Attachment disorganization and borderline personality disorder

An impressive range of theoretical inquiries, clinical studies, and controlled research studies suggest that attachment disorganization plays an important role in borderline psychopathology, although it is not a specific risk factor for BPD (for a review of these studies, see Liotti, 2014). These studies support the idea that the fundamental features of BPD can be explained by a developmental model based on attachment disorganization. Although we lack conclusive research evidence for the hypothesis that the developmental pathways leading to the disorder begin with early attachment disorganization in the majority of BPD cases (Levy, 2005), three controlled studies suggest that this may be the case (Carlson, Egeland, & Sroufe, 2009; Lyons-Ruth, Bureau, Holmes, Easterbrooks, & Brooks, 2012; Lyons-Ruth, Melnick, Patrick, & Hobson, 2007).

Reviewing and summarizing the studies that support the hypothesis of the key role of attachment disorganization in the genesis of BPD is beyond the scope of this chapter. Rather, we shall first provide readers with an overview of the nature and basic features of infant attachment disorganization and of its developmental sequelae, having in mind the final goal of using them in understanding the early developmental roots of painful incoherence in the experience of self in patients with BPD.

Disorganization of infant attachment

About 80% of infants’ attachments to the caregivers in low-risk samples can be reliably classified using the Strange Situation procedure (Ainsworth, Blehar, Waters, & Wall, 1978) into three main organized patterns (secure, insecure-avoidant, and insecure-resistant). Most of the remaining attachment styles, being characterized by lack of behavioral and attentional organization, are now classified in the disorganized category (Lyons-Ruth & Jacobvitz, 2008). In samples of families at high risk for psychopathology, the percentage of disorganized attachments may be as high as 80% (Lyons-Ruth & Jacobvitz, 2008).

Infants with disorganized attachment manifest bizarre and/or contradictory behavior when reuniting with their caregiver after a brief separation: bizarre behavior such as freezing, hiding, head-banging, abrupt lowering of the muscular tone ending sometimes with the baby collapsing to the ground, and contradictory behavior such as trying to approach the attachment figure with their head averted (Main & Solomon, 1990). Unresolved experiences of losses and traumas, as retrieved through the caregiver’s Adult Attachment Interview (AAI; Hesse, 2008), are a frequent precursor of disorganized attachment in the infants, and are significantly less frequent in the caregivers of infants with organized attachment patterns (for a meta-analysis of research on this topic, see Van IJzendoorn, Schuengel, & Bakermans-Kranenbourg, 1999). An important mediating factor between the caregiver’s unresolved state of mind and the infant’s attachment disorganization is that the infant’s fear is increased rather than soothed in the attachment-caregiving interactions. Studies of infant attachment disorganization describe parental behavior, linked by unresolved trauma and loss, that is either frightened and thereby indirectly frightening, or aggressive and directly frightening to the infant (Main & Hesse, 1990; Hesse & Main, 2006). Other adverse influences in the caregivers’ past attachment experiences, besides unresolved traumas or losses, have also been evidenced as antecedents of infant attachment disorganization. These antecedents are expressed in hostile and helpless states of mind concerning the attachment-caregiving interaction (Lyons-Ruth, Yellin, Melnick, & Atwood, 2003), and in the “abdication” of the responsibility of caregiving in the face of the infant’s expression of attachment needs (Solomon & George, 2011).¹

Although the type of interaction between the infant and the caregiver plays the key role in infant attachment disorganization, genetic influences exert a moderating influence (Gervai, 2009). Reflections on gene–environment interaction in infant attachment disorganization may contribute to reconciling genetic and attachment-based theories of BPD (Gunderson & Lyons-Ruth, 2008).

Attachment disorganization and dissociative processes

The term “dissociation” in psychopathology is used to signify both a diagnostic category and the pathogenic processes caused by traumatic experiences that, by hindering mental functions involved in the integration of experience and self-perception, generate the dissociative symptoms. These symptoms characterize the dissociative disorders and, according to Meares (2012a) and others (e.g., Farina & Liotti, 2013; Howell, 2008), also BPD. The higher-order integrative functions hindered by dissociative processes are consciousness, self-identity, memory, perception, and those involved in the control of bodily movements (Nijenhuis & Van der Hart, 2011). Many researchers and clinicians have extended the effects of dissociative processes to other integrative mental functions, typically altered in BPD patients, such as affect regulation, metacognitive monitoring, mentalization,² and the capacity for coherent autobiographic narratives (Carlson, Yates, & Sroufe, 2009; Farina & Liotti, 2013; Fonagy & Bateman, 2008; Meares, 2012a; Van der Hart, Nijenhuis, & Steele, 2006). This extended view of the potentially wide-ranging effects of trauma on higher mental functions can be seen as the heritage of Pierre Janet’s theory of posttraumatic désagrégation (Van der Hart & Dorahy, 2006) – Janet’s favorite synonym of the word “dissociation,” which can be rendered in English as “disintegration.”

Liotti’s (1992) hypothesis that dissociative processes underpin infant attachment disorganization is supported by two longitudinal controlled research studies (Dutra, Bureau, Holmes, Lyubchik, & Lyons-Ruth, 2009; Ogawa, Sroufe, Weinfield, Carlson, & Egeland, 1997). These studies provide robust evidence that children and adolescents who had been disorganized in their infant attachments are more prone to dissociative mental processes than are their peers who have histories of organized (secure, insecure-avoidant, and insecure-ambivalent) early attachments. The vehicle bringing the dissociative tendencies of infant disorganized attachment into adulthood is, according to attachment theory, the internal working model (IWM: Bowlby, 1969). Let us now have a look at how the kernel of the disorganized IWM can be conceived.

An attachment figure who is neglecting, helpless, frightened, or hostile and straightforwardly frightening to the infant creates a situation in which the source of potential comfort is also, at the same time, the source of fear, even when this caregiver’s behavior is not obviously maltreating. This situation has been called fright without solution (Main & Hesse, 1990), because infants cannot find relief from fear in flying from the caregiver nor in approaching her or him. The experience of fright without solution in infant disorganized attachment has been regarded as an early relational trauma (Schore, 2009) because, like other traumatic experiences, it involves powerlessness in the face of an unbearable and inescapable life-threatening situation. Trauma, by definition, involves overwhelming the capacity to cope successfully with a threat through fight or flight defensive responses (supported by the adrenergic arousal of the sympathetic system) so that the only response that remains available is the archaic vagal one – the vasovagal syncope evolved to provide an extreme attempt at self protection under inescapable life-threatening circumstances (Porges, 1997, 2001).³ The activation of this vagal response causes cataplectic immobility and a shutdown of higher brain connections affording, besides protection from pain, in the case of a predator’s attack, the last desperate defense from a predator’s attack in the form of feigned death (i.e., the only fugue when no other fugue is possible). The activation of the archaic vagal defense causes disconnection between the different functional levels of the mind, prevents the integration of traumatic memories, and causes discontinuity and fragmentation in the experience of self (Schore, 2009). Such a fragmentation is expressed by the bizarre and contradictory behaviors of disorganized infants in the Strange Situation procedure and, in the AAI, by the incoherence of thought and discourse characterizing the adult mental state related to attachment disorganization.

The clinical relevance of understanding the developmental trajectory that leads from infant disorganized attachment to adult BPD justifies theoretically informed speculations on the features of the contradictory and non-integrated representations stemming from the disorganized IWM. Liotti (2004) suggested that they are akin to the three basic roles of Karpman’s (1968) drama triangle: the powerful rescuer, the equally powerful but malevolent persecutor, and the powerless victim. Being at least potentially available and willing to help and comfort the infant, parents and other caregivers are perceived by children as rescuers. At the same time, when they are neglecting, subtly hostile, or prone to episodes of aggression, they are perceived as persecutors. Simultaneously, because they express their helplessness, fear, and suffering (caused by their own unresolved traumatic memories) while taking care of their infants, the parents of disorganized children are perceived as victims. These reciprocally incompatible representational prototypes are the base for construing the behavior of self and others during later attachment interactions. Being constructed during the first two years of life, these representations pertain to the non-verbal domain of inner representations – that is, they operate at the implicit level of self-knowledge (Amini et al., 1996). In other words, they are aspects of the ongoing implicit relational knowing that characterizes the early phases of personality development and persists throughout the life span (Lyons-Ruth, 1998). Therefore, throughout the developmental years, the multiple and non-integrated representations of the self and of the single caregiver manifest themselves in communication as intersubjective enactments rather than as explicit verbalized structures of memory (Ginot, 2007). No synthesis of these representations in semantic memory and in fully conscious narratives is therefore possible, at least not during childhood. The different, incompatible, simultaneous representations of self-with-other of the disorganized IWM tend to remain compartmentalized throughout the early phases of personality development. Compartmentalization, it should be remembered, is one of the two basic aspects of dissociation (Holmes et al., 2005), the other being detachment (expressed mainly in the manifold symptoms of depersonalization).

The compartmentalized representations of disorganized attachment, together with the dramatic re-experiencing of fear without solution during later attachment interaction, tend to hamper the higher (conscious and regulatory) mental functions during personality development, so that mentalization deficits, emotional dysregulation, and impulsivity may also follow infant attachment disorganization (Bateman & Fonagy, 2004). It should be emphasized that both dissociation among representations of self-with-other and mentalization deficits tend to occur, in people with disorganized attachment, during the experience of attachment needs and wishes rather than in moments where interpersonal behavior is motivated by systems different from attachment (e.g., the competitive, the sexual, the care-giving, or the cooperative systems: Liotti, Cortina, & Farina, 2008; Liotti & Gilbert, 2011).

There is some evidence that a disorganized IWM in infancy may be the first step in the developmental psychopathology of BPD. A longitudinal research study evidenced correlations between early disorganized attachment and adult BPD symptoms (Carlson et al., 2009). A controlled study comparing the AAI coding hostile/helpless (linked to infant attachment disorganization) in samples of dysthymic and borderline female patients shows the expected statistically significant difference (Lyons-Ruth et al., 2007). In the prospective study by Lyons-Ruth and her collaborators (2012), a relational antecedent of attachment disorganization (maternal withdrawal in infancy) significantly predicted borderline symptoms and suicidality/self-injury in late adolescence. Also in keeping with the hypothesis that early attachment disorganization plays a role in BPD are the findings of a neuroscience study evidencing a disintegrative effect of attachment memories on cortical EEG connectivity in adult patients with histories of chronic childhood trauma and a disorganized mental state related to attachment (Farina et al., 2014).

Developmental sequelae of infant attachment disorganization

Remarkably, disorganized attachment in infancy develops into rigid, controlling behavior in middle childhood (Lyons-Ruth & Jacobvitz, 2008). The controlling strategies seem to compensate for disorganization in the child–parent interactions: they allow for organized interpersonal exchanges with the caregivers, thus reducing the likelihood of dissociative processes during these exchanges (Liotti, 2011). There is evidence (Lyons-Ruth & Jacobvitz, 2008) that infants disorganized in their attachments can either become bossy children who strive to obtain dominance by exerting aggressive competitiveness toward the caregiver (controlling–punitive strategy), or become children who invert the attachment relationship and display precocious caregiving toward their parents (controlling–caregiving strategy). A major cause of the controlling caregiving strategy is the relationship with a vulnerable, helpless parent who encourages the child to invert the normal direction of the attachment–caregiving strategy. A parent who perceives the child as powerful and evil may be one particularly malignant condition for the development of a controlling–punitive strategy (for examples, see Hesse, Main, Abrams, & Rifkin, 2003).

The controlling strategies collapse in the face of events (e.g., traumas, pain, threats of separation), that stimulate intensely and durably the child’s attachment system (Hesse et al., 2003). During the phases of collapse of the controlling strategies, the child’s thoughts and behavior suggest that dissociative processes are at work, presumably because of the reactivation of the disorganized IWM (Hesse et al., 2003; Liotti, 2004, 2011). It is noteworthy, for our understanding of the developmental psychopathology of BPD, that children with a controlling–punitive strategy are more prone than other children to develop externalizing disorders characterized by impulse dyscontrol, while children with a controlling–caregiving strategy tend to develop internalizing disorders, characterized by anxiety and depression (Moss et al., 2006). It can be hypothesized that a controlling–punitive strategy mediates between infant attachment disorganization and adult cluster B personality disorders (including BPD), while a controlling–caregiving strategy is a risk factor for anxiety disorders, mood disorders, or cluster A personality disorders.

While there is no evidence for the hypothesized different pathways of developmental psychopathology toward personality disorders being laid open by the two disorganized/controlling strategies, in the above-quoted longitudinal study of Lyons-Ruth and her collaborators (2012) a disorganized/controlling strategy at age 8 contributed to the prediction of borderline symptoms independently of later traumatic experiences. Borderline personality functioning in adolescence, presumably including painful incoherence in the experience of self, are thus predicted both by disturbed interactions as early as 18 months of age and by later controlling strategies.

In summary, it is reasonable to conclude from the existing data that infant disorganized attachment may lead to adaptational vulnerabilities (e.g., a controlling–punitive strategy developed in middle childhood) which, especially as a consequence of further traumatic experiences, can cause BPD. However, because disorganized attachment can also be an antecedent of other disorders (Levy, 2005; Liotti, 2014), specific developmental pathways must lead to other types of adaptational vulnerabilities linked to adult disorders different from BPD.

Notes on the developmental psychopathology of borderline personality disorder

Borderline psychopathology is the likely consequence of multiple, intertwined relational and mental processes, and is probably influenced by specific temperamental factors. The disintegrative influence of infant attachment disorganization, especially if it is followed by later traumatic experiences, may explain the genesis of the core feature of DBP (painful incoherence in the experience of self), and also the roots of other features: emotional dysregulation, a mentalization deficit, coexisting fears of attachment loss and of affectional closeness (Van der Hart et al., 2006), and the shifting between idealized (attachment figure seen as a powerful rescuer) and aggressively destructive (the same figure perceived as a persecutor) representations of significant others.

Being linked to the disorganized IWM, all these dysfunctions are typically triggered by the relational context in which the patient’s attachment system is activated (Liotti et al., 2008; Liotti & Gilbert, 2011). Given this, we should study the peculiarities of these relational contexts for a better understanding of the specific developmental pathways that lead to BPD rather than to other mental disorders whose genesis is potentially linked to infant attachment disorganization. The different responses of the interpersonal environment to the frequent moments of dysfunctional behavior of children who have been disorganized in their infant attachments may influence whether the final result of early disorganized attachment is mental health, BPD or another mental disorder (Liotti, 1992, 2004). For instance, the child’s (and later the adolescent’s) shifting from the fear of attachment loss to fear of emotional closeness may be met by significant others with a sufficient degree of understanding (facilitating a healthy growth of the personality), by rage or violent aggression leading to severe trauma (a likely antecedent of dissociative identity disorder), or with a rather chaotic admixture of criticism, withdrawal, anger and sometimes oversolicitous attitudes that, in our clinical opinion, is more likely to be found in the personal history of patients with BPD. Such contradictory responses of significant others circularly confirm both the idealization of significant others and the fear of being intruded upon, abandoned, or maltreated – typical features of BPD.

The discovery of untoward ways to shut down the recurring mental suffering linked to the experience of painful incoherence – such as self-injurious behaviors (self-cutting, self-burning, and food or substance abuse) – can also pave the way to a full-blown BPD picture.