Show Summary Details
Page of

(p. 537) Conclusion 

(p. 537) Conclusion
(p. 537) Conclusion

Robbie Duschinsky

Page of

PRINTED FROM OXFORD CLINICAL PSYCHOLOGY ONLINE ( © 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


In Becoming Attached, Karen described the origins of Bowlby and Ainsworth’s ideas, and considered emergent lines of research by the end of the 1980s, including work by Main and colleagues at Berkeley, Sroufe, Egeland, and colleagues at Minnesota, and Shaver and colleagues at Buffalo/Davis. These have been some of the cornerstone research groups for the development of contemporary attachment research. At the time Karen was writing, leadership of attachment research was passing from Bowlby and Ainsworth to a second generation. Furthermore, attachment as an area of academic study was rapidly expanding as perceptions of the field’s credibility and relevance opened the door to research funding. The rapid growth of attachment research since the 1990s means that no book today about the history of the field could even hope to be comprehensive. Work is already underway on another book to address developments from the 1990s onwards. Yet even with attention to these later developments still to come, it is possible to draw together some reflections on the ideas, methods, and priorities of the early groups of attachment research considered in the present book. These have set the scene for subsequent attachment researchers in important ways. This concluding chapter will begin by reviewing three structural dynamics that have contributed to the present state of attachment research as a paradigm: the apparent accessibility of attachment concepts; the subtlety and complexity of its theory and methods; and the differentiation of a field of cumulative attachment research. The chapter will then examine the problems the intersection of these structural dynamics have caused for communication between researchers and practitioners, using the case of child welfare practitioners as illustration. Threats to the credibility and apparent relevance of attachment research from such breakdowns has led to claims that the paradigm is reaching exhaustion. The chapter will evaluate this claim and discuss three lively areas of research that build precisely on the benefits of the field’s long development and history. It will close by considering what comes next for attachment research.

Structural issues

Over the course of this book, we have seen how the circulation and the shape of attachment theory and research have been influenced by the remarkable and at times misleading accessibility of some of Bowlby’s basic ideas. This can be regarded as the first of three structural dynamics that in interaction have formed the basis for the strengths and weaknesses of attachment as a paradigm today. In part, the intimacy of Bowlby’s ideas stemmed from their emergence out of psychoanalysis. Bowlby was intent on developing a theory that was integrated with the scientific developments of his day in other disciplines and that could offer testable hypotheses, but that also retained a portion of the capacity of psychoanalysis to address intimate life, ugly feelings, and inner conflict. Attachment theory speaks to themes that are constitutive of human life and that readily absorb our attention: being a (p. 538) child, becoming an adult; closeness and distance; the difficulties posed both by loss and by belonging. However, a further contribution to the power of Bowlby’s attachment theory was his appeal to ordinary language. Bowlby’s language in the 1950s was built to be able to both circulate to widespread publics and to persuade both academic and clinic audiences. The originating language of attachment theory used terms—such as ‘mother’, ‘attachment’, ‘separation’, ‘love’, ‘anxiety’—that had a strong and emotive set of connotations in ordinary language. Bowlby overlaid these connotations with a set of technical meaning in his scholarly work. ‘Love’ proved unworkable in this regard and was abandoned. And the second generation mostly stopped using ‘mother’ to mean attachment figure. Other terms have, however, been retained over the decades. This grounding in ordinary language has contributed to the flexibility, urgency, and reach of ‘attachment’ discourses, allowing it to seem plausible and at times to catch at the heart. These qualities, together with the credibility provided by attachment research as an empirical paradigm, have contributed to the appeal of attachment to diverse audiences.1

A second structural dynamic that has shaped attachment research has been the way that theory has been enshrined in complex measures. This has been a mixed blessing. Over time, the Ainsworth Strange Situation has replaced the interactions in the home that it had been intended to capture and preserve. It likewise embodied but also supplanted discussions of the attachment behavioural system. The development of the Attachment Q-Sort by Waters and colleagues sought to preserve Ainsworth’s concern with naturalistic observation and with the capacity of caregivers to offer their child a secure base. However, this measure too stimulated no further discussion of the attachment behavioural system, which has increasingly functioned as a memento in the developmental tradition of attachment research rather than as an active object of further theoretical work. Its place has been taken by the ideas of ‘minimising’ and ‘maximising’ strategies, though without clarity about what exactly is being minimised or maximised. Main’s theory was that it was attention to attachment-relevant information. This position was the basis of her interpretation of the Ainsworth classifications and her group’s introduction of the disorganised attachment classification and the Adult Attachment Interview (AAI). The innovations by Main and her group gave the Ainsworth categories an apparently universal resonance, across cultures and across lifespan development. However, with Main’s attentional theory largely unrecognised, there has been little precision in discussions of how the four-category coding system can be applied cross-culturally.

Similarly, without clarity on what exactly the AAI measured, controversies were primed when a social psychological tradition of attachment research emerged, which also made appeal to the Ainsworth Strange Situation categories but as part of a self-report measure. Unlike the developmental attachment tradition, Shaver and Milkulincer have further interrogated the idea of behavioural systems and attempted to articulate the concept of attachment. The developmental and social psychological traditions of attachment research have come to a better coexistence over recent years, both contributing to the Handbook of (p. 539) Attachment and the journal Attachment & Human Development. However, dialogue between the two traditions has been profoundly hindered by their use of the same central vocabulary to mean quite different things.

An influential study by Brennan, Clark, and Shaver in 1998 found that existing self-report measures could be integrated on the basis of two latent dimensions: avoidance and anxiety. The Experiences in Close Relationships scale was developed using items that reflected these two latent constructs. However, the exact relationship remains uncertain between the anxiety construct and Ainsworth’s ambivalent/resistant category, given that the latter incorporates anxiety, anger, and passivity. And it is not clear how well the Experiences in Close Relationships scale handles experiences of security and trauma. Like the Strange Situation and the AAI, the Experiences in Close Relationships scale embodies a particular theory of individual differences relevant to attachment, but has also served to close down certain questions about that theory. Nonetheless, these questions have generated increasing discussion in recent years from some of Shaver and Milkulincer’s closest students.

A third structural dynamic made salient by taking a historical perspective on early attachment research groups has been the work of ‘field building’: the construction of attachment as a differentiated paradigm with its own characteristic dispositions, methods, theories, values, allies, enemies, and nodal institutions.2 The Strange Situation provided the initial basis for a cumulative research programme, capable of attracting a sustained flow of research funds and empirical findings. Whereas Bowlby was writing primarily for clinical audiences and the general public, the central audience for Ainsworth and the second generation of attachment researchers was the American community of academic psychologists. With the construction of attachment research as a differentiated empirical paradigm following Ainsworth and her immediate colleagues, attachment research became a ‘non-formative’ activity, in which reshaping parenting practices was not a focal concern.3

Certainly there are some second-generation attachment researchers for whom clinical relevance has been the guiding principle of their work. Peter Fonagy, Patricia Crittenden, Alicia Lieberman, and Karlen Lyons-Ruth are salient examples.4 And social historians and sociologists have sometimes extrapolated from Bowlby’s early writings and the reception of attachment ideas in welfare practice to assume that contemporary attachment researchers are oriented towards the pathologizing and disciplining of families.5 Yet, overall, what is striking on close examination is how little developments in the second generation of attachment research, especially in America, seem to have been responsive to the concerns of (p. 540) clinical and social welfare practice, and how much by the demands of institutional academic psychology. Indeed, with some exceptions, the structural articulation between research and clinical attachment discourses has frequently been what Foucault termed ‘feeble’ and ‘slack’.6 For instance, the developmental research community have generally looked askance at both the ‘attachment disorder’ diagnosis and the use of attachment measures in assessment of risk in a context in which clinical and welfare services and training are largely structured by diagnostic pathways and risk assessment. Main has never written about what parents should do, and she has emphasised in print that practitioners should not regard her tentative suggestions as authoritative for what they should do.7 The Minnesota Longitudinal Study of Risk and Adaptation did contribute to the development of STEEP. However, this was a sideline for the Sroufe and Egeland group at Minnesota, whose primary focus was the cohort study. Shaver and Mikulincer have focused their attention predominantly, though certainly not exclusively, on samples of university students. Their remarks about the implications of attachment anxiety and avoidance for clinical and child welfare practice are rare and underdeveloped.8

Meeting the demands of institutional academic psychology provided advantages. Whilst the labour-intensive measures of the developmental tradition generally constrained sample size,9 a cumulative programme could be pursued over decades by the second generation of attachment researchers. Based on a massive number of studies, a series of meta-analyses in the early 2010s confirmed the capacity of the Strange Situation to predict later social and mental health outcomes. The extent of this prediction is either moderate or very substantial—depending on how effect sizes are interpreted.10 In any case, Groh and colleagues observed that they are comparable or stronger than other psychological assessments of socioemotional development conducted in infancy.11 The dedication of attachment researchers to commensurate measures is a marker of important continuities over time. Attachment has appeared as a strangely stable theoretical paradigm, in a context in which psychological theories were (p. 541) in continual churn and psychological theory in general was in decline. Highlighting this continuity, Waters and colleagues could claim that ‘Bowlby’s attachment theory is the only current theoretical framework in developmental psychology that is cast in the grand theory model, widely accepted, and empirically productive’.12

To anthropologists, this continuity has appeared as an imperviousness to criticism among attachment researchers, contributing to a sense of frustration and of being ignored by critics of the paradigm (Chapter 2). Yet, as Latour has urged, social studies of science must see the stability of social institutions and practices not as merely the default state of things, but as ‘exactly what has to be explained by appealing to costly and demanding means’.13 In the case of attachment as a research paradigm, these costly and demanding means have included the labour-intensive Strange Situation and astonishing amounts of convergent effort. Yet the stability of attachment research as a unified paradigm has also been achieved through maintenance of a common terminology, the cost of which has been confusion and miscommunication. One site of such problems has been in communication between research groups. Likewise, the underpinnings of the AAI were obscured by Main’s choice to frame this methodological innovation in terms of Bowlby’s concept of internal working models, miscuing decades of subsequent discussion of the measure.

Attachment in child welfare practice

The apparent accessibility of attachment concepts, the subtlety and complexity of its theory and methods, and the differentiation of a field of cumulative attachment research have together contributed to many of the strengths and attraction of the paradigm today. However, their interaction has also contributed to serious and pervasive misunderstandings regarding the concepts that organise attachment theory and research. The language of psychological research often grows out of ordinary language, since it is from ordinary problems that psychology frequently takes its starting point. Yet attachment research has been especially vulnerable to problems of mistaken identity between technical and ordinary language.

Summarising themes from across the previous chapters, four factors may be identified as of special importance. First, Bowlby’s early vocabulary was drawn deliberately and strategically from ordinary language in order to support the popular appeal of his ideas. Second, the coding systems of the developmental tradition are complicated and information about how constructs were actually operationalised has been unduly limited in circulation to an oral culture. Third, the appearance of attachment as a unified paradigm has been maintained in part through the stretching of terminology, with concepts invested with some overlapping and some non-overlapping senses. The capacity for concepts to be used flexibly has allowed attachment to seem relevant in diverse areas and diverse ways. However, this has come at a price. Areas of practice that appreciate conceptual precision and clearly operationalised concepts have tended to reject the attachment paradigm.14 And communication by and among attachment researchers has been hindered by the way that its concepts serve as (p. 542) magnets attracting quite heterogenous investments. Fourth, then, it can be observed that attachment research has been distinctively ill-equipped in terms of infrastructures for pruning how concepts are used in technical discussions. It is to be hoped that the journal Attachment & Human Development, future editions of the Handbook of Attachment, and organisations like the Society for Emotion and Attachment Research and the International Association for the Study of Attachment might give greater attention to this issue.15

Yet, if there has been trouble in communication between researchers, the situation has been even worse for dialogue between researchers and practitioners. In 1999, Rutter and O’Connor described extensive ‘conceptual confusion’ in appeals to attachment theory and research by clinicians and social workers.16 As the previous chapters have shown, there have been several ways in which the attachment research community have made effective dissemination of their ideas harder than many other fields. If Bowlby demonstrated the dangers of populism, his successors have demonstrated the dangers of turning away from public engagement. Misunderstandings and simplifications based on Bowlby’s early and popularising texts have abounded, and continue to shape public perceptions of the paradigm today. The gap between researchers and their public, where it is not left empty, has often been occupied by professional trainers, whose distance from empirical research on attachment and the field’s oral culture opens varied opportunities for quick-fix misapplications. A recent example was commercial training provided for thousands of UK social workers in using disorganised attachment as seen in naturalistic settings as an indicator of child maltreatment.17

Such misunderstandings and simplifications should be recognised as predisposed by obstacles to clear understanding of the disorganised category. These included the restricted circulation of Main’s texts, her use of ordinary language terms in highly technical ways without definitions, and by the repetition of a superficial account of Main’s ideas and findings by other attachment researchers. The lack of psychometric scrutiny of the disorganised classification has helped naturalise its status as a category, contributing to its quasi-diagnostic appearance, especially when combined with Carlson’s finding—intended as exploratory—showing that the category was associated prospectively with a general index of mental illness. More generally, the reification of attachment and attachment classifications no doubt facilitated the assumption of the commercial trainers that behaviour seen at home by child welfare practitioners would have the same meaning as behaviour seen in the research laboratory.18

Recently, problems in the role of attachment theory within child welfare practice have been the topic of a dedicated book by White, Gibson, Wastell, and Walsh.19 White and (p. 543) colleagues observe that Bowlby was active in promoting his theory to social workers.20 And in turn, social workers found in attachment a knowledge base to help them claim professional status, in which cases could be interpreted in terms of a credible theory. Attachment offers welfare professionals a framework that appears to predict later risk to a child’s health and development from the child or parents’ observable behaviour.

Over the years, both social work academics and policy documents have encouraged welfare professionals to use the image of secure attachment as the point of comparison when making assessments of parenting capacity. For instance, in the Department of Health practice guidance Assessing Children in Need and their Families, published in 2000, practitioners were told that ‘secure attachments are so important in the early years. Where these attachments are absent or broken, decisions to provide children with new attachment figures must be taken as quickly as possible to avoid developmental damage.’21 Taken at its word, in the absence of a secure attachment the guidance seems to suggest that children need to be considered for separation from their parents. White and colleagues identify that this position seems underpinned by the idea that ‘attachment patterns, once formed, are stable and set forever’.22

In recent years, with cuts to supportive services for families, the balance of child welfare in the UK has tilted firmly towards a focus on statutory investigation of families.23 White and colleagues regard attachment discourses as complicit in this shift. They suggest that, at least in UK child welfare practice, attachment provides ‘a handy vocabulary, a diagnostic gaze, learned-sounding re-descriptions of messy relationships and often a foil for moral judgements’.24 They report from an extensive ethnographic study of social workers in a child protection service. They found that attachment was never mentioned by social workers in their discussions or personal reflections. Nonetheless, appeals to attachment appeared repeatedly in official reports when judgements were made about parenting capacity. This expressed a contradiction facing social work, stemming structurally from its dominated position compared to both legal and medical professionals. Social workers require access to a category-based and prognostic knowledge system in order to operate and make qualitative judgements as professionals, but are not regarded as qualified to make use of reserved clinical diagnoses. A quasi-diagnostic use of attachment language appeals as a workable, if imperfect, solution to this structural conflict.

Uses of attachment discourse in the reports observed by White and colleagues were generally vague, essentially functioning as synonyms for a poor child–parent relationship. Similar observations have been made by other researchers. Potter’s ethnographic observations and North’s qualitative interviews suggest that vagueness in the use of attachment language assessments may reflect the climate of child welfare services and the family courts, (p. 544) and ambiguities around the professional status of social workers. This would imply that social workers draw on attachment concepts in their reasoning, but are vague about this in their reports, so as to avoid claiming expert levels of knowledge where this might be challenged by higher-status professionals.25 One of North’s participants, Bryony, described this predicament:

Whereas although we can say we have concerns about the attachment, I don’t feel we’re qualified enough to say, you know [softly], ‘They’ve got an attachment issue, you know, they’ve got a dis … organised [almost inaudible] … attachment or whatever,’ because I don’t feel we’re qualified enough. I don’t feel qualified enough to say that.26

North observed that across the interview, Bryony ‘spoke softly when she referred to attachment, seemingly to emphasise her shame at not feeling competent to make theoretically informed judgments. This was a response echoed by many interview participants.’ In North’s analysis, ‘social workers often want to be more proficient in their application of attachment theory and in how they describe their utilisation of it in assessments. They also experience frustration at lacking the skills to explain effectively how they have used attachment theory to indicate the potentially harmful outcomes for a child of experiencing emotional abuse.’27

By contrast, White and colleagues adopt a more sceptical conclusion. They claim that attachment language is being drawn upon, post-hoc, in reports as authority to justify practitioners’ intuitions about adequate and inadequate care: ‘the social workers referred to “good” or “positive” attachments, rather than the scientific community concepts of secure and insecure, to describe a good quality relationship between a child and their parents’.28 However, White and colleagues felt that the attachment frame of reference, even if post-hoc, had negative effects. It helped legitimate ‘the narrow focus on the mother–child relationship, and the responsibility of the mother for this relationship’, directing attention away from the family socioeconomic context, the availability of social support, and potential neurodevelopmental issues experienced by the child.29 In this, White and colleagues suggest that practitioners’ use of attachment theory was shaped by the priorities of child protection practice in the UK, (p. 545) which focus on procedure and risk assessment at the expense of a rounded attention to the context of families over time.

The primary target of White and colleagues is not attachment research, but the way that in the UK child welfare practice ideas from attachment theory may at times be ‘used with a mixture of excessive credulity and zealotry, a cavalier heavy-handedness and unsophisticated reductionism’. In fact, White and colleagues claim that attachment theory and research has a lot to contribute to child welfare practice: ‘we agree with the proponents of the theory that policy and practice is not sufficiently attachment minded. The complexities and nuances of the concepts are hidden by simplification … at its best, attachment research has produced ideas that practitioners can use to understand the quality of child–carer relationships when the child is anxious, scared, or upset, and to guide them in their work to improve familial relationships.’30

White and colleagues give the Love Barrow Families project as an exemplary use of attachment theory and measures by helping professionals. Love Barrow Families is an innovative service-delivery model for families with multiple complex needs based in Cumbria, UK. The model was codesigned with local families, and local families sit on the steering group. Key principles include working with the whole family and support for families to integrate with their local community. Rather than multiple services working with families in crisis at once, one member of the Love Barrow Families team acts as a key-worker for each family. Decisions about how to direct and prioritise work with the family are supported through use of attachment assessments with family members, which feed into a whole-family formulation and plan. An evaluation undertaken by Vincent documented that, compared to a matched comparison sample, the Love Barrow Families project reduced the number of children taken into care and the number of children on child protection plans.31 White and colleagues express appreciation that Love Barrow Families does not use attachment measures or theory to diagnose or label, but to sensitise practitioners to both the strengths and the needs of the different members of family systems.

Love Barrow Families draws especially on the ideas and assessments developed by Patricia Crittenden, which have been influential in the UK.32 Love Barrow Families regards Crittenden’s work as offering a range of clinically relevant archetypes, and proposals for distinct forms of therapeutic approach to clients displaying different forms of behaviour.33 There have been recent debates among attachment researchers about whether the assessments developed by Crittenden have sufficient sensitivity and specificity for use in court assessments of parenting capacity.34 White and colleagues are generally hostile to the use (p. 546) of attachment measures in court-mandated parenting assessments, in contrast to the use of these assessments to identify how best to proceed in supporting a family. This seems to parallel recent work of Madigan and colleagues in their creation of a brief assessment, usable by practitioners in real-time, for identifying forms of frightening or disrupted parenting (AMBIANCE-Brief).35 Madigan and her collaborators anticipate that this assessment will be able to help professionals identify forms of parenting associated with adverse child socioemotional development and to target supportive interventions. However, the assessment has been designed in such a way as to minimise its likelihood of being used in court assessments of parenting capacity, for instance in avoiding the implication that certain forms of caregiving are in themselves pathological. Madigan and colleagues are now in the process of validating the AMBIANCE-Brief in terms of its capacity to predict Strange Situation classifications, and evaluating the measure’s reliability when used in clinical and welfare practice. One contributing factor supporting Madigan’s work has been the availability of funding in Canada for research–practitioner collaborative inquiry. There also seems to be a broader shift towards greater engagement with practitioners and publics associated with the third generation of attachment researchers.36

In their book, White and colleagues at times depict practitioners as enthralled to simplified and haughty guidance about attachment theory. Presumably this is intended as a counterweight to what the authors regard as the untroubled circulation of lazy assumptions about attachment. One qualification that can be made regarding their account, though, is that there is more variety among guidance for practitioners than White and colleagues suggest. The Circle of Security graphic of the caregiver as a secure base and safe haven has gone into wide circulation, and represents an admirably effective visual representation of Ainsworth’s concepts; the Circle of Security Intervention itself is used in many countries and there is initial evidence that it contributes to less judgemental forms of supportive work with families.37 There are several books that do an excellent job in representing the available research evidence and theory, whilst also offering reflections relevant to helping professionals and parents. The best available in English is perhaps either Understanding Attachment and Attachment Disorders by Prior and Glaser or Attachment in Therapeutic Practice by Slade and Holmes.38 However, it should not be thought that effective mediation (p. 547) of research and practice always entails the implementation of the former in the latter. An example of a book that offers thoughtful integration of theory and practice, in an area where there is little empirical attachment research, is Blood and Guthrie’s Supporting Older People Using Attachment-Informed and Strengths-Based Approaches.39 There are many works that take care to distinguish contemporary scholarly consensus from ‘allodoxia’—a cut-price, simplified account offered up as if it had the same meaning as the technical conclusions of researchers. For example, in Golding’s Nurturing Attachments Training Resource for delivering parenting groups, a flipchart of potentially ambiguous terms is used to help participants keep in mind these distinctions.40

Nonetheless, White and colleagues are unquestionably right that a great deal of guidance for clinicians and social workers is overconfident and poorly informed.41 Moreover, this guidance is disproportionately likely to be available for free or cheaply, in contrast to works that offer greater access to attachment research in its complexity, which are frequently more expensive or behind journal paywalls. Three features can be identified that especially characterise allodoxic guidance about attachment for clinicians and social workers and related policy discourses. First, these texts offer little or no explanation of Bowlby’s behavioural systems model, Ainsworth’s operationalisation of sensitivity, or how disorganised attachment is actually coded. Instead they conjure with the broad ordinary language connotations of attachment, security, sensitivity, and disorganisation.42 Second, empirical attachment research over recent decades is ignored in favour of statements made by Bowlby in popularising writings, as well as selected other statements from early attachment theory implying a massive and stable causal influence of early care on later social behaviour.43 Attachment researchers are treated as generally all much the same, which allows for changes over time to be ignored. For instance, the concept of ‘felt security’ is routinely attributed to Bowlby, rather than Sroufe and Waters’ critique of Bowlby. Third, no reference is made to the qualified findings reported in meta-analyses or the findings from these meta-analyses regarding moderators of the effect of caregiving on attachment, or attachment on later development. It should be acknowledged that allodoxic attachment discourse contributes, often helpfully, to recognition that child–parent relationships are very important to children and their socioemotional development. However, the impression of a basis in empirical attachment research is spurious or heavily overstated. And some topics, such as disorganised attachment and attachment disorders, are pervasively mischaracterised.

(p. 548) At first sight it seems curious that many hostile academic discussions of attachment research, such as those of Vicedo (Chapter 1) and anthropologist critics (Chapter 2), possess the same three qualities: they mistake technical for ordinary language (e.g. regarding the meaning of ‘sensitivity’); recent attachment research is ignored in favour of classic statements by Bowlby; and the findings of meta-analyses are neglected. However, these qualities are less mysterious in light of Keller’s reflection (Chapter 2) that such critiques of attachment research are, at least in part, a proxy for criticism of the uses of attachment discourse in child welfare contexts. In this regard, a limitation of the work of White and colleagues is that they lean on statements by Vicedo that homogenise and caricature attachment research and its applications. Though White and colleagues refer to ethnographic observations and scrutinise several court reports, the empirical basis for their claims about the uses of attachment remains relatively limited.

As mentioned in the Introduction to this book, colleagues and I have several studies underway exploring how attachment theory and measures are used by child welfare practitioners and clinicians working with children. Findings are just in, and results will be reported over the coming couple of years; this presentation of initial impressions should not be taken to pre-empt the formal analysis. However, already these initial impressions indicate more heterogeneity than suggested by Vicedo or White and colleagues. In a project led by Sarah Foster, we conducted interviews with 24 children’s services social workers, and asked them to offer responses to two vignettes of fictionalised cases drawn partly from Serious Case Reviews. This was part of a wider project comparing the responses of different professional groups: social workers, general practitioners, and clinical psychologists. What was especially striking from the interviews with the social workers was the significant individual variation in knowledge of and views on the potential value and actual use of attachment theory concepts and discourse in practice. This was despite the fact that all the social workers were employed by just two Local Authorities, both in the same region of the UK. For some of the social workers interviewed, they regarded attachment theory and discourses as having little role in their work, regardless of guidance suggestions that they take attachment into account in assessments of children. This stance was most frequently underpinned by a lack of confidence in any theories relevant to social work. Theory was seen as ‘high’ knowledge, beyond the reach of ordinary practitioners like themselves. However, some social workers drew on other theories but were specifically critical of attachment theory. For example, one social worker was concerned that use of attachment theory could risk drawing attention away from fully acknowledging different cultural family contexts.

However, many of the interviewed social workers felt attachment was directly important to their work with children and families. Practitioners reported using attachment to inform their thinking about a variety of matters, including in thinking about what a child’s repeated pattern of behaviour in the family context might mean, and how a parent’s own childhood experiences might be affecting how they are currently responding to their child. However, among practitioners who appreciated attachment theory and its relevance to their work, here again there was variation. Some practitioners were enthusiastic about attachment theory and research because it offered a lens on the relationships, emotional life, and socioemotional development of everyone. Other practitioners were enthusiastic but viewed attachment as ‘one aspect of everything that you take into consideration’ and only ever helpful alongside other theories and frameworks.

In direct contrast to the finding in White and colleagues’ research, there were multiple interviewees who said in interview that they used attachment theory and research extensively to inform their practice, whilst intentionally avoiding the use of attachment terminology in (p. 549) their reports. Explaining this stance, some of the social workers stated that identification of individual differences in attachment, such as disorganised attachment, required specialist assessment training which they did not have. Many expressed concern over being challenged on their expertise in court if they used such terms. Another set of concerns raised by interviewees was that many felt that using attachment terminology was inappropriate, as it may be inaccessible to families and unhelpful or misleading for other practitioners:

I try not to put labels on, this child is suffering disorganised or ambivalent attachment, I think in terms of your theory … you know what you’re looking for in a secure attachment and you know when it’s not good! But it’s focusing more on the behaviours and what the behaviours will be telling you as opposed to sticking a label, an attachment theory label on it. Because it’s just words, we need to understand what that means for these children and what we’re actually seeing.

Whilst much rarer amongst those interviewed, a few social workers stated that they did use attachment classifications in their written reports despite not having trained in conducting attachment assessments or carried out formal attachment assessments in these cases.

Practitioners drew on attachment language in responding to the two vignettes. There was a good deal of confused use of attachment terminology in these responses. Examples included: talking about the ‘strength’ of attachment, reasoning about ‘poor’ attachment, conflating attachment with love, using the terms ‘attachment difficulties’ and ‘attachment disorders’ interchangeably, and confusion about whether a child can be attached at all to a maltreating caregiver. This does support the concerns White and colleagues have raised about the use and understanding of attachment. However, what was also clear from the interviews was that many social workers were reflective about the limitations of their knowledge and of the version of attachment theory in circulation within child welfare contexts. We were told:

I think it [attachment theory] can be vague and I think it can be overused by people, or not always used appropriately. So I sometimes read reports or I hear professionals talking about bad attachment and good attachment, and I understand what they’re talking about but I don’t think it really explains anything.

One of the reasons I don’t like using that word [attachment] too much is because I think it’s widely kind of used in, if not, it might be harsh to say it’s misused but it’s used in lots of different ways and so you can’t really be confident that you’re talking about the same thing when you talk to people.

In another study, led by Barry Coughlan, we asked practitioners from various clinical services (including child and adolescent mental health, neurodevelopmental teams, and primary care) to discuss cases they perceived as ‘attachment-related’ versus ‘neurodevelopmental’. This was part of a broader study of how clinicians distinguish autism, ADHD, and attachment-related difficulties, drawing on interviews and examination of clinical records from children’s mental health services. Regarding neurodevelopmental conditions, clinicians tended to lean on psychiatric nosology and standardised assessments to buttress their formulations. Yet these frames of reference did not seem available or desirable to clinicians when thinking about attachment-related cases. Instead, practitioners tended to place greater weight on unstructured observations and the non-standardised taking of a developmental history. In interview, few clinicians used standardised attachment assessments. In fact, references to attachment literature were relatively sparse in the interviews, though several (p. 550) clinicians described keeping attachment ‘in mind’ alongside the array of other frameworks when making decisions about case conceptualisation and intervention planning. However, they tended to be vague about what behaviours would prompt them to think about attachment. Instead, attachment theory seemed to be drawn in when clinicians had other evidence of insufficient care or maltreatment, or where there was a discrete precipitating event such as a major separation. Most practitioners indicated a preference for the general phrase ‘attachment difficulties’ over the technical clinical term ‘attachment disorder’. But practitioners did not seem confident in articulating the difference between these terms. Practitioners were also unsure what interventions or support should be offered for families where attachment was identified as a relevant issue.44

Finally, in a third study, led by Helen Beckwith, we used Q-methodology to examine the perspectives held about attachment concepts and research findings among two groups: (i) 30 child mental health clinicians and (ii) 30 established attachment researchers. A variety of other self-report data were also collected. We have been fascinated to see both the convergences and divergences in the perspectives of these groups. Clinicians and researchers alike saw value in attachment theory as a framework for facilitating personalised care for children and for making decisions about fostering and adoption placements. Further, all agreed that ‘attachment theory could be used more precisely within mental health practice’, and that attachment research and measures were largely inaccessible to clinicians. Both researchers and practitioners were mindful of the difficulty of maintaining the integrity of attachment concepts and methods when applying them to practice.45 Interestingly, both researchers and clinicians were unsure whether callous and unemotional traits in children originate from their early attachment experiences, or what the role is of attention within attachment processes.

One line of difference was that attachment researchers held strongly that good quality care throughout childhood is a better predictor of future mental health than a child’s early attachment pattern. Clinicians were unsure. Researchers also felt that the most effective interventions to improve child–caregiver attachment are those that target security, presumably following the meta-analysis from Bakermans-Kranenburg and colleagues (Chapter 4). By contrast, clinicians appeared not to regard this as a practicable area of focus, and held that they could achieve improvements in child–caregiver attachment through other means. We were also interested that clinicians endorsed strongly the idea that early attachment experiences determine how the brain develops, whereas the research community offered neither agreement nor disagreement, presumably waiting on further evidence. This may reflect the role of allodoxia, perhaps intersecting with the wider discourse that if something influences the brain then it is more real.46

Clinicians were enthusiastic about the relevance of attachment language to helping them make sense of cases, especially in understanding dyadic processes. However, they were less (p. 551) positive about the relevance of existing attachment assessments. Researchers held that at root, attachment is a dyadic and relational phenomenon, for all that it can stabilise with development. Clinicians agreed in principle that attachment is dyadic in nature, but regarded this as rather immaterial in clinical services structured by individual diagnoses and labels. Though these are just indications of some initial findings, already they suggest some provisional conclusions. They indicate that researchers and practitioners have many important points of convergence; that divergences in perspective often express the specific and different demands of the research and clinical contexts; and that there is appetite from both sides to improve the precision with which attachment theory is used in clinical practice with children. One line of future collaborative work may be efforts to validate the use of attachment assessments in applied contexts, to see if they can demonstrate benefits compared to assessment as usual. Another may be efforts to make the findings of attachment research more readily available to practitioners and publics.

The exhaustion of attachment research?

As we saw in the above section, White and colleagues advocate for a better understanding of attachment research among practitioners, to help counteract the dangers of the distorted and deadened version currently in circulation, and especially the dependence on assumptions from Bowlby’s early writings. Yet, in their book, they offer little discussion of changes in attachment research over time or diversity among attachment researchers.47 Relevant to their concerns are discussions that have been taking place among attachment researchers themselves about what within attachment research remains lively and what is fading from view. These discussions have a long history, even if they have become especially salient in the past few years.

In the late 1980s, van IJzendoorn sought to characterise the contemporary state of attachment research in terms from the philosophy of science. He drew distinctions between research paradigms as characterised, by degrees, in terms of ‘formation’, ‘construction’, ‘saturation’, and ‘exhaustion’.48 In ‘formation’, research is starting up, exploring the possibility of a programme of work. In the ‘construction’ stage, research is focused on the verification of bold hypotheses, and social communication in a growing community is facilitated by papers and symposia. There is a dancing, exhilarating sense of excitement and discovery to the work, alongside endless hard work. In ‘saturation’, research is focused on developing consistencies and responding to inconsistencies in the theory, journals are established to formalise the field, and ideas from research get applied in practice in ways that do not feed back into further scientific developments. In this stage, the big discoveries seem to have been made, and empirical work becomes increasingly about filling in gaps and qualifying effect sizes. (p. 552) In ‘exhaustion’, a research paradigm becomes primarily concerned with defending established orthodoxies, and communication becomes rigidified in handbooks and training institutes. Alternative perspectives gain ground in achieving funding, recruiting new researchers, and securing institutional recognition. Van IJzendoorn characterised Bowlby’s development of attachment theory as the ‘formation’ stage for the paradigm and Ainsworth’s introduction of the Strange Situation procedure as the ‘construction’ stage, which lay the basis for a cumulative empirical research programme in which new findings continually inform the development of theory.49 However, looking about in the late 1980s, van IJzendoorn also identified symptoms of the saturation stage. There was a sense that the major discoveries had been made, and evidence that applications of attachment theory in psychotherapy, parent education, and policy were not feeding back into the priorities of research practice.

It is interesting to reflect today on van IJzendoorn’s characterisation of the field of attachment research in the 1980s. There seem to be aspects of the research paradigm that resemble construction and saturation—and exhaustion. Since van IJzendoorn’s analysis, attachment research has seen major developments, not least the extension of the paradigm to study attachment phenomena in adulthood. Today, some developments are still in the stage of ‘construction’. For instance, research on the relationship between individual differences in attachment and sexuality remains an area in which early hypotheses are still being validated (Chapters 3 and 5). By contrast, some developments seem to have achieved saturation. For instance, the relationship between caregiver sensitivity and infant attachment is now well established, and current discussions have been mostly concerned with qualifying effect-sizes and communicating with practitioners. However, there remains lively interest in other factors contributing to intergenerational attachment processes.50 And the proposal by Woodhouse, Cassidy, and colleagues (Chapter 2) to refine sensitivity to secure base provision may also reignite discussions in this area.

Over recent years, a few commentators have characterised attachment research as reaching exhaustion. Unquestionably, criticisms of the paradigm and of individual attachment researchers have gained ground in academic circles. Harkness has observed that ‘earlier critics of attachment theory have recently been joined by others; if not exactly a chorus of critics, there are now enough to form the basis for an organized, multi-referenced, and multi-faceted counter-offensive’.51 While popular and clinical interest in attachment remains high, it has become increasingly difficult to get funding for attachment research. Discussing the UK context, Fonagy recalls that ‘attachment theory had ten good years in the research community from about 1985–1995. We had excellent research grants from the Wellcome, the Economic Social Research Council and the European Union; we were invited speakers at British Psychological Society congresses, the British Association and all that. Attachment theory is still good, but research interest has moved to neuroscience.’52

(p. 553) In a watershed development in 2016, the National Institute of Mental Health in the USA removed the Ainsworth Strange Situation from its list of recommended procedures for publicly funded mental health research, citing the debt of attachment theory to psychoanalysis and its tendency to ‘reify … theoretical claims’ as essential problems with the paradigm.53 The framing of this decision was striking: in contrast to other judgements in the document, the exclusion of the Strange Situation was made without any justification in terms of any of the scientific criteria identified by the National Institute of Mental Health as relevant to recommendations, such as psychometric properties, longitudinal stability, standardised administration, and cultural specificity. Instead, it was simply asserted that the Strange Situation was an invalid measure. And a letter of query led by Lyons-Ruth was ignored by the National Institute of Mental Health.54 From the information available, it would appear that the Strange Situation has been rejected for public funding on the basis of the disfavour of attachment research among psychological researchers, rather than on the grounds the National Institute of Mental Health itself had established as criteria. Though raked by the claws of this development, attachment research in the USA has continued. The central preoccupations of contemporary developmental science—brains, genes, big samples, and clinical application—are all viable for attachment research.55 However, work in the USA using the Strange Situation has had to stow aboard grant applications that justify themselves in other terms. Whilst the situation is less severe in other countries, the National Institute of Mental Health announcement illustrates the challenge of credibility felt by attachment research more broadly.

Fonagy and Campbell have offered the view that, essentially, the party is over for developmental attachment research pursued in the manner of the second-generation researchers such as Main and Sroufe.56 Above all, Fonagy and Campbell claim that even if the research paradigm was built upon the categories from the Strange Situation, it is now coming to grief upon them. They regard categories as an outdated way of representing human differences. This presentation of this criticism seems to reflect both the psychometric criticisms discussed in Chapter 4 and wider contemporary discourses that treat categories for human beings as a set of oppressive and misleading expectations, inappropriate for societies characterised (p. 554) by fluidity and precarity.57 Fonagy and Campbell also argue that whilst correlations between caregiving, attachment, and mental health have been in the expected direction across thousands of studies, the associations have been less strong than Bowlby appeared to predict. Furthermore, the sciences on which attachment theory was erected have moved on.58 Fonagy and Campbell acknowledge that attachment research has brought valuable attention to the question of the evolutionary function of mental health and illness. However, they feel that since Main’s concept of conditional strategies, the dialogue between attachment research and evolutionary biology has generally petrified (though Belsky and Crittenden are cited favourably as exceptions).59 Fonagy and Campbell argue that developments in evolutionary biology in thinking about cultural or gene–culture co-evolution models seem to have passed by attachment research, which is looking rather old-fashioned.60

Ultimately, Fonagy and Campbell predict not a replacement of the developmental tradition of attachment research, but its supersession by an approach that incorporates its strengths.61 In the process, they anticipate that familiar aspects of attachment research, above all the labour-intensive Strange Situation and its lumbering categories, will—and should—be put aside.62 One tradition that may fill part of the remaining space is the social psychological approach. However, given that this approach remains without an adequately elaborated account of trauma, its clinical relevance remains limited. Fonagy and Campbell instead recommend their own assessments of mentalising. Mentalising is the capacity to imagine, perceive, and interpret human behaviour in terms of intentional mental states (e.g. needs, desires, feelings, beliefs, goals, and reasons).63 Fonagy and Campbell regard deficits in mentalising as associated with attachment-related processes since they are predisposed by problems within early family relationships. But they argue that it is mentalising and the capacity to learn from interactions with others rather than attachment that is important for the development of adult mental health and caregiving behaviour.

Fonagy and Campbell’s intent in their paper is clearly in part advocacy of their own theory and measures. The attendant predictions of the demise of attachment research also have (p. 555) something of the air of a challenge, more than simply a statement of beliefs about the future. Some similar points about the limitations of attachment research have been made by Pehr Granqvist, who sees them as areas for renewal and revision of the paradigm, rather than as evidence of exhaustion.64 Though not explicitly mentioned by Fonagy and Campbell or by Granqvist, a factor in the wider context of their appraisal of the state of attachment as a paradigm is the retirement of the second generation of attachment researchers. The infrastructure built by this generation has been inherited in recent years by a new generation of research leaders.65 Some examples, among many, in the developmental tradition include Sheri Madigan, Pasco Fearon, Carlo Schuengel, Chantal Cyr, and Glenn Roisman; in the social psychological tradition Chris Fraley, Omri Gillath, and Gurit Birnbaum. Granqvist pursues work in both developmental and social psychology, making use of each set of attachment measures. Patrick Luyten and others represent a younger cohort of researchers on mentalisation who likewise draw from both traditions.

If a new generation of attachment researchers can be identified, this is not defined by their chronological age. Nor, besides Roisman’s work at times,66 is there a sense of insurrection against an older generation. Rather the point is that a bundle of problems that could previously be ignored or postponed are coming due, at a point where long-time research leaders have retired, and the field is seeing a changing of the guard. As such, a third generation of attachment researchers have been constituted by the playing out of problems in theory and method at a time when a generation of researchers who had direct contact with Ainsworth are putting down their tools. This third generation has also been shaped by active efforts to guide its development: in the developmental tradition, many of the third generation have received mentorship from van IJzendoorn, who has encouraged cross-group, international collaboration in meta-analytic research and the adoption of new methodologies from medicine and biostatistics.67 Shaver and Mikulincer have likewise played a concerted role in mentoring a younger generation of rather ingenious and highly skilled social psychologists.

The first question the new generation of laboratory heads face is what must be preserved, altered, or rejected from the legacy of their teachers in responding to the field’s challenges and opportunities. This is the challenge pointed to by Fonagy and Campbell’s predictions of the supersession of attachment theory and methods. Inheritance is never a given; it is always (p. 556) a task of filtering and sorting, of deciding between alternative pasts and the different futures they might make possible. Inheriting effectively is hard and takes courage. In a sense, the third generation of leaders of attachment research need to figure out together what genre of story they are in. Is it a swashbuckler? Fonagy and Campbell seem willing to battle on deck. Is it a dystopian story of complicity with oppressive state power, as critics such as White and colleagues suggest? Is it a mystery, with new questions, methods, or interdisciplinary collaborations drawing the paradigm back towards the stages of construction and formation?

Will the story be an elegy, in which the present is oriented only towards the achievements of the past? It looks not. So far, the early chapters are of a coming-of-age story. Ainsworth’s immediate students remained loyal to the four-category system. Yet as we saw in Chapter 4, renegotiation of this position seems to have a symbolic role, as a banner marking the presence of some new leaders, who now hold the datasets and the research agenda.68 There are other changes too, coincident with the retirement of the second generation of attachment researchers. One is particular attention to moderators of stability or intergenerational transmission, including predictable forms of movement or cross-transmission between patterns of attachment. The third generation also seems more able to cross-pollinate methods and ideas between the developmental and social psychology traditions of attachment research. In particular, the psychometric rigour and priorities of the social psychological tradition have been adopted by younger researchers in the developmental tradition. However, there have also been a variety of other migrations. One of special interest is that the model of minimising and maximising strategies as elaborated by Shaver and Mikulincer has frequently been re-imported back into developmental psychology—though, at least to date, the wider behavioural systems theory of Shaver and Mukulincer, and their measures of minimising and maximising strategies in the domains of sexuality, caregiving, and dominance, have not been taken up by developmentalists.

In an attempt at stock-taking the current state of attachment research on children, Schuengel and colleagues reported from a bibliometric study. The results suggest that the past few years represent a period of transition. For instance, citations of Patterns of Attachment have been in decline since 2015, whereas citations of meta-analyses have gone from strength to strength.69 Among meta-analytic research on child attachment Schuengel found a comparative decline in focus on the concerns of Ainsworth and Main. He found a comparative increasing focus on mental health symptoms and supportive interventions with families, a growing concern with fathers, and rising attention to anxiety and avoidance conceptualised as the latent dimensions of individual differences in attachment.

Even if there are definitely some pulls towards ‘exhaustion’ of the developmental tradition, Schuengel’s findings indicate that, in general, attachment research is changing, not fading. There are significant strengths that mark its ongoing vitality. Three developments can be used as illustrations, though by no means are intended as exhaustive. These developments are: Individual Participant Data meta-analysis; security priming; and intervention research. (p. 557) Many other lively areas of research have been discussed in previous chapters. Two of particular note are the growth of attention to secure base scripts and observations of secure base/safe haven provision—with a cadre of talented younger researchers currently pursuing a programme of work very much in the ‘construction’ stage (Chapter 2). Other lively areas of research build from later research groups, to be considered in a later volume, and so would not make sense to discuss here. This includes extensive research using biomedical measures, in acknowledgement of the embodied aspects and underpinnings of attachment relationships. This was touched on in Chapter 4, discussing the work of Dante Cicchetti. In making a selection, Individual Participant Data meta-analysis, security priming, and intervention research have been chosen as illustrations because they capitalise precisely on the maturity of the field’s theory and methods discussed in this book, whilst also reflecting the changing priorities and methodological infrastructure of attachment research as an area of science inquiry.

Individual Participant Data meta-analysis

A first area of particular vitality in contemporary attachment research has made use precisely of the advantages of ‘saturation’. Over the decades, attachment research on young children has used the labour-intensive Strange Situation or measures adapted from it like the MacArthur preschool system70 or the Preschool Assessment of Attachment.71 This has permitted the accumulation of thousands of relatively small-scale studies that provide little pieces of the puzzle. Each on its own is unable to offer confident guidance for interventions with families, justify commissioning decisions, or inform policy to protect and support young children. Attachment researchers, particularly van IJzendoorn and colleagues, were early adopters of meta-analysis as a methodology for studying the overall effect sizes identified by these numerous small studies and for examining factors accounting for variance in effect sizes between studies.72 There were also attempts to pool data from different studies more directly in order to increase statistical power and explore potential moderating factors.73 This offered a richer array of variables and greater depth than reliance on the published record as in traditional meta-analysis. Data pooling permits investigations that are wider than single studies and deeper than traditional meta-analyses. However, data pooling efforts in attachment research foundered in the 1990s, in part because the statistical techniques for multilevel modelling were not available to developmental psychologists. Instead, the use of traditional meta-analyses flourished.74

(p. 558) One recent such meta-analysis was completed by Verhage and the Collaboration on Attachment Transmission Synthesis in 2016.75 The researchers examined the association between AAI for parents and Strange Situation classifications for dyads from 95 samples, with 4,819 participants in total. They found that the association between secure-autonomous state of mind on the AAI and secure attachment in the Strange Situation was r = .31 and for unresolved state of mind to disorganised infant attachment r = .21. This was very substantially down from the meta-analysis conducted by van IJzendoorn in 1995 (r = .47 and r = .31, respectively).76 Caregivers with dismissing states of mind were more likely to have avoidant attachment relationships and less likely to have secure attachment relationships, but they were no less likely than other caregivers to be part of dyads receiving an ambivalent/resistant classification. Similarly, preoccupied states of mind were associated with more ambivalent/resistant and fewer secure, but not with fewer avoidant attachment relationships. This suggests that classical and widely repeated assumptions since Main and colleagues regarding the mechanisms that lead to avoidant and resistant attachment might need additional thought, and especially the underarticulated concept of ‘inconsistent care’ inherited from Ainsworth. Such a conclusion aligns with proposals by diverse researchers, proposing ways that dismissing states of mind may also contribute to resistance, and preoccupied states of mind to avoidant attachment relationships.77

Another curious finding was that speakers with unresolved/disorganised states of mind regarding attachment were less likely to have secure and avoidant attachment relationships with their infant, but the dyads were no less likely to receive an ambivalent/resistant classification. Verhage and colleagues observed that this finding has pertinence to debates about how hard or how fluid the distinction is between resistance and disorganised attachment (Chapter 4). Around 25% of the relationship between AAI and Strange Situation classifications could be accounted for on the basis of assessments of caregiver sensitivity; after correction for test–retest reliability, this left a little under half of variance explained by sensitivity.78 Examining moderators, Verhage and colleagues reported that the association between the AAI and the Strange Situation was stronger in low-risk samples and weaker in at-risk samples. At first sight, the decline in the association since van IJzendoorn’s meta-analysis in 1995 could be thought to reflect the fact that early samples were generally with low-risk samples and subsequent samples have more frequently been with high-risk samples. However, Verhage and colleagues found that, in fact, risk status did not account for the effect of publication year on the reported effect size.

(p. 559) Subsequently, the Collaboration on Attachment Transmission Synthesis has pursued Individual Participant Data meta-analysis, in which data are pooled from studies using the same measures. The Collaboration on Attachment Transmission Synthesis received data from 58 research laboratories from around the world that had used the Strange Situation and the AAI (4,396 parent–child dyads); 67% of eligible studies contributed their data. The first analyses from this collaboration are just appearing, and already are offering important qualifications to existing theory and new insights, for instance for understanding how poverty and social adversity impact parenting. Adding to their previous traditional meta-analysis, Verhage and colleagues found that risk factors lowered the likelihood of a secure child–caregiver attachment if the caregiver has a secure-autonomous state of mind regarding attachment. The finding suggests that secure-autonomous states of mind do not offer as much resilience as previously expected by attachment researchers against contextual risks.79 However, low education and single parenthood had no effect on the strength of the association between the AAI and the Strange Situation. Neither did the gender of the caregiver. In a further study, Verhage found that contextual risk factors weaken the association between caregiver sensitivity and attachment, again illustrating the importance of the challenges faced by families in the development of attachment relationships.80

Another study by the Collaboration on Attachment Transmission Synthesis was led by Lee Raby and sought to examine the psychometric structure of the AAI. As we have seen, Main, Goldwyn, and Hesse initially offered a four-category model (Chapter 3). However, Roisman, Fraley, and colleagues have subsequently proposed instead a two-dimensional model modelled on the ECR, with avoidance and anxiety as latent dimensions (Chapter 4). Raby and colleagues found that the distribution of the data-points suggested that individual differences in attachment states of mind reflect differences in degree, not kind. However, the findings were not clear regarding what dimensions should be used. A two-factor model based on dismissing and preoccupation was found to be an adequate fit. But a three-factor model based on dismissing, preoccupation, and lack of resolution was also found to be an adequate fit. Raby and colleagues argued in favour of the two-factor model on the basis of parsimony.81 However, they did not examine which model might best serve to predict variables of interest such as Strange Situation classifications or other indices of child mental health. It is likely that predictive validity will be a key criterion for any potential overhaul of the dominant approach to reporting results from attachment research.

Over the coming years, it is possible that the social psychological tradition will adopt Individual Participant Data meta-analysis as a methodology. Pooled data would permit exploration of some of the issues faced by this tradition, such as the status of security and the implications of trauma. Individual Participant Data meta-analysis, much like the present book, is oriented by a sense that the past can open anew in response to curation and the questions of the present. It may be that many social psychology laboratories will have thrown (p. 560) away their scale scores. That has certainly been the case for some research groups in the developmental attachment tradition: a third of eligible laboratories did not contribute data to the Collaboration on Attachment Transmission Synthesis. But enough may have been preserved to make application of Individual Participant Data meta-analysis a feasible and fruitful possibility for the social psychological tradition and for answering other questions in the developmental tradition. Work is currently underway, for example, to assemble an Individual Participant Dataset for randomised control trials of attachment-based interventions. The ambition is to ask questions with greater power and precision than have been possible before, about what works for whom, and how, in supportive interventions with families.82

Security priming

A second area in which attachment research seems to be thriving and breaking new ground is in studies of priming and attachment-relevant states—as opposed to longer-term attachment patterns or styles. The concept of ‘attachment states’ originated in the developmental tradition of attachment research, introduced by Fonagy, Steele, and Steele in 1991. Fonagy and colleagues argued that Bowlby’s ‘internal working model’ construct may well not be a unitary whole. Instead it likely encompasses some components that are relatively unresponsive to environmental cues and other components that are relatively responsive to specific cues.83 This distinction was accepted by other researchers in the developmental tradition of attachment research.84 However, as discussed in Chapter 2, researchers in the developmental tradition have tended often, in practice if not in theory, to treat attachment as a trait. In line with this, state attachment fluctuations have not been much studied in childhood by developmentalists, since attachment has generally been treated as relatively stable.85

The differentiation between global adult attachment style and more particular states was developed strongly by the social psychology tradition, especially by Mark Baldwin.86 An important early study was reported by Baldwin and colleagues in 1996. Participants were asked to characterise their most significant relationships and then to indicate which attachment style best captured their feelings about each relationship. Baldwin and colleagues found that most participants reported having experienced secure, avoidant, and anxious experiences (p. 561) across their relationships. Despite this diversity, participants reported more relationships that fitted with their global attachment style. Baldwin and colleagues also found that this global attachment style was influential for shaping self-report of dating preferences. Priming secure, avoidant, and anxious experiences did not influence self-report global attachment style. But the researchers did find that priming altered reported dating preferences among their participants, for instance with a prime for avoidance decreasing the attractiveness of a partner offering an intimate and committed relationship.87

The findings suggested to Baldwin and colleagues that a person’s global attachment style is generally more accessible for informing semantic appraisals. However, other information about attachment experiences, which could inform alternative attachment styles, is potentially available in the hierarchical network, and specific cues in the environment may influence this availability. The qualities of an avoidant relationship, for example, may make available insecure generalised schemas even for someone with a secure attachment style. Following up on the work of Baldwin and colleagues, Mikulincer and Arad published a vignette-based study in 1999 of the effects of attachment style and security priming on accessibility for participants of incongruent information about their partner.88 They found that a secure attachment style was associated with greater openness to and better recall of incongruent information when it suggested that their partner was caring and available. And asking participants to think about a time when others were available to them for support and comfort—a security prime—made an independent additional contribution to openness to and recall of incongruent positive information. Attachment style did not moderate the effect of the prime. Mikulincer and Arad concluded that a security prime led to behaviour that resembled a secure attachment style by activating relevant parts of the hierarchical network of associations. However, curiously, this activation of the ‘attachment state’ appeared to operate through an independent process to the role of global attachment style in influencing behaviour, as there was no interaction between attachment style and the security prime in predicting behaviour.89

Over the subsequent decades, Mikulincer, Shaver, and colleagues in the social psychology tradition have regularly found that similar effects can be obtained on the basis of a secure attachment style or on the basis of a prime for security.90 Furthermore, they have demonstrated that the effect of the secure base prime cannot be accounted for in terms of variations in mood or self-esteem.91 Part of what has been electrifying about the new line of research (p. 562) on priming has been that, in many domains, a prime seems able to produce prosocial effects for all participants that would otherwise only be limited to those with a secure attachment style. For instance, Mikulincer and Shaver found that a secure attachment style and a security prime contributed to greater tolerance towards out-groups.92 Mikulincer, Shaver, Gillath, and Nitzberg reported that secure attachment style and a security prime contributed to greater willingness to help someone in need.93 And Shaver, Mikulincer, Lavy, and Cassidy found that a secure attachment style and a subliminal security prime could contribute to less defensive and more constructive responses to feeling hurt by a romantic partner.94

However, some research has documented an interaction between attachment style and a security prime. One suggestive finding reported by Mikulincer, Hirschberger, Nachmias, and Gillath was that attachment insecurity can moderate the effect of a security prime on positive mood. If a security prime is followed by visualisation of an experience of separation from attachment figures, only participants with a secure attachment style reported an increase in positive mood from the security prime. The effect disappeared for participants with an avoidant or anxious attachment style.95 Another relevant finding by Mikulincer, Shaver, Sahdra, and Bar-On was that a secure base prime could overcome mental depletion, allowing a participant to remain responsive to their romantic partner during a laboratory-based task.96 The prime also removed the negative effect of an avoidant attachment style on the extent of help offered. By contrast, there was no interaction between attachment style and the security prime in intention to help strangers.

Mikulincer and colleagues speculated that with any longstanding romantic partner there will be a rich network of associations, more so than for a stranger. The secure base prime prior to the couple interaction may have activated the ‘reservoir of positive feelings and memories’ within the hierarchical network of associations, knocking out the minimising strategy that might otherwise be enacted.97 Later research demonstrated, however, that this effect was specific to help with practical rather than emotional tasks. Mikulincer, Shaver, and colleagues observed participants responding to their romantic partner in two conditions: discussing a distressing problem and exploring personal goals. An avoidant attachment style was generally associated with lower supportiveness in both conditions. A security prime eradicated this effect in relation to the partner’s exploration of personal goals. However, the security prime had no effect on the condition where the partner was discussing a personal problem and needed emotional support.98

(p. 563) Yet perhaps the most curious set of interactions between attachment style and priming have been in relation to anxious attachment. Though, in general, security primes tend to foster comfort, happiness, and more positivelyregarded behaviour, the interaction between security priming and an anxious attachment style in several cases has had the opposite effect. For example, Taubman-Ben-Ari and Mikulincer examined the effect of attachment style and security priming on driving using a simulator. They found that an anxious attachment style was associated with higher willingness to drive recklessly. Furthermore, whereas for other participants a security prime reduced reckless driving, among participants with an anxious attachment style the security prime increased recklessness.99 The researchers wondered whether the security prime was being interpreted by their participants as signalling the availability of a secure base, contributing to overconfidence, without the feeling of being cherished that would otherwise reduce willingness to endanger oneself. Mikulincer and colleagues suspected that rumination may be another factor involved in the paradoxical effects of security priming on participants with an anxious attachment style. The security prime may cause participants to think about the unavailability of their attachment figures. Support for this conclusion came from a study by Mikulincer, Shaver, and Rom.100 The researchers found that an explicit security prime (recalling experiences of being cared about), but not a subliminal presentation of the names of attachment figures, was associated with less creativity and effectiveness in solving laboratory-based tasks for participants with an anxious attachment style. It was concluded that an explicit security prime provides a prompt for rumination, in contrast to a subliminal prime.

In opening priming effects as a domain of inquiry, the social psychology tradition has allowed the attachment paradigm as a whole to put out new branches. It has taken some time, but researchers in the developmental tradition have begun to explore priming techniques, such as in the collaboration between Bosmans, van IJzendoorn, and Bakermans-Kranenburg.101 One obstacle is that the AAI is not especially well adapted for repeated administration to assess attachment states or validate the effects of a prime. But neither is the ECR. The State Adult Attachment Measure has been developed by Gillath and colleagues (Chapter 5) to circumvent this issue in the social psychological tradition. There is likewise no intrinsic obstacle to the use of priming for researchers in the developmental tradition. For instance, repeated observation of caregiver–child interaction in free-play situations (not the Strange Situation) could be pursued by developmental researchers to examine the effects of a prime. The recent development of briefer assessments for the study of caregiver–child interaction, such as the brief Attachment Q-Sort or the AMBIANCE-Brief, may facilitate this by reducing the labour required to code observational data.102 There remain some concerns about research on attachment priming, since, in general, studies of priming in social (p. 564) psychology have an uneven record of replication.103 Nevertheless, so far, results regarding attachment priming specifically have been encouraging.

Intervention research

A third development has been the growing prominence of intervention research both for the sake of clinical relevance and in order to articulate potential causal mechanisms. Description of a variety of attachment-based interventions for families with children of different ages is offered in the recent Handbook of Attachment-Based Interventions, edited by Howard and Miriam Steele.104 A concern with interventions among attachment researchers is, of course, far from new. Attachment theory was initially developed in a clinical context by Bowlby, and many of Ainsworth’s graduate students pursued training as clinicians. The development of supportive interventions with parents based directly on attachment principles and evaluated using the Strange Situation began in the 1980s. The first such interventions were STEEP and, though perhaps more grounded in psychoanalytic than attachment principles, Lieberman’s child–parent psychotherapy (Chapter 4). The Circle of Security intervention, discussed in previous chapters, emerged out of a determinate attempt to translate the central principles of Ainsworth’s concept of sensitivity into a supportive intervention with families.

Over time, the evidence base has grown for interventions designed with attachment principles in mind and evaluated with attachment measures. A systematic review of this literature by Mohamed and colleagues identified 32 empirical studies of attachment-based interventions. The most frequent elements of the interventions were psychoeducation, increasing parents’ awareness of the functioning of the attachment system, and supporting the parent relationship. Other elements, more specific to some interventions, included the intervener ‘subtitling’ the child’s behaviours by indicating what the behaviours seem to be saying, video-feedback, and supporting the development of caregiver insight into their own behaviour.105 Among this growing literature, two attachment-based interventions for families with young children have generated both considerable research evidence and clinical interest: Video-feedback Intervention to promote Positive Parenting (VIPP) and the Attachment and Biobehavioural Catchup (ABC). Both have had significant, and growing, penetration into child health and welfare services.

Video-feedback Intervention to promote Positive Parenting

VIPP was developed by Femmie Juffer, Marian Bakermans-Kranenburg, and Marinus van IJzendoorn. It was first used in a randomised control trial in a sample of 130 families with infants who had been internationally adopted in their first weeks of life. When the children were 9–12 months, home visits were undertaken, during which sessions of video-feedback were used to support parenting sensitivity. Whereas video-based instruction to parents had (p. 565) proven ineffective or even counterproductive,106 it was anticipated that drawing on 10- to 30-minute clips of the parent’s own behaviour with the child in ordinary situations would be more relatable and empowering.

In the intervention, the intervener discusses these film clips with the caregiver, helping the caregiver to consider the baby’s signals, their meaning, and how the baby responds to the caregiver’s behaviours, especially when these show sensitivity. In a strategy inherited from STEEP, as well as Fraiberg’s approach to therapeutic work,107 the intervener interprets the baby’s behaviours as if giving them subtitles, translating the infant signals into a verbal form that the caregiver may find easier to reflect upon and consider. Positive film fragments are emphasised; the intervention builds precisely on parents’ own expertise, with caregivers serving as a reinforcing role model for themselves.108 An illustration is offered by Juffer and Bakermans-Kranenburg from a detailed case study:

Using the recordings, she pointed out how happy and proud Ava reacted to Noah’s compliments and how they were peacefully playing together afterward. The intervener explained that research has shown that it works much better to praise your child and give compliments when she does things well than to punish her when things go wrong. Sometimes it is even better to ignore naughty or difficult behavior because that way your child is not receiving attention for that type of behavior. The intervener summarized this message by saying: ‘Because you are so important for Ava, she really loves to have your attention and compliments, and that is why it works so well to praise Ava when you want her to listen to you or be compliant.’109

Juffer, Bakermans-Kranenburg, and van IJzendoorn anticipated that, compared to STEEP, a shorter intervention focused on specific behaviours would prove more manageable for parents and more targeted to supporting change in attachment-relevant processes.110 VIPP is conventionally six to eight sessions, which makes it short enough even for use in assessing caregiving capacity, as well as directly for parenting support.111 The first session of VIPP focuses on exploration and attachment behaviour; the second session helps the caregiver consider the meaning of the infant’s signals; the third session addresses the role of prompt and adequate response to these signals; and the fourth session emphasises shared emotions and affective attunement with the child. Two additional sessions are often used to review (p. 566) the information and feedback from the first four.112 Issues around sensitive discipline of the child are also addressed, helping the caregiver provide non-coercive discipline and positive reinforcement, to de-escalate tantrums, and achieve empathy and consistent limit-setting for the child.

In a first evaluation of VIPP with 130 families, the researchers found that compared to the control group receiving treatment as usual, the intervention enhanced sensitivity as measured by Ainsworth’s scale. The dyads were also less likely to receive a disorganised attachment classification on the basis of the child’s behaviour in the Strange Situation procedure.113 In a later trial with 237 families with toddlers with externalising behaviour problems, the mothers who completed VIPP showed more sensitive discipline, and their children showed fewer conduct problems at a later follow-up.114 A process evaluation revealed that neither mothers nor home visitors were able to anticipate the relative effectiveness of the intervention based on their first impressions of one another.115 Whereas advocates for STEEP had argued that behavioural-focused targeted interventions would be less effective in families facing multiple adversities and daily hassles (Chapter 4), in fact VIPP was found to be more effective in such circumstances.116 A meta-analysis of the first twelve randomised control trials of VIPP found an effect size of d = .47 for sensitivity, d = .36 for infant–caregiver attachment classifications, and d = .26 for reduced child behaviour problems, with no decrease in effect size resulting from length of follow-up.117 Across multiple samples, VIPP has been found to decrease disorganised attachment as assessed in the Strange Situation. However, why this is (p. 567) the case remains an outstanding question. VIPP focuses on increasing caregiver sensitivity, and caregiver sensitivity has only a very weak association with disorganised attachment.

One proposal put forward by Out, Bakermans-Kranenburg, and van IJzendoorn, building on Main and Hesse’s thinking, is that the intervention directs the caregiver’s attention to the child, which reduces absorption and the intrusion of unresolved memories and affects.118 This proposal remains to be tested. However, some supportive evidence is available: an intervention focused on supporting caregiver attention to the child’s experience—Minding the Baby—reduced rates of disorganised attachment without impacting PTSD symptoms.119 Such findings suggest that the behavioural and attentional focus on the here-and-now of the child supported by VIPP may reduce disorganised attachment because it helps caregivers avoid absorption or mental states that may prompt the intrusion of segregated systems. An alternative explanation, implied by the researchers working on Minding the Baby, is that Bakermans-Kranenburg and colleagues have underestimated the capacity of parenting interventions to reduce disorganised attachment by increasing the caregiver’s reflective functioning about attachment relationships.120

Attachment Biobehavioral Catch-up

Another widely studied attachment-based parenting intervention is Attachment Biobehavioral Catch-up (ABC), developed by Dozier and colleagues. ABC is a ten-session, in-home intervention to help families meet the needs of infants and toddlers.121 Holding the intervention in the home is anticipated to help the integration of learning into daily life, and to allow others in the home to see and gain benefits from the intervention. Part of what is intriguing about ABC is its parsimony. The focus of the intervention is extremely strict: the comments made by the coaches only seek to help parents achieve three things: (i) to be nurturing when the child is distressed; (ii) to follow the child’s lead, where possible taking cues from the child rather than dictating activities to them; and (iii) to avoid displaying intrusive, harsh, or frightening behaviours.122 Dozier and colleagues take these to be respectively the central proposals of Bowlby, Ainsworth, and Main and Hesse. Bowlby emphasised the importance of a child’s confidence in their caregiver when distressed, which Dozier and colleagues regard as facilitated by nurturing behaviour and tone. Ainsworth emphasised heeding and responding to the child’s signals, which she described as sensitivity; she (p. 568) anticipated that this would contribute to a child’s capacity for self-regulation.123 And Main and Hesse argued that alarming caregiver behaviour disrupts the capacity of the child to coordinate attention and behaviour in making coherent use of the caregiver, either directly or conditionally, as a secure base and safe haven.

As the first attachment-based intervention, STEEP had a broad set of goals and a broad repertoire of intervention strategies, as the relationship between attachment theory and intervention science was just getting started (Chapter 4). By contrast, the focus of ABC on only three concerns represents a fierce, hard-earned, and specific theoretical security. One consequence is that training in delivering the intervention can be completed in two days. Another consequence is that the large majority of those interested to train as parent coaches can do so and gain certification following a fidelity test: potential coaches are screened out only if they find it difficult to make in-the-moment comments to parents, or if they show a dismissing state of mind regarding attachment on a cut-down version of the AAI. Following Main’s theory (Chapter 3), Dozier and colleagues assume that a dismissing state of mind is incompatible with cognitive openness and comfort with nurturance.124

The sessions of ABC have a few components. A component early in the intervention is psychoeducation: parents are informed about some key findings from attachment research, such as Ainsworth’s findings regarding the effects of prompt response to infant crying (Chapter 2). These findings are used as an opportunity for discussion of parents’ expectations of themselves and their child, for instance beliefs about the potential for care to spoil an infant or for harsh care to help an infant by promoting toughness. Another component of ABC is the use of films of caregivers responding to infants in desirable and undesirable ways, which likewise serve as a basis of discussion with the parent. Parents are also asked to complete tasks with their child, such as playing with blocks or making pudding together. These are filmed, and clips may be brought by the caregiving coach to review with the parent in a subsequent session. During the intervention sessions, caregiving coaches provide frequent and specific comments to parents ‘in the moment’, around 60 times per hour. Dozier and colleagues have found that the frequency of coaches’ comments, as well as their quality, could predict the magnitude of change in parent sensitivity.125

Comments address exclusively the three areas of focus of the intervention, drawing parents’ attention to specific behaviours, their purpose, and what ensues from them. This intensity of focused commenting is one of the most distinctive aspects of ABC—especially when compared to more psychoanalytically inspired forms of parenting intervention, such as GABI (Chapter 4), where clinicians are advised to address a wider range of topics and practice reticence in making comments. Caron, Bernard, and Dozier have documented that 27% of caregivers leave ABC before the end of the intervention: families who dropped out tended to receive fewer comments in sessions and more comments on issues away from the (p. 569) three areas of focus of the intervention.126 A related aspect of ABC is the meticulous use of video-based supervision to ensure the fidelity of coaches to the goals of the intervention, ensuring that coaches do not miss opportunities to offer relevant comments.127 In the early sessions of ABC, comments focus on identifying positives in the parent’s behaviour. As the sessions progress, the coach may identify areas in which the parent may improve on being nurturing, following the child’s lead or avoiding intrusive, harsh, or frightening behaviours. Though the focus of ABC is on behaviour, in the later sessions the coach also works with the parent to recognise aspects of their past experience that may be proving an obstacle to providing sensitive and non-frightening care. These may not be the aspects of past experience that are most salient for the parent in their day-to-day life; the focus is rather on cognitive and procedural obstacles to the effective functioning of the caregiving behavioural system. At the end of the tenth session, the coach presents the parent with a montage of film clips from earlier sessions, celebrating occasions when their behaviour was nurturing or when they followed the child’s lead.

Bernard, Dozier, and colleagues conducted a randomised control trial of ABC with parents identified as at risk for neglecting their young children. The children were assessed in the Strange Situation. Classifications of disorganised attachment were less frequent in the ABC group (32%) than in the control intervention (57%), and classifications of secure attachment were more frequent (52%) relative to the control (33%).128 Mothers in the ABC group showed greater increases in sensitivity and decreases in intrusiveness than participants from the control intervention.129 Mothers in the intervention group also showed higher secure base script knowledge (Chapter 2) than parents assigned to the control group.130 In a recent study, Dozier’s team have reported that, contrary to expectations, no improvements were seen in frightened, frightening, disoriented, or role-confused parenting behaviours.131 However, base rates of the behaviours were low, given that the observations were exceptionally brief, which may have contributed to false negatives. Nonetheless, the researchers found that parents in the ABC intervention group were less likely to display withdrawing behaviour towards their child’s attachment behaviours; this difference accounted for around 19% of the effect of ABC on the proportion of child–caregiver dyads receiving a disorganised attachment classification.

In a follow-up, children from the ABC intervention group showed less externalising behaviours than the control group in response to the frustrating problem-solving task developed for the Minnesota study (Chapter 4).132 Such intervention effects have been replicated (p. 570) at multiple sites, with diverse population groups.133 A large-scale implementation of ABC by the State of New York has recently reported successful impacts on caregiver sensitivity and reduced frightening behaviours.134 Further trials are underway in Australia, Germany, South Africa, and Russia. There is emergent endocrinal, physiological, and neurological evidence that converges with behavioural findings to suggest that the ABC group display greater skills at emotion regulation in the face of challenges, even several years after the intervention.135 Dozier and colleagues regard ABC as a targeted intervention for parenting, ideally to be delivered in the community alongside other targeted support for families depending on their needs, whether caregiver depression, substance use, housing, or other difficulties. It remains to be adequately explored, however, whether contextual factors moderate the success of the ABC intervention. One finding recently reported by Berlin was that the impact of ABC on caregiving behaviour is moderated by self-reported attachment style. The intervention was more successful for participants self-reporting a less avoidant attachment style, but actually reduced caregiver sensitivity among participants endorsing an avoidant attachment style. Berlin and colleagues suggest that ‘it may be that more avoidant mothers experienced the intervention model’s emphasis on nurturance, along with the parent coaches’ frequent comments, as dissonant or even aversive, which, in turn, led to iatrogenic effects’.136

There remains further work to be done to maximise the benefits of attachment-based interventions. Attachment-based interventions are seeing increasing take-up by child welfare and clinical services. However, there remains much we do not yet know about their cost-effectiveness and, perhaps more to the point, their capacity to survive in cash-strapped and continually restructuring services. Further research is also needed to understand how attachment-based interventions can best be delivered with different families and communities, with sensitivity to signals regarding their needs and to the ways of working they find acceptable and satisfying.137 Attachment-based interventions offer a fruitful possible point (p. 571) of connection between researchers and practitioners. They do so in the context of serious obstacles to mutually satisfying and informative communication. Nonetheless, they form one of several valuable bases for such a conversation, especially in the context of the turn to greater engagement and collaboration with practitioners among the third generation of attachment researchers. There is certainly a new wind blowing through attachment research at the moment, with the windows held wider than before to the field’s different audiences. Given the mutual benefit of mutual learning between research and practice, it is heartening to see attachment researchers taking steps to facilitate this two-way dialogue.

Concluding remarks

The need for attachment research to be open to different audiences and to the future was a central theme in the keynote address to the International Attachment Conference in 2019 by Carlo Schuengel. In reflecting on the state of the field, Schuengel described a number of questions that colleagues in attachment research have been raising with one another. These include pinning down further: the origin of individual differences in child attachment, the meaning of attachment disorganisation, the attachment-specific elements of interventions and how they work, and the cognitive or symbolic aspects of attachment.138 For instance, complacent and vague appeals to the internal working model concept need to be superseded by a more detailed account of the cognitive or symbolic components of attachment, their parameters and features, their interrelations, and links with perceptions and actuators.139 Schuengel also urged further scrutiny of the developmental mechanisms studied by attachment research, the psychometric properties of its measures, the scalability of its interventions, and its broader relevance to child welfare and clinical practice.

Thompson, Simpson, and Berlin have likewise taken the opportunity of 50 years since the publication of Attachment, Volume 1 to consider the fundamental outstanding questions for attachment research.140 To Schuengel’s list they have added several further questions. Among these, they feel that more work is needed to understand what kinds of relationships qualify as attachment relationships. They urge concern with how attachment processes are manifested in different cultures, and how culture manifests itself in attachment processes. They feel attachment research needs a better conceptual model of what domains of later behaviour early security of attachment should predict and what domains it would not be anticipated to predict. And they emphasise that there is much more to be done to understand how attachment theory and research might best inform services for children and families, including divorce and custody proceedings, home visiting, education, and foster care.

(p. 572) If such questions as those posed by Schuengel and by Thompson and colleagues form some of the present horizon of empirical attachment research, they do so knowingly on the basis, in part, of the work of research groups considered in this book. The critical historical consideration of these groups, with their different perspectives and ambitions, has aimed to put more firmly at the disposal of the present a portion of the resources of the past in their liveliness and diversity, and in their potential relevance for attachment research during and after this period of transition. If attachment research has faced pressures that contort and abbreviate its ideas in their circulation, it is hoped that a historical perspective has the potential to exert some contrary force, whilst at the same time offering observations relevant to the wider history of science. In this regard, Cornerstones has sought to add to existing historical studies, which to date have not considered developments after Ainsworth. Those interested to understand, use or change attachment research must walk with and negotiate with the past. In addition to Bowlby and Ainsworth, this entails the important legacy from a second generation of attachment researchers.


1 Ziv and Hotam have stated that ‘the “attachment language” is undeniably unique, coherent, rich, and complex enough to interest professional academic audiences, to be clinically meaningful to practitioners, and to be straightforward enough to attract laypersons’. Despite their reference to coherence, Ziv and Hotam would likely agree that these audiences do not encounter a unitary discourse, but at least in part encounter different discourses that have overlapping vocabularies. Terms such as ‘attachment’ and ‘separation’ (Bowlby) or ‘security’ and ‘sensitivity’ (Ainsworth) may be used by academics, clinicians, and laypersons alike—but with little overlapping meaning. Ziv, Y. & Hotam, Y. (2015) Theory and measure in the psychological field: the case of attachment theory and the Strange Situation procedure. Theory & Psychology, 25(3), 274–91, p.279.

2 On field differentiation see Bourdieu, P. (1989) The conquest of autonomy: the critical phase in the emergence of the field. In The Rules of Art, trans. S. Emanuel (pp.47–112). Cambridge: Polity Press.

3 On the concept of non-formativeness see Wood, M. (2009) The nonformative elements of religious life: questioning the ‘sociology of spirituality’ paradigm. Social Compass, 56(2), 237–48.

4 E.g. Lyons-Ruth, K. & Spielman, E. (2004) Disorganized infant attachment strategies and helpless-fearful profiles of parenting: integrating attachment research with clinical intervention. Infant Mental Health Journal, 25(4), 318–35. Crittenden, P.M. (2015) Raising Parents, 2nd edn. London: Routledge; Fonagy, P., Luyten, P., Allison, E., & Campbell, C. (2017) What we have changed our minds about: Part 2. Borderline personality disorder, epistemic trust and the developmental significance of social communication. Borderline Personality Disorder and Emotion Dysregulation, 4, 9. Van IJzendoorn and Bakermans-Kranenburg are also partial exceptions, albeit primarily in the past decade, for instance co-authoring a popular book aimed at child welfare professionals: van IJzendoorn, M.H. & Bakermans-Kranenburg, M.J. (2010) Gehechtheid en Trauma. Diagnostiek en Behandeling voor de Professional. Amsterdam: Hogrefe.

5 E.g. Kanieski, M.A. (2010) Securing attachment: the shifting medicalisation of attachment and attachment disorders. Health, Risk & Society, 12(4), 335–44; Garrett, M.P. (2017) Wired: early intervention and the ‘neuromolecular gaze’. British Journal of Social Work, 48(3), 656–74.

6 The sociological literature discussing attachment research has tended towards the assumption that the psy-disciplines always tend to work in the same way. However, this prefabricated ‘critical’ account of psychological discourse misses important heterogeneity. Foucault urged that it is important to register and to study not only effective medical and professional power/knowledge relations, but also ‘a domination that grows feeble, poisons itself, grows slack’, and not to mistake one for the other. Foucault, M. (1971, 1994) Nietzsche, genealogy, history. In J.D. Faubion (ed.) Aesthetics, Method, and Epistemology, Essential Works of Foucault 1954–1984. New York: The New Press, p.381. In turn, it may be gently acknowledged that the prefabricated discourse of ‘critical’ social science is not itself free from imbrication with forms of domination, for instance as a capital-accrual strategy within academia.

7 The only sustained consideration appears in Main, M., Hesse, E., & Hesse, S. (2011) Attachment theory and research: overview with suggested applications to child custody. Family Court Review, 49(3), 426–63. Here Main and Hesse firmly situated themselves as non-experts on professional practice, offering reflections as best they could in response to a direct request rather than because custody decisions were central to their work as researchers.

8 What remarks are offered seem to generally stem from work led by Shaver’s wife Gail Goodman, rather than reflecting the priorities of Shaver and Mikulincer’s research agenda, e.g. 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.

9 The central early exception here was NICHD Early Child Care Research Network (1997) The effects of infant child care on infant–mother attachment security: results of the NICHD Study of Early Child Care NICHD Early Child Care Research Network. Child Development, 68(5), 860–79. A larger exception was Jaddoe, V.W., van Duijn, C.M., Franco, O.H., et al. (2012) The Generation R Study: design and cohort update 2012. European Journal of Epidemiology, 27(9), 739–56.

10 Funder, D.C. & Ozer, D.J. (2019) Evaluating effect size in psychological research: sense and nonsense. Advances in Methods and Practices in Psychological Science, 2(2) 156–68.

11 Groh, A.M., Fearon, R.P., van IJzendoorn, M.H., Bakermans-Kranenburg, M.J., & Roisman, G.I. (2017) Attachment in the early life course: meta-analytic evidence for its role in socioemotional development. Child Development Perspectives, 11(1), 70–76.

12 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, p.81.

13 Latour, B. (2005) Reassembling the Social. Oxford: Oxford University Press, p.35.

14 E.g. Bolen, R. (2000) Validity of attachment theory. Trauma, Violence & Abuse, 1, 128–53; Bosmans, G. (2016) Cognitive behaviour therapy for children and adolescents: can attachment theory contribute to its efficacy? Clinical Child and Family Psychology Review, 19(4), 310–28.

15 One set of definitions has been offered by Schuengel, C., de Schipper, J.C., Sterkenburg, P.S., & Kef, S. (2013) Attachment, intellectual disabilities and mental health: research, assessment and intervention. Journal of Applied Research in Intellectual Disabilities, 26(1), 34–46.

16 Rutter, M. & O’Connor, T. (1999) Implications of attachment theory for child care policies. In J. Cassidy & P.R. Shaver (eds) Handbook of Attachment: Theory, Research and Clinical Applications (pp.823–44). New York: Guilford Press, p.823.

17 Granqvist, P. (2016) Observations of disorganized behaviour yield no magic wand: response to Shemmings. Attachment & Human Development, 18(6), 529–33; Granqvist, P., Sroufe, L.A., Dozier, M., et al. (2017) Disorganized attachment in infancy: a review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 19(6), 534–58.

18 Reijman, S., Foster, S., & Duschinsky, R. (2018) The infant disorganised attachment classification: ‘patterning within the disturbance of coherence’. Social Science & Medicine, 200, 52–8.

19 White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press.

20 E.g. Association for Psychiatric Social Workers (1955) Presentation at the Annual General Meeting 1955: Dr John Bowlby on preventative activities. Modern Records Centre Warwick University, MSS.378/APSW/P/ 16/6/19-20.

21 E.g. Department of Health (2000) Assessing Children in Need and their Families: Practice Guidance. London: TSO.

22 White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press.

23 Bilson, A. & Munro, E.H. (2019) Adoption and child protection trends for children aged under five in England: increasing investigations and hidden separation of children from their parents. Children and Youth Services Review, 96, 204–11.

24 White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press.

25 Potter, A. (2019) Judging social work expertise in care proceedings. In D.S. Caudill, S.N. Conley, M.E. Gorman, & M. Weinel (eds) The Third Wave in Science and Technology Studies (pp.71–85). London: Palgrave. Further findings from Potter’s doctoral research are forthcoming. See also Shemmings, D. (2018) Why social workers shouldn’t use ‘attachment’ in their records and reports. Community Care, 28 June 2018.

26 North, G. (2019) Assessing for bruises on the soul: identifying and evidencing childhood emotional abuse. Journal of Social Welfare and Family Law, 41(3), 302–320, p.313.

28 Similar observations have been made by Wilkins, D., Shemmings, D., & Shemmings, Y. (2015) A–Z of Attachment. London: Palgrave: ‘We have also seen what can only be described as quite crude and, in all likelihood, mistaken applications of attachment theory in practice. For example, we have read reports by contact supervisors observing that a child stays very physically close to their attachment figure throughout the session, with this then interpreted as a sign of a “positive attachment”; we see conclusions being drawn about a child’s attachment relationships based on one, short observation; descriptions of parents being “attached” to their babies; of a child’s attachment relationship with their father being almost completely overlooked and of social workers describing young children as always happy and content without reflecting on whether this is a “good thing” or not. Additionally, we often hear workers speak of “strong attachments” … We have become increasingly worried that attachment theory and research are often used “against” families: to highlight “problems” and “gaps” in parenting or caregiving relationships—particularly when writing reports for the courts or in relation to child protection procedures’ (xv–xvi).

29 White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press.

31 Vincent, S. (2017) ‘The Magic Is in the Co-Production’: Summary Report from the Evaluation of the Love Barrow Families Project. Newcastle: Northumbria University.

32 Crittenden, P.M. (2015) Raising Parents, 2nd edn. London: Routledge. See also Baim, C. & Morrison, T. (2014) Attachment-Based Practice with Adults. Brighton: Pavilion: ‘There is a general truism about attachment theory, which is that the more you learn about attachment theory the more cautious you become about attaching labels to people. This includes labels such as “securely attached”, “reactive attachment disorder”, “disorganised attachment” and so on, which tend to be over-used and under-defined’ (23).

33 See also Baim, C. (2019) DMM vs ABC+D—a controversial discussion. DMM News, 32. ‘The DMM community attempts to find a way of assessing and informing treatment that is more accurate and useful because it is focused on the function of behaviour rather than labelling symptoms—which puts people into boxes.’

34 Spieker, S.J. & Crittenden, P.M. (2018) Can attachment inform decision-making in child protection and forensic settings? Infant Mental Health Journal, 39(6), 625–41; Van IJzendoorn, M.H., Bakermans, J.J., Steele, M., & Granqvist, P. (2018) Diagnostic use of Crittenden’s attachment measures in Family Court is not beyond a reasonable doubt. Infant Mental Health Journal, 39(6), 642–6.

35 Madigan, S. (2019) Beyond the academic silo: collaboration and community partnerships in attachment research. Paper presented at International Attachment Conference, Vancouver, 20 July 2019; Haltigan, J.D., Madigan, S., Bronfman, E., et al. (2019) Refining the assessment of disrupted maternal communication: using item response models to identify central indicators of disrupted behavior. Development & Psychopathology, 31(1), 261–77.

36 This transition may be placed in the broader context of the rise of discourses, since the late 1990s, questioning whether or how academic research meaningfully contributes to addressing the challenges faced by practitioners and/or the public good. Irwin, A. (2001) Constructing the scientific citizen: science and democracy in the biosciences. Public Understanding of Science, 10(1), 1–18; Gunn, A. & Mintrom, M. (2016) Higher education policy change in Europe: academic research funding and the impact agenda. European Education, 48(4), 241–57.

37 Powell, B., Cooper, G., Hoffman, K., & Marvin, B. (2016) The Circle of Security Intervention. New York: Guilford; McMahon, C., Huber, A., Kohlhoff, J., & Camberis, A.L. (2017) Does training in the Circle of Security framework increase relational understanding in infant/child and family workers? Infant Mental Health Journal, 38(5), 658–68.

38 Prior, V. & Glaser, D. (2006) Understanding Attachment and Attachment Disorders: Theory, Evidence and Practice. London: Jessica Kingsley Press; Slade, A. & Holmes, J. (2017) Attachment in Therapeutic Practice. London: SAGE. The best available in Dutch is likely van IJzendoorn, M.H. & Bakermans-Kranenburg, M.J. (2010) Gehechtheid en Trauma: Diagnostiek en Behandeling voor de Professional. Amsterdam: Hogrefe. Another thoughtful work is Page, T. (2017) Attachment theory and social work treatment. In F. Turner (ed.) Social Work Treatment: Interlocking Theoretical Approaches, 6th edn (pp.1–22). Oxford: Oxford University Press. Beyond the published literature, some trainings offered to clinicians and foster carers are also exemplary in effective characterisation of the technical aspects of attachment theory and research, and its relevance. Those by John Sands and Lydia Fransham can be mentioned, though as yet neither has published their training materials.

39 Blood, I. & Guthrie, L. (2018) Supporting Older People Using Attachment-Informed and Strengths-Based Approaches. London: Jessica Kingsley Press.

40 Golding, K. (2017) Nurturing Attachments Training Resource: Running Parenting Groups for Adoptive Parents and Foster or Kinship Carers. London: Jessica Kingsley Press.

41 One example is Pearce’s A Short Introduction to Attachment and Attachment Disorder, which is admirably accessible but contains many outright errors alongside confused oversimplifications. Pearce, J. (2009) A Short Introduction to Attachment and Attachment Disorder. London: Jessica Kingsley Press. Even in otherwise good works there are major errors, such as the conflation of controlling-punitive/controlling-caregiving with Reactive Attachment Disorder in Howe, D., Brandon, M., Hinings, D., & Schofield, G. (1999) Attachment Theory, Child Maltreatment and Family Support: A Practice and Assessment Model. London: Palgrave, pp.135–6. A decade later, these problems were corrected in Howe, D. (2011) Attachment Across the Lifecourse: A Brief Introduction. London: Palgrave.

42 E.g. Perry, B. (2001) Attachment: the first core strength. Early Childhood Today, 16(2): ‘Attachment is the capacity to form and maintain healthy emotional relationships’ (28).

43 E.g. Marshall, N. (2014) The Teacher’s Introduction to Attachment: Practical Essentials for Teachers, Carers and School Support Staff. London: Jessica Kingsley Publishers.

44 See also Alexander, S.L., Frederico, M., & Long, M. (2018) Attachment and children with disabilities: knowledge and views of early intervention professionals. Children Australia, 43(4), 245–54; Morison, A., Taylor, E., & Gervais, M. (2019) How a sample of residential childcare staff conceptualize and use attachment theory in practice. Child & Youth Services, 1(25).

45 See also Oppenheim, D. & Goldsmith, D.F. (2007) Attachment theory in clinical work with children: bridging the gap between research and practice. Journal of Canadian Academic Child and Adolescent Psychiatry, 16(4), 186–7.

46 See also Wastell, D. & White, S. (2017) Blinded by Science: The Social Implications of Epigenetics and Neuroscience. Cambridge: Policy Press.

47 White and colleagues do praise the work of Fonagy and colleagues for advocating a shift ‘away from instinct and into interaction, and the social and psychological circumstances in which the mother and infant may find themselves’. White, S., Gibson, M., Wastell, D., & Walsh, P. (2019) Reassessing Attachment Theory in Child Welfare. Bristol: Psychology Press. However, the characterisation of attachment research as focused on instincts is half a century out of date (Chapter 1). And Fonagy gives no more attention to interaction and social and psychological circumstances than many attachment researchers.

48 Van IJzendoorn, M.H. & Tavecchio, L.W.C. (1987) The development of attachment theory as a Lakatosian research program. In L.W.C. Tavecchio & van IJzendoorn, M.H. (eds) Attachment in Social Networks: Contributions to the Bowlby–Ainsworth Attachment Theory (pp.3–31). New York: Elsevier Science, Table 1.

49 Ibid. p.16.

50 E.g. Bailey, H.N., Tarabulsy, G.M., Moran, G., Pederson, D.R., & Bento, S. (2017) New insight on intergenerational attachment from a relationship-based analysis. Development & Psychopathology, 29(2), 433–48.

51 Harkness, S. (2015) The strange situation of attachment research: a review of three books. Reviews in Anthropology, 44(3), 178–97, p.179.

52 White, K. & Schwartz, J. (2007) Attachment here and now: an interview with Peter Fonagy. Attachment, 1(1), 57–61, p.57. Thompson likewise has marked the standing of attachment research in this period, commenting in 2000 that attachment has been ‘the dominant approach to understanding early socioemotional and personality development during the past quarter-century of research’. Thompson, R.A. (2000) The legacy of early attachments. Child Development, 71(1), 145–52, p.145.

53 National Institute of Mental Health (2016) Behavioral assessment methods for RDoC constructs.

54 Lyons-Ruth, K., Belsky, J., Booth-LaForce, C., et al. (2016) Letter to Joshua Gordon—Director of NIMH and Sarah Morris—Acting Director of the RDoC Unit, responding to the behavioral assessment methods for RDoC constructs. Unpublished letter shared by Karlen Lyons-Ruth.

55 The work of Cicchetti, Rogosch, Toth, and colleagues (Chapter 4) provides one illustration. The work of van IJzendoorn, Bakermans-Kranenburg, and colleagues offers another. See Kok, R., Thijssen, S., Bakermans-Kranenburg, M.J., et al. (2015) Normal variation in early parental sensitivity predicts child structural brain development. Journal of the American Academy of Child & Adolescent Psychiatry, 54(10), 824–31; Bakermans-Kranenburg, M.J. & van IJzendoorn, M.H. (2015) The hidden efficacy of interventions: gene × environment experiments from a differential susceptibility perspective. Annual Review of Psychology, 66, 381–409.

56 Fonagy, P. & Campbell, C. (2015) Bad blood revisited: attachment and psychoanalysis, 2015. British Journal of Psychotherapy, 31(2), 229–50. This paper represents thinking with a wider group of colleagues also reported in Fonagy, P., Luyten, P., Allison, E., & Campbell, C. (2017) What we have changed our minds about: Part 2. Borderline personality disorder, epistemic trust and the developmental significance of social communication. Borderline Personality Disorder and Emotion Dysregulation, 4(1), 9. In the USA, there have likewise been claims that the field of attachment research is moving from saturation to exhaustion. Waters and colleagues have reported a growing perception of attachment research among some American colleagues that ‘There is some great work there. But it is a mature field now and I think all the big studies have been done’. Waters, E., Petters, D., & Facompre, D. (2013) Epilogue: reflections on a special issue of attachment & human development in Mary Ainsworth’s 100th year. Attachment & Human Development, 15(5–6), 673–81.

57 See e.g. Bauman, Z. (2000) Liquid Modernity. Cambridge: Polity Press; Weeks, J. (2015) Beyond the categories. Archives of Sexual Behavior, 44(5), 1091–7.

58 See also Fonagy, P. & Target, M. (2007) The rooting of the mind in the body: new links between attachment theory and psychoanalytic thought. Journal of the American Psychoanalytic Association, 55(2), 411–56: ‘Advances in the sciences to which Bowlby’s ideas are coupled dictate a reconsideration’ (420).

59 Fonagy, P. (2016) The role of attachment, epistemic trust and resilience in personality disorder: a trans-theoretical reformulation. DMM News, 26.

60 An exception is Simpson, J.A. & Belsky, J. (2016) Attachment theory within a modern evolutionary framework. In J. Cassidy & P.R. Shaver (2018) Handbook of Attachment: Theory, Research, and Clinical Applications, 3rd edn (pp.91–116). New York: Guilford. However, Simpson and Belsky themselves readily acknowledge that attachment research has been slow to absorb developments in modern evolutionary theory. This issue is also discussed in Granqvist, P. (2020) Attachment, culture, and gene-culture co-evolution: expanding the evolutionary toolbox. Attachment & Human Development, 2 January, 1–24.

61 Fonagy, P. & Campbell, C. (2015) Bad blood revisited: attachment and psychoanalysis, 2015. British Journal of Psychotherapy, 31(2), 229–50, p.230.

62 See also Fonagy, P. (1999) Points of contact and divergence between psychoanalytic and attachment theories: is psychoanalytic theory truly different. Psychoanalytic Inquiry, 19(4), 448–80: ‘Attachment theory, far closer to empirical psychology with its positivist heritage, has been in some ways method-bound over the past 15 years. Its scope was determined less by what fell within the domain defined by relationship phenomena involving a caretaking-dependent dyad and more by the range of groups and behaviors to which the preferred mode of observation, the strange situation, the adult attachment interview, and so forth, could be productively applied. This sheltered the theory from a range of ideas’ (472).

63 Fonagy, P., Gergely, G., & Target, M. (2007) The parent–infant dyad and the construction of the subjective self. Journal of Child Psychology and Psychiatry, 48(3–4), 288–328, p.288.

64 Granqvist, P. (2020) Attachment, Religion, and Spirituality: A Wider View. New York: Guilford. Granqvist admits that, compared to the past, ‘I am less optimistic about the prospect of attachment theory … This is because of attachment theory’s conceptual boundaries, its rudimentary defense mechanisms, and the attachment research habit of “cross-tabulating” people into types (secure versus insecure and organized versus disorganized), despite no individual being reducible to a type. Although these features have indisputably contributed to attachment theory’s prosperity as an empirical research program, the attachment framework remains somewhat schematic and impoverished.’ Unlike Fonagy and colleagues, Granqvist is firmly of the view, however, that these matters can be resolved through innovations within attachment theory and research, and especially from renewed cross-fertilisation with developments in evolutionary theory.

65 On the concept of infrastructures see Star, S.L. (1999) The ethnography of infrastructure. American Behavioral Scientist, 43, 377–91; Berlant, L. (2016) The commons: infrastructures for troubling times. Environment and Planning D: Society and Space, 34(3), 393–419.

66 E.g. Haltigan, J.D. & Roisman, G.I. (2015) Infant attachment insecurity and dissociative symptomatology: findings from the NICHD Study of Early Child Care and Youth Development. Infant Mental Health Journal, 36(1), 30–41.

67 Van IJzendoorn is perhaps best regarded as generation 2.5. He learnt the Strange Situation from Brian Vaughn and Mary Main and first met Ainsworth only at an AAI training with Main in the late 1980s. As well as his tendency to be an early adopter of new technologies and quantitative methodologies, some ‘third-generation’ characteristics to van IJzendoorn’s work may stem from the closeness of his collaboration with his former students Marian Bakermans-Kranenburg and Femmie Juffer. He has also mentored and in turn learnt from collaborations with Madigan and Fearon.

68 See e.g. Raby, L., Verhage, M.L., Fearon, R.P.M., et al. (2019) The latent structure of the Adult Attachment Interview: large sample evidence from the Collaboration on Attachment Transmission Synthesis. Paper presented at Biennial Meeting of the Society for Research in Child Development, Baltimore, 21–23 March. On the symbolic work demanded in the constitution of a generation see Eyerman, R. & Turner, B.S. (1998) Outline of a theory of generations. European Journal of Social Theory, 1, 91–106; Purhonen, S. (2016) Generations on paper: Bourdieu and the critique of ‘generationalism’. Social Science Information, 55(1), 94–114.

69 Schuengel, C., Verhage, M., & Duschinsky, R. (2019) Representing attachment through meta-analyses: a move to the level of collaboration, manuscript under review.

70 Cassidy, J. & Marvin, R., with the MacArthur Network on Attachment in the Preschool Years (1992) Attachment organisation in pre-school children: procedures and coding manual. Unpublished manuscript.

71 Crittenden, P.M. (1981) The Pre-School Assessment of Attachment Coding Manual. Miami: Family Relations Institute.

72 Van IJzendoorn, M.H., Goldberg, S., Kroonenberg, P.M., & Frenkel, O.J. (1992) The relative effects of maternal and child problems on the quality of attachment: a meta-analysis of attachment in clinical samples. Child Development, 63(4), 840–58. See also van IJzendoorn, M.H., Bakermans-Kranenburg, M.J., & Alink, L.R.A. (2011) Meta-analysis in developmental science. In B. Laursen, T.D. Little, & N.A. Card (eds) Handbook of Developmental Research Methods (pp.667–86). New York: Guilford.

73 Lamb, M.E., Sternberg, K.J., & Prodromidis, M. (1992) Nonmaternal care and the security of infant mother attachment: a reanalysis of the data. Infant Behavior & Development, 15(1), 71–83.

74 Another factor may have been that attitudes among researchers were not yet favourable to data sharing—a situation that has, by degrees, shifted in recent years. Tenopir, C., Dalton, E.D., Allard, S., et al. (2015) Changes in data sharing and data reuse practices and perceptions among scientists worldwide. PLoS One, 10(8), e0134826.

75 Verhage, M., Schuengel, C., Madigan, S., et al. (2016) Narrowing the transmission gap. Psychological Bulletin, 142(4), 337–66.

76 Van IJzendoorn, M.H. (1995) Adult attachment representations, parental responsiveness, and infant attachment—a meta-analysis on the predictive validity of the Adult Attachment Interview. Psychological Bulletin, 117(3), 387–403.

77 Crittenden, P.M., Partridge, M.F., & Claussen, A.H. (1991) Family patterns of relationship in normative and dysfunctional families. Development & Psychopathology, 3(4), 491–512; Shah, P.E., Fonagy, P., & Strathearn, L. (2010) Is attachment transmitted across generations? The plot thickens. Clinical Child Psychology and Psychiatry, 15(3), 329–45; Kondo-Ikemura, K., Behrens, K., Umemura, T., & Nakano, S. (2018) Japanese mothers’ prebirth Adult Attachment Interview predicts their infants’ response to the Strange Situation procedure. Developmental Psychology, 54(11), 2007–2015.

78 See also Bernier, A., Matte-Gagné, C., Bélanger, M.-È., & Whipple, N. (2014) Taking stock of two decades of attachment transmission gap: broadening the assessment of maternal behavior. Child Development, 85(5), 1852–65; van IJzendoorn, M.H. & Bakermans-Kranenburg, M.J. (2019) Bridges across the intergenerational transmission of attachment gap. Current Opinion in Psychology, 25, 31–6.

79 Verhage, M.L., Fearon, R.P., Schuengel, C., et al. (2018) Examining ecological constraints on the intergenerational transmission of attachment via Individual Participant Data meta-analysis. Child Development, 89(6), 2023–37.

80 Verhage, M.L., Fearon, R.P., Schuengel, C., et al. (2019) Does risk background affect intergenerational transmission of attachment? Testing a moderated mediation model with IPD. Paper presented at Biennial Meeting of the Society for Research in Child Development, Baltimore.

81 Raby, L., Verhage, M.L., Fearon, R.P.M., et al. (2019) The latent structure of the Adult Attachment Interview: large sample evidence from the Collaboration on Attachment Transmission Synthesis. Paper presented at Biennial Meeting of the Society for Research in Child Development, Baltimore, 21–23 March.

82 Schuengel, C., Verhage, M., & Duschinsky, R. (2019) Representing attachment through meta-analyses: a move to the level of collaboration. Manuscript under review.

83 Fonagy, P., Steele, H., & Steele, M. (1991) Maternal representations of attachment during pregnancy predict the organization of infant–mother attachment at one year of age. Child Development, 62(5), 891–905, p.902.

84 E.g. Thompson, R.A. & Raikes, H.A. (2003) Toward the next quarter-century: conceptual and methodological challenges for attachment theory. Development & Psychopathology, 15, 691–718.

85 Attention to state attachment variation in childhood has been a concern of Guy Bosmans, whose work has increasingly spanned the divide between developmental and social traditions of attachment research, e.g. Bosmans, G., Van de Walle, M., Goossens, L., & Ceulemans, E. (2014) (In)variability of attachment in middle childhood: secure base script evidence in diary data. Behavior Change, 31, 225–42; Vandevivere, E., Bosmans, G., Roels, S., Dujardin, A., & Braet, C. (2018) State trust in middle childhood: an experimental manipulation of maternal support. Journal of Child and Family Studies, 27(4), 1252–63.

86 E.g. Baldwin, M.W., Carrell, S.E., & Lopez, D.F. (1990) Priming relationship schemas: my advisor and the Pope are watching me from the back of my mind. Journal of Experimental Social Psychology, 26(5), 435–54. Attachment styles have been conceptualised by researchers in the social psychological tradition as having trait-like and state-like qualities, e.g. Fraley, R.C. & Roberts, B.W. (2005) Patterns of continuity: a dynamic model for conceptualizing the stability of individual differences in psychological constructs across the life course. Psychological Review, 112(1), 60–74.

87 Baldwin, M.W., Keelan, J.P.R., Fehr, B., Enns, V., & Koh-Rangarajoo, E. (1996) Social-cognitive conceptualization of attachment working models: availability and accessibility effects. Journal of Personality and Social Psychology, 71(1), 94–109.

88 Mikulincer, M. & Arad, D. (1999) Attachment working models and cognitive openness in close relationships: a test of chronic and temporary accessibility effects. Journal of Personality and Social Psychology, 77, 710–25.

89 In a recent study, Hudson and Fraley reported intriguing findings that repeatedly priming attachment anxiety over time contributed to reduced anxiety on the ECR over time, no less than repeated priming attachment security. The researchers concluded that repeated priming of attachment anxiety or attachment security offered participants the opportunity to reflect on their overall network of associations, contributing to change in the global attachment style. Hudson, N.W. & Fraley, R.C. (2018) Moving toward greater security: the effects of repeatedly priming attachment security and anxiety. Journal of Research in Personality, 74, 147–57.

90 Mikulincer, M., Gillath, O., Halevy, V., Avihou, N., Avidan, S., & Eshkoli, N. (2001) Attachment theory and reactions to others’ needs: evidence that activation of the sense of attachment security promotes empathic responses. Journal of Personality and Social Psychology, 81(6), 1205–224; 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.

91 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; Gillath, O. & Karantzas, G. (2019) Attachment security priming: a systematic review. Current Opinion in Psychology, 25, 86–95.

92 Mikulincer, M. & Shaver, P.R. (2001) Attachment theory and intergroup bias: evidence that priming the secure base schema attenuates negative reactions to out-groups. Journal of Personality and Social Psychology, 81(1), 97–115.

93 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.

94 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 (pp.92–119). Cambridge: Cambridge University Press.

95 Mikulincer, M., Hirschberger, G., Nachmias, O., & Gillath, O. (2001) The affective component of the secure base schema. Journal of Personality and Social Psychology, 81(2), 305–321, Study 7.

96 Mikulincer, M., Shaver, P.R., Sahdra, B.K., & Bar-On, N. (2013) Can security-enhancing interventions overcome psychological barriers to responsiveness in couple relationships? Attachment & Human Development, 15(3), 246–60.

97 Ibid. p.257.

98 Mikulincer, M., Shaver, P.R., Bar-On, N., & Sahdra, B.K. (2014) Security enhancement, self-esteem threat, and mental depletion affect provision of a safe haven and secure base to a romantic partner. Journal of Social and Personal Relationships, 31(5), 630–50.

99 Taubman-Ben-Ari, O. & Mikulincer, M. (2007) The effects of dispositional attachment orientations and contextual priming of attachment security on reckless driving. Transportation Research Part F, 10, 123–38.

100 Mikulincer, M., Shaver, P.R., & Rom, E. (2011) The effects of implicit and explicit security priming on creative problem solving. Cognition and Emotion, 25(3), 519–31.

101 E.g. Verhees, M.W., Ceulemans, E., Bakermans-Kranenburg, M.J., van IJzendoorn, M.H., De Winter, S., & Bosmans, G. (2017) The effects of cognitive bias modification training and oxytocin administration on trust in maternal support: study protocol for a randomized controlled trial. Trials, 18(1), 326. Granqvist has made extensive use of priming in his research, though in studies using measures from the social psychological tradition, and not in his research using measures from the developmental tradition.

102 E.g. Haltigan, J.D., Madigan, S., Bronfman, E., et al. (2019) Refining the assessment of disrupted maternal communication: using item response models to identify central indicators of disrupted behavior. Development & Psychopathology, 31(1), 261–77; Cadman, T., Belsky, J., & Fearon, R.M.P. (2018) The Brief Attachment Scale (BAS-16): a short measure of infant attachment. Child: Care, Health and Development, 44(5), 766–75.

103 Locke, E.A. (2015) Theory building, replication, and behavioral priming: where do we need to go from here? Perspectives on Psychological Science, 10(3), 408–414.

104 Steele, H. & Steele, M. (eds) (2018) Handbook of Attachment-Based Interventions. New York: Guilford.

105 Mohamed, A.R, Sterkenburg, P., van Rensburg, E., & Schuengel, C. (2019) The development of a coding system to examine the effective elements of attachment-based interventions. Paper presented at International Attachment Conference, Vancouver, 19 July 2019.

106 Lambermon, M.W. & van IJzendoorn, M.H. (1989) Influencing mother–infant interaction through videotaped or written instruction: evaluation of a parent education program. Early Childhood Research Quarterly, 4(4), 449–58.

107 Carter, S.L., Osofsky, J.D., & Hann, D.M. (1991) Speaking for the baby: a therapeutic intervention with adolescent mothers and their infants. Infant Mental Health Journal, 12(4), 291–301.

108 Juffer, F. & Steele, M. (2014) What words cannot say: the telling story of video in attachment-based interventions. Attachment & Human Development, 16(4), 307–314, p.311.

109 Juffer, F. & Bakermans-Kranenburg, M.J. (2018) Working with Video-feedback Intervention to promote Positive Parenting and Sensitive Discipline (VIPP-SD): a case study. Journal of Clinical Psychology, 74(8), 1346–57, p.7.

110 Van IJzendoorn, M.H., Juffer, F., & Duyvesteyn, M.G. (1995) Breaking the intergenerational cycle of insecure attachment: a review of the effects of attachment-based interventions on maternal sensitivity and infant security. Journal of Child Psychology and Psychiatry, 36(2), 225–48; Bakermans-Kranenburg, M.J., van IJzendoorn, M.H., & Juffer, F. (2003) Less is more: meta-analyses of sensitivity and attachment interventions in early childhood. Psychological Bulletin, 129(2), 195–215.

111 Bakermans-Kranenburg, M.J., Juffer, F., & van IJzendoorn, M.H. (2019) Reflections on the mirror: on video-feedback to promote positive parenting and infant mental health. In C. Zeanah (ed.) Handbook of Infant Mental Health, 4th edn (pp.527–42). New York: Guilford.

112 Juffer, F., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (eds) (2008) Promoting Positive Parenting: An Attachment-Based Intervention. New York: Psychology Press.

113 Juffer, F., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2005) The importance of parenting in the development of disorganized attachment: evidence from a preventive intervention study in adoptive families. Journal of Child Psychology and Psychiatry, 46(3), 263–74.

114 Klein Velderman, M., Bakermans-Kranenburg, M.J., Juffer, F., Van Ijzendoorn, M.H., Mangelsdorf, S.C., & Zevalkink, J. (2006) Preventing preschool externalizing behavior problems through video-feedback intervention in infancy. Infant Mental Health Journal, 27(5), 466–93. However, both maternal discipline practices and child behavioural problems seem to have been independently affected by the intervention. Analysis did not suggest effects of maternal discipline on child behaviour. See Mesman, J., Stoel, R., Bakermans-Kranenburg, M.J., et al. (2009) Predicting growth curves of early childhood externalizing problems: differential susceptibility of children with difficult temperament. Journal of Abnormal Child Psychology, 37(5), 625–36, p.633.

115 Stolk, M.N., Mesman, J., van Zeijl, J., et al. (2008) Early parenting intervention aimed at maternal sensitivity and discipline: a process evaluation. Journal of Community Psychology, 36(6), 780–97.

116 Van Zeijl, J., Mesman, J., Van IJzendoorn, M.H., et al. (2006) Attachment-based intervention for enhancing sensitive discipline in mothers of 1 to 3 year-old children at risk for externalizing behavior problems: a randomized controlled trial. Journal of Consulting and Clinical Psychology, 74(6), 994–1005. Subsequent research has complicated this picture: Euser, S., Alink, L.R., Stoltenborgh, M., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2015) A gloomy picture: a meta-analysis of randomized controlled trials reveals disappointing effectiveness of programs aiming at preventing child maltreatment. BMC Public Health, 15(1), 1068. ‘The “less is more” effect in attachment-based interventions found by Bakermans-Kranenburg and colleagues seems only partly applicable to programs aimed at reducing or preventing child maltreatment. We found a curvilinear association with program duration and number of program sessions. Programs with a moderate duration (6–12 months) or a moderate number of sessions (16–30) yielded significantly higher effect sizes compared to shorter or longer programs and programs with fewer or more sessions’ (11). Additionally, initial research suggests that parent learning disabilities and child physical disabilities may moderate the capacity of VIPP to increase caregiver sensitivity or sensitive discipline, though other benefits were identified from the intervention. Platje, E., Sterkenburg, P., Overbeek, M., Kef, S., & Schuengel, C. (2018) The efficacy of VIPP-V parenting training for parents of young children with a visual or visual-and-intellectual disability: a randomized controlled trial. Attachment & Human Development, 20(5), 455–72; Hodes, M.W., Meppelder, M., de Moor, M., Kef, S., & Schuengel, C. (2018) Effects of video-feedback intervention on harmonious parent–child interaction and sensitive discipline of parents with intellectual disabilities: a randomized controlled trial. Child: Care, Health and Development, 44(2), 304–311.

117 Bakermans-Kranenburg, M.J., Juffer, F., & Van IJzendoorn, M.H. (2019) Reflections on the mirror: on video-feedback to promote positive parenting and infant mental health. In C. Zeanah (ed.) Handbook of Infant Mental Health, 4th edn (pp.527–42). New York: Guilford.

118 Out, D., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2009) The role of disconnected and extremely insensitive parenting in the development of disorganized attachment: validation of a new measure. Attachment & Human Development, 11(5), 419–43, p.438.

119 Slade, A., Holland, M.L., Ordway, M.R., et al. (2019) Minding the Baby®: enhancing parental reflective functioning and infant attachment in an attachment-based, interdisciplinary home visiting program. Development & Psychopathology, 32(1), 123–37.

120 See also Tereno, S., Madigan, S., Lyons-Ruth, K., et al. (2017) Assessing a change mechanism in a randomized home-visiting trial: reducing disrupted maternal communication decreases infant disorganization. Development & Psychopathology, 29(2), 637–49.

121 Dozier, M. & Bernard, K. (2017) Attachment and Biobehavioral Catch-up: addressing the needs of infants and toddlers exposed to inadequate or problematic caregiving. Current Opinion in Psychology, 15, 111–17.

122 Dozier, M. & Infant Caregiver Project (2016) Attachment and Biobehavioral Catch-up for Infants Who Have Experienced Early Adversity (ABC-1). Intervention Manual. Unpublished manuscript; Dozier, M. & Bernard, K. (2019) Coaching Parents of Vulnerable Infants: The Attachment and Biobehavioral Catch-up Approach. New York: Guilford. For the adaptation of ABC for toddlers the focus on avoidance of intrusive, harsh, or frightening behaviours is partially supplanted by, partly incorporated within, a focus on encouraging parents to collaborate with their children in regulating difficult emotions.

123 Bernard, K., Meade, E.B., & Dozier, M. (2013) Parental synchrony and nurturance as targets in an attachment based intervention: building upon Mary Ainsworth’s insights about mother–infant interaction. Attachment & Human Development, 15(5–6), 507–523. Another important influence for Dozier and colleagues was Raver, C.C. (1996) Relations between social contingency in mother–child interaction and 2-year-olds’ social competence. Developmental Psychology, 32(5), 850–59.

124 Caron, E.B., Roben, C.K., Yarger, H.A., & Dozier, M. (2018) Novel methods for screening: contributions from Attachment and Biobehavioral Catch-up. Prevention Science, 19(7), 894–903.

125 Caron, E.B. Bernard, K., & Dozier, M. (2018) In vivo feedback predicts parent behavior change in the Attachment and Biobehavioral Catch-up intervention. Journal of Clinical Child & Adolescent Psychology, 47(1), 35–46.

127 Meade, E.B., Dozier, M., & Bernard, K. (2014) Using video feedback as a tool in training parent coaches: promising results from a single-subject design. Attachment & Human Development, 16(4), 356–70.

128 Bernard, K., Dozier, M., Bick, J., Lewis-Morrarty, E., Lindhiem, O., & Carlson, E. (2012) Enhancing attachment organization among maltreated children: results of a randomized clinical trial. Child Development, 83, 623–36.

129 Yarger, H.A., Hoye, J.R., & Dozier, M. (2016) Trajectories of change in Attachment and Biobehavioral Catch-up among high-risk mothers: a randomized clinical trial. Infant Mental Health Journal, 37(5), 525–36. An important finding has been that the effects of ABC on caregiver sensitivity appear to be just as strong when implemented by welfare organisations as when applied as part of a randomised clinical trial. This is quite unusual. Effect sizes generally drop when interventions are disseminated in the community. Dozier, M. & Bernard, K. (2019) Coaching Parents of Vulnerable Infants: The Attachment and Biobehavioral Catch-up Approach. New York: Guilford, p.177.

130 Ibid. p.143.

131 Yarger, H.A. (2018) Investigating longitudinal pathways to dysregulation: The role of anomalous parenting behaviour. Unpublished doctoral dissertation, University of Delaware.

132 Lind, T., Bernard, K., Ross, E., & Dozier, M. (2014) Intervention effects on negative affect of CPS-referred children: results of a randomized clinical trial. Child Abuse & Neglect, 38(9), 1459–67. See also Lind, T., Raby, K.L., Caron, E.B., Roben, C.K., & Dozier, M. (2017) Enhancing executive functioning among toddlers in foster care with an attachment-based intervention. Development & Psychopathology, 29(2), 575–86.

133 Grube, W. & Liming, K. (2018) Attachment and biobehavioral catchup: a systematic review. Infant Mental Health Journal, 39(6), 656–73.

134 Chapin Hall at the University of Chicago (2019) Strong families New York City. Final evaluation report, June 2019.

135 Bernard, K., Hostinar, C.E., & Dozier, M. (2015) Intervention effects on diurnal cortisol rhythms of child protective services-referred infants in early childhood: preschool follow-up results of a randomized clinical trial. JAMA Pediatrics, 169(2), 112–19; Tabachnick, A.R., Raby, K.L., Goldstein, A., Zajac, L., & Dozier, M. (2019) Effects of an attachment-based intervention in infancy on children’s autonomic regulation during middle childhood. Biological Psychology, 143, 22–31; Dozier, M. & Bernard, K. (2019) Coaching Parents of Vulnerable Infants: The Attachment and Biobehavioral Catch-up Approach. New York: Guilford.

136 Such findings also stand in thought-provoking contrast to a trial of the Circle of Security intervention, which led to change in children’s attachment classification only for mothers higher in attachment avoidance (Chapter 5). Berlin and colleagues speculate that Circle of Security is a gentler intervention than ABC, and so less likely to be aversive to caregivers with strong assumptions that devalue nurturance and emotional needs. Berlin, L.J., Martoccio, T.L., & Jones Harden, B. (2018) Improving early head start’s impacts on parenting through attachment-based intervention: a randomized controlled trial. Developmental Psychology, 54(12), 2316–27. See also Cassibba and colleagues who found iatrogenic effects of discussions with caregivers about their attachment representations, if these began secure/autonomous. Cassibba, R., van IJzendoorn, M.H., Coppola, G., et al. (2008) Supporting families with preterm children and children suffering from dermatitis. In F. Juffer, M.J. Bakermans-Kranenburg, & M.H. van IJzendoorn (eds) Promoting Positive Parenting: An Attachment-Based Intervention (pp.91–110). New York: Psychology Press.

137 For work to date see Klein Velderman, M., Juffer, F., Bakermans-Kranenburg, M.J., & van IJzendoorn, M.H. (2008) A case study and process evaluation of video feedback to promote positive parenting alone and with representational attachment discussions. In F. Juffer, M.J. Bakermans-Kranenburg, & M.H. van IJzendoorn (eds) Promoting Positive Parenting: An Attachment-Based Intervention (pp.23–36). New York: Lawrence Erlbaum; Hodes, M.W., Meppelder, H.M., Schuengel, C., & Kef, S. (2014) Tailoring a video-feedback intervention for sensitive discipline to parents with intellectual disabilities: a process evaluation. Attachment & Human Development, 16(4), 387–401; Scourfield, J., Allely, C., Coffey, A., & Yates, P. (2016) Working with fathers of at-risk children: insights from a qualitative process evaluation of an intensive group-based intervention. Children and Youth Services Review, 69, 259–67.

138 Schuengel, C. (2019) Representing attachment: the future is open. Paper presented at International Attachment Conference, Vancouver, 20 July 2019.

139 Schuengel, C. & Tharner, A. (2020) Patterns of parenting: revisiting mechanistic models. Attachment & Human Development, 22. See also Petters, D.D. (2019) The attachment control system and computational modeling: origins and prospects. Developmental Psychology, 55(2), 227–39.

140 Thompson, R.A., Simpson, J.A., & Berlin, L. (2020) Introduction: synthesizing the fundamental questions and issues in attachment theory. In R.A. Thompson, J.A. Simpson, & L. Berlin (eds) Attachment: The Fundamental Questions. New York: Guildford.