unlike Mikulincer, they found few or weak associations from attachment avoidance. They found stronger associations from attachment anxiety with all three clusters of PTSD symptoms (intrusive, avoidance, and hyperarousal). The reasons for these different results may stem from the different sample. It may be that an avoidant attachment style was still a viable strategy for participants subject to long-term missile threat, in contrast to the settlers studied by Mikulincer. Besser, A., Neria, Y., & Haynes, M. (2009) Adult attachment, perceived stress, and PTSD among civilians exposed to ongoing terrorist attacks in southern Israel. Personality and Individual Differences, 47(8), 851–7.

473 Mikulincer, M., Dolev, T., & Shaver, P.R. (2004) Attachment-related strategies during thought suppression: ironic rebounds and vulnerable self-representations. Journal of Personality and Social Psychology, 87(6), 940–56: ‘Deactivating strategies can be inadequate and overwhelmed, resulting in a marked decline in functioning and what Horowitz (1982) called “avoidance-related” posttraumatic symptoms (e.g., psychic numbing, behavioral inhibition)’ (941).

474 Shaver, P.R. & Mikulincer, M. (2002) Dialogue on adult attachment: diversity and integration. Attachment & Human Development, 4(2), 243–57, p.247. Shaver and Mikulincer did not flesh out the conditions under which failure of hyperactivating strategies might be expected. It is also not fully clear whether failure would entail the flooding of anxiety in some internal sense, or the breakdown of the capacity to strive for closeness with attachment figures, or both. The former appears to be described in a characterisation of the hyperactivating strategy early in Shaver and Mikulincer’s work together: Mikulincer, M. & Shaver, P.R. (2003) The attachment behavioral system in adulthood: activation, psychodynamics, and interpersonal processes. In M.P. Zanna (ed.) Advances in Experimental Social Psychology, Vol. 35 (pp.53–152). New York: Academic Press: ‘Intrapsychically, amplification of threat appraisals heightens the chronic accessibility of negative thoughts and makes it likely that new sources of distress will mingle and become confounded with old accessible ones … the person experiences an endless and uncontrollable flow of negative thoughts and moods, which in turn may lead to cognitive disorganization and, in certain cases, culminate in psychopathology’ (82–3).

475 These concerns may also be placed in other discussions at the time in social psychology regarding the potential for different forms of personality disorder to reflect extreme forms of attachment system activation or deactivation, especially in the context of loss, trauma, or chronic adversities. See Bartholomew, K., Kwong, M.J., & Hart, S.D. (2001) Attachment. In W.J. Livesley (ed.) Handbook of Personality Disorders: Theory, Research, and Treatment (pp.196–230). New York: Guilford.

476 E.g. Mikulincer, M., Gillath, O., & Shaver, P.R. (2002) Activation of the attachment system in adulthood: threat-related primes increase the accessibility of mental representations of attachment figures. Journal of Personality and Social Psychology, 83(4), 881–95: ‘It would be interesting to explore further how avoidant persons’ inhibitory processes work, what they are designed to accomplish (e.g., protection from a potentially angry, punitive attachment figure; reduction of the attachment figure’s tendency to threaten abandonment or decrease support if a particular separation is resisted), and when they arise—either in the course of development or in the course of a particular long-term relationship’ (893–4).

477 Mikulincer, M., Shaver, P.R., & Pereg, D. (2003) Attachment theory and affect regulation: the dynamics, development, and cognitive consequences of attachment-related strategies. Motivation and Emotion, 27, 77–102, pp.97–8.

478 Ibid.: ‘Researchers should examine the conditions under which secondary attachment strategies seem to work sufficiently well to avoid severe psychopathology. We still do not know why some insecurely attached individuals function within the normal range whereas others require clinical intervention’ (100).

479 Solomon, Z., Dekel, R., & Mikulincer, M. (2008) Complex trauma of war captivity: a prospective study of attachment and post-traumatic stress disorder. Psychological Medicine 38(10), 1427–34.

480 Ein-Dor, T., Doron, G., Solomon, Z., Mikulincer, M., & Shaver, P.R. (2010) Together in pain: attachment-related dyadic processes and posttraumatic stress disorder. Journal of Counseling Psychology, 57(3), 317–27, p.321, Table 1.

481 Solomon, Z., Dekel, R., & Mikulincer, M. (2008) Complex trauma of war captivity: a prospective study of attachment and post-traumatic stress disorder. Psychological Medicine 38(10), 1427–34, p.1431. See also Mikulincer, M., Ein-Dor, T., Solomon, Z., & Shaver, P.R. (2011) Trajectories of attachment insecurities over a 17-year period: a latent growth curve analysis of the impact of war captivity and posttraumatic stress disorder. Journal of Social and Clinical Psychology, 30(9), 960–84. A related finding was reported by Ghafoori and colleagues in their study of US military veterians. Ghafoori, B., Hierholzer, R.W., Howsepian, B., & Boardman, A. (2008) The role of adult attachment, parental bonding, and spiritual love in the adjustment to military trauma. Journal of Trauma & Dissociation, 9(1), 85–106.

482 Bowlby, J. (1974) A guide to the perplexed parent. New York Times, 2 March 1974.

483 Granqvist, P. (2020) Attachment, Religion, and Spirituality: A Wider View. New York: Guilford: ‘The principal ideas about attachment-related individual differences largely derive from, and are intertwined with, the methods used to measure such individual differences.’ See also Verbeke, W., Belschack, F., Bagozzi, R.P., Pozharliev, R., & Ein-Dor, T. (2017) Why some people just ‘can’t get no satisfaction’: secure versus insecure attachment styles affect one’s ‘style of being in the social world’. International Journal of Marketing Studies, 9(2), 36–55: ‘In using general attachment scales to study why attachment styles are so pervasive in affecting people’s relationships, we take an ontological stance (philosophical study of the nature of being) in order to look at how individual differences in attachment styles reflect people’s sense of being in the social world’ (37).

485 Developmental cognitive psychology can here be identified as an exception, since the focus here is often less relational. There are certainly exceptions, though, such as the study of babies’ recognition of human faces, e.g. Farroni, T., Menon, E., Rigato, S., & Johnson, M.H. (2007) The perception of facial expressions in newborns. European Journal of Developmental Psychology, 4(1), 2–13.

486 On the history of social psychology see Danziger, K. (2000) Making social psychology experimental: a conceptual history, 1920–1970. Journal of the History of the Behavioral Sciences, 36(4), 329–47; Greenwood, J.D. (2004) What happened to the ‘social’ in social psychology? Journal for the Theory of Social Behaviour, 34(1), 19–34; Jahoda, G. (2007) A History of Social Psychology: From the Eighteenth-century Enlightenment to the Second World War. Cambridge: Cambridge University Press; Kruglanksi, A.W. & Stroebe, W. (eds) (2012) Handbook for the History of Social Psychology. Bristol: Psychology Press; Pettigrew, T.F. & Cherry, F. (2012) The intertwined histories of personality and social psychology. In M.R. Leary & R.H. Hoyle (eds) Handbook of Individual Differences in Social Behavior (pp.13–32). New York: Guilford. On distinctions between trends in American and European social psychology see Schruijer, S.G. (2012) Whatever happened to the ‘European’in European social psychology? A study of the ambitions in founding the European Association of Experimental Social Psychology. History of the Human Sciences, 25(3), 88–107.

487 See e.g. Jahoda, G. (2016) Seventy years of social psychology: a cultural and personal critique. Journal of Social and Political Psychology, 4(1), 364–80. On the role of ongoing boundary-work in the construction of legitimate and illegitimate exceptions to predominant trends in a discipline see Good, J.M. (2000) Disciplining social psychology: a case study of boundary relations in the history of the human sciences. Journal of the History of the Behavioral Sciences, 36(4), 383–403.

488 Stainton Rogers, R., Stenner, P., Gleeson, K., & Stainton Rogers, W. (1995) Social Psychology: A Critical Agenda. Cambridge: Polity Press; Danziger, K. (1992) The project of an experimental social psychology: historical pespectives. Science in Context, 5, 309–328. See also Spini, D., Elcheroth, G., & Figini, D. (2009) Is there space for time in social psychology publications? A content analysis across five journals. Journal of Community & Applied Social Psychology, 19(3), 165–81. Allport’s ‘official’ definition of social psychology as the study of how ‘the thought, feeling, and behavior of individuals are influenced by the actual, imagined, or implied presence of others’ is illustrative: the individual is influenced by others, but is registered already as an individual. Allport, G.W. (1954) The historical background of modern social psychology. In G.L. Lindzey (ed.) Handbook of Social Psychology, Vol. 1 (pp.3–45). Reading, MA: Addison-Wesley, p.5. Nonetheless, continuities in social psychology over time should not be overemphasised. See Lubek, I. & Apfelbaum, E. (2000) A critical gaze and wistful glance at Handbook histories of social psychology: did the successive accounts by Gordon Allport and successors historiographically succeed? Journal of the History of the Behavioral Sciences, 36(4), 405–428. In particular, the rise of relationship research in social psychology since the 1980s has roots in earlier work, but should be recognised as a distinct development. See Reis, H.T. (2012) A brief history of relationship research in social psychology. In A.W. Kruglanski & W. Stroebe (eds) Handbook Online Dating (pp.363–82). Bristol: Psychology Press.

489 Mikulincer, M. & Shaver, P.R. (2012) An attachment perspective on psychopathology. World Psychiatry, 11(1), 11–15, p.14. See also Shaver, P.R. & Mikulincer, M. (2012) Attachment processes in relationships: reply to commentaries. Journal of Family Theory and Review, 4, 311–17: ‘If interpersonal relations are somewhat like dances or doubles tennis games, it is reasonable to expect that some of what is going on in the dance or in a match is “in” the individual dancers or players (e.g., their skills, their muscle development, their history of training and performance), some of it is in the interpersonal dynamics of a particular dyad, and some of it is in the context of a particular performance (e.g., the other players, a particular piece of music, the audience)’ (313).

490 Lemma, A., Target, M., & Fonagy, P. (2011) Brief Dynamic Interpersonal Therapy: A Clinician’s Guide. Oxford: Oxford University Press, p.92.

491 Fonagy, P., Luyten, P., Moulton-Perkins, A., et al. (2016) Development and validation of a self-report measure of mentalizing: the Reflective Functioning Questionnaire. PLoS One, 11(7), e0158678.

492 See also Pickering, A. (1995) The Mangle of Practice: Time, Agency, and Science. Chicago: University of Chicago Press.