Introduction

Over the past 12 years child and adolescent mental health services (CAMHS) (community outpatient clinics) in South Australia have been delivering, researching and refining a structured dyadic psychotherapy, parallel parent and child therapy (P-PACT) to high-risk mothers and their children (aged 3–12 years), experiencing severely distressing, maltreating relationships. The P-PACT process targets the mother–child relationship, reflecting its initial development within an attachment theory framework (Chambers et al. 2006).

The mothers offered P-PACT have experienced chronically injurious relationships with their mothers and other important caregivers, as infants/growing children. They continue to be affected by these experiences, attracting trauma related mental health diagnoses such as complex posttraumatic stress disorder (PTSD), dissociative disorders and borderline personality disorder. These presentations are complicated by chronic anxiety, dysthymia and depression. The children exhibit severe mixed disturbances of conduct and emotion at home and at school. The mother–child relationships are complex and changeable. The mothers sometimes attempt to control their children with physical and emotional aggression and harsh punishments, imposing strict rules. At other times the mother’s hide from their children in fear, give in to children’s demands, disengage from them, or use them as adult-like confidants. Their children move between aggressive, controlling, oppositional behaviours; symptoms of hyperarousal such as poor sleep, separation anxiety and extended periods of un-smoothable distress; and appearing over-helpful and solicitous. The mother and child are both extremely distressed by the ongoing struggles in their relationship, equally wanting to flee the relationship and longing to be together in a way that is more joyful.

Attachment theory has been powerful in understanding the range of relationship patterns that develop between mother’s and their infants and growing children. When developing attachment theory Bowlby drew on ethology, using detailed descriptions of patterns of behaviour in birds and mammals to understand the motivation (or instinct) underlying the close emotional bonds between mothers and their infants. He described an inbuilt biological motivational system, which he referred to as the attachment system. Bowlby hypothesised that this was just one of a number of motivational systems that would become important in understanding human behaviour (Bowlby 1969/1982). Bowlby explained these patterns of behaviour as adaptations which promoted survival in man’s “environment of evolutionary adaptiveness”, the environment in which instinctive behavioural systems of survival would have developed over the first two million years of human existence (Bowlby 1969/1982, p. 59). In this way Bowlby highlighted the survival advantages of the enduring emotional bond between mothers and their infants/growing children.

Recent extensions of attachment theory have documented specific attachment-caregiving relationship patterns in high-risk mothers and infants/children, systematically coding videotaped relationship-based interactional patterns. These patterns include inverted caregiving relationships in early childhood (Macfie et al. 2008); controlling hostile-punitive and controlling caregiving behaviors in school age children (Main and Cassidy 1988); and hostile-helpless states of mind in mothers, based on the coding of adult narrative descriptions of early attachment relationships (Lyons-Ruth et al. 2005). Key features of these patterns are disturbances in the natural hierarchy between adult caregivers and their children, and the struggles for control and dominance: also characteristics of our mother–child dyads. Whilst the survival advantages of close and attuned relationships are clear, what is less clear is whether these coercive relationship patterns can also be considered adaptive when viewed from an evolutionary perspective.

The aim of this paper is to investigate this question. The paper is divided into 6 sections. In “Using Attachment Theory to Understand High-risk and Maltreating Mothers and their Children” we summarise our use of attachment theory to understand high-risk mother-child relationships, focusing in particular on our earlier conceptualization of the pathways and mechanisms that lead to ongoing maltreatment. In “Gaps and Inconsistencies in the Attachment and Motivational Systems Driven Model of Maltreatment" we identify the gaps and inconsistencies in our understanding of this pathway to maltreatment, particularly how it relates to other theoretical accounts. In “Modes of Group Living in Primates: Implications for Human Relationships” we describe key findings from ethological research regarding modes of group living (agonic and hedonic) in primates and how these findings inform our understanding of human relationships. In “Application of Agonic Mode and Hedonic Mode to Maltreating Mothers and their Children” we use the findings to refine our understanding of the relationship dynamics in maltreatment. In “Clinical Implications" we explore the clinical implications of this work. In “Discussion” we finish with a discussion of the value of these ideas for reconciling theoretical inconsistencies in the literature, and point to the potential for these ideas to inform other clinical areas (e.g. domestic violence) and understanding adult survivors of traumatizing attachment relationships in childhood.

Using Attachment Theory to Understand High-risk and Maltreating Mothers and Their Children

In a secure care seeking, care-giving mother-infant relationship the mother is the first, most important resource for her infant. A security-generating mother responds in a sensitive manner to her infant’s needs by providing protection and material, social and intellectual resources in a predictable and timely way. For example, a mother helps her infant regulate the fear and distress related to the infant’s physical and emotional needs such as hunger and loneliness, when they arise and as signaled to her by her infant (Sroufe et al. 2005).

In a previous publication we described a model of the pathways leading to the maternal maltreatment of children aged 3–12 years, integrating perspectives from attachment research, learning theory, and models of trauma and dissociation (Amos et al. 2011). To support the arguments presented in this paper, the model will be summarised here. A complete description of the model is beyond the scope of this article and readers are referred to Amos et al. (2011) for further explication.

Over 80 % of infants in maltreating mother-infant dyads, where the mother is both frightened of and/or frightening to the infant develop a disorganised style of attachment (Cyr et al. 2010). In this situation, the mother is both the source of comfort and protection and the source of threat for the infant. The infant must seek proximity to the same person that they must also defend themselves against, giving rise to contradictory needs to approach and flee from the same person. In response, the infant experiences incompatible, distressing somatic-emotional and behavioral responses to the mother that cannot be integrated into a coherent internal working model of how to get help when fearful and distressed. This leaves the infant in an enduring state of fright without solution (Hesse and Main 2000). Drawing on Van der Hart et al. (2006; Amos et al. 2011) we suggested that these contradictory and incompatible attachment-related experiences become isolated (or structurally dissociated) from the rest of the emerging personality, persisting into adulthood in implicit memory, in a relatively unchanged form. We further argued, following Liotti (2008, 2011), that other biological motivational systems are activated in the place of the attachment system to manage distress, in particular the motivational systems referred to by Liotti (2008, 2011) as the “competitive ranking” and “caregiving” motivational systems. In the final part of the model we discussed how the structurally dissociated somatic-emotional experiences could be re-activated when the traumatized adult enters motherhood. We proposed that the intense emotional needs of a new infant reawaken the mother’s unmet needs for closeness, care and protection, and her agonizing attempts to find solutions to the unsolvable fear in her own infancy (Amos et al. 2011). During this traumatic reactivation the mother experiences her infant as the source of her fear and distress (source attribution error) rather than understanding the distress as an emotional memory of her own infancy and early childhood. Maltreatment, in terms of aggression towards or withdrawal from the infant, becomes the means for the mother to resolve her distress (Amos et al. 2011). As a result, the mother becomes an actual threat to the infant. Together the infant and mother face the dilemma of being both terrified of and inextricably tied to each other: the mother through social and maternal responsibility and the infant as a matter of survival.

Gaps and Inconsistencies in the Attachment and Motivational Systems Driven Model of Maltreatment

On reflection we considered that the explanation for the controlling caregiving and hostile punitive behaviours in school age children adopted in our model was inadequate. First, the explanation did not help us understand the possible survival advantages of these behaviours, which would explain why these behavioural strategies are widely adopted by high-risk mothers and children. Second, the theoretical work of Liotti (2008, 2011) and the empirical and theoretical work of Lyons-Ruth et al. (2006), which we drew on in developing the model, embody inconsistencies about the internal working models that underpin the observed behaviours. Liotti theorizes that the controlling hostile-punitive and controlling caregiving behaviors seen in children correspond to the activation of the competitive social ranking system and the caregiving system respectively. His theory embraces the idea that additional motivational systems may be involved in the genesis of other observable behaviors in such children; (for example that the sexual pair bonding system may explain sexualized behaviors) (Liotti 2008). Liotti’s position thus infers the persistence, from infancy into childhood, of multiple incompatible internal working models of relationship. In contrast, Lyons-Ruth’s empirical research has led her to hypothesize the formation of a single internal working model of relationship during early childhood, which is characterised by two opposing poles, hostile and helpless. She argues this explains the inverted caregiving and hostile behaviours in children, the hostile helpless states of mind in adults and the maternal emotional communication errors leading to attachment disorganization in infants (Lyons-Ruth et al. 2005).

In an effort to increase the internal consistency of our model and to deepen our understanding of the mother–child dyad’s adaptations to severe ongoing distress, we re-examined historical and recent ethological research. Our focus of study was threat management in group-living primates, as models for interpersonal behaviours in humans in their primary attachment relationships. The results of this research are reported in this paper.

Modes of Group Living in Primates: Implications for Human Relationships

Access to food and shelter, to relationships with caregivers, to sexual partners, to peers and allies, and safety from predators are potent drivers for survival (and thus natural and sexual selection). One strategy for increasing safety from predators and securing resources is by living in groups and working collectively. Group living, however, invariably leads to rivalry and conflict between group members. Handling tension between individuals using individual defensive strategies (such as withdrawal to isolated territories typical of solitary species) is not viable for group-living species, as a key requirement is for groups to remain intact. Group membership offers subordinates the protection and resources that may be provided by the dominant and benefits the dominant by preserving kin and numbers required for group stability. Thus specific strategies that reduce conflict, foster group cohesion and allow individuals to remain in close proximity, whilst sharing scarce resources have had to evolve to support group-living (Hawley 1999; Cantor 2005).

Modes of Group Living in Primates

Michael Chance employed ethological research techniques of “non interventionist observation and narrative description” to investigate the social behaviours of primates used in laboratory research (Kirk 2009, p. 13). Chance documented two distinct modes of relationship employed by adult members of higher primate troupes such as Macaques and Baboons, to manage the tension related to group living. He referred to these alternative ways of managing threat whilst maintaining group cohesion as agonic and hedonic modes (Chance 1986; Gilbert 1992).

Chance observed that while many primate groups are organised around a dominant male, the relationships in hedonic and agonic modes have a different quality. Relationships in hedonic mode are cooperative, egalitarian, and mutually supportive; group members are relaxed and non-defensive, and can focus their attention beyond the relationships within the group (Chance 1986). By comparison, relationships in agonic mode are chronically tense. Instead of being egalitarian, relationships reflect a dominant, subordinate polarity. Individuals are highly defensive, vigilantly focusing their attention on the more dominant adult group members. Individuals remain braced and ready to respond to rising tension with displays of ritualized aggression and submission to maintain rigid hierarchies (Chance 1986; Gilbert 1992).

Chance also identified a behavioural strategy referred to as “reverted escape” which allows weaker or subordinate animals to remain part of the group (Chance 1986, 1988). After being attacked or threatened, subordinate individuals return to the dominant animals, in submission. Both dominant and subordinate animals engage in affiliative behaviors such as hugging and grooming to facilitate reconciliation. This reinforces the emotional bond between them, both increasing group cohesion and reinforcing the hierarchy (Chance 1986).

Cantor and Price (2005, 2007) have extended this work, focusing more broadly on the use of the mammalian defensive strategy of appeasement. Appeasement involves the submissive pacification of an often dominant other, to deter further attacks or threats from the dominant. Appeasement behaviors in mammals take many forms. “Reverted escape” (the seemingly paradoxical retreat or return to the oppressive dominant in the place of true escape), is one form of appeasement. Other forms include infantile mimicry, cowering, bowing, kneeling; prostration, sickness behavior and self-destructive behavior all of which involve signaling subordinate status. Sexual behaviors constitute another approach to appeasement that is not necessarily orientated to signaling inferiority.

Kortmulder and Robbers (2005) are ethologists who have re-examined the concepts of agonic and hedonic modes, proposing new operational definitions to facilitate research into the application of agonic and hedonic modes to species other than higher primates (including birds and fish).

Understanding how Kortmulder and Robbers define tension in animal systems is integral to understanding their definitions of agonic and hedonic modes. In their description, tension results from combinations of “uncertainty and insecurity” (Kortmulder and Robbers 2005, p. 22). Uncertainty refers to a situation where it is “subjectively impossible to know what is coming” and insecurity, to an “absence of togetherness and safety” (Kortmulder and Robbers, 2005, p. 22). From this standpoint, the agonic and hedonic modes are understood as ways to either; increase safety and togetherness, or reduce uncertainty. Their definitions of agonic and hedonic mode also cover the range and flexibility of behaviour; the breadth of attention; the flow of information within the system and, symmetry versus rigid hierarchy.

In hedonic mode relationships are characterized by varied interpersonal roles and behaviours where movement between roles is fluid and flexible, and interactions are typically playful and tender, fostering trust and togetherness. Attention is broad and inclusive and information flows freely. Group members, regardless of their place in the hierarchy, exhibit a similarly wide range of behaviours. Groups in hedonic mode are said to demonstrate “higher symmetry” (Kortmulder and Robbers 2005 p. 24).

Agonic modes predominate in contexts of danger and threat. Behaviors are constricted in range following predictable and ritualized sequences. Attention is narrow and threat focused and the flow of information is similarly reduced. When tension builds between individuals, one participant aggressively asserts their dominance, establishing or re-establishing an asymmetrical dominance and submission hierarchy. Dominants and subordinates exhibit different sets of behaviours and relationships are therefore asymmetrical in terms of the available behavioural responses. Maintenance of this hierarchy depends on the participants adhering to the rigid and restricted roles and responses; reducing tension by increasing predictability and reducing uncertainty. However, as behavior is never entirely predictable, a state of inter-individual relaxation is never fully achieved and vigilant distrust becomes the basic currency of the group’s interactions. The cost associated with agonic mode is the loss of “the intensity and continuity of social contact” (Kortmulder and Robbers, 2005, p. 1). Kortmulder and Robbers also identify a state that they have named agon, an unstable, high-tension transitional state between agonic and hedonic mode.

Application of Agonic and Hedonic Modes to Human Beings and their Relationships

Chance’s work on agonic and hedonic modes has made a highly significant contribution to the understanding of human behaviour but has not achieved the same prominence as attachment theory. In a series of papers Chance (1984, 1986, 1988, 1992) made a number of interesting observations. He noted that the quality of the interaction between mother and infant influences which mode later dominates an individual’s relationships; that children as young as 2 years of age can be classified as hedonic leaders or as agonic (tending towards aggression or escape manifest in social withdrawal or over compliance); that children in agonic modes hold a continuous internal tension between conflicting needs for proximity and escape and that being in agonic mode in childhood is the foundation of what Chance refers to as the adult agonic personality.

Chance describes the agonic adult personality as manifesting in two different forms; those occupying a dominant status in society, who exert authority, but are also authoritarian and seek to over control or overawe others: and “those whose reactions are dominated by escape tendencies, becoming over-dependent, exaggeratedly obedient or withdrawn from reality” (Chance 1984, p. 151). We note that this characterization is remarkably similar to descriptions of the hostile/helpless mothers (Lyons-Ruth et al. 2005) and controlling hostile punitive and controlling caregiving children (Main and Cassidy 1988).

Cantor, utilizing Chance’s work has hypothesised that PTSD not only represents an enduring over-activation of in-built mammalian defense systems but also an enduring switch to a reliance on agonic mode (Cantor 2005). In addition Cantor and Price postulate that an over reliance on the defense of appeasement, helps explain some of the interpersonal behaviours observed in complex PTSD (which results from chronic interpersonal trauma), Stockholm syndrome (the outwardly paradoxical attachment between kidnappers and their hostages) and domestic abuse (Cantor and Price 2007).

Kortmulder and Robbers’ (2005) revised definitions of the modes, combined with the work of Chance and that of Cantor and Price offer a promising lens through which to examine children’s adaptations to maltreating mother–child relationships.

Application of Agonic Mode and Hedonic Mode to Maltreating Mothers and their Children

Lyons-Ruth et al. (2006) have stressed the importance of contradictory maternal emotional communications with infants in the earliest months of life, in the genesis of attachment disorganization. They consider that these disrupted emotional communications are underpinned by maternal hostile/helpless states of mind with regards to attachment, which in turn are outgrowths of the controlling hostile and caregiving strategies in school age children (Lyons-Ruth et al. 2006). The formulation presented in this paper equates the mother’s pervasive, contradictory emotional communications with her infant, with the use of agonically driven, dominance and submission hierarchies and the defense of appeasement. We further hypothesize that where the mother continues to be both needed and threatening during the first years of their child’s life, children (subordinates) need a way of maintaining proximity to the mother (dominant) in the face of the ongoing fear that they experience and that they also fall back on the solution of ritualized hierarchy formation. These strategies act as replacements for attachment-caregiving-based safety and togetherness as the way of maintaining much needed relationship cohesion.

We further propose that the enduring switch to agonic mode is the fundamental adaptation and that agonic mode underpins the activation of any motivational system activated in mother and child, at times of rising tension in their relationship. Logically it follows that the complex, contradictory responses exhibited by mothers whose infants become disorganised (e.g. failing to respond to an infant’s distress, withdrawing from the infant, hostility and intrusion, seductive sexualisation and tender sympathy (Lyons-Ruth et al. 2006; Lyons-Ruth 2008) and the similarly complex and contradictory behaviours seen in these infants as they develop into children, can all be understood as expressions of dominance and submission and appeasement.

Accordingly, the hierarchical structure of the mother–child relationship and the four facets of the definitions proposed by Kortmulder and Robbers most usefully define these distressing mother–child relationships regardless of the motivational system activated. The implications for the child’s development and for the relationship between the mother and child are now explored.

Asymmetrical Hierarchy and Appeasement

The natural hierarchy in the good-enough mother infant relationship has been succinctly captured in the description of security-promoting mothers, as “bigger, stronger, wiser and kind” (Cooper et al. 1998). The mother is dominant (big and strong and the holder of resources) and mostly operating in hedonic modes (wise and kind). The infant is subordinate, and in need of resources such as nurturance and protection.

In agonic relationships, the mother’s behavior is less nurturing and more competitive in orientation. Her behaviour does not match with her role as dominant protector, provider and nurturer; but is constricted by an over-riding focus on her own needs. She thus fails to provide the wise and considered care that her infant/child needs. She may be helpless, overwhelmed, frightened and withdrawn, overtly hostile or may switch unpredictably between both types of response (Solomon and George 2011). Even when the mother’s behavior toward the child is placating, pleading and submissive, the child must appease the mother in whatever way is possible to remain in close proximity to her and feel some sense of safety. The child with the abdicating mother may behave in a contemptuous, dominant-hostile fashion, taking charge and dispensing advice. The child with a distressed and overtly anxious mother may find that caregiving behavior appeases the mother’s anxiety and the child of the depressed mother may clown and entertain (Main and Cassidy 1988).

Constricted Attention and Constricted Behavioural Repertoires

In agonic mode the price the child pays to belong in the relationship is the need to pay constant, vigilant and anxious attention to the mother’s moods and behaviors, in order to remain alert to shifts in the hierarchy, which require efforts at re-stabilization. The child’s externally focused attention supports the activation of the motivational system (and attendant behaviours) required to develop and keep the rigid asymmetrical hierarchy functioning.

It has been long held that the earliest foundations of an authentic self are to be found in infancy (Winnicott 1958). An infant who is able to ‘relax’ in the presence of an available yet non-intrusive mother and await the arrival of a sensation/emotion from within will experience the impulse to act and any subsequent action as their own. By contrast, a pervasive external focus will disrupt the infant/growing child’s sense of a personal self with agency in the world. This leads to enmeshment with the mother. The mother’s identity, developed in her own traumatizing childhood, also depends on paying the same vigilant attention to others in her environment, contributing to the enmeshment. A recent empirical study investigating boundary disturbances in families supports this idea. Enmeshed family relationships of the hostile, controlling and disengaged types were positively correlated with depressive symptoms, somatization, anxiety and ADHD at age 7 years (Jacobvitz et al. 2004). All, except ADHD, are consistent with forms of appeasement described by Cantor and Price (2007).

Another aspect of focusing attention solely on threat-based cues is that the child learns that the intentions of others towards them will be threatening; a notion that has been confirmed in research with behaviorally disturbed children (Orobio de Castro et al. 2002). In addition, the child will interpret any perceived decrease in predictability in their environment as a signal to intensify their efforts to become either dominant or submissive. Thus, in a hedonic group, where flexibility is the norm and roles and behaviors change readily, the agonically orientated child will feel threatened by the lack of predictability. Their responses will be incomprehensible to others in the hedonic context and may draw forth, familiar hostile or helpless responses, which will paradoxically increase their sense of safety but alienate them from the group (Dozier and Sepulveda 2004).

Stagnation of the Flow of Information

Another aspect of agonic mode is the stagnation of the flow of information, internally for each, and interpersonally within the dyad. Siegel (2012) has noted that attention controls the flow of information in the brain and mind. When attention is broad as in hedonic mode, multiple aspects of experience enter awareness; including an individual’s own bodily sensations, emotions and thoughts as well as varied aspects of other individuals and the wider environment. In agonic mode where the focus is on dominance and submission, this breadth and flow is lost. Unless the internal or external cues are directly relevant to the maintenance of asymmetry and the management of interpersonal threat, the cues will not influence the dyadic system, or the minds of the mother or the child. This precludes the incorporation of and accommodation to new information, making the system relatively impervious to change, except perhaps through experiential processes.

It seems possible that this stagnation of the flow of information within the system corresponds in some way to the idea of the lowered integrative capacity described as a mechanism leading to structural dissociation of the personality (Van der Hart et al. 2006); an outcome of chronic interpersonal traumatization in primary caregiving relationships (mother–child relationships in our context). The model predicts that in this situation, a child would develop different forms of asymmetrical relationship in response to different maternal behaviors, related to the activation of different motivational systems within the mother and corresponding motivational systems in the child. These different sets of balancing behaviors then become self-organizing systems within the personality characterized by distinctive emotional and psychological features (Liotti and Gilbert 2011).

Clinical Implications

This application of agonic and hedonic modes, hierarchy and appeasement to understanding mothers and children trapped in situations of intergenerational relational trauma predicts two primary tasks that should guide the development of effective treatment for this population: (1) the task of helping the dyad to move from agonic mode to the hedonic mode (2) the task of facilitating the emergence of an authentic sense of a personal ‘I’ in mother and child.

As discussed below, we suggest that a combination of established psychotherapeutic principles, and some novel ones, will be required to facilitate the movement from agonic to hedonic modes and allow each member of the dyad to find their authentic, individual self in the context of the mother–child relationship.

Movement from Agonic to Hedonic Mode

The movement from agonic mode to hedonic mode can only be achieved by de-stabilizing an established pattern of relationship that serves to reduce tension and conflict within the dyad. If the asymmetrical dominance-submission hierarchy is disrupted, activation of severe distress in mother and the child will inevitably follow; leading to the unstable and highly distressing state of agon (Kortmulder and Robbers 2005). This activation is likely to be intense, volatile and frightening for the mother and the child (and clinicians), and also potentially dangerous. The structure of an effective treatment must encourage this activation to occur in the clinic, with clinicians who are skilled in safely managing these moments.

Since reducing uncertainty is a key to increasing safety for these dyads maintaining the therapeutic frame is vital. This frame includes providing therapy at the same time each week, with sessions being the same length. Both mother and child must understand the structure within sessions, and the therapy space must have relatively unchanging physical characteristics, something that has long been understood in psychodynamically orientated treatments. The response of the therapist must also have elements of consistency. Since distrust is the foundation of relationship in agonic modes, a non-judgmental stance, with clear boundaries and clarity about mandatory child abuse reporting requirements and the limits of confidentiality will be essential. These elements, combined with a commitment to continue the process, even when it is uncomfortable, will support the development of some basic trust.

In addition, therapists must operate as far as possible in hedonic mode remaining cooperative and egalitarian and avoiding control, dominance or submission, in the face of sometimes-intense pressure to join the agonic asymmetrical relationship structures. Since hedonic modes involve play and tenderness, the therapy needs to expand the focus of attention of both mother and child beyond immediate threat to include these elements. The therapy must also foster an ability to understand the motivations of others as often benign, actively supporting this recognition during the therapeutic hour.

Facilitating the Emergence of a Differentiated Sense of Self

The second related goal of treatment is to facilitate differentiation in the relationship, supporting the child to develop a personal I, rather than being alienated from the “awareness of inner (distinctive) experience” by virtue of pathological accommodation to the mother (Brandschaft 2007, p. 667). One vital experience for a developing infant/young child is that of being alone in the presence of another (Winnicott 1958). This will have failed to occur where the mother has been over intrusive, abdicating or has moved between these contradictory positions changeably (Solomon and George 2011). The challenge for therapy with severely distressed mothers and their children is to find a way that the child can be in the presence of the mother actually or symbolically without focusing their attention exclusively on the mother or adjusting their behaviour to maintain the established hierarchy. This enables the child to begin to act from his/her own subjective experience. But, for this to be safely achieved requires the mother’s own emotional and behavioural responses to have changed, creating a technical challenge for a successful therapeutic model. In addition, without the self-definition provided by the external focus of attention and continued accommodation to the needs of the mother and the mother–child relationship, the child will experience a terrifying sense of not existing or not knowing who they are or how to act. The therapy must provide adequate support for the child to safely encounter these terrifying experiences without being provided with an alternative source of external cueing.

The mother will also have struggled to establish a sense of being a differentiated self, and will typically lack personal agency. The therapy must therefore also provide support for the mother to begin to discover who she is as a person in her own right differentiated from her child and her identity as a mother.

The Risks of Not Understanding the Impact of Hierarchy and Appeasement

If treatments reflect a flawed conceptual model and an incomplete understanding of the underlying purpose of certain observed behavior it is likely that treatment will be ineffectual (Segal et al. 2012). For example, in the situations that we are discussing, clinicians may interpret children’s “caregiving” behaviors as evidence of developing empathy and an ability to see the perspective of the other. Such qualities, if genuine, are signs of progress and should be encouraged, but in the presence of maltreatment, they are more likely to represent evidence of asymmetry, narrowed focus of attention and a loss of a sense of the personal ‘I’. Thus encouraging these behaviours is likely to be detrimental to the child. Where a child is acting in a way that is submissive and deeply appeasing of the mother, the treatment needs to support the child to declare himself or herself in a more complete, authentic and powerful way to the parent. Where the child is dominant and controlling, they need to be encouraged to take on the child role, allowing the mother to accept their role as protector, provider and nurturer. Each significant adjustment made by the child, requires a corresponding adjustment in the mother if lasting change is to occur, another challenge that needs to be addressed by any effective psychotherapy for this population.

Discussion

Two areas of enquiry emerged from the construction of our original model of the pathways to the ongoing maternal maltreatment of children aged 3–12 years (Amos et al. 2011). The first was to explore whether the coercive interactions seen in these traumatising mother–child relationships could be understood as adaptive, when considered from an ethological perspective. The second was to address internal inconsistencies in our model arising from contradictory perspectives put forward by two of the authors whose work informed the model.

The analysis presented in this paper argues that the dominant and submissive behaviours seen in traumatising mother–child relationships appear to represent the activation of biologically conserved adaptive behavioural strategies for the management of inter-individual threat in group-living mammals. We conclude that these behaviours in our mothers and children can be understood as serving the vital, adaptive function of supporting proximity by increasing predictability, but at the expense of ongoing intimacy.

Our exploration also helps to integrate the differing perspectives of Liotti and Lyons-Ruth. As Lyons-Ruth has observed (Lyons-Ruth et al. 2006) elements of Liotti’s formulation are inconsistent with those of her team. Liotti proposes a multi-motivational system explanation of the observed behaviors where-as Lyons-Ruth suggests a single motivational system explanation. Our formulation offers a way to integrate these two perspectives. Relationships in agonic mode rely on the asymmetry of dominance and submission, which is consistent with Lyons-Ruth’s single motivational system explanation. But our formulation also supports the involvement of multiple motivational systems in determining the exact nature of the behaviors involved in the asymmetrical relationship structures, a multi-motivational systems perspective, consistent with Liotti (2008, 2011). Furthermore, focusing on the structural qualities of the relationship, not the differing behavioral expressions, can explain the complexity and individual variations in clinical presentations. (For example research has shown that somatization; behavioral and cognitive inflexibility, depression, anxiety, and a sense of personal badness and shame are common outcomes of chronic interpersonal trauma (Brady 1997; Platt and Freyd 2012). All of these behaviours can be accounted for by our model as forms of appeasement, entailing as they do submissive behavior).

We anticipate that our model (and particularly the role of agonic and hedonic modes) will have other applications: notably in understanding intimate partner violence and in psychotherapy with adult survivors of early maltreating relationships. This analysis of disturbed intimate relationship patterns based in the activation of conserved mechanisms of threat management in endangering relationships, offers an alternative viewpoint to the dominant understanding of intimate partner violence based in the societal imbalances of power and control. Our model also offers a relational understanding of hostile, coercive and submissive and appeasing responses of adult survivors of chronic interpersonal trauma to therapists, as automatic, unaware, attempts to enhance the therapeutic alliance in the face of emerging fear and distress. This formulation is likely to have implications for how therapists respond to these difficulties when they are encountered in the therapeutic relationship. These areas are beyond our own research focus and require further exploration by others with the relevant expertise.