Most psychoanalysts seem to see attachment theory in much the same way as traditional doctors view homeopathy — as a potentially destructive dilution of what may be valuable in one's clinical approach. John Bowlby was seen as a reductionist who failed to understand the importance of unconscious fantasy, simplified the nature of psychic conflict out of existence, ignored the formative nature of the infant's experience of his or her physical body, and saw only one etiological variable: physical or psychological separation.
Opposition to Bowlby's views was probably one small area of common ground for the three warring factions of the British Psychoanalytic Society. Not that Bowlby was any more balanced in his view of psychoanalysis. He ruled out the relevance of psychoanalytic observations to the understanding of psychopathology on epistemological grounds, believing them to be hopelessly flawed by their subjectivity and retrospective nature. Perhaps this was why, unlike most of his creative psychoanalytic contemporaries, Bowlby's ideas failed to generate another psychoanalytic theoretical school. Instead they led to a line of empirical research in developmental psychopathology that distanced attachment theory further and further from psychoanalysis. Now the gap has become not just a novel orientation to understanding clinical cases but an incompatible epistemology.
Bowlby's interests in observation and the representation of the real — as opposed to the reality of the representation — put him out of bounds for all but the most unorthodox of psychoanalysts. The influential British psychiatrist Jeremy Holmes put it poignantly: “Bowlby's theory and its tremendous ramifications for clinical work were for decades virtually airbrushed out of the psychoanalytic record — rather like some dissident in Stalinist times.”
Nor did the controversy end with Bowlby's death in 1990. The pertinence of attachment research has become hotly disputed among early development researchers. The explanatory value of attachment theory is championed by some psychoanalytic researchers (e.g., Daniel Stern, Robert Emde, Louis Sander, Mary Main, and Karlen Lyons-Ruth) and just as firmly opposed by others (e.g., Peter Wolff and Andre Green).
“A careful study of attachment and psychoanalytic theories reveals that the relationship between these two attitudes may be more complex than adherents of either community generally recognize. There are many points of contact.”
In the pro-attachment camp are those whose clinical theory focuses on the development of the self or the evolution of inter-subjective understanding. Their beliefs about the importance of attachment theory for the understanding of human development are reinforced by observations from several sources, including infants' receptivity to interpersonal interactions from the first moments of life, the confirmed role of the early social environment as a predictor of subsequent development, and the developmental significance of the infant's agency from the earliest days.
To those in the anti-attachment camp, whose clinical explanations are couched in terms of the formative nature of infantile sexuality and the adaptive value of accommodation to destructive impulses, infancy research is not just an irrelevance. Some reject its prepsychoanalytic depiction of a “baby being” whose social interactions are untarnished by aggressive or sexual fantasies.
While their theoretical allegiances clearly predict their positions, the combatants rarely declare these interests. Consequently, the arguments tend to be esoteric, and are frequently dominated by philosophical debates about the nature of psychoanalytic knowledge.
As frequent contributors to this kind of discourse, we have had ample opportunity to persuade colleagues that knowledge of early development is indeed helpful to us as analysts, particularly in trying to understand primitive states of mind. Such knowledge is easier to obtain in experimental laboratories or in systematic observational studies of normal development than solely through the clinical study of adult pathology. We hope that we might have persuaded some. However, our preaching might have been most successful when addressed to the previously converted.
Of course, as so often in psychoanalytic debates between two opposing groups, our discussions constantly risk mutually assured destruction of both positions. The debates get simpler and simpler, more and more stereotypical, as the subject matter is increasingly stripped of its complexity.
A careful study of attachment and psychoanalytic theories reveals that the relationship between these two attitudes may be more complex than adherents of either community generally recognize. There are many points of contact. For example, Melanie Klein's description of a paranoid-schizoid mode of mental function shares much with attachment researchers' descriptions of certain types of insecure attachment. There is reasonable evidence to indicate that infants who deal with the overwhelming distress of separation by avoidance may be those most likely to develop narcissistic traits in adulthood. It is also probable
Peter Fonagy, Ph.D., F.B.A., is Freud Memorial Professor of Psychoanalysis and director of the Sub-department of Clinical Health Psychology at University College, London. He is director of the Child and Family Center at the Menninger Foundation in Topeka, Kan., and director of research at the Anna Freud Centre in London. Mary Target, Ph. D., is a senior lecturer in psychoanalysis at University College and an associate member of the British Psychoanalytic Society. She is deputy director of research at the Anna Freud Centre.
that infants who fail to develop successful strategies in the face of their mothers' unpredictable frightened or frightening actions are most at risk for diagnosable psychiatric disorder in middle childhood and adolescence.
Attachment theory has also forged close links with modern object relations theories of psychopathology by providing detailed descriptions of the evolution of affect regulation. These complement Bion's ideas about alpha function and containment and are compatible with Winnicott's descriptions of a holding environment as well.
Around the very same concepts where we may detect convergence, we may also identify profound divergence, potentially of immense relevance to attachment theorists. A key discovery of modern Kleinian theory has been the constant oscillation between paranoid-schizoid and depressive positions. The dialectic whereby depressive anxieties constantly lead us to abandon the sense of responsibility we might momentarily feel for our actions, and throw us back on to more primitive modes of mental functioning, is one of the key drivers of psychic life. Modern attachment theory, on the other hand, is static. The achievement of attachment security theoretically precludes the re-emergence of insecure states. Naïve attachment theory asserts a developmental determinism that is poorly supported by research data.
A desirable theoretical development in attachment theory would embrace the developmental dialectic, highlighting the fact that secure attachment might unconsciously represent a threat to some individuals. This could explain why secure attachment is by no means inevitably stable and why longitudinal studies frequently show that it sometimes gives way to insecure attachment with no evidence of environmental influence. Under certain circumstances — for example, environments that enforce separation — an avoidant dismissing attachment style may be indeed more adaptive than a secure one.
On the other hand, psychoanalytic clinicians could benefit from better understanding of the contribution that attachment theory has made toward identifying certain developmental determinants of the regulation of emotion. We are particularly impressed by what attachment theory has taught us about the role that the other's interpretation plays in the subject's experience of affects.
In summary, we make a plea here for a type of cross-fertilization between two distinguished and independent traditions that would compromise neither but lead to an enrichment of both.