Attachment theory seems to be the most recent in a long line of psychotherapy crazes promising to finally provide the magical solution to reliably heal suffering. Unfortunately, like other predecessors such as recovered memory and primal scream therapy, the evidence seems to indicate that attachment theory overpromises and under-delivers. Let's take a look at why.
Briefly, for those unfamiliar, attachment theory was developed by psychoanalyst-ethologist John Bowlby who was influenced by the work of Conrad Lorenz on baby geese in which they became attached to the first thing they saw rather than the mother goose herself. From there, subsequent researchers such as Mary Ainsworth observed the behavior of small infants in relation to their caretakers.
Ainsworth in particular developed what is called the "strange situation." In this experiment, caregivers would bring in 1-year-old infants into a room full of toys, then leave. A stranger would then enter, and finally, the stranger would leave again and the caregiver would return. Based on the reactions of the infants, Ainsworth developed three main categories of attachment: secure, anxious-ambivalent, and avoidant, and later on a fourth, disorganized. The secure babies would cry when the caregiver left but were easily soothed when she (and it was usually a she, returned). The anxious children would also cry but had trouble being soothed, while the avoidant children betrayed no reaction, although heart rate and blood pressure monitors demonstrated that they too were experiencing a high degree of stress.
Ainsworth concluded that secure children had all of their needs closely met by the caregiver, that anxious children experienced inconsistent caregiving, while the avoidant children experienced some form of neglect. Further research discovered that children tended to have the same attachment styles as their parents. Using a test called the Adult Attachment Interview (AAI), other researchers concluded that attachment was relatively stable across the lifespan.
As such, attachment research using the AAI or Experiences in Close Relationships (ECR) has been utilized to gain further insight into various populations being studied. Both the AAI and ECR are valid and reliable tests and so attachment is a real concept that can be measured. So far so good.
But this is where the train starts to head off the rails. In clinical settings, attachment has been used to seemingly explain everything seen in psychotherapy, from psychopathology to relationship disturbances to sexual proclivities, and has often been the main focus of clinical interventions. In its extreme form, practitioners of attachment-based therapies believe they can alter the attachment style of the client through the crucible-like effect of the therapeutic relationship.
But to buy into this belief system, we have to accept a few implicit assumptions. First, attachment is entirely environmentally constructed. In other words, the child is born a blank slate, and the behavior of the caregiver, whether they were consistent or inconsistent (what does that mean and how is that even measured?) were the sole causative effects of an anxious or avoidant attachment style. This belief only leads to nervous parents fretting about the grievous implications of every single "mirroring" opportunity missed, and the creation of an entire style of parenting called attachment parenting.
This viewpoint is beginning to be debunked. A recent article in the Psychological Bulletin found that attachment "transmission" from caregiver to child is confounded by genetic variables of heritability. Attachment style is heritable just like IQ, personality, and a whole host of other things, such as political leanings and spirituality. (Yes, political leanings and spirituality, I'm not kidding, click on the links.)
Second, attachment theory seems to have posited that attachment is some kind of a monolithic relational mind map that applies globally, but recent research shows that individuals can be attached in different ways to different people. Indeed the child can have a secure attachment to his mother, but an avoidant attachment to his father, and an anxious attachment to an aunt, and so on. Indeed, the research I link to above found that "the relationship between father-child and mother-child attachment was not significant" and "parents’ time away from their child was not a significant predictor of attachment." If attachment is context-dependent, variable between relationships, and not influenced by separation, then what really is its utility?
Which leads to the third and most important issue. What relevance does attachment theory hold for clinical practice? On a basic level, it is often helpful for clients to understand common themes and dynamics in their relationship history, but that doesn't require a deep dive into attachment. The most common application is that attachment will somehow be "healed" through the therapeutic relationship. Sure, after a number of sessions over months or years, the client can and will grow to trust the therapist and may practice and rehearse a number of behaviors within the context of the relationship. And so, obviously, it comes to pass that the client feels securely attached to the therapist. I'm not knocking this; sometimes this is exactly what the client wants and needs. But what then? Does that mean that the client is now securely attached to other people?
The indications seem to point to No. Take a look at this recent study that explored the relationship between childhood attachments and relational attachments. The researchers concluded "the variable of attachment to parent, compared to the attachment of a romantic partner had a medium effect size.... (while) the variable comparing attachment to parent, and the relationship quality and satisfaction had a small to medium effect size."
Small to medium effect size? Doesn't sound too inspiring. So, with these findings in mind, what is the purpose of such therapy, besides building a sense of intimacy between client and therapist? How does such sense of "security" help the client's relationships the other 167 hours of the week? I would argue that obviously, it can be quite helpful to have a trusting ear to lean on, but based on the research, I would temper expectations for that therapeutic relationship to readily translate to other relationships.
Finally and most perniciously, the use of attachment theory has often been used to pathologize non-normative sexualities. Under the guise of relational "normality" established by a social construction of what a secure relationship looks like, folks have been labeled as insecurely attached simply for desiring sexual variety or displaying an interest in fetishistic activities. Hence, attachment-based therapists, especially those influenced by Freudian concepts, will often take on the stance that the security of their stable, long-term therapeutic relationship will not only change the client's attachment style, but in this way also transform the "sexual deviant" into a stable, sexually conforming member of society. Look, attachment theory is often a helpful research tool, (and I have used it myself in my own research), and can often also be useful as another lens through which to better understand the client, but let's take its most pernicious aspects out of the therapy office so it can no longer impose arbitrary, moralistic societal standards on relational and sexual desires.