7 Attachment and women’s mental health

In this chapter we will present a brief account of attachment theory and its evolution, then describe the contribution of attachment theory to the understanding of some key disorders that are more frequently seen in women.

Attachment theory is identified with John Bowlby, a child psychiatrist and psychoanalyst and Mary Ainsworth (née Salter), a research psychologist, amongst others. Before completing his medical degree Bowlby had worked in a school for maladjusted children. This introduced him to the lasting effects on personality development of deficient early family relationships (Bretherton 1992). Bowlby’s first published study was of the characteristics and family backgrounds of juvenile thieves. In this study he drew attention to the most prolific thieves, clinically described as affectionless characters, who uniformly had backgrounds characterized by combinations of complete or prolonged separation from the mother between 12 months and 4 years, anxious or ambivalent mothers or foster mothers, fathers who hated them, a family history of mental illness or had experienced recent trauma. (Bowlby 1944a, 1944b).

Bowlby used the perspective of ethology (the scientific study of animal behaviour) to see attachment to caregivers as a primary instinctual behavioural system. Bowlby hypothesized that evolutionary pressure had led to the development of an instinctual system of attachment to caregivers, and a complementary system of caregiving by parents or kin to infants, to ensure the survival of infants and hence of the clan and species. Bowlby argued that the evolutionary environment in which attachment developed was that of nomadic hunter-gatherer tribes and that the specific evolutionary pressure driving attachment was protection from predation. The goal of the attachment system is then to maintain proximity to caregivers who would provide safety from danger. As such, attachment is of equal biological importance to feeding and mating for survival of the species (Bowlby 1957, 1969).

Bowlby also took from ethology the idea of critical periods in development during which a specific behavioural system is particularly sensitive to the environment. Bowlby demonstrated that the attachment system is particularly sensitive to the environment during the first three or four years of life. Appropriate interactions with a caregiver during this period will allow the development of normal attachment behaviour. Outside of the critical period the attachment behaviour system will not be so responsive to the environment. If appropriate interactions do not occur sufficiently during the critical period, or development is disturbed for some other reason, then the attachment system will not develop normally. Later experiences have less impact than experiences during the critical period (Bowlby 1969).

In three seminal papers Bowlby argued that the infant’s attachment behaviour is made up of a number of component instinctual responses, clinging, sucking, and following, plus smiling and crying which had the purpose of eliciting care from caregivers. These appear in the first year of life and become increasingly focused on the mother during the second six months of life (Bowlby 1958). He argued that the intensity of attachment and the degree of distress aroused by separation from the mother could not be explained by previous theories (Bowlby 1960a). In his third paper he argued that grief and mourning are present in children whenever attachment behaviour is activated in the child and the attachment figure is not available, and that repeated disruption of attachment relationships during the critical period of the early years may result in the individual being unable to develop the capacity for relationships in depth (Bowlby 1960b). Bowlby theorized that attachment behaviours are encoded in mental representations of the individual in relation to others; he called these internal working models (Bowlby 1969).

Ainsworth conducted longitudinal naturalistic studies of mothers and infants less than two years of age observed at home in rural Uganda and of mothers with newborn babies in Baltimore, United States. These observations were complemented by observations in a laboratory procedure know as the strange situation. The strange situation was designed to measure levels of exploratory behaviour in infants of 12 months of age under circumstances of high and low stress. In the strange situation mother and baby are observed in a laboratory playroom, then an unfamiliar woman joins them. After three minutes the mother leaves the baby with the stranger for three minutes before she returns and the stranger leaves the mother and baby together. The mother then leaves again and the baby is alone for a further three minutes. Finally mother returns to the baby.

An unexpected observation made in the strange situation experiment was the range of infant behaviour on reunion with the mother. The majority of infants displayed attachment behaviour, sought proximity, interaction, or contact with their mother on reunion, and thus reassured returned to playing. These were described as secure (B). A second group of infants appeared to suppress attachment behaviour and ignored their mother on her return, even though they had searched for her in her absence. These were described as insecure-avoidant (A). A third group were surprisingly angry with their mothers and showed high levels of attachment behaviour that didn’t terminate—they wouldn’t settle even after she returned. These were described as insecure-ambivalent (C) (also called insecure-resistant). Just over half the infants were classified as secure with the remainder divided between the two insecure categories (Ainsworth et al. 1978).

It was shown that the reunion behaviour in the strange situation correlated with the quality of the mother–infant interaction at home. The infants with more harmonious relationships with their mothers showed secure attachment in the strange situation, while those with less harmonious relationships showed insecure-avoidant or insecure-ambivalent behaviour (Ainsworth et al. 1978).

In addition to describing the infants’ behaviour the attachment classification is also related to the infants’ perception of the availability of the caregiver. Securely attached infants have internalized the expectation of an adult who will respond to their distress, when they feel threatened they are able to direct attachment behaviours (crying, approaching, seeking contact) towards the caregiver, and they are able to take comfort from their caregivers’ reassurances. In this way the parents of securely attached infants promote their infants’ exploration of the world. This is not the case with infants who are insecurely attached. Insecurely attached infants do not expect a caregiver to be reliably responsive.

Insecure attachments develop when the infants do not have a mental representation of a responsive caregiver in times of need, such as when they feel fearful or helpless. These infants develop different strategies to gain access to their caregiver in order to survive.

In Ainsworth’s original study, the mothers of insecure-avoidant infants showed aversion to physical contact when the infants sought it and expressed little emotion in interaction with their infants. Insecure-avoidant attachment is therefore an adaptive strategy on the part of the infants of defensively turning away from their distress and minimizing their attachment behaviour in order to maintain proximity to a caregiver who is aversive to contact and is rejecting. These infants therefore have less access to their own feelings and have lower expectations of help being available.

The mothers of infants later classified as insecure-ambivalent were found to be less sensitive to their babies and more interfering with their children’s behaviour. Thus insecure-ambivalent attachment is a strategy of maximizing attachment behaviour to gain the attention of a caregiver who is inconsistently available to the infant. Because these infants are so enmeshed in issues of caretaker availability they are unable to make accurate appraisals of the level of threat and the availability of help (Ainsworth 1978). Insecure attachments are therefore considered to be adaptive in that they offer an organized strategy for gaining protection from the caregiver.

Subsequent researchers reported difficulty classifying some infants’ strange situation behaviour in samples where the children had been maltreated. The behaviour of infants from maltreated samples was often found to be contradictory or disorganized, they seemed to lack an organized strategy for dealing with reunion with their parent in the strange situation and did not fit neatly into any of the secure/insecure categories. This led to the addition of a fourth attachment category, disorganized (D). Behaviour during reunion that is described as disorganized includes contradictory, misdirected, or fragmented behaviour patterns, e.g. approaching and moving away from the parent, crying against the wall, stereotypies, freezing or slowing, or disoriented behaviour (Main and Solomon 1990).

Disorganized attachment behaviour arises when infants are overwhelmed by separation and can be induced when toddlers are given conflicting signals, confusing signals, or exposed to inescapable situations. Conceptually, disorganization is explained by the coexistence of two contradictory urges: the urge to seek safety from the attachment figure and the wish to flee from the attachment figure because the attachment figure is the source of distress or fear.

In addition to the strange situation, which assesses attachment behaviour, a number of tools have been devised to measure and classify attachment representations in older children, adolescents, and adults. For a review see Prior and Glaser (2006). The Adult Attachment Interview is widely used to assess the state of mind of adults with regard to attachment. It consists of an hour-long interview in which adults are asked to describe early attachment relationships and experiences and to evaluate the effects of these on their personality and functioning (Main et al. 1985) The transcripts of the interviews are then classified on the basis of form and content into Secure/autonomous (F), Dismissing (Ds), Preoccupied (E); subsequently a fourth category has been added, Unresolved/disorganized (U/d). The categories use similar constructs to the classification of infant behaviour in the strange situation.

About 50–80% of maltreated infants show disorganized attachment, and 15–25% of children in low-risk non-maltreated samples also demonstrate disorganized attachment (van IJsendoorn et al. 1999).

Dissociative phenomena in adults are of particular interest for several reasons. Unresolved/disorganized attachment on the Adult Attachment Interview is conceptually linked to dissociative phenomena arising whilst discussing loss and trauma themes. Unresolved/disorganized attachment in caregivers in the Adult Attachment Interview predicts disorganized attachment in their offspring in the strange situation. Dissociative phenomena in the strange situation support disorganized attachment classification. Disorganized infants have been shown to go on to display and report raised levels of dissociative behaviour and symptoms at 16 and 19 years (Ogawa et al. 1997; Weinfield et al. 1999).

Once dissociation is established as a defence then it tends to occur at lower thresholds in the future. During the pre-school period, children who have been maltreated show increasing levels of dissociation compared to non-maltreated peers who show decreasing levels of dissociation (Macfie et al. 2001).

A number of explanations exist for disorganized attachment in low-risk samples. A number of the children in low-risk samples will have experienced maltreatment. A proportion of infants may be more sensitive to maternal insensitivity due to genetic variation, specifically variations in a dopamine receptor subtype (Gervai 2007).

A hypothesis that has generated much research is that mothers who show frightening or frightened behaviour unknowingly place their infants in situations of inescapable fear, the infant is frightened by the parent’s behaviour and thus wants to flee, and yet the infant also wants to approach the parent as the person with whom the infant wishes to seek safety.

Main and Hesse (1990) found a strong association between parents being in the unresolved/disorganized category on the Adult Attachment Interview (ie unresolved or disorganized when discussing their own experiences of loss or abuse), and their infants demonstrating disorganized attachment in the strange situation. Over 90% of unresolved/disorganized mothers had disorganized infants. Only 16% of mothers who had experienced loss, but did not show unresolved/disorganized attachment, had disorganized infants. This finding has been replicated in a number of studies (Fonagy et al. 1991; van IJzendoorn 1995).

Based on behaviours shown in the Adult Attachment Interview that suggested that the parent was experiencing some degree of dissociation while recounting traumatic events, Main and Hesse have described a scale of frightening (FR) behaviours shown by parents. These include direct indices of dissociation, anomalous and inexplicable threatening behaviour, timid/deferential (role inverting) behaviour, sexualized behaviour towards the infant, and disorganized/disoriented behaviour compatible with disorganized infant behaviours. They hypothesize that unresolved, partially dissociated traumatic experiences intrude into the parent’s mind and are manifest in degrees of dissociated behaviour, and that it is these states of dissociation in the parent that the infant finds frightening, and which lead to disorganization of their infant’s attachment. Further evidence supporting the link between frightening (FR) parental behaviour and disorganization of infant attachment has been reported in high-risk samples. For a review see Hesse and Main (1999).

The ability to think about or reflect on one’s own mental states (intentions, wishes, feelings, and beliefs) or on the mental states of others is known as reflective function or mentalization. It is a largely preconscious activity. The social biofeedback theory of parental affect-mirroring suggests that the infant learns to recognize its own affects through seeing them contingently mirrored in its parent and over time is thus able to internalize the representation of its own affect states, which leads to the sense of having a mind. This requires a caregiver who is sensitively attuned to their child’s mental states. In this way the acquisition of mentalization is intimately linked to the quality of attachment relationships, and the development of a coherent or integrated sense of self and others.

Infants who have caregivers who are insensitive or poorly attuned to their mental states will have less opportunity to learn about their own mental states through contingent affect mirroring. Infants who experience their caregivers as frightened or frightening will avoid contemplating the contents of their caregivers’ minds. Infants whose attachment is disorganized do not develop an integrated image of themselves or their caregiver. A dissociated and unintegrated sense of self and other disrupts the acquisition of reflective function and leads to emotional disability. In this way, attachment disorganization limits mentalization. It is beyond the scope of this chapter to describe in detail the development of mentalization. For a detailed account see Fonagy et al. (2004).

There is evidence that attachment patterns are transmitted from mothers and fathers to children. The attachment status of adults prior to the birth of their children is highly predictive of their unborn children’s future attachment classification in the strange situation at one year (Fonagy et al. 1991).

Role reversal is one aspect of disorganized attachment that has been studied in depth. In role reversal, parents and children swap roles; a parent looks to their child to fulfil their own emotional needs, and a child may attempt to meet the parent’s unmet emotional needs in order to sooth the parent’s distress and thus achieve a greater feeling of security in that relationship. Mothers who reported role reversal with their mothers during the Adult Attachment Interview engaged in higher levels of role reversal with their toddler-aged daughters. Furthermore, when fathers reported role reversal with their mothers during the Adult Attachment Interview, then their female partners tended to engage in higher levels of role reversal with their toddler-aged sons, indicating that assertive mating allows the transmission of mother–son role reversal through the son’s choice of partner (Macfie 2005).

Bowlby argued that early attachment experiences are critical for developing a sense of mastery, emotional self-regulation, and interpersonal effectiveness. He argued that the internal working models of relationships that determine behaviour are adaptable to changing environments, but that they can become defensively distorted and that this will lead to rigidity and a failure of adaptation to the changing environment. The principal cause of defensive distortions and splits is the need to keep an internal working model of a good helpful parent separate from an internal working model of a frightening parent.

Infants whose caregivers are sensitive and responsive learn that they can get their needs met and be effective in the world. Infants whose caregivers are insensitive or inconsistent do not develop this sense of autonomy (Bowlby 1973).

A number of large longitudinal studies of attachment have been running for many years. The largest and longest running study of attachment is the Minnesota Longitudinal Study of Parents and Children (Sroufe et al. 2005), which began in 1975 with 267 pregnant first-time mothers; they were below the official poverty line and as such it is a high-risk sample. It studied the mothers before and after the birth, and the children throughout their lives and into (at the time of writing) their 28^th year. It has demonstrated a number of important continuities between infant attachment and subsequent behaviour, as discussed in the following sections.

Children who were securely attached in infancy show greater social competence. They were shown to be more cooperative with adults, more appropriate in their help-seeking behaviour from teachers, and more confident in their play when compared to insecurely attached children. Securely attached children were better able to modulate their negative feelings, expressed less anger and aggression during play with their peers, and were not found to be either bullies or bullied. Children with insecure-avoidant attachment showed more bullying behaviour towards their peers during play, and insecure-resistant children were more likely to be the victims of insecure-avoidant children. Children with secure attachment histories were rated as more socially competent by teachers throughout school and adolescence. Secure children were better at adhering to group norms and showed more leadership ability. Girls with a history of secure attachment were rated as having relationships that were more intimate. For a review of the findings of the Minnesota Longitudinal Study see Weinfield et al. (1999).

The Minnesota Study found that early insecure-avoidant attachment was significantly associated with aggression in boys, and in adolescence with childhood-onset antisocial behaviour (Egeland and Carlson 2004). Insecure-avoidant five-year-olds were more likely to bully insecure-ambivalent peers (Weinfield et al. 1999).

Warren et al. (1997) found in the Minnesota sample that after controlling for other key variables, anxious-ambivalent attachment was significantly associated with anxiety disorders among teenagers.

Disorganized attachment appears to be most important in terms of predicting later poor functioning and disturbance. Disorganized attachment is related to increased cortisol secretion in response to stress, indicating the infants’ lack of an effective strategy for dealing with stress. It is associated with mild delay on the Bayley Scales of Infant Development. Disorganized attachment is strongly associated with later controlling and role reversal behaviour of children with their parents (Main and Cassidy 1988). Disorganized children perform less well in tests of reasoning and delay-of-gratification task, even after controlling for self-esteem and intelligence quotient. Disorganized children are less competent in their play with peers, have higher rates of peer hostility and aggression, and are at greater risk of developing behaviour problems and oppositional defiant disorder (Greenberg 1999). Disorganized children and adolescents display and report raised levels of dissociative behaviour and symptoms at 16 and 19 years (Ogawa et al. 1997; Weinfield et al. 1999). For detailed review of disorganized attachment in childhood the reader is directed to Lyons-Ruth and Jacobvitz (1999) and Hesse and Main (2000).

Research on the relationship between attachment and mental health in adulthood largely depends on two strategies: assessing the concurrent attachment status of the adult using the Adult Attachment Interview, or using recall of parental behaviour and recall of adverse childhood experiences that can allow suppositions about the likely nature of the early attachment relationship.

The Adult Attachment Interview has been used in a wide number of different clinical and non-clinical groups. Generally, secure/autonomous attachment is underrepresented in clinical groups and appears to act as a protective factor. This is in line with Bowlby’s hypothesis that individuals with secure attachment view themselves as competent and have an expectation that help will be available when required, and that they will be able to use it.

The central thesis of attachment theory is the primary organizing influence of attachment to caregivers. It is to be expected, therefore, that the loss through death of the primary caregiver in childhood should have far reaching consequences. Brown and Harris (Brown et al. 1977; Brown and Harris 1978) demonstrated in a series of studies of women with depression, in both a community and hospital setting, that the death of the mother before the age of 11 years was the major factor determining vulnerability to depression in response to further losses in later life. Death of the mother before 11 years also determined the severity and symptomatology of the depression. Women whose mothers had died before they were 11 had much higher rates of melancholic depression (characterized by retardation and other somatic features). Drawing on the ideas of learned helplessness (Seligman 1975) and hopelessness in depression (Beck 1967) the authors showed that depressed women who had lost their mother before the age of 11 years were also less likely to seek treatment for their depression than other depressed women. They hypothesized that the early loss of the mother permanently lowered the women’s self-esteem and sense of mastery, and increased the chance that the women would passively endure future trauma, rather than expecting and seeking a solution.

Parkes (1991) studied an outpatient sample of adults with complicated bereavement. He used basic trust in self or others as a proxy measure for good early attachment experiences. He found that people without basic trust were more vulnerable to developing complicated bereavement. In his sample the patients with basic trust and complicated bereavement had all experienced particularly difficult bereavements that were either multiple or unexpected. Among the patients low in basic trust, only a minority had experienced multiple or unexpected bereavements; 58% had experienced a single, expected death. This indicates that people with basic trust only developed complicated bereavement after multiple or unexpected bereavements, whereas people lacking in basic trust were vulnerable to developing complicated bereavements even after a single death which they had had time to prepare for. Grief reactions complicated by feelings of ambivalence towards the dead person were significantly more likely to arise following the death of a parent.

The direct research evidence on attachment styles among depressed adults is inconsistent. In part this arises from samples being drawn from different populations—out patients, inpatients, or patients in tertiary referral centres. Furthermore, inclusion and exclusion criteria vary between the different studies; some control for comorbid personality disorder, others do not. Some research reports only the three-way classification (Secure/autonomous, Dismissing, Preoccupied), others also use the four-way classification (Secure/autonomous, Dismissing, Preoccupied, Unresolved/disorganized). Finally, the diagnostic category of depression is in itself probably heterogeneous with regard to attachment behaviours, including depressed individuals with insecure-dismissive (externalizing) strategies and depressed individuals with insecure-ambivalent (internalizing) strategies. Generally among depressed patients there are lower levels of Secure/autonomous attachment and higher rates of Preoccupied and Unresolved/disorganized attachment (Dozier et al. 1999).

Postnatal depression is more common in mothers who have insecure attachment, specifically insecure-anxious (enmeshed and fearful) attachment styles (Bifulco et al. 2004). Toddler offspring of depressed mothers have significantly more insecure attachments and behavioural problems than do toddlers with non-depressed mothers (Cicchetti et al. 1998).

Parental, and particularly maternal, depression is associated with an increased risk of childhood emotional and behavioural disorders in their children. As the children of depressed parents grow up, there is a two- to threefold increase in rates of depression, anxiety, substance misuse, and physical illness when their children reach adolescence and adulthood. Effective treatment of maternal depression leads to a reduction in emotional and behavioural disturbance in their children (Weissman et al. 2005, 2006a, 2006b).

Anxiety disorders are also heterogeneous and often comorbid with other disorders, particularly depression. Focusing on generalized anxiety disorder, the prospective studies of Warren et al. (1997) showing that insecure-ambivalent attachment in the strange situation at one year was associated with anxiety disorders at 17 years. Fonagy et al. (1996), in a sample of psychiatric inpatients at the Cassel Hospital (Richmond, UK), a tertiary referral centre for people with particularly severe and complex difficulties, found that using the four-way classification, unresolved/disorganized attachment status was present in 7% of controls, 75% of patients, and 86% of patients with a diagnosis of anxiety. Due to their severity these patients are unlikely to be representative of all those with anxiety.

Adults who experience clinically significant dissociation are most likely to be diagnosed as having chronic post-traumatic stress disorder, dissociative disorders (multiple personality disorder, dissociative identity disorder, depersonalization), or borderline personality disorder.

Dissociative identity disorder is, by definition, a disorganized state of the self. Ninety per cent of people with dissociative identity disorder have been found to have a history of childhood sexual abuse (Fonagy and Target 1995).

Post-traumatic stress disorder, by definition, is a state of being unresolved with respect to trauma, and therefore of unresolved/disorganized attachment (de Zulueta 2006). Diagnostic criteria for post-traumatic stress disorder emphasize the experiences of people who have experienced circumscribed traumatic events. However, it is recognized in DSM IV that children exposed to repeated, protracted interpersonal trauma such as childhood sexual or physical abuse or neglect, exhibit characteristic symptoms, including impaired affect modulation, dissociative symptoms, somatic complaints, feelings of ineffectiveness, shame, guilt, despair, hopelessness, feeling permanently damaged, hostility, social withdrawal, feeling constantly threatened, and impaired relationships with others (American Psychiatric Association 1997). This constellation of symptoms has been referred to as complex post-traumatic stress disorder (Herman 1992, van der Kolk et al. 2005). It is very likely that children exposed to this early environment would have disorganized attachment.

Borderline personality disorder is a disorder of personality development characterized by: instability in affects, identity, and relationships; feelings of emptiness; strenuous efforts to avoid feelings of abandonment that lead to feelings of psychic disintegration. It is used to describe a wide range of difficulties and behaviours. The backgrounds of people diagnosed with borderline personality disorder are characterized by reports of neglect, abuse, and early separations (Gabbard 2005). As a consequence, people with borderline personality disorder show high rates of unresolved/disorganized attachment. In the Cassel Hospital sample, 89% of inpatients with borderline personality disorder had unresolved attachment status (Fonagy et al. 1996).

An argument has been made to link some eating disorders with dissociation on the basis that the serious distortions of body image are related to dissociation (Liotti 1999). In the case of a woman with anorexia and a severe distortion of her body image, the dissociation might occur between two internal working models of herself, one of a little girl who is starving hungry for love and approval from an emotionally unavailable mother, and an incompatible internal working model of a good kind mother with a very bad greedy little girl, who must control her greediness. However, eating disorders and dissociation are considered separately here.

Attachment research in people with eating disorders shows that insecure attachment is the norm with a mixture of insecure-ambivalent/preoccupied attachment and insecure-avoidant/dismissive attachment; among more disturbed populations unresolved classifications predominate.

Ringer et al. (2007) using the Adult Attachment Interview found that all of their patients with eating disorders were insecurely attached with half being preoccupied and the rest mainly using a combination of preoccupied and dismissing strategies. Using a self-report questionnaire (the Reciprocal Attachment Questionnaire), Ward et al. (2000) in a mixed inpatient and outpatient sample of women at a tertiary referral centre found that patients scored higher than controls on most scales but particularly Compulsive Care-Seeking and Compulsive Self-Reliance, creating a ‘push–pull’ dilemma in their attachment relationships. Troisi et al. (2005) using the Attachment Style Questionnaire found that eating disorder patients had higher rates of insecure-anxious attachment but no elevation in insecure-avoidant scales. Later Troisi et al. (2006) found that childhood separation anxiety and preoccupied attachment best predicted negative body image. Latzer et al. (2002) using the Adult Attachment Scale found that eating disorder patients had higher scores on both insecure-avoidant and insecure-ambivalent attachment than controls.

In the Cassel inpatient sample Fonagy et al. (1996) using the Adult Attachment Interview found that patients with eating disorders on the three-way classification had high rates of insecure-preoccupied attachment (64%) and on the four-way classification all but one (93%, 13/14) were unresolved. Interestingly, the eating disorder patients differed from the other patient groups in having significantly elevated scores for idealization of parents; this sample is probably more severely disturbed than the other groups, coming from an exclusively inpatient sample with high levels of comorbidity at a national centre.

The attachment system is an instinctual behavioural system that evolved to reduce predation of the young. The early years are a critical period for the development of attachment. Attachment is patterned by parental sensitivity, responsiveness, and availability. It is disrupted by trauma, separation, or loss. Attachment styles tend to endure in the absence of marked social environmental change. Attachment styles are mediated by the subject’s internal working models—these are mental representations of how relationships work. Internal working models determine how individuals perceive and act in their social environment. Internal working models can become defensively structured in which case they become less adaptive to later environmental change. Secure attachment is associated with a sense of competence and a belief that help will be available if needed; it is associated with emotional resilience under stress. Insecure attachment is associated with vulnerability to emotional trauma. Insecure dismissive attachment is associated with reduced help seeking behaviour. Disorganized attachment arises when caregivers are frightened or frightening or show dissociative behaviour. Individuals with disorganized attachment lack an organized strategy for obtaining security from an attachment figure, who is seen as both a source of fear and of safety. Individuals with disorganized attachment display a lack of integration of self and other, they are most vulnerable to develop later psychiatric illness. Attachment styles are transmitted between generations, directly through parenting behaviour and indirectly by assortive mating.