All of us, from the cradle to the grave, are happiest when life is organised as a series of excursions, long or short, from the secure base provided by our attachment figures.
(Bowlby, 1988a)
In this and the following chapter we shall outline the main features of Attachment Theory, starting with the first of the two great themes described poetically by Bowlby as the ‘making and breaking of affectional bonds’.
Bowlby was in some ways, like Freud, a late starter. Although he had a substantial body of related work behind him, it was not until around his fiftieth year, in a series of papers published between 1958 and 1963 (Bowlby, 1958d; 1960d; 1961c), that he began to formulate the main outlines of Attachment Theory. Perhaps psychological theorising, like novel writing, unlike poetry or mathematics, requires a certain maturity; perhaps, like his hero Darwin, Bowlby's revolutionary spirit was combined with a cautiousness that meant he needed to be absolutely certain of his ground before attempting to challenge the heavens. Bowlby had always felt some unease about the scientific status of psychoanalysis: his discovery of ethology in the 1950s provided him with the scientifically secure base from which to make his conceptual advance: ‘The time is already ripe for a unification of psychoanalytic concepts with those of ethology, and to pursue the rich vein of research which this unification suggests’ (Bowlby, 1953c).
Bowlby's own account of the origins of Attachment Theory was typically modest: ‘in 1956 when this work was begun I had no conception of what I was undertaking’ (Bowlby, [1969b] 1971). This ‘undertaking’ turned out to be no less than a new paradigm, with implications for child development, psychiatry, parent–infant research, child rearing practices, and psychoanalysis. Thomas Kuhn, father of the ‘paradigm shift’, describes how: ‘the usual prelude to changes of this sort is … the awareness of anomaly … that does not fit existing ways of ordering phenomena’ (Kuhn, 1962). Bowlby's Kuhnian anomaly was that ‘secondary drive’ models, in which the parent–infant bond was seen psychoanalytically as a by-product of feeding (‘cupboard love’), or behaviourally as the result of associative propinquity, failed to account for the facts. For Bowlby the drive to relate – to hold, to cling, to play, to explore, to provide safety – was an entity in its own right, needing new theories and research. Bowlby's earlier work had shown that separated or bereaved children experienced, no less than adults, intense feelings of mental pain and anguish: yearning, misery, angry protests, despair, apathy and withdrawal. He had shown too that the long-term effects of these separations could sometimes be disastrous, leading to neurosis or delinquency in children and adolescents, and mental illness in adults. In separating parent from child, a delicate mechanism had been disrupted, a fundamental bond broken linking one human being to another. What is the nature of that bond, and how does it develop? This was the question Bowlby set out to answer.
He had at his disposal two sets of theories. The first was psychoanalysis which, as we have seen, he had embraced and struggled with for the preceding twenty years. The second was the emerging discipline of ethology (Lorenz, 1952; Tinbergen, 1951; Hinde, 1982b, 1987) whose naturalistic methodologies and evolutionary theorising he immediately saw were applicable to the mother–infant relationships which were his focus of interest. Other important influences were Kenneth Craik (1943) who, like Bowlby, was a product of the Cambridge Psychology Department and had developed the idea of psychological ‘maps’ of the world, and Ian Suttie (1935), already mentioned, whose book The Origins of Love and Hate was influential in the thirties and had contributed to Bowlby's views on social psychology.
Psychoanalysis offered two different accounts of the infant–mother bond: drive theory and Object-Relations Theory. Both of these were, in Bowlby's eyes, seriously flawed. The first, ‘classical’, drive-theory account came from Freud's early formulations. Here the bond which links mother to infant is libido, or psychical energy. The newborn infant lives in a solipsistic world of ‘primary narcissism’ and experiences a build-up of tension – the need to feed, to suck the breast as an expression of his infantile sexuality. The mother provides the vehicle for the discharge of this libido. If she, or her breast, is absent, tension arises due to undischarged libido, experienced by the infant as anxiety. The baby learns to love the mother because she feeds him, and so reduces this inner tension, which, with the advent of language, comes to be described as anxiety. Bowlby called this the ‘cupboard love’ theory of relationships.
In Inhibitions, Symptoms, and Anxiety, Freud (1926) changed this theory of anxiety from ‘dammed-up libido’ to the idea of signal anxiety. Here anxiety is felt whenever there is actual or threatened separation ‘from someone who is loved and longed for’. The basis of this love, however, remains satisfaction of physiological need:
The reason why the infant in arms wants to perceive the presence of its mother is only because it already knows by experience that she satisfies all of its needs without delay. The situation, then, which it regards as a ‘danger’ and against which it wants to be safeguarded is that of non-satisfaction, of a growing tension due to need, against which it is helpless.
(Freud, 1926)
Despite this retention of a physiological substratum to relationships, in this later formulation Freud emphasises that ‘it is the absence of the mother that is now the danger’ (my italics). This shift towards regarding anxiety as based on object-loss, rather than damned-up drive was a decisive move towards the Object-Relations viewpoint that has become the predominant psychoanalytic paradigm (Greenberg and Mitchell, 1983). For Melanie Klein, the infant is linked psychologically as well as physiologically to the mother, and her ‘breast’ (actual and/or symbolic), from birth. She sees an intimate link between the physiological processes of feeding and elimination, and the beginnings of mental and ethical structures in the mind of the infant. The satisfying, nourishing, comforting breast is the prototype of the ‘good object’; the absent, withholding, empty breast is the ‘bad object’, containing not only the actual failures and unresponsiveness of the mother, but also the infant's reactions to those failures, projected into and attributed to the ‘bad breast’.
For Bowlby, both Freud and Klein failed to take the all-important step of seeing attachment between infant and mother as a psychological bond in its own right, not a secondary phenomenon derived from feeding or infant sexuality, but sui generis:
The young child's hunger for his mother's love and presence is as great as his hunger for food…. Attachment Theory provides a language in which the phenomenology of attachment experiences is given full legitimacy. Attachment is a ‘primary motivational system’ with its own workings and interface with other motivational systems.
(Bowlby, 1973a)
Thus Bowlby based his new theory of attachment partly on the findings of ethology and partly on his theoretical critique of psychoanalysis.
As a keen naturalist, Bowlby had been particularly struck by the phenomenon described by Lorenz (1952) of following responses in some avian species. Newly hatched goslings follow their mother (or a mother-surrogate), and exhibit analogues of ‘anxiety’ (cheeping, searching) when separated from her, despite the fact that she does not directly provide them with food. Here bonding seems to be dissociated from feeding. The con-verse example is provided by Harlow's (1958) famous ‘wire mother’ studies, apprearing around the time Bowlby was publishing his first papers on Attachment Theory. In his tongue-in-cheek work entitled ‘The nature of love’, Harlow (1958) described how he separated infant rhesus monkeys from their mothers at birth and reared them with the help of surrogate ‘wire mothers’. In one series of experiments the infant monkeys were presented with a choice between a wire-only ‘mother’ to which a feeding bottle had been attached, and another ‘cuddly mother’, without a feeding bottle but covered instead with soft ‘terry-nappy’ (diaper) material. The infant monkeys showed a clear preference for the ‘furry’ mother, spending up to eighteen hours per day clinging to her (as they would with their real mothers) even though they were fed exclusively from the ‘lactating’ wire mother. Harlow used this finding to argue as forcibly against a behavioural ‘derived drive’ theory of bonding as Bowlby had against the psychoanalytic ‘secondary drive’ hypothesis. As Harlow (1958) put it, the finding: ‘is completely contrary to any interpretation of derived drive in which the mother form becomes conditioned to hunger–thirst reduction’.
Geese demonstrate bonding without feeding; rhesus monkeys on their ‘wire mothers’ show feeding without bonding. Thus, argues Bowlby, we must postulate an attachment system unrelated to feeding, and which makes sound evolutionary and developmental sense.
By thinking in terms of primary attachment and bringing the ideas of neo-Darwinism to bear on psychoanalysis, Bowlby identified what he saw as some fundamental flaws in psychoanalytic metapsychology. First, it overemphasises internal dangers at the expense of external threat. The original biological purpose of the attachment system, Bowlby argued, was protection from predators. This would have been a vital necessity in the environmental conditions in which early man evolved. Infants and small children need to stay close to their mothers at all times, and to signal separation if they are to remain safe from predation. It can be argued that this predation-protection was bent later in our evolutionary history to the need for social support, learning, and ‘mind-reading’ or mentalising of our con-specifics (Holmes, 2010). Suttie (1935) brought these two aspects together in his prescient observation of the ‘innate need for companionship which is the infant's only way of self-preservation’. Bowlby criticised psychoanalysts for their over-civilised view of man in which they discount environmental threat, and over-emphasise the projection of ‘internal’ dangers (feelings of rage and hatred, for example) onto a neutral or benign environment. In reality, even in an urban setting, external dangers are far from negligible and children who are victims of injuries in the home or from traffic accidents and sexual attacks are highly likely to be unprotected and unaccompanied.
Second, Bowlby was critical of the psychoanalytical picture of personality development in which each ‘phase’ – oral, anal, phallic and genital – succeeds one another in a linear fashion. He questioned the idea that normal development can be derived from considering pathological states, and was unhappy with the idea of ‘regression’ to fixation points as an adequate way of thinking about psychological illness vis-à-vis normality. He contrasted Freud's ‘homuncular’ model in which each stage is predetermined according to some pre-existing plan of development, with an ‘epigenetic’ model (Waddington, 1977) in which several lines of development are possible, the outcome of which depends on an interaction between the organism and its environment. Thus, although the developing child has a propensity to form attachments, the nature of those attachments and their dynamics will depend on the parental environment to which he or she is exposed.
Also, the development of the attachment dynamic can be considered as a process in its own right independent of other motivational systems – for example, sex or feeding – just as the different organs of the body develop relatively independently of one another. Attachment can be thought of as an ‘organisational perspective’ (Sroufe and Waters, 1977), whose level of analysis is not that of specific behaviours, or postulated intra-psychic forces, but patterns of interaction between individuals, and between those dyads and the surrounding interpersonal milieu.
Bowlby also rejected the teleological ‘Lamarckian’ view in which the ‘purpose’ of psychological functions can be determined by an a priori goal: the ‘purpose’ of attachment is not the reduction of physiological need, but to increase the evolutionary fitness in those possessed of it, protecting them from predators and enhancing their interpersonal skills.
Finally, Bowlby was critical of ‘hydraulic’ models of drive-discharge, seeing human behaviour in terms of control theory whose aim is the maintenance of homeostasis, that is, physiological and psychological stability. Infant monkeys separated from their mothers respond with a rise in pulse rate and a fall in body temperature (Suomi, 1997); a rat pup's physiology is intimately entrained with that of their mother's (Hofer, 2006). In humans, Brazelton and Cramer (1991) showed that mothers returning to work within a year after giving birth show higher levels of physiological disturbance than those who are able to stay with their babies, and that there is a correspondingly higher incidence of infection in the infants. Secure attachment provides an external ring of psycho-physiological protection which maintains the child's metabolism in a stable state, comparable to the internal homeostatic mechanisms of blood-pressure and temperature control.
There was one group of analysts to whom Bowlby felt a strong affinity. This was the ‘Hungarian School’, especially Ferenczi (1955) and Michael Balint (1964). Ferenczi, originator of the famous phrase ‘it is the physician's love which cures the patient’, had fallen out with Freud over his emphasis on Freud's insistence on the ‘real’ (as opposed to transferential) nature of the relationship between patient and therapist, and his rather dubious propensity to kiss and hug his patients when he felt it necessary. Balint, his pupil, together with his first wife Alice, had postulated a ‘primary love’ and a primitive clinging instinct between mother and child independent of feeding. Bowlby also saw links between his ideas and those of Fairbairn (1952) who, like Bowlby, had jettisoned drive-theory in favour of primary object-seeking, and who refused to see adult dependency as a relic of orality, but rather conceived of development as a movement from infantile to mature dependence. So-called ‘relational psychoanalysis’ (Mitchell, 1993) had yet to be born, but was and still is sympathetic to, and influenced by, Attachment Theory (Safran, 2012b).
As we have seen, the reaction of the analytic world to Bowlby's ideas was, on the whole, unfavourable. The Kleinians (Grosskurth, 1986) saw him as having betrayed analytic principles, contaminating psychoanalysis with behaviourism, trying to expunge the heart of psychoanalysis – its account of the inner world of phantasy. Anna Freud (1960) and her supporters could hardly fail to notice that the Oedipus complex and infantile sexuality – for them, the cornerstones of the psychoanalytic edifice – played virtually no part in Bowlby's writings. What started out as Bowlby's attempt to modernise psychoanalytic metapsychology, and to find a sound biological underpinning for Object-Relations Theory, began increasingly to look like a new psychological paradigm. As we shall see, recent developments in ‘post-Bowlbian’ research have opened out the possibility of reconciliation; but first we must focus in more detail on the nature of Attachment Theory in its original formulations.
Attachment Theory is in essence a spatial theory, both literally and metaphorically: when I am close to my loved one I feel good, when I am far away I am anxious, sad or lonely. The child away from home for the night plays happily until she hurts herself or bedtime approaches and then feels pangs of homesickness. The mother who leaves her child with a new baby-minder thinks endlessly about her baby and misses her dreadfully. Attachment is mediated by looking, hearing and holding: the sight of my loved one lifts my soul, the sound of her approach awakes pleasant anticipation. To be held and to feel her skin against mine makes me feel warm, safe and comforted, with perhaps a tingling anticipation of shared pleasure; but the consummation of attachment is not primarily orgasmic – rather, it is, via the achievement of proximity, a relaxed state in which one can begin to ‘get on with things’, pursue one's projects, and explore the world, ‘out there’, and the inner world of feelings.
The umbrella term ‘attachment’ is used in a number of ways. It is useful to distinguish between the interrelated concepts of attachment style, attachment behaviours, and the ‘internal working models’ of attachment.
‘Attachment style’ refers to the state and quality of an individual's attachments. As we shall see, these can be classified as secure or insecure, the latter being further sub-divided into avoidant, ambivalent and disorganised. The word ‘attachment’ carries both experiential and theoretical overtones. To feel attached is to feel safe and secure. By contrast, an insecurely attached person (to take the example of the ‘ambivalent’ or ‘resistant’ pat-tern) has a mixture of feelings towards their attachment figure: on the one hand, intense love, need and dependency, on the other, fear of rejection, irritability and/or vigilance. A lack of felt security arouses a simultaneous wish to be close and the angry determination to punish the attachment figure for the minutest sign of abandonment. It is as though such people are saying to themselves: ‘cling as hard as you can to your attachment figures – they are likely to abandon you; if you hang on to them and hurt them if they show signs of going away, then they may be less likely to do so’.
Attachment behavior is defined simply as being ‘any form of behaviour that results in a person attaining or retaining proximity to some other differentiated and preferred individual’ (Hinde, 1982a). Attachment behaviour is triggered by separation, or threatened separation from an attachment figure. It is terminated or assuaged by proximity, which, depending on the nature of the threat, may vary from merely being in sight, to physical closeness and soothing words without touching, to being tightly held and cuddled.
Both attachment style and attachment behaviour are based on an ‘internal working model of attachment’, implying a blueprint of the world in which the self and significant others, and their interrelationships, are represented, and which encodes the particular pattern of attachment peculiar to that individual. The ambivalently attached person described above might have a working model of others as desirable but unreachable, of themselves as unworthy of support and love, and/or of an unreliable and rejecting attachment figure triggering a protesting, attacking self.
Weiss (1982) usefully defines attachment in terms of three key aspects.
As parents of toddlers well know, small children have a maddening propensity to follow their attachment figures wherever they go. The dis-tance at which the child feels comfortable depends on such factors as age, temperament, developmental history, and whether the child feels fatigued, frightened or ill. The three latter features all enhance attachment behaviour. In addition, recent separations lead to greater proximity seeking, or ‘mummyishness’, as Robertson's (1952) film so beautifully demonstrates. The extent of the proximity required will also depend on circumstances. A three year-old being picked up from playgroup after her first day may rush up to the parent and bury her head in his lap and want to be held and cuddled for a long time. A month later she may be content to slip her hand quietly into that of her collecting parent and continue chatting to her friends as she walks down the road. Central to Attachment Theory is the notion that attachment is to a discriminated figure (or small group of figures). Bowlby originally explained this by analogy with the phenomenon of imprinting, in which in prosocial species, young birds will follow any mobile figure to which they are exposed at the ‘sensitive period’ in their development. Studies on primates suggest that imprinting does not occur in the same way as in birds, and that attachment, rather than being based on a brief temporal ‘window’ in the first few days of life, results from a gradual process of genetically programmed development and social learning (Rutter, 1981; Bretherton, 1991b).
The fact that attachment is, in Bowlby's word, ‘monotropic’ – that is, occurs with a single figure, most usually the mother – has profound implications for psychological development and psycho pathology throughout the life cycle.
It is because of this marked tendency to monotropy that we are capable of deep feelings, for to have a deep attachment to a person (or a place or a thing) is to have taken them as the terminating object of our instinctual responses.
(Bowlby, 1988a)
However, monotropy is by no means absolute: a small child's attachments can best be thought of as a hierarchy, usually, but not invariably, with the mother at the apex, closely followed by the father, grandparents, older siblings, aunts, god-parents, child-minders and so on. Inanimate objects such as transitional objects (Winnicott, 1971) can also be important as comforters and alleviators of attachment distress.
Attachment Theory accepts the customary primacy of the mother as the main care-giver, but there is nothing in the theory to suggest that fathers are not equally likely to become principal attachment figures if they happen to provide most of the child care (Bretherton, 2010).
Three-person psychology enters into Attachment Theory via separation and loss. Growing children have to learn, painfully, that the figure to whom they are attached is not their exclusive ‘possession’ and will at times be unavailable. Mother must also be shared with her sexual partner and other siblings, which forms the basis for the Oedipal situation, making separation and loss an inherent part of the attachment dynamic, over and above quotidian separations (when the parent is in the toilet, asleep, at work, ill and so on) or traumatic ones (when the parent is dead or abandoning).
For Melanie Klein (1986), the ‘depressive position’ represents the realisation that the loved and gratifying breast/mother and the hated and rejecting breast/mother are one and the same. For Winnicott (1971) the ‘stage of concern’, whicih may apply as much to siblings as to parents, represents the child's guilty realisation that the person they love, but at times hate, can be harmed or damaged, and needs protection. For Bowlby, the human dilemma turns on the central importance of an attachment that can never be entirely reliable, must perforce be shared, and will be lost eventually (and sometimes prematurely, before the child is ‘ready’). The capacity to separate from attachment figures and to form new attach-ments represents the developmental challenge of adolescence and young adulthood. The cycle repeats itself as these children-now-parents attach themselves to their own children only gradually to let them go as they reach adolescence and early adulthood. Finally, as the death of one's loved ones, and one's own death approaches, the ‘monotropic’ bond to life itself has gradually to be relinquished. Thus the invisible temporality of the life-cycle is inherent in the attachment dynamic.
Mary Ainsworth (1982), co-founder of Attachment Theory, coined the phrase ‘secure base’ to describe the role of the attachment-provider. The essence of a secure base (an abstract notion, but always ultimately derived from a relationship with a person) is the sense of safety that is a springboard for curiosity and exploration. When danger threatens we cling to our attachment figures. Once danger passes, their presence enables us to work, relax and play – but only if we are sure that the attach-ment figures will be there if we need them again. We can endure rough seas if we are sure of a safe haven.
Anderson (1972) made a naturalistic study of mothers and their toddlers in a London park. The mothers sat on the park benches, reading or chatting while their children toddled and played on the surrounding grass. He found that each child had an invisible radius – a Maginot line – beyond which it would not venture. When the toddler neared the limit it would begin to look anxiously towards the mother. Attachment exerted an invisible but powerful pull on the child, just as heavenly bodies are held together by gravitational forces, but unlike gravity, attachment makes its presence known by a negative inverse square law: the further the attached person is from their secure base, the greater the pull of attachment. The ‘elastic band’, or ‘hidden regulator’ (Hofer, 2006), which constitutes the attachment bond is slack and imperceptible in the presence of a secure base. If the secure base becomes unreliable or the limits of exploration are reached, the bond tugs at the heart-strings.
The example of the mother who leaves her child with the child minder and then worries about and misses her dreadfully suggests that there is a reciprocal care-giving dynamic that parallels care-seeking (Solomon and George, 1996). In addition, Heard and Lake (1986) extend the secure base concept in their model of an adult attachment dynamic, in which they postulate a fundamental need for ‘companionable interaction’ based on ‘preferred relationships in the attachment network’. Waters (2008) likewise argues that the security-providing aspects of attachment have been over-emphasised in comparison with the need for supportive companionship. Where no such secure base exists – externally in the case of small children, internally in the case of adults – defensive, ‘secondary attachment’ (see Mikulincer and Shaver, 2007) manoeuvres come into play: dissociation of fear and anger in order to minimise the pain of separation anxiety, or exaggerated helplessness coercively to manipulate support, at the expense of truly reciprocal companionship.
Jennifer, a successful painter, was forty when she entered psychotherapy. Her complaint was that she found it impossible be her ‘real self’ in close relationships. In social situations she could be jolly and cheerful and was well liked; by herself she often felt depressed and anxious, but could cope, especially when she was painting. In her marriages (she had had two) she never felt at ease, unable to share feelings openly or to feel relaxed with her husbands. She had rather desperately sought some affirmation of herself through affairs, but in the end these left her feeling empty and valueless. Naturally enough these patterns were repeated transferentially in therapy and she bent her best efforts towards trying to please, seduce and sometimes to exclude her therapist, the latter via ‘projective identification’ a desperate attempt perhaps to communicate to him how lonely she felt inside.
She dated the shattering of her secure base to an incident where her much-feared father (who had been away at the war for the first three years of her life) was playing with her older brother and sister when Jennifer was about four. She tried to gain his attention but was ignored; she pinched his leg harder and harder until suddenly and terrifyingly he threw her across the room. From that day (similar episodes were repeated in various ways throughout her childhood) she could only get attention, playfulness, support from others by means of pleasing them, controlling them, or vicariously caring for herself through her care for them (this characterised her relationship with her mother, herself chronically depressed).
Jennifer's struggles illustrate in an extreme form a typical family pattern of absent-father/depressed mother that so often underlies the lack of a secure base, and leads to defensive postures by the children who grow up in such an atmosphere. Negative affect and its acceptance, the hallmark of attachment security, were met with violent repudiation (her father) or an overwhelming sense of guilt (mother). Progress in therapy only began when this woman had tested her therapist again and again for his reliability and had, inevitably, found him wanting, but still felt safe enough to reveal the extent of her disappointment and rage towards him.
Try to prise a limpet away from its rock and it will cling all the harder. The best test of the presence of an attachment bond is to observe the response to separation. Bowlby identified protest as the primary response produced in children by separation from their parents. Crying, screaming, shouting, biting, kicking – this ‘bad’ behaviour is the normal response to the threat to an attachment bond, and presumably has the function of trying to restore it, and, by ‘punishing’ the care-giver, of preventing further separation. Observing, theorising, and classifying separation protest were the starting point for Ainsworth's when she devised the ‘Strange Situation’, the basic tool used for classifying the quality of attachment in children (Ainsworth, 1979). Clinically, analysis of patient responses to weekend and holiday ‘breaks’ are a basic theme in analytic psychotherapy.
A remarkable feature of attachment bonds is their durability. The persistence of attachment in the face of maltreatment and severe punishment has enormous implications for child and adult psychopathology. Harlow's monkeys clung ever more tightly to their cloth ‘mothers’ even when ‘punished’ by them with sudden blasts of compressed air (Rutter, 1980). It is hard to explain this phenomenon on the basis either of the intra-psychic psychoanalytic ‘cupboard love’ theory, or of stimulus-response reward-reinforcement learning theory. A cybernetic, organisational, systemic, or dyadic level of analysis is required. An ethological approach suggests that stress can lead to an unassuaged enhancement of attachment behaviours when the source of that stress is the attachment figure itself (see Hesse and Main, 2000). The battered wife who continues to return to her abusive husband, or the ‘frozen watchfulness’ of the physically abused child are both eloquent proof of the phenomenon Disorganised Attachment, its inhibition of normal exploration and playfulness, and the overwhelming need for a secure base, however maladaptive such clinging may be.
The human infant is born in a state of great immaturity (a consequence, evolutionary biologists suggest, of the need to get the huge human brain through the pelvic floor before it is too late!). It is not surprising therefore that, unlike in ducks, monkeys and other animals, the human attachment system takes several months to develop. Only after six months does the baby begin to exhibit the full attachment triad of proximity seeking, secure base effect and separation protest that we have described. The ontogeny of the attachment system can be conveniently divided into four phases.
Babies recognise and turn towards the smell of their mother's milk within 48 hours of birth. Newborn babies are highly responsive to human contact, and imitate facial movements, such as tongue protrusion, within a week or two of birth (Meltzoff and Moore, 1989). Central to this process is the sight of the human face, which evokes intense interest. The onset of the smiling response around four weeks marks the beginning of the cycles of benign interaction that characterise the relationship between babies and caregivers. The child's smile evokes a mirroring response in the mother; the more she smiles back the more the baby responds, and so on. Gergely (Fonagy et al., 2004) suggests that this contingent and marked mirroring (that is, responsive and slightly exaggerated) feeds back to the child a picture of their internal affective stage and forms the nucleus of internal models of one's emotions. Maternal sensitivity and appropriate responsiveness is a key determinant of the quality of attach-ment as development proceeds (Ainsworth, 1973).
This model is consistent with Winnicott (1971): ‘What does the baby see when he or she looks at the mother's face? I am suggesting that ordinarily what the baby sees is him or herself.’ He goes on to suggest that what happens in psychotherapy is ‘a long term giving the patient back what the patient brings. It is a complex derivative of the face that reflects what is there to be seen’ (Winnicott, 1971).
Daniel Stern (1985), from a perspective of developmental psychology, and Kenneth Wright (1991; 2010), from a psychoanalytical viewpoint, both see mutual gazing between mother and baby as a key element in the development of an internal world in which attachment can be represented and regulated. The invariability of the mother's face, the recognition of it as a pattern, give the baby a primitive sense of history, of continuity through time that is integral to the sense of self. To evoke her smile also provides a sense of agency and effectiveness. Her mirroring response is the first link between what is perceived out there, and what is felt in here.
For Wright the mother's face is the first symbol; her face is not part of the self and yet, because it is responsive, feels intimately connected to the self (Holmes, 2012c). In the Kleinian account of the origin of symbol formation – based on Freud's idea of hallucinatory wish-fulfilment – images are thought to arise as a consequence of loss or absence: ‘no breast; so imagine a breast’, thinks the Kleinian infant. Wright proposes a more harmonious theory in which the separation is simply spatial: the face is over there, held off and so is available for thinking about, contemplation, meditation. To watch a 3-month-old baby at the breast is to get visible proof of the rhythm of feeding and mutual gazing that constitutes the mother-child relationship at this stage. Freud, in his discussion of Leonardo (Freud, 1910), conceptualises looking as visual incorporation – a drinking-in with the eyes – rather than a modality of relating with its own dynamic. The complexity and specificity of the visual world, as opposed to the gustatory world, is what makes vision so important to attachment: ‘Wine comes in at the mouth, love comes in at the eyes.’ The world is mapped through the visual system: the mother's face is imaged on the retina and visual cortex before it is imagined in the inner world. We shall consider later some of the implications of the failure of this mirroring process.
As with looking, so with ‘holding’ and ‘containment’, terms used by Winnicott (1971) and Bion (1978), respectively. In his phrase ‘the holding environment’ Winnicott denotes not just physical holding, but the entire psycho-physiological system of protection, support, caring and containing that envelops the child, without which it would not survive physically or emotionally (see Polan and Hofer, 2008). The reliability and responsiveness of the holding environment form a springboard for emergent attachment patterns as the child begins the process of separation-individuation.
In the second half of the first six months the beginnings of an attachment relationship start to be evident. The baby becomes more discriminating in his looking. He listens out for and responds to his mother's voice; cries when she departs; greets her differently from all others; begins to put his arms up towards her in a request to be picked up. She in turn responds to the physiological and social cues from her baby in a way that leads to the establishment of a regulatory or homeostatic system. An interactive matrix is established, felt as a mutual ‘knowing’ of each other that is the hallmark of a secure mother-infant relationship.
In the second half of the first year, several developmental changes occur which mark the onset of attachment proper. Children removed from foster homes into permanent adoptive homes before 6 months show little distress, whereas after that watershed they show increased initial crying, clinging, apathy, and feeding and sleep disturbance (Bretherton, 1985). Around 7 months the baby will begin to show ‘stranger anxiety’, becoming silent and clingy in the presence of an unknown person (Spitz, 1950).
These changes coincide with the onset of locomotion in the child, which entails a much more complex system of communication if the baby is to remain in secure contact with the mother. The immobile baby is bound to remain where he is. The mother of the mobile baby needs to know that the child will move towards her at times of danger, and the child needs to be able to signal protest or distress when necessary to a mother who now feels she can put him down for a few minutes.
Bowlby conceives the attachment system at this stage as being based on ‘set-goals’, which he compares to the setting on a thermostat, maintained by a system of feedback control. The ‘set-goal’ for the infant is to keep ‘close enough’ to the mother: to use her as a secure base for exploration when environmental threat is at a minimum, and to exhibit separation protest or danger signalling when the need arises.
Figure 3.1 attempts to summarise the features of the attachment system at this stage. Several points need to be made about this diagram. First, attachment behaviour, although usually discussed from the point of view of the attached person, is a reciprocal relationship. The parent is simultaneously offering complementary care-giving behaviour that matches, or should match, the attachment behaviour of the child. For example, when put in a new situation the child will, through social referencing, make eye contact with the mother, looking for cues which will sanction exploration or withdrawal. Second, and as a consequence, parent-child attachment systems can be seen in terms of continuously monitored distance-regulation (Byng-Hall, 1980), with many opportunities for problematic variants. Over-anxious parents may inhibit children's exploratory behaviour, making them feel stifled or smothered; conversely, the neglectful parent may inhibit exploration by failing to provide a secure base, leading to feelings of anxiety or abandonment.
Third, inherent in the model is the notion of an internal map or ‘internal working model’ which represents the relative whereabouts of the self and attachment figure. To the analytically minded psychotherapist this may seem like a rather uninteresting, predominantly ‘cognitive’ map, but this would be mistaken. What is stored in the ‘internal working model’ is not so much an ordnance survey picture but an affect-regulation model (see Schore and Schore, 2008) which, if it could be translated into words, might be along the lines of ‘I feel tense when my mummy goes out of the room so I must keep a good eye out for her and scream if necessary’, or ‘when my mummy comes so close to me while I am playing I feel uncomfortable, so I'll try to move away a bit, without discouraging her so much that she loses interest’ (see, Beebe and Lachmann, 1988).
We have moved from set-goals which keep toddler and parent in eyesight and earshot of each other, to the idea of a relationship, and to a consideration of what internal processes might regulate it. This brings us to the attachment system in its fully-fledged form which, Bowlby maintains, is established by the third birthday and persists from then on throughout life.
As Bowlby first conceived it, the attachment system in the toddler was something like a ‘homing device’ in which the child was programmed to focus on the parent with the ‘set-goal’ of maintaining proximity. With the advent of language and the expanding psychological sophistication of the three- to four-year-old a much more complex pattern arises that cannot be described in simplistic behavioural terms. The child now can begin to think of his parents as separate people with their own goals and plans, and to devise ways of influencing them. If the mother is going to leave the child for the evening he may plead, bribe, charm or sulk in an attempt to maintain attachment, rather than crying or clinging as he would have done a year or two earlier. Attachment Theory at this point merges into a general theory about relationships (or ‘affectional bonds’, as Bowlby liked to call them) and how they are maintained, monitored, and may become problematic.
A key Bowlbian concept here is that of the ‘internal working model’. This was Bowlby's way of describing the internal world of the psychoanalysts, but couched in characteristically practical terms. The idea of an internal ‘model’ of the world derives from Kenneth Craik's (1943) influential The Nature of Explanation, in which he argues that:
Thought models, or parallels reality … the organism carries a ‘small-scale model’ of external reality and its own possible actions within its head which enable it to react in a fuller, safer, and more competent way to the emergencies which face it.
Wright (1991) remarks how, pre-Winnicott, the work ethic dominated the language of psychoanalysis: working through, getting the patient to work on their problems, forming a working alliance, and Bowlby's internal working models. Wright sees Winnicott as representing the female, maternal influence, a reaction against the paternal force of Freud. As suggested, Bowlby in turn was in part reacting against the powerful women who had trained him, his analyst Joan Riviere, and supervisor, Melanie Klein. The idea of a ‘working model’ implies a practical mechanism, a down-to-earth title which he claimed ‘allows for greater precision of description and provides a framework that lends itself more readily to the planning and execution of empirical research’ (Bowlby, 1981c).
Although derived from the psychoanalytic perspective, the idea of internal working models is perhaps closer to that of cognitive therapy (Beck et al., 1979), itself also a development of and a reaction against the psychoanalytic paradigm. The developing child builds up a set of models of the self and others, based on repeated patterns of interactive experience. These can be described as ‘basic assumptions’ (Beck et al., 1979); ‘representations of interactions that have been generalised’ (RIGS, Stern, 1985); or ‘self-other schemata’ (Horowitz, 1988). Whatever nomenclature is used, it is clear that the developing child forms relatively fixed representational models, used to predict and relate to the world of intimate relationships in which he finds himself.
A securely attached child will store an internal working model of a responsive, loving, reliable care-giver, and of a self that is worthy of love and attention, and will bring these assumptions to bear on all other relationships. Conversely, an insecurely attached child may view the world as a dangerous place in which other people are to be treated with great caution, and see himself as ineffective and unworthy of love. These assumptions are relatively stable and enduring: those built up in the early years of life are particularly persistent and are relatively impervious to subsequent experience, however disconfirmatory. Thus an avoidant person who encounters love and support may discount this as mere manipulation or seductive self-servingness on the part of the other.
Bowlby wanted to recast psychoanalytic theory in terms of a systems approach, in which feedback loops are a key element. They underlie the ‘epigenetic’ stability of psychological phenomena. These are the benign circles of healthy development, and the vicious circles of psychopathol-ogy in which negative assumptions about the self and others become self-fulfilling prophecies.
Attachment is classified into one of two main types: secure and insecure. The latter is then further subdivided into three variants: avoidant, ambivalent and disorganised. Bowlby used the notion of internal working models to account for these different patterns of attachment which Mary Ainsworth's research had established. Secure attachment manifests itself in a ‘primary’ attachment pattern typified by healthy protest on separation followed by relaxed, often collaborative, exploration. The basic problem for the anxiously attached child is how to maintain contact with an unpredictable or rejecting care-giver. Here ‘secondary’ attachment strategies are needed in order to achieve a modicum of security, even though it means accommodating to the vagaries of the care-giver. The two basic strategies are those of avoidance or adherence, leading to classifications of avoidant or ambivalent attachment (see Figure 3.2).
In avoidant attachment the child tries to minimise his needs for attachment in order to forestall rebuff, while at the same time remaining in sufficient contact with the care-giver for a modicum of safety. The latter's rejection, like the care-seeker's own neediness, is often non-conscious. Based on Dixon's (1971) concept of perceptual defence, Bowlby calls this ‘defensive exclusion’, the elimination from consciousness of painful affect such as fear or anxiety for the sake of a greater good, that is, a degree of security. The ambivalent strategy, by contrast, involves clinging to the care-giver, often with excessive submissiveness.
Avoidance and ambivalence are described as ‘organised’ forms of insecure attachment, in that a functioning relationship with the care giver is maintained. The third sub-type of insecure attachment, ‘insecure disorganised’, is less common than the first two but associated with much more severe pathology. Here, when faced with a ‘frightened or frightening’ care-giver the child relies on ‘autistic’ self-soothing strategies when threatened such as dissociation or self-harm. Disorganised (‘D’) children, as they grow older, may opt for ‘role reversal’ in which the care-giver is cared for rather than vice versa, thereby at least vicariously providing a degree of contact and connection. In all three cases, feelings of anger at rejection and fear of abandonment are conspicuously subject to defensive exclusion.
Although these strategies have the function of maintaining attachment in the face of difficulties, a price has to be paid. The attachment patterns so established are clearly restricted and, if repeated in all relationships, will be maladaptive in that they inhibit exploration and unguarded emotional expression. Defensive exclusion also means that models cannot be updated in the light of new experience.
Thus the drawback of defensive exclusion is that it deprives the child of the opportunity for emotional processing of painful affect. This is particularly evident in pathological mourning, which leads to the persistence of primitive feelings of hate and abandonment and restricts emotional growth and development.
This consideration of internal working models has been a necessary diversion in our discussion of attachment across the life cycle. It is through internal working models that childhood patterns of attachment are carried through into adult life (see Mikulincer and Shaver, 2007) and transmitted to the next generation.
As children grow older, and especially in adolescence, they are able to tolerate increasing periods of separation from their parents. Does this not mean that the attachment ‘phase’ has been outgrown, to be superseded by, say, ‘adult genitality’? According to Bowlby's ‘epigenetic’ model, emphatically not! Attachment and dependency, although no longer evident in the same way as in young children, remain active throughout the life cycle. From an attachment perspective, dependency is not outgrown, but transformed from infantile and childhood patterns to mature dependency, and remains a ‘hidden regulator’ (Hofer, 2006), available at times of stress and distress. For adolescents the parental home still remains an important anchor point, and the attachment system will become reactivated at times of threat, illness or fatigue (Allen et al., 1996). The turbulence of adolescence can be seen in Bowlbian terms as springing from the complexity of detachment and re-attachment which the adolescent must accomplish. To disengage from parental attachments, to mourn that loss, to move on, via the transitional phase of peer group attachment, to the pair-bonding of adult life can be no easy task.
Bowlby saw marriage, and ‘romantic attachments’, as the adult manifestation of attachment whose companionship provides a secure base allowing for work and exploration, and a protective shell in times of need. Like Fairbairn (1952), but unlike Freud (1929) for whom affection was conceptualised as sublimated or ‘aim-inhibited sexuality’, Bowlby saw bodily pleasure as a ‘signpost to the object’. In the attachment model, sexuality and attachment are separate, albeit closely related, psychobiological systems (see Diamond et al., 2007). In addition, operation of the attachment system via internal working models, often in non-conscious ways, plays an important part in the choice of marital partner and relationship patterns in marriage.
The Christian marriage rubric of ‘in sickness and in health’ is a reminder that a central aspect of marriage is to provide a secure base and an attachment system available in times of need. As we shall see, current conceptualisations of the role of attachment go far beyond Bowl-by's original conception of it as a security-providing system. Schore and Schore (2008) see affect regulation and companionable interaction as vital further aspects of what the attachment bond provides. To the extent that attachment relationships help regulate affect, marriage (or intimate relationships generally) are a primary means by which feelings, both joyful and pain-laden, are articulated, worked through, and resolved.
A young woman developed multiple phobic symptoms soon after the birth of her first baby. At first her fears were of harming the baby; later she became severely agoraphobic, and took to phoning her mother several times a day for reassurance. She insisted on moving house so as to be within easy reach of her mother. Her mother ‘helped’ by looking after the baby for much of the day, and would herself telephone frequently to check if the baby was ‘all right’. When the patient told her mother of a dream in which her son had fallen under a lorry, the mother (who was unlikely to have read Freud) told her that this meant that she wanted to kill her son! As a child the patient had lacked a secure base with this mother, whom she felt neglected her in favour of two younger sisters, one of whom had been chronically ill with kidney disease, while the other was epileptic, and to whom she had devoted all her attention.
In an initial phase of individual therapy she was able to link her fears of harming the baby with aggressive feelings towards her younger sisters and her angry dependency on her mother, but her symptoms persisted. Marital therapy was then offered. At the first session she proudly announced that her husband – unlike herself – was afraid of nothing. He accepted the compliment rather diffidently, but confirmed that he had been more or less self-reliant since the age of ten, when his parents had divorced and he and his younger brother had been left to fend for themselves on the rough estate where they lived. He worked as a scaffolder on high buildings. When asked if it was true, as his wife believed, that he was frightened of nothing, he confessed that he had slipped on a plank that morning and had been very scared, and that since the birth of the baby he had been much less of a daredevil. His wife seemed surprised at this revelation, but visibly relaxed and perked up. He then admitted that he saw it as his task to conceal his fears and worries from his wife because of her ‘illness’. For example, he resented his mother-in-law's intrusions into their family life, but was petrified by the idea of confronting her. Given the task of answering the phone when she rang, and explaining that his wife was too busy to speak, he became quite shaky and said that he would much rather be asked to go up a chimney-stack in a high wind! The patient was asked to rehearse him in this by role-playing her mother, and the session ended in laughter, with a much less anxious patient and subsequent good clinical improvement. Here then the wife had chosen a husband who, in his avoidance, superficially at least counterpoised her own ambivalent attachment insecurity, but his apparent insouciance served only to reinforce her feelings of fear and inadequacy. It was only when his fearfulness came into view and could be mutually ‘regulated’ that a more equal and secure marital balance began to be established.
This example shows how attachment patterns are represented as internal working models. The patient saw herself as uncared for, unworthy of care and therefore unable to care for her baby, whom she perceived, in a sense correctly, as in danger of neglect or attack. Lacking a secure base inside herself she was unable to provide one for others, and her anger and frustration about this lack of care interfered with her role as a mother. The intrusiveness of her own mother and the detachment of her husband (due in turn to their own insecure attachment patterns) underlined her sense of an unreliable secure base. Her demandingness and dependency represented desperate efforts to generate an ideally safe attachment, and her protest about the lack of it. Giving her an opportunity to nurture her husband made her feel better about herself, and the affective release of anger and laughter in the sessions enabled her to revise her internal working models towards a more realistic assessment of her capacities.
Parenthood re-awakens a person's own attachment history. As the life cycle unfolds, adults are presented with new challenges to their capacity to hold, respond to, attune with and ultimately release their children. With increasing age, the depth and strength of attachment bonds increases. At the same time losses begin to accumulate. Divorce, and separation and death begin to take their toll. We shall see in the next chapter how Attachment Theory provides a schematic map of this painful terrain of depression, disappointment and bereavement.
Bowlby's original mission was to find links between ‘life events’ such as parental loss or neglect and the development of psychiatric symptoms in children and adults. In ‘Forty-four juvenile thieves’ he linked such disruptions with the two major psychiatric disorders of childhood: conduct disorders and phobias. He anticipated that there would be connections between the vicissitudes of attachment in childhood and adult conditions such as depression, agoraphobia and psychopathic disorders. He made a fundamental distinction between secure and anxious attachment, seeing the latter as the precursor of developmental difficulty and adult psychiatric illness.
In his early formulations he saw anxious attachment as resulting from gross disruptions of parenting such as parental death or divorce. He also incriminated qualitative difficulties in parenting which included depression and unresponsiveness on the part of a parent; threats of suicide directed at the child; threats to send or give the child away; and role reversal in which the child is expected to ‘mother’ the parent.
Bowlby the systematiser and theoretician relied greatly on collaborators to provide the experimental evidence upon which his ideas rested. James Robertson's (1952) research and films confirmed his ideas about maternal deprivation. Ainsworth (1982, 1989) developed ways of measuring and conceptualising patterns of secure and insecure attachment, as well as extending and modifying Bowlby's original ideas. Today the focus has shifted away from gross disruptions of care such as bereavement, which do not in themselves necessarily result in psychopathology if conditions are otherwise favourable. The contemporary emphasis is much more on the subtleties of parent-child interaction contributing to the qualitative features of the attachment bond. Maternal responsiveness and the ability to attune to her child are now seen as key features in determining the security or otherwise of attachment bonds.
A further contemporary consideration concerns Bowlby's implicit assumption that the mother will be the primary care-giver in the first year of life. Changing patterns of family life mean that this is not necessarily the case. The evidence, scant in comparison with mothers though it is (Brazelton and Cramer, 1991; Bretherton, 2010), suggests that fathers are as capable of providing responsive attunement as mothers. Indeed for the purposes of the arguments presented in this and subsequent chapters mothers and fathers should generally be considered as interchangeable, but there are subtle differences in the ways in which fathers and mothers provide security for their children. Mothers are more likely to offer an extended containing ‘envelope’ for infant activity, while fathers tend to interact more intensely but for shorter periods, during which can be found the beginnings of organised play as development proceeds.
These and other research findings relevant to Attachment Theory will be reviewed in Chapter 6. But we turn now to the second of Bowlby's great themes, the breaking of affectional bonds (as opposed to their making which has been our theme in this one), and the need for affective processing so as to mitigate the psychological impact of separation and loss.
