In this posting I will summarize the presentations of Marjorie Beeghly and Ed Tronick on the subject of Attachment Theory (AT). Marjorie’s presentation explained and offered video demonstrations of the strange situation paradigm, and Ed’s presentation challenged some of the tenets of the theory. This is a long post, and I apologize, but I wanted to complete the description of the IPMH weekend.
Although we don’t know much about how attachment mechanisms get transmitted from one generation to the next, attachment research is a burgeoning field. The research has generated lots of controversy – sort of like religion or politics. The main idea underlying Attachment Theory is that if the baby learns to trust the mother to comfort him when he is in trouble, he begins to learn that he can trust other people later in life.
It used to be that the mother-child relationship was not thought of very much. Harry Harlow, in the late ‘50’s and early ‘60’s, designed a study with a monkey fed on a wire mother. In spite of the fact that the nourishment derived from the wire mother, the baby spent most of his time on the cloth mother. The tactile comfort of the cloth mother dominated.
WWII – war orphanages. Renee Spitz. The description of infants deprived of their mother for 5 months is haunting. It sounds like Autistic Spectrum Disorder except for the motor retardation. It also look like pictures of children from Romanian orphanages – face vacuous, stereotypic finger movements, etc.
Attachment is bio-behavioral, wired in. when you separate infants from the caregiver, especially in the second half of second year of life – the baby responds with anger, and then sadness. Attachment is only activated under conditions of uncertainty, threat, danger, fatigue, or illness. There are other competing biobehavioral systems, such as exploration and affiliation.
Bowlby’s 4 attachment phases
1. Pre-attachment (birth) – mother’s sensitivity; people are drawn to want to take care of babies .
2. Attachment in the making – 2-6 mo. Lot of brain development, infants start having differential behaviors (crying) to different caregivers, social smile, etc.
3. Clear cut attachment – 6-12 mo. Separation anxiety, motility.
4. Goal corrected part (reciprocal relation) 2 plus years. Bob Marvin researched this stage. Internal working models, felt security gets internalized. Erikson also talked about this idea.
Mary Ainsworth did not design the Strange Situation to measure attachment but instead to study how the three biobehavioral systems interchanged (attachment, affiliation, exploration). She did research in Uganda, and her Uganda findings confirmed in the Strange Situation. It is true that the n of her studies were very small.
Often, adherents of Attachment Theory are critical of Freudian theory. Although S Freud’s theory of infancy did have a “virtual infant” in that it derived from reconstructions of the childhoods of his adult patients, Freud was also an acute observer of children – such as his grandson who threw a spool out of his crib and pulled it back and then repeated it, saying “Fort. Da.” (Gone. There.) – when his mother had left the room, thinking that the infant was representing and attempting to master the separation (Freud, S, Beyond the pleasure principle, The Standard Edition,1920). Also his coaching of the analysis of 5-yo Little Hans (Analysis of the phobia of a five year old boy, The Standard Edition) in 1909 demonstrated sensitivity to children in that he told the father to follow Little Hans’ lead in the conversation (it wasn’t really play), he emphasized tolerance and affirmation in father’s responses, and he was keenly aware of the body-centered focus of children’s inner representational life. Also, Anna Freud with her friend Dorothy Burlingham observed in the war nurseries and were marvelous observers. In one paper that is not well known today they describe the powerful bonds developed by children kept in the same children’s home that lasted years later after they had been separated and adopted (Alpert, A (1945). Infants without families: By Anna Freud and Dorothy T. Burlingham, The Psychoanalytic Quarterly, 14:236-238).
Mary Main described “Disorganized Attachment” (Hesse, E, & Main, M (2000) J Amer Psychoanal Assoc, 48:1097-1127). This is a heterogeneous category, characterized by the lack of a coherent strategy to help the child cope with the stress of separation. It is hypothesized that in this case the source of a secure base is also the source of fear. This attachment category is associated with trauma in mother’s history and also with maternal psychopathology.
In Attachment Theory, you have to have special training to score kids.
In videos of the strange situation, the key is that the child is comforted by the mother after the separation and then able to “explore”, reengage with the toys in the mother’s presence. In avoidant behavior, when the mother returns, the child avoids her, a snub. Some babies ignore the mother. The last clip, from a study of cocaine exposure, was hard to score. The child can’t maintain the calming, plus there is a weird resistant behavior. In high risk populations, the relationship between attachment status and outcome is not so clear; there are so many other factors at play.
Is maternal sensitivity the best or the sole predictor of adjustment in life? DeWolfe and van Ijzendoorn conducted a meta-analysis of 65 studies. Their results confirm that maternal sensitivity is the best predictor but the effect size is small. Temperament: Jerry Kagan.and Nathan Fox, etc. suggest that attachment style is predicted by temperament. Most studies indicate that temperament affects attachment behavior but does not predict attachment status. Temperament can alter the type of subcategory.
Grazyna Kochanska studies MRO (mutually responsive orientation), a dyadic factor. She looked at fearfulness and looked at whether MRO predicted attachment status. Kochanska found that secure children showed less fear at 33 mos and were more joyful. Ambivalent kids were the most fearful and least joyful. Avoidant kids showed increasing negative emotions (counter to temperament). The kids who were angriest later were the kids who were classified as D at one year.
Instead of using AT to characterize infant-caregiver relationships, Ed proposes the following ideas, following a non-linear systems meta-theory (Tronick, E. (2007) The Neurobehavioral and Social-Emotional Development of Infants and Children, New York and London: WW Norton Press):
(1. Infant and adult are active. (2. Infant and adult are intentional. (3. Process of reparation is messy. (4. Form of reparation is unpredictable. (5. Process is co-creative. (6. Specific fittedness – must fit to intentions of infant and adult.
Dynamic systems include engagement with others and with oneself. DST (dynamic systems theory) includes an experiential component for us as humans (Tronick, 2009). When you are successfully gaining information you have an experience of pleasure, expansion, and you seek connection. It is in the co-creation of meaning that this happens. You are flexible and dyadic. When you fail, you have anxiety, withdraw, and are unhappy. Meaning making is a fundamental way of regulating ourselves. It can be in words and also in the body. When it goes wrong, things get disorganized and move towards entropy. To preserve ourselves, we become rigid, defensive, aggressive and hostile, use projection, etc. Kids who are compulsively one way or another are staving off disorganization. They are holding on for dear life (the Spitz kids) and staving off disorganization. Kids with behavior problems are also rigidly organized – they go from one state to another and the pattern in which they go from here to there is unvarying. Bullying, for example, is a rigid pattern; kids who bully gain coherence and a defense against anxiety. Homeostasis – there is a dynamic homeostasis over time, but real homeostatic systems are rigid and not dynamic systems. There are some states that must stay pretty much the same, but we bring to bear a lot of other systems to keep them stable, such as body temperature.
Normal development involves change. Normal disorganization is regulated by the parent-infant system. The infant goes from crawling to walking and disorganizes the crawling system to the walking system; this involves anxiety and takes energy. Change is costly and this is true in part because it is unpredictable, and because you are taking apart the old organization, which is anxiety provoking. The clinician’s work is to maintain the organization while the patient allows for the disorganization of the previous organization and takes the risk of trying something new.
The Jim Coan did a study with the MRI and holding hands. Being alone per se is stressful. You could take Coan’s study as a model of the need to be in the presence of another. In this study, the MRI demonstrated that pain can be mediated by holding the hand of a person with whom the subject was in a relationship better than by holding the hand of a stranger (Coan, J. A., Schaefer, H. S. & Davidson, R. J. (2006). Lending a hand: Social regulation of the neural response to threat. Psychological Science, 17, 1032-1039). ?
Getting attached has to do with forming a relationship and creating a new way of being together, a feeling of “connection”. Bowlby focused on a process going on over time that was related to emotions and making a connection in relationships. We have since somewhat lost focus on the process of Bowlby’s theory, through our attention to the categories. The process of meaning making leads to a different take on attachment relationships. When new meanings are co created they generate a variety of emotions and qualities. This may sound like attachment but we confuse ourselves with this overly broad use of the term. Attachments are primarily about safety, feelings of security and the reduction of fear. We have gone on from these ideas to what may be an unrealistic expansion of the theory. In fact, relationships involve lots of ways of being together. Attachment security and the myriad qualities of relationships can be dissociated. Moreover, relationships contain contradictory emotions. For example, intimate relationships may or may not be secure. An abused child may love the abuser but may not feel secure with him. You don’t only have positive feelings towards someone. In fact, to the extent that you have positive feelings about someone, you also have negative feelings such as the fear of loss of that person. What therapists sometimes have to do is to provide the hope or scaffolding to allow for change. For an individual to hold in mind these contradictory possibilities – love and hate – in the service of something you do not know will happen for sure, threatens organization. To the extent you can simplify it, you resolve many issues, so you protect yourself from having to hold it all at the same time.
A problem with Attachment Theory is that it is often used reductively. For example, just because someone is depressed does not mean that they are disorganized. There are also probably borderline personalities that have secure attachments. Which information would be more valuable to have – an assessment that this child has an insecure attachment or the information that from the time he was 5-yo he was in 9 different foster homes? We have elevated Attachment status to a level that carries too much meaning. People will give a patient a disorganized Attachment status and then go on to mention all the other traumatic features of their lives. In AT models, the Attachment status and clinical diagnosis can be reversed, such that instead of beginning with insecure attachment that leads to depression, you could start with depression and end up with an assessment of insecure attachment. These are all correlation studies. When predictions are made, it suggests causality, but in this case the issue of causality is not at all clear. Take for example, the development of the capacity to reach for an object. All infants at about 15 mos of age, end up reaching for an object, yet every infant gets there through a unique process. We have some understanding of what goes into that process – including gravity – but you can’t predict the particular path an individual child will take to get there (Thelan and Smith, 2001)). You you can predict the outcome but not how the process by which the outcome is reached.
A more variegated process is needed to account for the varieties of normal and abnormal, and Ed proposes the process is meaning making. Meanings come in infinite and multi-leveled forms, all of which are aimed at increasing complexity or coherence (according to the meta-theory of dynamic systems theory). When successful, relationships can take on infinite forms and qualities, intensities and rhythms, and ways of being together in emotional, sexual, and body domains because each of these domains is a domain of meaning making.
Most of the research in AT fails to test alternative hypotheses accounting for long-term stability. Take for example the AAI (Mary Main’s Adult Attachment Inventory). Suppose we consider the AAI to measures of self esteem, ego function, and sense of self. Self- esteem is a mid-level concept. You could think of self actualization, or of Erikson’s theories of development. If we compared Erikson’s theory of development and the AAI, which helps most? The basis for the legitimacy of the AAI is that it indicates the type of attachment but in fact it covers a whole lot of other things that are not discussed. Whereas the AAI proposes to reflect the coherence of thought in relation to the security of attachment, in fact it also reflects coherence of thought in other domains such as self-esteem. There have not been studies about this. Why is it not equally legitimate to state that if you have good self esteem, you are more securely attached? The AAI is really talking about a coherent narrative. Mary Main studied with Liz Bates, a linguist; she originally studied psycholinguistics. So the idea of coherent narrative came to her from this route. What if that one thing we call security picks up on many different things, but security is the only thing we are addressing? For example, when Marjorie talked about the child who had secure attachment, she pointed out shared attention and a lot of things and related them to secure attachment, but these behaviors were not in the coding scheme. It is similar to CBT studies in that if you do studies in CBT you study the manual but you do not study all the other things that are going on such as in the relationship. We do not have process research, but only outcome research. CBT may make you better, but we don’t know what it is about CBT that makes you better. When we take AT our of the laboratory and into the consulting room, we use the term AT as if it had the imprimatur of science, when actually we are only going from our observations about the relationship. We invoke a biologic scientific idea to explain what you should do in the clinical situation. It would be similar to using the strange situation paradigm to help guardian ad lidems make decisions about the child’s custody.
AT also forgets about what predictability means. Bowlby talked about process, but we have reified the status. We forget everything going on between one year and five years of age. If you are in a fairly stable environment, it is likely you will be in the same environment five years from then. What if you get an epigenetic change and it dissipates but then the environment comes in and says, let’s put that gum on the light switch again?
There is also the issue of problematic anchors – the end of the scale for sensitivity is not trauma and neglect. If you use scales, people often avoid using the first and last point. When Ed was taught scaling, he was taught to put down an anchor and scale from there. When people talk about attachment, there is a subtle movement from normal caregiving to neglect and abuse, and the neglect and abuse becomes the convincing anchor. It is the child with neglect and abuse that turns out to have problems at 5-yo. Ed bets that if you took a child with abuse and neglect, the abuse and neglect would be a better predictor of problems at 5, than attachment status. In other words, the “insensitive” parent is more likely to abuse or neglect. The sensitivity scale is problematic – sensitivity is a problematic way of talking about it. Also in most contexts if you are good enough to take care of a child’s fear, chances are that you take care of other things they need as well. We have taken one concept and cast a big net with it and end up with a score. Ed studied the Gusii tribe, who do not play with their infants and habitually do not respond empathically to the infant’s expression of emotion, and he guesses that there are lots of secure Gusii infants. The Gusii are sensitive about regulation but not about emotions.
With AT the marker is that the child is willing to stay with a person who can regulate them and make them feel comfortable, but not everyone has this opportunity. You first have to feel secure and physiologically stable in order to do anything. Lou Sander called these things pre-emptive. Although AT does describe illness or other biological sources of distress such as hunger as factors that may interfere with the development of secure attachment, the importance of these factors is frequently overlooked in the AT literature. One can be reduced to fighting and fleeing. Steve Porges said you first have to make the child feel secure, safe, physiologically well regulated. You can do that and you can stop, or having done that you can do a whole bunch of things – socialization, language, motor skills, etc. If you have been ill, all your resources are going into being not fearful, and you don’ want to be social.
Thelen E, Schoener G, Scheier C, Smith L (2001). the dynamics of embodiment: A field theory of infant perseverative reaching, Behavioral and Brain Sciences 24(1):1-34.