No More Learning

One can imagine that sensation is, in the early stages of life, a shared activity between parent and child as the experiences of holding, seeing, feeding and touching are presented to the growing child. As Garland (1991) argues, traumatic events overwhelm the 'stimulus barrier' so that, although perceived, they cannot be sensed. The subject is paralysed by them and cannot actively present them to themselves, while the parent or protector who might help to do so is inevitably absent. The task of therapy then is to 'represent' these traumatic events - via a narrative transformation from 'semantic' to 'episodic' memory - in such a way that they can be sensed, and therefore, by definition, made conscious. This process could possibly be described as 'omnipotent' in so far as any representation or map, including the cerebral 'map' of feelings, is 'omnipotent'. Thus a grain of sand could be said omnipotently to 'contain all heaven'. Here is an example of such emotion recollected in (comparative) tranquillity:
The tonsillectomy
A man in his thirties entered therapy because of his feelings of depression and a failed marriage. His relationships were characterised by avoidant attachment. He was always seemingly throwing away the very things that he wanted. He knew what he did not want, but not what he wanted. Whenever his career threatened to take off he would leave his job. A similar pattern affected his relationship with his partner: the closer they became the more likely there was to be a violent argument. He was an only child whose father had been killed in the war, and the origins
Attachment Theory and psychotherapy 167
of this pattern seemed to go back to his mother, on whom he was very dependent, but whom he experienced as intrusive and interfering.
One winter's day as he was waiting for his therapy session he saw the therapist through the closed window breathing steam into the cold air. He found himself worrying that the therapist might have something wrong with his lungs. Suddenly a flood of memories returned a tonsillectomy he had undergone when he was five. Visiting was restricted (these were normal regimes in those pre-Bowlbian days), but he was able to see his mother through a glass window twice a week (it may not have been that bad - this was how he recalled it). He remembered his fury at not being able to go home with her, throwing the toys she had left for him, shouting 'I want my mummy . . . '. As the memories returned so he began to cry profusely. This session was a turning point, enabling him to move from a position of 'I don't want . . . ', to 'I want . . . '. The traumatic separation had been re-experienced in the therapy, and no longer needed to be enacted via projective identification (doing to his employers and girlfriend what as a child he had felt had been done to him by his mother) but could be symbolised and so become part of the therapeutic narrative.
Therapists out of touch?
Attachment Theory throws an interesting light on the dilemma posed by the problem of touch in therapy. Bowlby emphasises the importance of real attachment of patient to therapist. Because attachment needs are seen as distinct from sexual or oral drives there is no intrinsic danger of gratification or seduction. Attachment provides a quiet background atmosphere of security within which more dangerous feelings can be safely explored. The patient who asks to touch the therapist, to hold a hand or be hugged, is wanting to get hold of the 'environment mother' who let him down or was absent in childhood, and it may be legitimate in certain circumstances, and with appropriate ethical safeguards (Holmes and Lindley 1989) for the therapist to respond to such a request (Balint 1968). In 'Attachment and new beginning', Pedder (1986) describes how a patient who had been separated from her mother for 6 months in infancy
buried her head in the pillow, extending her arms out loosely to either side of the pillow. Her hands moved around restlessly,
168 Imlications
reaching silently in my direction for some ten minutes. Eventually I said I thought she wanted me to take her hand, though she felt unable to say so, and then I did.
This seemed an important new beginning and she was later able to say how she had been terrified of being too demanding in asking me to hold a hand, fearing I might not trust her and might have mistaken her wish to be held as sexual.
(Pedder 1986)
Secure attachment to the therapist may be part of a 'new beginning' for certain patients, and some physical expression of this can be helpful. But - and here is the dilemma - pain and anguish of separation also need to be re-experienced if the patient is to feel safe enough to form new attachments, secure in the knowledge that, should things go wrong, the loss can be mourned and that he will not be left feeling permanently bereft.
Winnicott's view that trauma needs to be brought 'within the patient's omnipotence' is echoed by Casement (1985) in his discussion of another case in which the patient had asked to hold her therapist's hand. This was a woman who had been badly burned as a child and whose mother had fainted while holding her hand when the burn was being operated on under local anaesthetic. After initially agreeing, Casement later decided not to accede to the patient's request. This withdrawal led to fury and near-psychosis in the patient, but once this had been survived she began rapidly to improve, and it seemed that the uncanny repetition in the transference of the mother's holding and then letting go of the patient, while remaining in a therapeutic context that was basically secure, had contributed to this breakthrough. Casement quotes Winnicott:
the patient used the analyst's failures, often quite small ones, perhaps manoeuvred by the patient. . . . The patient now hates the analyst for the failure that originally came as an environmental factor, outside the area of omnipotent control, but that is now staged in the transference. So in the end we succeed by failing - failing the patient's way. This is a long distance from the simple theory of cure by corrective experience.
(Winnicott 1965)
Bowlby the scientist was always parsimoniously trying to devise a 'simple theory' with which to explain the enormous complexity of intimate human relationships. Attachment Theory, while in
Attachment Theory and psychotherapy 169
general being unworried by physical contact between patient and therapist, does provide a clear rationale for exercising extreme caution in dealing with patients who have been abused in childhood, as the next example illustrates:
Safe breathing: secure base
Sarah, of the 'ums' and 'aahs' discussed above, was increasingly distressed as her elderly mother became ill. This coincided with her therapist having to change the time of her appointments. She started to sob and shake and overbreathe during the sessions. She wrote a poem in which she longed for a pure and childlike intimacy with her therapist. She wanted him inside her, breathing him in through her lungs, rather than taking him in through her mouth or genitals which she saw as sullied and contaminated. She wanted desperately to hold his hand, but he intuitively felt that this would be wrong.
When patient and therapist looked at this together they realised that this was because, as well as being the secure-base mother she so longed for, he also represented the abusive father whom she feared and loathed. Had he held her hand this would have repeated the typical abusive vicious circle in which the child clings ever more tightly to her abuser: the abuse creates a terrible anxiety which leads to attachment behaviour, which provokes more abuse and so on. By holding his hand she would have remained an object, albeit one in need of protection, whereas her greatest need was to become the subject of her own life, even though this meant subjecting herself to intense pain and fear. In the end she soothed herself with the idea that if she could feel that she belonged for a while in his consulting room, things would be all right. Like Oliver Twist (see Chapter 3), she needed first to find a place to which she could become attached, before she could begin to own her story.
4 COGNITION IN THERAPY
We have argued in the previous chapter that Bowlby's concept of internal working models acts as a bridge between psychoanalysis, which conceives of an internal world populated with objects and their relationships, and cognitive science, which acknowledges internal models of the world
170 Imlications
in the form of mental representations. Psychoanalysis is concerned with affect-laden sensations which act as a distorting prism as we confront the world; cognitive therapy, with the perceptions and constructions which we put on those sensations and erroneous assumptions which follow from them. Psychoanalysis aims to make the unconscious conscious; cognitive therapy starts from conscious thoughts but then reveals the unexamined assumptions that underlie them. Bowlby provides a bridging language between the two approaches. He sees the neurotic patient as basing his relationship to the world on outdated assumptions; for example, that he will be ignored or let down by people, or that his feelings will be dismissed or ridiculed. While these are, in his view, fairly accurate reflections of the way the person has been treated as a child, they do not necessarily bear any relation to current reality, and can lead to poor adaptation in the form of avoidant or ambivalent relationships.
Two factors are at work in maintaining these outmoded models. The first is defensive exclusion of painful emotions which can be overcome by the kind of affective processing advocated in the previous section. The second, related, phenomenon is the need to preserve meaning and to order incoming information from the environment in some kind of schema, however inappropriate.
Liotti (1987; Bowlby 1985) sees these schemata as 'superconscious' (rather than unconscious) organising principles 'which govern the conscious processes without appearing in them', rather as computer programmes determine what appears on the VDU screen without themselves being apparent. An important part of the task of therapy, whether cognitive or psychoanalytic, is to elicit and modify these overarching mental schemata. Given that the patient is likely to become closely attached to the therapist, it is assumed that his assumptions, preconceptions and beliefs will be brought into play in relation to the therapist, and the therapist will re-present them, as they become visible, for mutual consideration. This is Bowlby's version of the phenomenon of transference.
Attachment Theory and psychotherapy 171
Always too considerate
Rose was in her fifties when she asked for help after splitting up with her second husband. She felt panicky and depressed and did not see how she could cope with being on her own. She had broken the marriage when she suddenly realised how she was compulsively deferential to her husband, and one more unreasonable request from him was the final straw.
As a child her life had changed dramatically when, at the age of seven, her father had walked out. She had been his favourite and every morning had sat on his lap while he fed her titbits. Now he had a new wife and family and she was relegated to occasional weekend visits where she slept in a cold and undecorated room, surrounded, as she saw it, by inaccessible luxury. At the same time her mother became profoundly depressed and developed an hysterical paralysis. When she recovered she had numerous boyfriends, one of whom she eventually married, and who resented Rose and her sisters' presence and insisted they went to bed at five o'clock every evening. Rose soon learned to suppress her own needs and disappointments and discovered in her teens that charm, good looks and compliance were a heady brew and she was able to attract powerful and successful men.
In her early psychotherapy sessions she announced that the last thing she wanted was any long-term commitment, merely a few sessions to 'sort her out'. She was grateful and dutifully took up any tentative suggestion from the therapist - that she might look at her dreams, or anger - with apparent enthusiasm. As the final scheduled session drew near she looked sad and tentative, but insisted that she was 'fine' and that everything was now going well. When challenged, however, she admitted that she did feel nervous about the end of therapy and really wanted to go on, but had 'assumed' that the therapist was far too busy to be bothered with her for more than a few meetings. In this example of ambivalent attachment she had reproduced with the therapist the very pattern of suppression of need, compliance and role reversal (she looking after the therapist) that characterised her relationship with her mother. She carried over into therapy the cognitive assumption 'I will only be loved if I look after others and please them'. This had served her well as an organiser of experience and a way of avoiding painful
172 Imlications
disappointment and frightening rage, but also acted as a barrier to her achieving what she really wanted and deprived her of feelings of intimacy and ease.
5 COMPANIONABLE INTERACTION
Attachment Theory sees exploratory and attachment behaviour as reciprocal behavioural systems. The securely attached infant feels safe to explore the environment; if danger threatens, exploration is abandoned in favour of proximity-seeking to an attachment figure. In adults, attachment can be differentiated from affiliation (Weiss 1982; Sheldon and West 1989). Affiliative relationships are typically with friends, best 'mates' (an interesting non-sexual use of the term) and comrades and are usually based on mutual exploration of shared interests. Attachment relationships, unlike affiliation, typically provide protection from danger, including the dangers of painful feelings. Thus, as we shall discuss further in the next chapter, Brown and Harris (1978) found that women experiencing loss who had a close confiding relationship with a spouse were protected from depression, while single mothers, even if they had close affiliative-type friendships, were not.
The relevance of this to psychotherapy lies in the likelihood that Heard and Lake's (1986) companionable interaction - synonymous with affiliation - is likely to be a feature of the psychotherapeutic relationship, although it is rarely considered as such by theorists. Freud's early 'training analyses' consisted of a few walks around the Wienerwald (Roazen 1976). A friendship bond undoubtedly does develop in some psychotherapeutic relationships. The tension between the patient's need to see the therapist as a friend, and the professional parameters of the relationship may provide useful transferential material.
Contrasting opening moves
Sarah and Peter, described earlier in the chapter, provide good examples of this point. Sarah would start each session in a bright and breezy way, referring to the weather or to current events as she entered the consulting room. The therapist instinctively did not respond in kind - in a way that would, from the point of view of affiliation, seem almost rude. It was clear from her history
Attachment Theory and psychotherapy 173
that she had always managed to avoid intimacy through group living, and by making sure she was the 'life-and-soul' in any gathering, but always keeping her real self well hidden. Her problem was with one-to-one attachments, not affiliation.
Peter similarly would start his sessions with talk about current politics or sport, but in his case the therapist was prepared to join in, in a limited way, again without this being a thought-out strategy. Eventually, when this was discussed in therapy, what emerged was his desperate need to be liked, and his fear of being an outsider, an emotional orphan whom everyone ignored.
In Sarah's case the therapist was adjusting the therapeutic space so that she could get far enough from him to look at what was going on between them; in Peter's he was encouraging him to affiliate enough for some therapeutic interaction to begin.
In most therapies there is an interplay between attachment and affiliation - which might in different terminology be seen as the interplay between transference and the working alliance. The sensitive therapist, like the good-enough parent, is always alert to the patient's need for security in the face of painful affect on the one hand, and, on the other, their wish to explore in a playful, humorous or companionable way.
The issue of affiliation is even more evident in group and family therapies. Affiliation to group members helps demoralised patients feel that they are of some value and importance, and to overcome isolation. Attachment in group therapy is to the group 'matrix' (derived from the word for mother) that holds its members securely and allows for exploration and affective processing. The family group is an affiliative as well as an attachment system, and much of the effort of systemic therapists is directed towards encouraging family members to do more things together and have more fun (while retaining their individuality and separateness). This chapter concludes, therefore, with a brief consideration of Bowlby's contribution to family therapy.
BOWLBY AND FAMILY THERAPY
In all his vast output Bowlby only published one purely clinical - as opposed to theoretical or research - paper. This was 'The study and reduction of group tensions in the family' (Bowlby 1949a). In it he describes his treatment at the Tavistock Clinic of a disturbed young adolescent boy who was destructive and difficult
174 Imlications
and failing to reach his potential at school. After two years of individual therapy Bowlby felt he had reached an impasse: there was no improvement, and the boy was becoming increasingly resistant to the therapy. In desperation he took the innovative step of arranging a joint meeting with the boy and his parents, together with a social worker. The meeting lasted two hours. The first hour consisted of a painful reiteration by the parents of their frustrations and disappointments with the boy. Bowlby countered this by suggesting that their nagging had contributed to his behaviour, but suggested that this had to be understood in the context of their own unhappy childhoods:
After 90 minutes the atmosphere changed very greatly and all three were beginning to have sympathy for the situation of the others . . . they found themselves co-operating in an honest endeavour to find new techniques for living together, each realising that there was a common need to do so and that the ways they had set about it in the past had defeated their object. This proved the turning point in the case.
(Bowlby 1949a)
One senses that here at last Bowlby was allowing himself free rein to do what he really wanted, a process which began in the 1930s when he first began to chafe at the Kleinian bit. Based on Bion's ideas about group therapy he conceptualised the processes involved in family therapy as analogous to individual therapy in which the warring parts of the personality are enabled to communicate more freely with one another and to reach compromise and accommodation. The social optimism of the period (with perhaps also a nod towards Bowlby's surgeon father) is contained within his remark that, once painful and angry feelings are openly expressed,
the recognition of the basic fact that people really do want to live happily together and that this drive is working for us gives confidence, much as a knowledge of the miraculous healing powers of the body gives confidence to the surgeon.
(Bowlby 1949a)
The paper ends with a section entitled 'Circular reactions in family and other social groups', which is thoroughly systemic in its outlook. Bowlby points out the vicious circles of neurosis in which
Attachment Theory and psychotherapy 175
'insecure parents create insecure children, who grow up to create an insecure society which in its turn creates more insecure parents', and contrasts this with the virtuous circles of health and the need for 'one great therapeutic endeavour: that of reducing tensions and of fostering understanding co-operation between groups of human beings'.
Although Bowlby did not specifically return to family therapy as a topic after this, he must be credited with having introduced the technique of seeing families together at the Tavistock Clinic, and therefore, alongside Gregory Bateson's Palo Alto group (Bateson 1973), with being the originator of family and systemic therapy which was to become such an important therapeutic mode over the ensuing decades.
Bowlby's ideas have been developed in Britain particularly by John Byng-Hall (1991c), Dorothy Heard (1982) and Robin Skynner (1976). Byng-Hall has addressed the spatial aspect of attachment, which can be illustrated by Schopenhauer's porcupine metaphor as an image for 'too near-too far' dilemmas within families:
A number of porcupines huddled together for warmth on a cold day in winter; but, as they began to prick one another with their quills, they were obliged to disperse. However the cold drove them together again, when just the same thing happened. At last, after many turns of huddling and dispersing, they discovered that they would be best off by remaining at little distance from one another.
(Quoted in Melges and Swartz 1989)
Byng-Hall (1991a), from a child psychiatry perspective, sees the symptomatic patient in a dysfunctional family behaving like the buffer zone between parental porcupines: when the parents start to drift apart the child will develop symptoms which bring them together, and if they start to get dangerously close he will insinuate himself between them, thereby alleviating the imagined dangers of intimacy. Byng-Hall (1985) sees the presuppositions and assumptions which partners bring from their 'families of origin' into their 'families of procreation' in terms of 'family scripts'; namely, patterns of interaction or 'dance' (Minuchin 1974), which an individual expects of himself and those close to him. The distinction made by Minuchin et al. (1978) between enmeshed
176 Imlications
and disengaged families (the former tending to occur in anorexia, the latter in behaviour disorders), can be equated in Attachment Theory terms with ambivalent and avoidant insecure attachment based on the parents' experiences as children and now reproduced with their own offspring.
Perhaps as a counter-balance to Bowlby and Winnicott's emphasis on mothers, Skynner (1976) highlights the role of the father in family attachment patterns. In the early stages of infancy the father's job is to protect the mother-child dyad, to allow attachment to develop and for the mother's 'primary maternal preoccupation' (Winnicott 1965) to flower. Later, he needs to intrude on the intimacy of mother and child, partly in order to make his own relationship with the child and to promote attachment to himself, but also to encourage the process of healthy separation from the mother. The child needs to be able to go off with the father, knowing that he can return to the secure base of the mother when he needs to. Without this Oedipal paternal function the mother will be more likely actively to reject the child, using threats of sending him away or even suicide, which Bowlby sees as a particularly dangerous breeding ground for insecure attachment.
The family therapy perspective shows how attachment patterns perpetuate themselves through the life cycle, event scripts being the psychological equivalent of the genome, or, in Dawkins' (1977) neologism, the 'meme'. The basic aims of psychotherapy - the need to provide a secure base, to help people express and come to terms with anger and disappointment (both of which can be seen in terms of separation protest), to achieve integration and coherence within themselves and their families - represent an attempt to intervene in this cycle, altering not so much an individual personality as a pattern of relating so that good experiences lead, by benign rather than vicious circles, to yet more good experiences, and so on. In this way a healthy social mutation will have occurred and Bowlby's vision of psychotherapy as preventive medicine will, to some degree at least, have been realised.
Chapter 9
Attachment Theory and psychiatric disorder
Many of the most intense of all human emotions arise during the formation, the maintenance, the disruption and the renewal of affectional bonds . . . in terms of subjective experience, the formation of a bond is described as falling in love, maintaining a bond as loving someone and losing a partner as grieving over someone. Similarly, threat of loss arouses anxiety and actual loss causes sorrow; whilst both situations are likely to arouse anger. Finally the unchallenged maintenance of a bond is experienced as a source of security, and the renewal of a bond as a source of joy.
(Bowlby 1979c)
Social psychiatry is concerned with the ways in which the environment influences the origin, course and outcome of psychiatric disorders. In his last, and one of his greatest papers, 'Developmental psychiatry comes of age', Bowlby (1988c) bemoans the 'kidnapping' of the label 'biological psychiatry' by those concerned with biochemical and genetic factors in mental illness. Theories of psychological development, if based on sound ethological and evolutionary principles, are no less 'biological' than is research in neurotransmitter chemistry. As the quotation above implies, a key feature of Attachment Theory is its attempt to combine the psychological and subjective with the biological and the objective. In Chapter 3 we suggested that psychotherapy could be seen as a branch of social psychiatry. The integration of psychodynamic ideas into psychiatry has always been bedevilled by the difficulty in translating the language of the inner world into the quantifiable terms of scientific psychiatry. The aim of
178 Imlications
this chapter is to explore the meeting points between social psychiatry research and the recent developments in Attachment Theory discussed in Chapter 6. Out of this encounter there is beginning to emerge the possibility of a more psychologically meaningful psychiatry, and a more scientifically based psychotherapy.
Bowlby compares the role of Attachment Theory in psychiatry with that of immunology in medicine. The comparison is apt, not just because both are concerned with the integrity and security of the individual, but also because immunology, as well as being concerned with specific disorders of the immune system, has a contribution to make to the understanding of a wide variety of medical conditions. Similarly, Attachment Theory has its 'own' disorders to which it is particularly applicable - abnormal grief, neurotic depression, agoraphobia - but can also inform many other aspects of social psychiatry.
Psychoanalytic theorising about the relationship between childhood experience and psychiatric illness - for example, Freud's linking of repressed homosexuality and paranoia - or even more recent speculations about childhood 'theories of mind' (Fonagy 1991) and borderline personality disorder have found disfavour in psychiatric circles for two main reasons. First, psychiatrists tend to use much more tightly defined categories of mental illness than do psychotherapists, for whom, for example, a term such as 'psychotic' is often used in an overinclusive and arcane way. Second, it is very difficult to specify the presence or absence of a category such as 'repressed homosexuality' in a way that lends itself to research. Bowlby's strategy for getting round these difficulties was to concentrate on external, uncontroversial events such as separations. But here too the attempt to relate adult psychological disorder to single events such as childhood separation has been found to be an oversimplification. Apart perhaps from post-traumatic stress disorders there is no one-to- one link between environmental trauma and psychiatric illness. Indeed, given the complexity of psychological development, the variety of experience, and fluidity of meanings by which experience is comprehended, it would be surprising if this were so. A more subtle, if less attractively simple, model of stress, vulnerability and buffering is required.
Attachment Theory is a theory about relationships, based on the idea that human beings evolved in kinship groups and that in
Attachment Theory and psychiatric disorder 179
the original 'environment of evolutionary adaptedness' (Bowlby 1969b) survival was increased by the maintenance of secure bonds between their members, primarily, but by no means exclusively, between parents and children. The theory, fundamental to social psychiatry, suggests that relationships and their difficulties might influence psychiatric disorder in three distinct but interrelated ways. First, the breaking or disruption of bonds is likely in itself to be a cause of disturbance. Second, the internalisation of disturbed early attachment patterns may influence subsequent relationships in a way that makes a person both more exposed and more vulnerable to stress. Third, a person's current perception of their relationships and the use they make of them may make them more or less vulnerable to breakdown in the face of adversity. We shall briefly consider each of these points, and then proceed to discuss a number of selected psychiatric disorders in the light of them.
Loss
There is strong evidence of the relationship between acute loss and increased vulnerability to psychiatric and physical disorder. Widows and widowers are more likely than non-bereaved people to die themselves from a coronary in the year following the sudden death of their partners from a heart attack. Among depressed patients 60-70 per cent have had an unpleasant loss event (usually involving the loss of or threat to an attachment relationship) in the year preceding their illness, as opposed to only 20 per cent of non-depressed controls. Schizophrenic relapse is often brought on by loss or unexpected change. People who commit suicide or attempt suicide are similarly more likely to have experienced loss than those who do not.
However, as we discussed in Chapter 3, for loss to be pathogenic it has to be in the context of other important variables. Not all those who experience bereavement succumb to depression. Those for whom the loss was sudden and untimely, who had a dependent relationship with the person they have lost, or felt ambivalent towards them, and who lack a supportive relationship and network of friends, are much more vulnerable.
A similar story appears to hold for the long-term effects of childhood loss. Early speculation suggested that childhood bereavement was an important factor in adult depression. While
180 Imlications
recent research on this point has been contradictory (Tennant 1988: Harris and Bifulco 1991), it does seem clear that the lack of good care that is so often a result of childhood bereavement is a vulnerability factor for depression, and that there are important additive effects, so that loss in adult life, in the presence of vulnerabilities in the personality, makes a person much more likely to become depressed than in their absence.
Attachment styles and vulnerability to psychiatric disorder
We presented in Chapter 6 the evidence that infant attachment patterns persist well into middle childhood, and the Adult Attachment Interview (AAI) data suggest a further continuity of these patterns into adult life. This means, in Western countries at least, that about one-third of adults are likely to have relationships which are characterised by anxious attachment, and this could constitute a major vulnerability factor for psychiatric illness when faced with stressful life events. Using postal questionnaires, Shaver and Hazan (1988; Hazan and Shaver 1987) surveyed a college freshman population and a middle-aged sample about 'romantic attachments' and found remarkable parallels with the Bowlby- Ainsworth classification of infant attachment in the Strange Situation. Of their respondents 56 per cent demonstrated a secure attachment pattern, describing themselves as finding it relatively easy to get close to others, to depend on them, and not worrying about being abandoned or about being intruded upon. Twenty- five per cent showed an avoidant pattern, with difficulty in trusting their partners, and often feeling that their partners wanted more intimacy than they felt able to provide.
The remainder (19 per cent) were anxious-ambivalent, often worrying that their partners didn't really love them, and aware that their great neediness and possessiveness often drove potential partners away.
Attachment research on children has shown correlations between attachment styles and social competence. Similar connections can be demonstrated in college students (Kobak and Sceery 1988): those classified as secure on the AAI were rated by their peers as more ego-resilient, less anxious and hostile, and as having greater social support than the anxious-dismissives and anxious-preoccupieds who were less resilient, less supported and more hostile or anxious respectively.
Lake (1985) has pointed to the discrepancy between the
Attachment Theory and psychiatric disorder 181
frequent invocation of the notion of ego-strength as a mark of mental health, and the lack of a satisfactory definition and operational criteria for its presence. For him ego-strength comprises the ability to form mutually satisfying intimate relationships, the capacity to cope with change, good self-esteem, and a sense of competence. In a similar vein, Holmes and Lindley (1989) define 'emotional autonomy' as the key to mental health and a central goal of psychotherapy:
Autonomy, in the context of psychotherapy, implies taking control of one's own life . . . emotional autonomy does not mean isolation or avoidance of dependency. On the contrary, the lonely schizoid individual who preserves his 'independence' at all costs may well be in a state of emotional heteronomy, unable to bear closeness with another person because of inner dread and confusion. A similar state of emotional heteronomy affects the psychopath who is unaware of the feelings of others. The emotionally autonomous individual does not suppress her feelings, including the need for dependence, but takes cognisance of them, ruling rather than being ruled by them.
(Holmes and Lindley 1989)
Attachment research shows how the psychotherapeutic constructs of ego strength and emotional autonomy have their origins in early familial relationships, and how in turn they affect relationships in adult life. Social psychiatry makes the links between disordered relationships and psychiatric illness, but, as we have seen in Chapter 3, these links are not as straightforward as Bowlby's original analogy between the effects of vitamin deficiency and those of maternal deprivation would imply. Epictetus' doctrine that 'men are troubled not so much by things as by their perception of things' is a reminder that environmental difficulty is mediated by a person's state of mind, and that mental set may powerfully influence how a person responds to stress.
Autobiographical competence
Loss and attachment style affect vulnerability to psychiatric disorder by way of the effect on the personality of past difficulty. But a person's current relationships - the support available from family, friends, and neighbours - seem likely also to be important as a source of buffering against the impact of stress. Henderson
182 Imlications
and his colleagues (Henderson et al. 1981) undertook a major study of the relationship between social networks and neurotic disorder in Canberra.
Inspired by Bowlby, Henderson set out to test the 'social bond hypothesis' that deficiency in social relationships, or 'anophelia', is a causal factor in the onset of neurosis. He devised the Interview Schedule for Social Interaction (ISSI) as a way of measuring the adequacy of a person's actual and perceived social support both in the past and in their current situation. Using a General Practice community sample (that is, one with relatively low morbidity), they failed to confirm their original hypothesis, finding no association between morbidity and impairment of present or past social relationships. What they did find, to their surprise, was that a person's perception of the adequacy of their relationships did, in the face of adversity, have a big impact on whether or not they succumbed to anxiety and depression. In their epidemiological study it was not possible to tease out whether this perception was an accurate reflection of their performance, whether it was a manifestation of a 'complainant attitude' on the part of the affected individual, or whether there was a self-fulfilling pattern in which people who see their relationships as inadequate evoke unsatisfactory responses from their intimates. They conclude that 'the causes of neurosis lie much more within the person than within the social environment', and suggest, rather despairingly, that the attempt to provide good relationships for potential patients is unlikely to be an effective strategy in preventive psychiatry.
Attachment Theory suggests that this pessimistic viewpoint is unwarranted. First, we have seen that secure attachment is associated not so much with the absence of childhood disruption and trauma, as with 'autobiographical competence' - that is, the ability to give a balanced account of difficulty and the capacity for emotional processing of painful events in the past. Second, the evidence is that the 'social environment' does influence neurosis, but further back in the causal chain than Henderson was able to look, via the internalisation of childhood attachment patterns. Third, if perception of inadequate relationships is the crucial issue, rather than the relationships themselves, then any psychotherapeutic technique which can alter that perception, whether directly as in cognitive therapy, or indirectly as in analytic and systemic therapies, is likely to be helpful.
Attachment Theory and psychiatric disorder 183
Armed with this optimism, let us look now at a number of different psychiatric disorders from the perspective of attachment theory.
ABNORMAL GRIEF
In his early work, Bowlby was keen to establish the reality of childhood mourning in the face of those who disputed whether children were able to experience the same full gamut of emotions as adults (Bowlby 1960d). The fact that adults do grieve is in itself evidence for the continuing importance of attachment throughout life. Parkes (1975; 1985; Parkes and Weiss 1983) has shown how the quality of the relationship broken by the death influences the course of mourning. Pathological grief can be divided into four distinct patterns. First is the unexpected grief syndrome: major losses which are unexpected or untimely, characterised by shock and disbelief and a persisting sense of the presence of the dead person. In the face of major trauma, securely attached people are as vulnerable as the less secure, and Parkes et al. (1991) found that 100 per cent of those referred with abnormal grief to his clinic whose capacity to trust themselves and others was good, had had sudden, unexpected or multiple bereavements. In delayed grief, seen typically in people with an avoidant attachment style, the patient characteristically lacks emotional response to the loss, feels numb and unable to cry, and cannot find any satisfaction in relationships or distractions. In the ambivalent grief syndrome, the previous relationship was stormy and difficult, often with many quarrels and much misery. Initially, the bereaved person may feel relief, and that they have 'earned their widowhood'. Later, however, intense pining and self-reproach may follow, with the sufferers blaming themselves in an omnipotent way for the death of their partners, based on the earlier unconscious or semi-conscious wishes that they would die. In chronic grief the sufferer becomes locked into a state of despair from which there seems no escape. These people have usually shown lifelong dependency on parents and partners. Often such dependency may mask ambivalence, and the unearthing of negative feelings can be the chink through which new life begins to appear.
184 Imlications
MRS W: I can't bear to look
Mrs W, a fifty-year-old housewife, had been in a state of chronic grief since the death of her grandmother three years previously. She was unable to carry on looking after the house or caring for her twenty-year-old daughter, herself handicapped with agoraphobia. She was tearful and apathetic, had failed to respond to antidepressants, and her husband and GP were at their wits' end. Referred for psychotherapy, she described how she had to avert her gaze on going past her grandmother's house, tried to avoid going near it although this often meant inconvenient diversions, and could not possibly visit her uncle who still lived there.
When she was a child her father had been away in the war, but on his return when she was four, her mother promptly went off with another man, and she had had no contact with her since. She was brought up by her maternal grandmother to whom she felt close, but who ruled with a rod of iron. When she was eleven, her father remarried and she was summoned to live with him and her stepmother. She was never happy with them, and she spent her teens oscillating between her grandmother and father. At eighteen she left home, made two disastrous marriages, and eventually met her present husband, twenty years her senior, who was very 'good' and 'understanding', but, she felt, was unable to understand her grief and was intolerant of her tears.
Offered brief therapy based on 'guided mourning' (Mawson et al. 1981), she brought photographs of her grandmother which, initially, she could only look at with great difficulty. Mixed with her reverence and awe towards her grandmother, a new theme began to emerge - anger at the way her mother had been 'written off' and had become a forbidden subject not to be mentioned in the grandmaternal home. With therapeutic prompting, Mrs W made enquiries about her mother, found that she had died and visited her grave. Then she happened to bump into her maternal uncle at the local supermarket and was able to talk to him for the first time since her grandmother's death. She then went to the house, at first just looking at it from the outside, later going inside. When therapy came to an end after eight sessions her depressive symptoms had lifted and she felt better 'than for years' although she remained overinvolved with her daughter.
Attachment Theory and psychiatric disorder 185
DEPRESSION
Attachment Theory has made an important contribution to current thinking about the social causes of depression. Freud's (1917) speculation about the relationship between current loss and melancholia has been repeatedly confirmed by studies showing how adverse life events can precipitate depression. His linking of depression with childhood loss has also been confirmed, although not without controversy. The balance of evidence (Brown and Harris 1978; Tennant 1988) suggests that early loss of their mother, especially if accompanied by disruption and lack of care, makes a person more vulnerable to depression when faced with adversity in adult life. Harris and Bifulco (1991) have tracked the interweaving of social and psychological variables in their Walthamstow study of a group of women who had lost their mothers in childhood. They found, as predicted, that this group of women had significantly raised rates of depression compared with non-bereaved women: one in three versus one in ten. The strand of social causation starts with early loss of mother, whether through death or separation, leading to lack of care in childhood. This is linked, in the teens of the patient-to-be, with high rates of pre-marital pregnancy. This in turn leads to poor choice of partner, so that when these women, often living in disadvantaged circumstances and therefore prone to large amounts of stress, experience loss they are more likely to have unsupportive or nonexistent partners, and so to develop depression.
Harris and Bifulco's 'Strand 2', the psychological, centres on a sense of hopelessness and lack of mastery in both the childhood and current circumstances of the depressed patient. As children their depressed patients had not only lost their mothers, but also felt utterly helpless - unable to protest or grieve or retrieve or be comforted, like Bowlby's little patient who, at the age of nine, on the day when his mother died, was told to go and play in his nursery and not to make such a fuss (Bowlby 1979c). When they were adults the feeling of helplessness persisted: when they became pregnant, they coped badly with it. Their perception of their current relationships played a big part in determining whether or not they became depressed; the more helpless
186 Imlications
they felt, the greater the chance of depression, and when they felt some degree of effectiveness they were protected from it.
Harris and Bifulco (1991) distinguish between a general sense of hopelessness and lack of mastery and what they call 'vulnerable attachment styles' - that is, difficulty in interpersonal relationships. Depression was much more likely in those who showed evidence of poor relating and especially interpersonal hostility. It seems that it is the interpersonal aspect of hopelessness (as opposed to things like managing money and housework) that matters most. We have seen that it is precisely this interpersonal dimension that is formative in insecure attachments: mothers who had difficulty in attuning to their infants and who showed unpredictable hostility were more likely to have anxiously attached children.
Brown and Harris (1978) see self-esteem as the key psychological variable in the genesis of depression. As Pedder (1982) points out, to have good self-esteem is to have internalised a two-person relationship in which one bit of the self feels good about another. This is the good internal object of psychoanalytic theory, arising out of the responsiveness of the mother - the mother who not only feeds, but recognises one as a person, is sensitive to one's feelings and moods, whom one can influence, and with whom one can, through play, create and re-create, in the 'present moment' (Hanh 1990), the spontaneity of love.
Brown's group have also suggested a relationship between the age at which the mother is lost, the circumstances of the loss, and subsequent symptom formation. The earlier and more sudden the loss, the more likely the chance of depression, and the greater the chance that the depression will be psychotic rather than neurotic in character. Pedder relates this to the Kleinian notion of the 'depressive position' (see Chapter 5). Children who have not yet developed an internal image of a whole, good mother, safe from destruction by angry attacks, will, when depressed, be more likely to despair and feel overwhelmed with depression. Older children, who do have some sense of a whole mother, or who have had at least an inkling that loss is imminent, will react to her loss with anger and attempts to retrieve her through suicidal gestures or psychosomatic illness. Pedder (1982) relates this to
several particular clinical situations that must be familiar to many psychotherapists which reflect this protesting state of affairs and make mourning for the lost person very difficult. One is when a parent absents themselves by suicide; another
Attachment Theory and psychiatric disorder 187
when a marital partner is left unwillingly by the other; or when a psychotherapist abandons a patient without due warning. In all such cases there is a special problem to internalise any good version of the departing person.
(Pedder 1982)
Bowlby (1980) suggested there were three typical patterns of vulnerable personality arising out of anxious attachment: ambivalent attachment, compulsive care-giving and detachment. The Walthamstow study confirmed the importance of the first two, but found, contrary to expectation, that detachment actually protected against depression. There are two possible explanations for this. One is that their measures were not sensitive enough to distinguish between healthy autonomy (which is a form of mastery) and compulsive detachment (which is not). The second is that detachment may be connected more with borderline personality disorder than depression, a possibility we shall consider below.
Harris and Bifulco (1991) were studying only a small sub-group of depressed patients: although people who have been bereaved in childhood appear to be more vulnerable to low self-esteem and so to depression in later life, the majority of depressives come from intact homes. Parker's Parental Bonding Instrument (Parker 1983) is an attempt, via retrospective accounts, to reconstruct the family atmosphere in patients' childhoods, searching for qualitative features of parenting which may predispose to depression. Parker isolates a particular combination of low care and overprotection which he calls 'affectionless control' that is especially corrrelated with neurotic depression: in one study it was present in nearly 70 per cent of patients but in only 30 per cent of controls. Affectionless control conjures up a childhood in which the potential patient lacks a secure parental base, and at the same time is inhibited in exploratory behaviour, thereby reducing the two ingredients of self-esteem: good internal objects and a feeling of competence and mastery.
One of the strengths of Attachment Theory is that it brings together past and present influences, the social and the psychological, providing a comprehensive picture of the varied factors which result in the development of a psychiatric disorder. Bowlby (1988c) gives a vivid picture of this epigenetic process. There is
188 Imlications
[a] chain of adverse happenings. For example, when a young woman has no caring home base she may become desperate to find a boyfriend who will care for her. That, combined with her negative self-image, makes her all too likely to settle precipitately for some totally unsuitable young man. Premature pregnancy and childbirth are then likely to follow, with all the economic and emotional difficulties entailed. Moreover, in times of trouble, the effects of her previous adverse experiences are apt to lead her to make unduly intense demands on her husband and, should he fail to meet them, to treat him badly. No wonder one in three of these marriages break up.
Gloomy though these conclusions are, we must remember that a disastrous outcome is not inevitable. The more secure an attachment a woman has experienced during her early years, we can confidently predict, the greater will be her chance of escaping the slippery slope.
(Bowlby 1988c)
AGORAPHOBIA
In Separation (1973a), Bowlby puts forward a theory of agoraphobia based on the notion of anxious attachment. He sees agoraphobia, like school phobia, as an example of separation anxiety. He quotes evidence of the increased incidence of family discord in the childhoods of agoraphobics compared with controls, and suggests three possible patterns of interaction underlying the illness: role reversal between child and parent, so that the potential agoraphobic is recruited to alleviate parental separation anxiety (this may well have happened with Mrs W's daughter in the case described above); fears in the patient that something dreadful may happen to her mother while they are separated (often encouraged by parental threats of suicide or abandonment, Bowlby believed); and fear that something dreadful might happen to herself when away from parental protection.
Central to the theory and treatment of phobic disorders is the idea that painful feelings and frightening experiences are suppressed and avoided rather than faced and mastered. In what Bowlby first described as 'the suppression of family context' (Bowlby 1973a) and later 'on knowing what you are not supposed to know and feeling what you are not supposed to feel' (Bowlby 1988a), he hypothesised that the potentially phobic adult has first been exposed to trauma - such as witnessing parental suicide attempts, or being a victim of sexual abuse - and then
Attachment Theory and psychiatric disorder 189
subjected to intense pressure to 'forget' what has happened, either by the use of overt threats, as often happens in sexual abuse, or by denial - as, for example, when a grandmother brings up her daughter's illegitimate offspring as one of her own, and the child is led to believe that her true mother is her older sister. The use of denial means that the child does not have the experience of emotional processing of painful affect, and so cannot, as described in Chapter 6, achieve the autobiographical competence that is a hallmark of secure attachment. Liotti (1991) sees in phobic disorders a dissociation between the physiological concomitants of anxiety and the 'meaning structures' that go with them. The events which might make a child anxious cannot be linked up into mental schemata which would enable that child to face and overcome them. When, as adults, such individuals experience shock or conflict, they focus merely on the symptoms of panic, and not on the events which triggered them. He advocates an exploratory form of cognitive psychotherapy which does not merely require exposure to the feared stimulus, but also encourages self-exploration so that emotions and the relationships which evoke them can begin to be linked together in a meaningful way.
Morbid jealousy and agoraphobia
David was a fifty-year-old ex-taxi-driver who developed panic attacks whenever he was separated from his wife, even for half an hour, and could not go out of the house unaccompanied. Her life was made increasingly miserable by his possessiveness, and his ceaseless questioning of her when she returned from brief excursions to visit their daughter. During David's attacks he was convinced that he would die and frequently was rushed to hospital casualty departments with suspected heart attacks. He initially described his childhood as 'all right', that he had few childhood memories, and that 'what's past is past'. Then, in the second session, when asked again about his childhood he began to cry and talked about his terrors on being left alone by his mother who was a night-club 'hostess', about never having known his father, and his misery and confusion about the different men with whom she lived. When it was gently suggested that he must have felt very jealous of these men, and that there might be some connection between this and his present attitude towards his wife, he became extremely distressed and recounted how at the age of twelve he had attacked one of these men with a knife and was
190 Imlications
taken to a remand home as a result. In subsequent sessions he began to reveal his depression much more openly, and was gradually able to tolerate being on his own for increasing periods of time.
ATTACHMENT STYLES AND EXPRESSED EMOTION IN SCHIZOPHRENIA
It has repeatedly been stressed that Bowlby's early ideas of a simple relationship between, for example, childhood bereavement and depression, maternal deprivation and psychopathy, or anxious attachment and agoraphobia, have had to be modified into much more complex causal models in which early experience, current life situation, adverse events, personality, and mental set all contribute to outcome. It is unlikely that there is a simple relationship between particular attachment patterns in infancy and specific psychiatric diagnoses in adult life.
In considering psychoses, this multifactorial approach has to be further extended to include genetic and biochemical or even infective influences. Nevertheless, social psychiatry has firmly established the importance of the environment in determining the course of schizophrenic illness (Left and Vaughn 1983). Patients living in families in which there is high 'Expressed Emotion' (EE) - especially high levels of hostility or overinvolvement - are much more likely to relapse than those who live with calmer, less hostile, less overinvolved relations. The effect of EE is not specific to schizophrenia, and also influences, for instance, the course of manic-depression, Alzheimer's disease and diabetes. The prevalence of high EE in the general population is unknown, but in families of schizophrenic patients about one- third are high in EE. It seems at least possible that there is a relationship between EE and anxious attachment, which also affects about one-third of the population. The two main patterns of high EE, hostility and overinvolvement, correspond with those found in anxious attachment; that is, avoidant and ambivalent attachment. The mothers of avoidant infants, it will be recalled, tend to show hostility and to brush their children aside when they approach, while the ambivalent mothers are inconsistent and intrusive. Both patterns can be understood in terms of boundaries. The avoidant mothers feel invaded by their children and tend to maintain a rigid boundary around themselves, and this may lead
Attachment Theory and psychiatric disorder 191
to hostility when confronted with a mentally ill, and therefore in some ways child-like, grown-up child or spouse. Conversely, ambivalent parents cannot separate themselves from their children, and, if one becomes mentally ill as an adult, the pattern will repeat itself. Such parents cannot draw a firm boundary between themselves and their offspring because of overwhelming feelings of guilt.
Too many telephone calls
Mr P felt intensely guilty when his son Richard developed a severe schizophrenic illness at the age of twenty-two. He blamed himself for being so heavy-handed during Richard's teens, and, as a psychiatric nurse, felt from his reading of Laing and others that he must be a 'schizophrenogenic father'. He tolerated in an almost saint-like way very difficult behaviour from Richard, who would come into his parents' bedroom throughout the night asking for constant reassurance that he was not going to die, on one occasion brandishing a knife. Occasionally Mr P would flip from excessive tolerance into furious outbursts at his son, and then feel even more guilty. When Richard was admitted to hospital and moved later to a hostel, Mr P felt even more guilty, especially as Richard insisted that he hated the hostel and his only wish was to return home to his parents and brothers and sisters (of whom he showed in fact considerable jealousy).
Mr P had himself been an anxious child and had found separations from his mother very difficult, running away from his boarding school where he was sent at the age of nine on several occasions. Therapeutic attempts to create a boundary between Richard and his family were made very difficult because every attempt to do so was immediately interpreted by Mr P as a criticism of his parenting, and as carrying the implication that he was a negative influence on his son. But when it emerged that Richard would phone home from his hostel with unfailing regularity just when the family were sitting down to tea, Mr P was asked to take the phone off the hook for that half hour each evening. With much misgiving and strong feelings that he was rejecting his son, he agreed, without disastrous results, and with a general lightening of the relationship between Richard and his parents. Through this small change the family seemed to have come to accept that a firm boundary can be a mark of loving attachment rather than rejection.
192 Imlications
BORDERLINE PERSONALITY DISORDER
Patients with borderline personality disorder (BPD) form an increasing proportion of specialist out-patient psychotherapy practice, and comprise a significant part of the work of in-patient psychiatry, often consuming time and worry disproportionate to their numbers. Despite debate about its validity as a distinct nosological entity (Rutter 1987), BPD is, for the psychodynamically minded, an indispensable concept. It is defined in the American Diagnostic and Statistical Manual as comprising a constellation of symptoms and behaviours which include unstable interpersonal relationships, with violent swings between idealisation and devaluation; unstable mood states; self-injurious behaviour, including deliberate self-harm and drug abuse; angry outbursts; identity disturbance with uncertainty about goals, friends, sexual orientation; and chronic feelings of emptiness and boredom. In short, there is an atmosphere of 'stable instability' (Fonagy 1991) about these patients with which most clinicians are familiar.
Empirical studies suggest that these patients have been subjected to high levels of emotional neglect and trauma in childhood, although neither is of course confined to BPD. Bryer et al. (1987) found that 86 per cent of in-patients with a diagnosis of BPD reported histories of sexual abuse, compared with 21 per cent of other psychiatric in-patients, and Herman et al. (1989) found in out-patient BPDs that 81 per cent had been subjected to sexual abuse or physical abuse or had been witness to domestic violence, as compared with 51 per cent of other out-patients. Of those who had been traumatised in this way under the age of six, the figures were 57 per cent for BPD and 13 per cent for other diagnoses.
Psychoanalysts working with these patients (for reviews, see Fonagy 1991; Bateman 1991) have emphasised the extensive use of projective identification that arises in the transference-counter- transference matrix. The therapist is, as it were, used as a receptacle for the patient's feelings and may be filled with anger, confusion, fear and disgust in a way that, for the inexperienced, is unexpected and difficult to tolerate. The patient treats therapy in a very concrete way, and may become highly dependent on the therapist, seeking comfort in fusion with a rescuing object who is, at other times, felt to be sadistic and rejecting. These latter aspects emerge
Attachment Theory and psychiatric disorder 193
especially at times of breaks, or when the therapist lets the patient down, as inevitably he will through normal human error and the pressure of counter-transference.
With an approach to these patients from the perspective of Attachment Theory two issues stand out. The first concerns the oscillations of attachment (Melges and Swartz 1989) that are so characteristic of BPD, and the related question of why they persist in relationships with their families and partners (and sometimes with their 'helpers') that are so destructive. Here we are reminded of the behaviour seen in rhesus monkeys brought up on wire mothers who, when subjected to physical trauma, cling all the more tightly to the traumatising object (Harlow 1958). According to attachment theory, a frightened child will seek out their attachment figure, and if he or she is also the traumatising one a negative spiral - trauma leading to the search for security followed by more trauma - will be set up.
A second, more subtle conceptualisation of the borderline predicament has been proposed by Fonagy (1991). He suggests that the borderline experience can be understood in terms of the lack in these patients of what he calls a 'mentalising capacity'. By this he means that they lack adequate internal representation of their own or others' states of mind, especially in relation to emotions. A similar idea is contained in Main's (1991) notion of deficits in 'metacognition', the ability to think about thinking. The work of Stern and the post-Bowlbian attachment researchers suggest that maternal responsiveness is internalised by the growing child so that he or she begins to build up an idea of a self that is responded to and understood, and, reciprocally, to be able to understand and take another's point of view. Where there are difficulties in responsiveness, the child is faced with levels of excitation and pain which cannot be soothed and shaped and contained by the parent (perhaps through their own depression or inability to mentalise). Also, to represent to oneself the idea that one's parent might want to hurt or exploit one would in itself be deeply painful. Deprived of the capacity for symbolic representation of their unhappiness, and therefore the opportunity for emotional processing or transcendence, the traumatised child resorts to projective identification in which the intolerable feelings of excitation and pain are 'evacuated' into those to whom he or she is attached. For the child this is the abusing parent who is clung to with 'frozen watchfulness'; for the adult patient it is their intimates, including
194 Imlications
the therapist. The patient is temporarily relieved of mental pain, at the price of a feeling of emptiness and boredom, to be followed, as the projections are returned or further trauma arises, by yet more episodes of intolerable discomfort leading to more projection.
These speculations are given some substance by a recent Attachment Theory-inspired study by Hobson and his colleagues (Patrick et al. 1992), in which they compared a group of twelve borderline psychotherapy patients with a similar number of depressives. They were given Parker's Parental Bonding Instrument (PBI), mentioned above, and Main's Adult Attachment Interview (AAI), described in Chapter 6. Both groups showed Parker's 'affectionless control' constellation of low parental care and overprotection, with the BPD group demonstrating this even more clearly than the depressives, a result also found by Zweig-Frank and Parris (1991). If these retrospective accounts of childhood reflect not just a person's perception of what happened but what actually took place - and there is evidence to suggest that they do (Mackinnon et al. 1991) - a picture emerges of parents who were anxious but unable to respond accurately to their children, and, from the child's perspective, of an attachment figure to whom one clings, but who does not assuage one's insecurity (Heard and Lake 1986), with resulting inhibition of exploration.
Even more interesting were the results of the AAI. It will be recalled that this is a psychodynamic snapshot of a person's attachments and reactions to loss in childhood. Based on the coherence and emotional tone of the transcript, the interview is scored not so much for actual trauma as for the way a person describes it - and so is a measure of autobiographical competence (Holmes 1992). There are four possible categories: secure; insecure-dismissive; insecure-preoccupied or - enmeshed; and a fourth category, recognised after the AAI was first developed, unresolved/disorganised/disoriented, which is judged when the subject is talking about past trauma and is rated in parallel to the other categories. Thus someone who can be quite coherent for most of their narrative can still receive an unresolved classification if their story becomes incoherent when they talk about trauma. The results showed that none of the BPD group was secure, and all were classified as enmeshed, while in the depressive group four were enmeshed, six dismissive and two secure. Only two of the depressives were unresolved/ disorganised, but nine of the BPDs were so classified.
Attachment Theory and psychiatric disorder 195
The combination of enmeshment with disorganisation in relation to trauma suggested that BPD patients were wrestling with an inability to find a way of describing overwhelming mental pain - implying exactly the sort of deficit in mental representation postulated by Fonagy and Main. By contrast, several of the depressives had also been traumatised, but the effect on their linguistic coherence was much less marked.
It is clear from these studies that no one diagnostic entity can be correlated with a particular childhood constellation. 'Affectionless control' occurs in both depressive and BPD; some accounts of BPD stress avoidance, others enmeshment as childhood precursors. But the evidence in general that insecure attachment is an important developmental precursor of psychopathology is increasingly strong. Herman et al. (1989) suggest that qualitative differences may relate to different diagnostic outcomes, with the most severe forms of childhood trauma and parental unresponsiveness being linked to multiple personality disorder, less severe forms with BPD, and yet milder types linked to neurotic depression and anxiety. This would be consistent with Pedder's (1982) suggestion along Kleinian lines that the greater the difficulty in integrating a parental good internal object, the greater the likelihood of severe pathology.
The Bowlbian perspective on BPD has several implications for treatment. The patient will lack a sense of a secure base. Extreme forms of avoidance or ambivalence are likely. The patient may resist any emotional involvement in therapy as a defence against the trauma that close relationships have entailed in the past, leaving the therapist with the uncomfortable feeling that he is inflicting therapy on an unwilling subject. Alternatively, the patient may cling to the therapy for dear life, leaving the therapist feeling stifled and guilty about the need to lead their own life.