Attachment theory, despite being one of the most established theories of human development, continues to be subject to debate and criticism. Whilst it is certainly not new or unusual – nor particularly unwelcome – for scientific theories to be debated and criticised, it is somewhat surprising, to me at least, that some of the strongest criticisms have been made from within the social work profession. The reason this seems surprising to me is because attachment theory is generally accepted to be very popular amongst child and family social workers  , because it has for many years formed a central part of the social work curriculum  and because it will almost certainly continue to do so in future   . Thus, where criticism is made of the ways in which social workers (mis) use attachment theory in practice, in part this may reflect poorly on the way social workers have been taught – but beyond criticisms of the use of attachment theory in practice, some have asserted that the fundamentals of the theory itself are flawed. None of this is to suggest that attachment theory should be beyond debate – of course, it must not be – but it did prompt me, as someone who regularly attempts to apply attachment theory in practice, to consider afresh what the fundamentals of the theory actually are.
What observations does attachment theory seek to explain?
Before outlining these fundamentals, it may provide some helpful context to first set out what a theory is. A theory, in the scientific meaning, is a model that has stood up to repeated scrutiny and that provides an explanation as to why we observe the things we observe. In this sense, a scientific theory has a more complex and ‘higher’ level of meaning than a scientific law. For example, the law of gravity (F = G m1m2/r^2) describes the observation that mass attracts mass and the strength of the attraction but it tells us nothing about why mass attracts mass. For this, we need a scientific theory such as General Relativity and / or particular aspects of quantum theory. Thus, attachment theory is a model for explaining certain observations, an answer to the question of why we observe what we observe. But what observations does attachment theory seek to explain?
John Bowlby is often credited with the development of attachment theory and it is his observations that we may initially consider. However, it is worth noting that although Bowlby is often referred to as ‘the father of attachment theory’, it would be mistaken to consider his writings as the best source of information and evidence regarding contemporary attachment theory (in much the same way that although Charles Darwin is rightly credited with developing the theory of evolution by natural selection, it would be a mistake to consider the work of Darwin himself as the best source of information and evidence regarding the modern evolutionary synthesis). Nevertheless, it is from Bowlby’s observations that attachment theory developed, so what was it that Bowlby saw that he wanted to explain?
In his early career, Bowlby worked in a school for ‘maladjusted children’ . During this time, Bowlby met a very isolated, remote, affectionless teenager and a second, younger child described as ‘very anxious’, who tended to follow Bowlby around to the extent that he was given the nickname of ‘John’s shadow’. Both of these children had been separated from the care of their birth families some time earlier and Bowlby wondered about the effect of this separation and of their early family relationships upon their subsequent development. Bowlby later trained as a child psychiatrist and was exposed to the ideas of Melanie Klein, who held that children’s emotional (and behavioural) problems derived almost entirely from a conflict between the (‘phantasy’ of the) child’s aggressive and libidinal drives and not as a result of events in the real world. Bowlby found this explanation, as many others did, to be unsatisfactory. Instead, Bowlby wondered about his own experiences of working with carers and how it seemed possible to help the child by helping the carer. Of course, Bowlby’s anecdotal observations and personal experiences are not the basis of attachment theory but they are (part of) what Bowlby hoped to explain, in the same way that Darwin famously observed “an entangled bank, clothed with many plants of many kinds, with birds singing on the bushes, with various insects flitting about, and with worms crawling through the damp earth” (p. 425) and reflected on how this complexity could have been produced by natural processes.
In addition, Joyce and James Robertson, colleagues of Bowlby, spent time observing the behaviour of children in hospital (including their own daughter, Jean) At the time, children admitted to hospital were often separated from their close carers, with visiting also being much discouraged. Together, the Robertsons observed how this situation – of relatively prolonged separation from their close carers – greatly distressed many of the children and how they could not be comforted by the provision of care from an alternative adult and how the children appeared to lose all interest in playful, exploratory behaviour. In other words, although these children were by any objective measure physically well cared, they were nevertheless distressed for much of the time. This suggested that the idea of children being relatively uninterested in who cared for them was incorrect and that children need something more from their close carers than the provision of ‘simple’ physical care.
In summary, the observations that Bowlby sought to explain via the development of attachment theory related to (1) the emotional and behavioural problems exhibited by some children, (2) the positive impact upon the child’s emotional and behavioural development that could result from therapeutic or other parenting work with the child’s close carers and (3) the emotional distress of (particularly younger) children when separated from their close carers, even when provided with suitable alternative care by another adult. That some children do exhibit emotional and behavioural difficulties is, surely, undeniable, as it is that undertaking therapeutic or other parenting work with carers does (in many cases) prove helpful for the child. It is less obvious as to exactly what duration and kind of separation distresses children of different ages and those with different cultural experiences of being cared for but nevertheless, much as we can observe Darwin’s entangled bank or the fact that mass attracts mass, we can relatively easily make these observations of children and their close carers – the task then is to devise a theory that explains these observations (or as many of them as possible).
Briefly, before considering how attachment theory seeks to explain the observations outlined above, it will no doubt be obvious that I am using the term ‘close carer’ when referring to those who might care for the child, potentially encompassing a wide range of people including (but not limited to) mothers, fathers, grandparents, wider family members and non-family carers such as childminders and foster carers. Bowlby, on the other hand, tended to use the term ‘mother’ – and was understandably criticised for doing so. Nowadays no self-respecting writer on attachment (or in any field of child development) would use ‘mothers’ because there is simply a much greater awareness of the roles that many different adults have in taking care of children.
How does attachment theory seek to explain these observations?
Based on his observations and those of others, Bowlby theorized as follows – that human babies are born with an innate need to form a bond with a relatively small number of adults. Bowlby referred to this as an attachment bond. Bowlby further theorized that such a bond might have an evolutionary explanation, based on the extreme helplessness of human babies and their survival-related need for the close attention of adult members of the species. Bowlby reasoned that for the baby, forming a small number of attachment bonds with a small number of adults would be an effective survival strategy whilst for the adults, straightforward natural selection would (usually) explain why they would be interested in caring for the baby.
Bowlby also theorized that in order to maximize the potential for this attachment bond to provide for the child’s immediate and longer-term needs, the child would ‘learn’ to adapt his or her behaviour to the particular adults with whom they were attached. In other words, depending on how the adults behaved towards the child, particularly at times of heightened stress or anxiety, the child would learn how to prompt, provoke and respond to the adults so as to maximise the chances of receiving a caring response in return. Bowlby also suggested that these early relational experiences would in some important ways guide how the child responded to other relationships in future, as they grew older, and in time these experiences could also influence how they responded to their own children. Therefore, attachment theory seeks to explain the emotional and behavioural difficulties of some children as being related to the nature and quality of their attachment relationships, particularly when the child has been separated from their close carers or when the child’s close carers are either incapable of meeting the child’s needs or when they actively harm the child.
Some predictions of attachment theory
In addition to providing explanatory models, scientific theories enable us to make predictions. Indeed, testing the accuracy of the predictions made by a theory is one of the best ways of testing the theory’s validity (its fealty with the real world). For example, the theory of evolution by natural selection predicts that a mechanism must exist by which heritable traits can be modified and passed from parent to offspring (a prediction validated with the discovery of RNA and DNA).
Similarly, attachment theory enables us to make predictions about families, carers and children. For example (and in no particular order):
- That a carer’s ‘state of mind’ with regards to attachment will predict the nature of the child’s attachment relationship with that carer (it does) 
- That secure attachment relationships in childhood tend to be associated with greater levels of self-reliance and independence in later life (they do) 
- That relational (or parenting) difficulties, rather than wider difficulties per se, will tend to predict the security of infant attachment (they do) 
- That attachment ‘classifications’ are not fixed, individual properties but are amenable to change (they are) 
- That negative changes in the caregiving environment (life events and circumstances) can lead to changes in the nature of the individual’s attachment relationships (or ‘state of mind’ with regards to attachment) and in the predictable direction of increased insecurity (they do)
- That secure attachment relationships in infancy tend to be associated with greater social competency in later life (they do) 
- That as attachment has evolved in the context of relatively physically helpless but relatively cognitively developed offspring, attachment will also exist in non-human primates (it does)
- That children from different cultures will form attachment bonds with their close carers, even when the child is cared for in significantly varied ways (when compared with most children in the UK and USA), including when the child is cared for in a communal setting (they do) 
- That maltreated children will still form or maintain an attachment with a close carer who is maltreating them although the nature of this attachment relationship is very likely to be insecure rather than secure (they do and it is) 
The relative accuracy of these predictions suggests that attachment theory does, at the very least, have a degree of validity although clearly, as with any theory of such a complex phenomena as human development, there remains much to learn (both about attachment but also about other aspects of relationships and development).
One additional point worth making is that it is sometimes argued that attachment theory lacks predictive validity when applied to individuals because we cannot know what life events (positive or negative) they may encounter in future. In other words, that security of attachment in childhood does not really predict more competent social functioning in adulthood because, for example, something terrible might occur during the child’s teenage years (or vice versa with regards to insecurity or disorganisation of attachment). And of course, this is true, in much the same way that whilst we can predict with a high degree of certainty how Newtonian physical systems will develop in future, we cannot account for unforeseen changes. For example, if you hit a cue ball on a snooker table, given enough time and the right kind of data about the initial setup and your cue stroke, it is perfectly possible to predict where all of the balls will land – but not if your opponent suddenly decides they’ve had enough and puts their cue down on the table.
This means that whatever the nature of a child’s attachment relationships in early childhood, we cannot say with certainty what the future will hold for their development – and attempts to do so without nuance and sufficient caution are unwarranted – but we can certainly note the association between certain kinds of experiences (such as maltreatment), certain kinds of attachment relationships or behaviour and the potential likelihood of certain future outcomes if nothing changes.
How to disprove attachment theory
The final aspect of a scientific theory to consider, in addition to offering an explanation for what we observe and enabling us to make predictions, is that it is falsifiable. Unless it is possible to disprove a particular theory, it cannot be considered as genuinely scientific. For example, it has been said (somewhat tongue-in-cheek) that the theory of evolution by natural selection could be falsified by the identification of a Precambrian rabbit fossil. With regards to attachment theory, there are many ways in which it might be falsified, or at least severely challenged, including (but not limited to):
- Finding that children without neurological disorders and without extremely neglectful upbringings do not form attachments with a small number of adults.
- Finding that positive changes in the caregiving environment led to greater insecurity in the child’s attachment relationships or, vice versa, that negative changes in the caregiving environment led to greater security in the child’s attachment relationships.
- Finding that the majority of maltreated children formed secure attachment relationships with their abusive or neglectful carers.
- Finding that children formed similar attachment relationships (secure or insecure, avoidant or ambivalent, etc.) with different carers, despite those carers treating them very differently.
- Finding that young children are just as likely to form attachment relationships with other children (of a similar young age) or with animals or inanimate objects as they are with their adult carers.
- Finding that the way a child behaves towards their close carers is dependent entirely on genetic factors and / or entirely independent of the caregiving environment.
This is not to say, of course, that attachment theory explains every dynamic of close human relationships or development (a Grand Unified Theory of human development would seem much more difficult and much less likely than a Grand Unified Theory of physics), nor that every child with emotional and behavioural difficulties must have experienced attachment-related difficulties (claims to the contrary are mistaken and demonstrate a significant misunderstanding of attachment theory and child development). We know, for example, that some neurological disorders lead to emotional and behavioural difficulties and that emotional and behavioural difficulties arise for some children in the absence of both neurological and attachment-related problems. In other words, that attachment is but one facet of human development.
Conclusion – social work and attachment theory
More broadly, there are some critics of attachment theory who claim that, regardless of whether it offers a good model for understanding human development or not, within the context of social work practice, it focuses on ‘the wrong things’, namely individual and relational difficulties, requiring individual and relational solutions, rather than on social difficulties, requiring social solutions. However, there is no evidence to suggest that it must be a case of ‘either-or’ and indeed many definitions of social work suggest that social workers should aim to address both social and personal problems together. Indeed, some would claim that our ability to do just that is part of what makes social work unique.
Thus, whilst a thorough consideration and understanding of the social difficulties that many families in the UK routinely experience is hugely important for social work and for individual practitioners, so too is an understanding of individual and relational difficulties and as such, social work surely needs individual theories for practice (including but not limited to attachment theory) as much as it needs social theories of practice.
 Beckett, C. (2006). Essential Theory for Social Work Practice. London: SAGE Publications (see p. 49).
 Barth, R., Crea, T., John, K., Thoburn, J. and Quinton, D. (2005). Beyond attachment theory and therapy: Towards sensitive and evidence-based interventions with foster carers and adoptive families in distress. Child and Family Social Work, 10(4), 257 – 268 (see p. 257).
 Module information from 3 of the top 5 Universities for Social Work in 2014 – University of Sussex, University of Bath and University of Leeds.
 Croisdale-Appleby, D. (2014) Re-visioning social work education (see p. 28).
 Munro, E. (2011a) Munro Review of Child Protection: Interim Report – the Childs Journey (see p. 96).
 Narey, M. (2014) Making the education of social workers consistently effective (see p. 10).
 Bretherton, I. (1992) The Origins of Attachment Theory: John Bowlby and Mary Ainsworth. Developmental Psychology, 28(5), pp. 759 – 775.
 van IJzendoorn, M. (1995) Adult Attachment Representations, Parental Responsiveness, and Infant Attachment: A Meta-Analysis on the Predictive Validity of the Adult Attachment Interview. Pyshcological Bulletin, 117(3), pp. 387 – 403.
 Sroufe, L. (2005) Attachment and Development: A prospective, longtitudinal study from birth to adulthood. Attachment and Human Development, 7(4), pp. 349 – 367.
 Easterbrooks, A. and Graham, C. (1999) Security of Attachment and Parenting: Homeless and Low-Income Housed Mothers and Infants. American Journal of Orthopsychiatry, 69(3), pp. 337 – 346.
 Egeland, B. and Farber, E. (1984) Infant-mother attachment: Factors related to its development and changes over time. Child Development, 55(3), pp. 753 – 771.
 Waters et al (2000) The Stability of Attachment Security from Infancy to Adolesnce and Early Adulthood: General Introduction. Child Development, 71(3), pp. 678 – 683.
 Sroufe, L. (2005) Attachment and Development: A prospective, longtitudinal study from birth to adulthood. Attachment and Human Development, 7(4), pp. 349 – 367.
 Suomi, S. (1995) ‘Influence of Attachment Theory on Ethological Studies of Biobehavioural Development in Nonhuman Primates’ in S. Goldberg, R. Muir and J. Kerr (eds) Attachment Theory: Social, Developmental and Clinical Perspectives. The Analytic Press, Hillsdale, NJ.
 Mizuta et al, (1996) A Cross-cultural Study of PreSchoolers’ Attachment: Security and Sensitivity in Japanese and US Dyads. International Journal of Behavioural Development, 19(1), pp. 141 – 159.
 Morton, N. and Browne, K. (1998) Theory and observation of attachment and its relation to child maltreatment: a review. Child Abuse and Neglect, 22(11), pp. 1093 – 1104.