Parents are only as happy as their least happy child
Parents are only as happy as their least happy child(traditional saying)
A fundamental postulate – perhaps the fundamental postulate – of attachment theory is that children, when threatened, tired, or sick, seek out an ‘older, wiser’ Secure Base for protection and succour (Bowlby 1971; Holmes 2013). Reciprocally, parents are programmed to respond to their infants’ and children’s distress with nurturing behaviours such as physical contact, hugging, comforting, warming, cleansing, and feeding.
In addition to this ‘emergency’ aspect of the attachment relationship, there is a ‘non-emergency, ordinary’ (Waters 2008) everyday aspect, in which the care-giver attends to the child’s developmental needs by close attention, providing scaffolding and supporting the child’s agency and independence. A crucial component in both is ‘affect regulation’ (Mikulincer et al. 2003) in that the child’s distress is modulated by a variety of parental manoeuvres such as containment, soothing, distraction and/or stimulation, but also that positive affects are contained and modulated, and responded to with pleasure, thereby instilling the capacity for mutual fun and joyfulness. As development proceeds, these become internalised by the child, who learns both to manage her own feelings when she can, and to recruit others for help when needed. Authors such as Waters (2008) and Schore (2004) see both dimensions as integral to the attachment dynamic, although it can be argued that the inter-subjective developmental aspects of early relationships are distinct from the security-providing attachment component.
Staying with the ‘emergency’ vector, the sequence of stress/distress followed by soothing/recovery proceeds fractally throughout development, at micro and macro levels. In the early weeks and months of life, several times per day, there will be mini-stresses and disruptions to be contained, focused on and overcome. At a macro level, children can be expected to negotiate a series of developmental hurdles – separation and stress (parents away or ill; starting school; bullying; fears, phobias and frightening experiences, etc.), as well as physical illnesses and injuries, usually self-limiting. The cycle of stress, mental or physical pain, soothing and recovery, when things go well, fosters resilience, The developing child is strengthened emotionally, with a growing sense that adversity can, with appropriate help, be faced and overcome (Fonagy et al. 2004). A comparable maturation occurs for parents, who gain confidence in their capacity to provide effective nurture and support for their children, as, together, they face the quotidian vicissitudes of family life.
Chronic illnesses, physical or mental, pose particular difficulties for both parents and children within this model. Unlike self-limiting illnesses or physical trauma, there is often no clear recovery point. Rather than complete cure, optimal outcomes are likely to be those of adjustment, acceptance, and courage in the face of adversity. These considerations apply equally to ill parents and parents of ill children.
Most psychotherapeutic practice relies on hunches and heuristics, permeated by a mild ‘cult of personality’, rather than established facts. In establishing the attachment paradigm, Bowlby’s self-effacingness, and strict adherence to scientific principles of evidence and refutation, mitigated against both: one of the great strengths of attachment theory is its empirical base, eschewing appeal to authority.
A significant breakthrough came in the 1990s when a number of researchers found evidence of continuity between attachment patterns in the parental generation and their children in the next. The classic Fonagy, Steele and Steele (1991) study measured attachment dispositions in a non-clinical sample of ‘pregnant parents’ using the Adult Attachment Interview (AAI); then, based on the Strange Situation procedure (SSP) (Mary Ainsworth’s first formal measure of infant/mother attachment) the resulting offspring were assigned attachment categories some 15 months later. For a detailed discussion of the Strange Situation assessment measures see The Routledge Handbook of Attachment: Theory (Holmes & Farnfield 2014) or The Routledge Handbook of Attachment: Assessment (Farnfield & Holmes 2014).
Parents whose narrative styles were ‘free/autonomous’ were statistically more likely to have children who were securely attached, while those with dismissing narrative styles tended to have children who were classified as insecure-avoidant in the SSP.
This seminal study opened up two decades of study of how attachment patterns might be transmitted from one generation to the next. The current view rests on the prevailing gene × environment perspective. Given the inherent polymorphism of the human genome, some children are relatively robust and can achieve a degree of attachment security and psychological health notwithstanding adverse parenting styles or trauma. By contrast, infants with ‘plasticity’ genes are powerfully affected by their parents’ handling, for good or ill (Steele & Siever 2010). The ongoing parent–child dynamic has to be conceptualised in terms of a subtle set of self-maintaining developmental pathways to which genetic and environmental factors and the relationship itself contribute.
The early attachment literature saw the differences between secure and insecure children in terms of parental ‘sensitivity’ (Ainsworth 1973). Fonagy and coworkers now specify aspects of this quality under the rubric of ‘mentalising’ (Fonagy & Luyten 2009; Holmes 2010). Mentalising is a complex group of interand intra-personal skills which include the ability to: (a) view oneself and others as sentient beings with desires, projects and motivations, (b) reflect on one’s own and others’ motivations, and (c) understand how beliefs about self and others, because filtered through the mind, are inherently error-prone, and in need of negotiation and correction.
In the original Fonagy/Steele study, mothers able to mentalise (with high levels of ‘reflexive function’, as it was then called), even when their own upbringing had been troubled, tended to have secure offspring. Those without that capacity were more likely to have insecure children, especially if their own developmental history had been problematic. Parents who can see their children as separate sentient beings, who can differentiate their own feelings from those of the child, see the child in a developmental context, and recruit appropriate help when needed, are likely to be rated as ‘sensitive’ and therefore able to foster secure attachment in their care-providing practices.
Neuroscience techniques help to further illuminate these findings. In a recent study Strathearn et al. (2009) identify two features which differentiate secure from insecure mothers as measured on the AAI. When exposed in an fMRI setup to smiling or sad pictures of their infants’ faces, brain patterns in the two groups were significantly different. Compared with their insecure counterparts, mothers with secure attachment dispositions showed increased activation of mesocorticolimbic reward brain regions on viewing their own infant’s smiling face. They also showed an increased peripheral oxytocin response while interacting with their infants, which was correlated with activation of oxytocinergic and dopamineassociated reward processing regions of the brain (hypothalamus/pituitary and ventral striatum). Thus it might be said that secure mothers find interactions with their babies more rewarding than the insecure.
Even more remarkable was the finding that ‘striking differences in brain activation were seen in response to their own infant’s sad facial affect’ (Strathearn et al. 2009: 2662). Securely attached mothers showed greater activation in reward processing regions of the brain, whereas ‘insecure/dismissing’ mothers showed increased activation of the anterior insula, a region associated with feelings of unfairness, pain, and disgust.
For securely attached mothers, infant cues (whether positive or negative) reinforce and motivate responsive maternal care. By contrast, insecure mothers, it seems, when exposed to negative cues in their infants, react with withdrawal and aversion when the expected positive feedback from their child is absent. This in turn evokes negative feelings in the mother, including disgust. This finding is consistent with the typical dampening-down of feeing seen in avoidant/deactivating children: to manifest separation distress runs the risk of further alienating the care-giver, and exposing the child to danger as parental Secure Base provision is withdrawn. In such sub-optimal care-giving environments affect suppression maintains proximity, and therefore a degree of safety, but at an emotional cost. This formulation is consistent with the finding that negativity is a more salient feature in insecurely attached children than in the securely attached (Belsky et al. 1996). But negative affect, although suppressed, remains subliminally present, adversely colouring the child’s mental universe.
Insecure attachment in children is not itself problematic. At least 30 per cent of non-clinical populations demonstrate insecure attachment. This may, indeed, be an adaptive response to sub-optimal care-giving environments (which are of course widespread, and reflect prevailing social and political structures). It is important that professionals respect such resilience, and eschew as far as possible imposing their own ‘middle class’ social and ethical norms, both in research and clinical contexts.
However, the case of Disorganised attachment (labelled D in attachment theory), especially when combined with severe patterns of hyperactivating insecurity is somewhat different. Disorganisation is highly predictive of later psychological difficulties including poor peer adaption, disruption in school, and externalising and dissociative disorders in middle childhood and adolescence. It is these children who are most likely to come to the attention of educational, medical, social care, and criminal justice agencies (see Shemmings in Holmes & Farnfield 2014).
Disorganised attachment as measured in the SSP between 12 and 18 months predicts psychopathology in later childhood and early adulthood (Van IJzendoorn et al. 1999). Lyons-Ruth has pioneered studies looking at the characteristics of mothers of D children, and on the basis of a number of ‘communication errors’, classifies them into two broad groups: frightened/withdrawn and hostile/intrusive (Lyons-Ruth et al. 1999). Both groups report significant amounts of childhood trauma in their own development, the withdrawn group more likely to have been sexually abused, the hostile group physically abused.
In the withdrawn group the mother appears unable to respond to her child’s distress, leaving the infant to find some means of self-soothing, however uncoordinated – appearing dissociated, rocking, head banging, etc. In the hostile-intrusive pattern the mother typically attacks and blames the child for being distressed. ‘You’re only doing this to wind me up’ would be a characteristic hostile-intrusive response to an infant’s attachment-seeking behaviours. The relevance of this to later responses to chronic illness might be either a helpless ‘giving up’ on the part of the parent, or ‘blaming the victim’ where the child’s illness is seen as self-inflicted, leading often to futile attempts to get the child to behave ‘normally’. Neither the child’s nor the parent’s (herself often a victim of abusive or insensitive parenting) attachment needs are met in these self-defeating interactions.
Such dysfunctional patterns were originally conceptualised primarily from the child’s point of view as ‘fear without solution’ (Main & Hesse 1990). The child is presented with an insoluble paradox, reminiscent of the supposed ‘schizophrenogenic’ ‘double bind’ (cf. Holmes 2010) in which the very person who would alleviate attachment arousal is also the source of the threat, and the child thus faces an unresolvable approach-avoidance dilemma.
Within the context of this chapter, Disorganised attachment needs also to be seen from the point of view of the parent. From this perspective, the care-giver herself feels either disempowered and helpless, or suffused with unassuaged rage. The parent’s care-giving attachment dynamic is activated, but cannot be assuaged – the child remains unwell, despite the parent’s best efforts at alleviation. The withdrawn ‘given up’ posture is an expression of helplessness in the face of an unresolvable situation; conversely the hostile-intrusive pattern can be seen in terms of a desperate parent, longing for diminution of attachment arousal, verbally (or sometimes physically) attacking the child in the hope that rage will somehow alleviate her unbearable feelings.
Further, children in Type D are typically found beyond infancy to be ‘controlling’, often with role-reversal between themselves and their parent (O’Connor et al. 2011). This can be understood from the child’s point of view as a means of locating their own vulnerability in the other and thus feeling a measure of freedom from fear and mental pain. From the parent’s perspective this can be disempowering, reinforcing feelings of helplessness and incompetence. An analogous pattern may exist with chronically ill children, for example with eating disorders or chronic fatigue syndrome (CFS), where parents may cow-tow and pander to their child’s every need in the hope of recovery, but end up feeling baffled and frustrated.
The discussion thus far has focused on children designated as Disorganised in the Strange Situation, the parental handling associated with that classification, the impact on the parent, and their long-term consequences for child mental health. In this section we will look at the converse: the implication for their children of mothers with mental health difficulties, especially Borderline Personality Disorder (BPD).
Murray and co-workers (Murray et al. 2011) have studied the implications of maternal depression on attachment patterns in children, showing that depressed mothers are more likely to have insecurely attached children, but that when the mood disorder is successfully treated, the child’s attachment pattern often becomes secure. Personality Disorders are, by definition, more long-lasting than so-called ‘Axis 1’ illnesses such as depression. The implications for the child of being brought up by a mother suffering from BPD are thus highly significant.
From the point of view of parenting, important features of BPD include affective instability; an unstable sense of self; substance abuse; and episodes of deliberate self-harm. All of these are likely to compromise effective and sensitive parenting. Difficulties with mentalising are at the heart of the Fonagy-Bateman model of BPD (Fonagy & Luyten 2009), and their interventions are targeted around efforts to enhance it. If parental mentalising is the key to the fostering of secure attachment in children, then clearly the children of BPD-suffering parents will be at risk. A person’s ability to help regulate their child’s feelings is likely to be compromised if they themselves are subject to unstable moods, and their mentalising capacities deficient if they lack a secure sense of self which provides a vantage point from which to view their own needs and those of others, including their children. In addition parental intoxication and deliberate self-harm (DSH) are potentially traumatic for children. In the absence of a mentalising care-giver, such children will suffer not only from the trauma itself, but, given the absence of an affect-regulating parent, lack of someone with whom to process the feelings it engenders.
A number of studies have looked at the impact of BPD on developmental processes in sufferers’ offspring. Feldman et al. (1995) confirmed clinical impressions that children aged 11–18 of BPD sufferers are at risk of psychopathology, showing more delinquency, aggression and depression than comparable offspring of non-BPD sufferers. Hobson and colleagues have studied the possible developmental antecedents of these clinical findings. Children of BPD mothers have high levels of Disorganised attachment at 18 months (Hobson et al. 2005). The origins of this are illuminated by studies which show that such mothers show more intrusiveness and insensitivity, are more likely to be withdrawn and fearful when their infants become distressed, and are limited in their capacity for reflexive functioning (i.e. mentalising) compared to children whose mothers are not so diagnosed (Crandell et al. 2003; Hobson et al. 2009). We see here the possible routes by which BPD may be transmitted from one generation to the next, and also the developmental difficulties that children of BPD mothers are likely to encounter.
Interventions can be broadly classified under three headings. First, preventive measures aiming to reduce the likelihood of Disorganised attachment in the offspring of at-risk parents, whether due to low socio-economic status or physical or mental illness. Second, there are treatment programmes aimed at helping mothers once an ‘at-risk’ child is identified, especially with Disorganised attachment. Third, parents who are themselves ill, especially those with BPD, can be helped with the aim to improve not just their mental health but, indirectly, that of their children.
Integral to the attachment perspective is the hypothesis that infant security is causally linked to, rather than merely correlated with, parental sensitivity. If so, interventions to enhance parental sensitivity should increase security of attachment in their infants. In an early meta-analysis, Van IJzendoorn et al. (1999) found that this was the case, but there were a number of paradoxical effects. First, short-term interventions appeared to be more effective than longer ones. Second, increasing parental sensitivity did not always impact on infant attachment security and, conversely, attachment security could be improved without changes in parents’ sensitivity. Third, causal links, although demonstrable, were weak and did not account for the total variance.
Van IJzendoorn’s group dub this the ‘transmission gap’. Recent genetic findings mentioned above go some way to explain this – if only a proportion of target infants have ‘plasticity genes’ (see above), the overall impact of interventions will be thereby lessened (Bakermans-Kranenburg & Van IJzendoorn 2007). A subsequent review (Berlin et al. 2008) summarised a number of intervention programmes, varying from simple behavioural measures such as giving mothers a soft sling with which to carry their babies around, to more sophisticated and prolonged psychoanalytically informed programmes aiming to enhance mentalising. Their findings suggest that relatively brief, sensitivity-focused interventions in high-risk groups, where the base-line level of Disorganised attachment is greatest, show the most gains. However, the evidence for the benefits in terms of maternal sensitivity and child attachment security of long-term psychoanalytic child psychotherapy interventions is rather weak. The authors concede this may be due to the broad-spectrum impact of this approach, with general rather than specific benefits, and that there may be ‘sleeper effects’ which would reveal themselves with longer term follow-up.
The studies summarised above are based mostly on attachment-informed interventions in the early years, and many are mother–infant, rather than parent-only programmes. In this section I shall comment from a research-informed clinical perspective on two types of problem: working with parents suffering from BPD, and working with parents of children suffering from chronic illnesses.
Both groups may benefit from a ‘pedagogic’ psycho-educational exposition of the attachment model. It can be a relief to hear that constant worry and inability not to think about an ill child is a normal biologically driven aspect of the attachment dynamic. One parent with an eating-disordered daughter confessed guiltily that while taking a deserved break away for a night in a health spa, she found that she had forgotten to think about her daughter for at least 10 minutes! Such guilty feelings are common in the bereaved; having an ill child could be seen as a species of bereavement, but one with no definite end-point. This is consistent with Bowlby’s (1980) formulation of bereavement in terms of irreparable separation. Equally BPD parents can be helped with the thought that some of their feelings of fear and helplessness when confronted with a needy child are normal responses, not signs of intrinsic inadequacy or madness.
Another aspect is the attachment view that attachment behaviour and exploration are mutually incompatible (Holmes 2010). As a result, parents who are chronically aroused with worry about their children, may find it difficult to ‘accept’ – i.e. to think about and experiment with – the help that is being offered, since their thoughts and feelings are so dominated by the search for, and seeming impossibility of finding, a solution to their child’s difficulties. In one case, after apparently helpful sessions in which strategies for living with a depressed teenaged son were discussed, his mother would repeatedly, as she made for the door, blurt out: ‘Well, will he get better or not? I want and need an answer – why won’t you give me one?’
Bakermans-Kranenburg et al. (2005) see preventive interventions as having one or more of three objectives: enhancing parental sensitivity, improving attachment security in the IP (identified patient) child, and using the therapeutic relationship as a model for secure parent–child interaction. As mentioned, ‘sensitivity’ remains a somewhat mysterious capacity.
Ainsworth’s early formulations focused on the rapidity and ease with which mothers responded to and assuaged their infants’ distress. More recent approaches emphasise ‘affect regulation’ and ‘mentalising’, and the idea that to the extent that these are available in the parent–child relationship in infancy and early childhood, they become gradually internalised by the growing child.
Affect regulation entails the ability of the parent to identify a child’s emotional states and to mobilise a regulatory appropriate response. A crucial aspect of ‘mentalising’ is the parent’s ability to see her child not as an extension of herself and her own needs, but as an autonomous sentient being with his or her own projects, desires and wishes. Mentalising is thus contra-narcissistic, and implies the ability to find an internal vantage point from which to view and reflect on one’s own and others’ thoughts, emotions and actions.
If we postulate a ‘parallel process’ (a concept borrowed from the supervisortherapist/therapist-patient constellation, cf. Holmes 2012) between therapist–parent and parent–child relationships, then the therapist’s ‘sensitivity’ to the presenting parent will be a mutative ingredient in helping the parent to be more sensitive to her child. Therapists need to tune into and identify verbally parents’ feelings, to contain and hold them, to acknowledge and soothe where appropriate. A further aspect entails including the parent’s partner, if there is one. An ill child often drives a wedge into a family, prising couples apart. The needs of the other siblings are often overlooked. Working with the parent of the IP will also mean focusing on the couple relationship, and helping the parent to recruit her partner (usually the father) as a Secure Base to whom she can turn with her worries and troubles, and who will also take over when she herself is miserable or exhausted.
The neuroimaging data on the ‘disgust’ reaction on the part of the parent when failing to elicit an expected positive response from their child, contrast with secure mothers who can accommodate and encompass their children’s unhappiness. In insecure relationships the parent might say something like, ‘Look, I do everything for that child, drive him/her to school, supply his meals, bail him out when he’s got no money, cook wash and clean – and all I get is sulks and monosyllables. It’s intolerable! He’s got to get his act together’. This might be a typical avoidant child–parent pattern; the child gets a degree of protection, but with diminished parental ‘sensitivity’. The price paid is the suppression or diverting of distress, rather than its resolution, with both parent and child’s affect unassuaged. All this goes on non-consciously: avoidant children show physiological evidence of arousal in the SSP as manifest by elevated cortisol levels, and tachycardia (Dozier & Rogers Kobak 1992) although outwardly they appear equable.
Therapeutic work informed by this would concentrate: (a) on the therapist acknowledging and accepting parents’ worry, misery, and ‘disgust’, (b) helping the parent similarly to ‘name and contain’ their children’s difficult feelings. A typical interaction with, say, an eating-disordered teenager might be:CHILD:
I hate myself, I’m fat, fat, fat …PARENT:
You’re NOT fat – you’re thin! If you just put on some weight you’ll stop feeling so bad about yourself”.
An attachment-guided alternative response (which could be tried out in roleplay) might be:PARENT:
It sounds like you’re feeling really terrible and unhappy today … shall we have a look at those diet websites together and see if we can work something out? …
An interesting finding that emerged from the Bakermans-Kranenberg et al. (2005) meta-analysis was that video-feedback was particularly effective in enhancing both maternal sensitivity and infant security. Observing oneself on video is inherently mentalising, if the latter is viewed as the ‘ability to see oneself from the outside and others from the inside’ (Holmes 2012). Since arousal is inimical to mentalising, the stressed parents of ill children are often too agitated to be able to reflect on the impact they are having on the child, let alone the past traumas the child’s illness may be activating from their own childhood.
Quietly observing oneself on video, however painful it might be, provides such an opportunity for observation and learning. In one-to-one therapy, ‘role-play’ with a therapist can work in similar ways. The therapist might move from conventional dialogue into role-plays in which the therapist alternates between enacting the parent’s role, with the mother ‘being’ the child, and vice versa. The same scenario can be repeatedly played this way, with participants alternating: playing it ‘as it was’, and then with a different hoped-for outcome. It is important, however, to observe the dictum that when working with parents, a therapist should never make them feel that he or she makes a better job of being a parent than they do. 1
A final point comes not from attachment research but from a systemicpsychoanalytic perspective. It concerns the ‘location’ of the presenting problem. A systemic point of view sees the family as a system, with individual members playing parts in a drama where the target dysfunction is the family itself, rather than the individuals who make it up (cf. Lock et al. 2010).
This is consistent with the psychoanalytic notion of ‘projective identification’, in which feelings, conflicts and desires may be transferred unconsciously from one member of an intimate system to another, typically in the Kleinian model from baby to mother, but equally, and especially when problematic, vice versa. Seen this way, ‘badness’, and especially negative affect, is no longer located in any one individual, but in the system as a whole. Thus when a patient refers to the needs of her children – for instance is worried about ‘child care’ for her baby – it is a useful heuristic to think that it is the patient’s ‘inner child’ that needs therapeutic attention. Working with parents of ill children, part of the therapeutic task is to help the parent grasp and identify her own un-met childhood needs, manifest in her child as she perceives him.
According to maestro (a degree of ‘cult of personality’ is admissible!) Daniel Barenboim (Barenboim & Said 2004) ‘music’ arises at the intersection of two axes: vertical, instantaneous and harmonic; horizontal, temporal and melodic/rhythmic. Comparably, families have their own music. At any given moment there are harmonies and discords whose overtones reverberate throughout the system. At the same time family members are bound together in an ongoing story or ‘script’ (Byng-Hall 1999) which is transmitted and modified from generation to generation. Both melody and harmony have attachment overtones.
Early psychoanalytic theorising tended to see the developmental process in terms of two generations – the IP, and his or her parents. Fraiberg’s (1980) ‘ghosts in the nursery’ acknowledged that at any clinical situation a minimum of three generations are involved – the patient, the patient’s parents, the parents’ parents, and if the patient is a parent, their children. Lacan (Leupnitz 2009) similarly shows how at the moment of birth, a child has a pre-assigned ‘place’ in the family and parental system/psyche, and that the child, and his or her difficulties, must be seen as a link in a family chain stretching back into the past and forwards into the future.
Therapists tend to be highly attuned to these self-perpetuating patterns of pathology, but sometimes miss an opposing trend – the inherent self-righting capacity of both individuals and families. A parental generation plagued by abuse might, by the third generation, be once more on a more secure track. The key to this selfrightingness is the capacity of parents to mentalise their own difficulties, and thus to hold them in check as they work for the best for the next generation. It is the task of the science and art of psychotherapy, and especially the mentalising approach based in the attachment paradigm, to foster this self-righting process, and to study those conditions that make it more, or less, likely to be achieved.
The author often gets round this by comparing watching an ‘action replay’ of a missed open goal in football on the television – easy to get it right from the comfort of one’s armchair; equally easy to miss in the real-life heat of the moment.