Introduction

Given that mentalizing is, as we have described in Chapter 1, rather difficult and prone to going wrong, and that it can cause us considerable psychological distress when it does go wrong, why do we do it so much? Humans are a social species and, in order to support our survival, natural selection has provided us with the tools necessary to interact with and manage the hugely complex terrain of our social world. Understanding the thoughts, intentions, and mental states of our fellow humans is the key to the creation of human societies and our attainment of dominance over other species on our planet. We constantly infer thoughts, feelings, and beliefs in the minds of other people, and much of our mental life is spent engaging with the task of processing social information [Reference Meyer, Davachi, Ochsner and Lieberman1]. Many writers and researchers believe that the evolutionary advantages of living in relatively large social groups were made by possible by humans developing the capacities to interpret, explain, and predict each other’s behavior, and using these capacities to share and accumulate experiences, to plan and collaborate, and to attain joint goals that they would not have been able to achieve as individuals or as members of smaller groups [Reference Bloom2–Reference Tomasello4]. There are excellent neuroscientific experimental studies which provide ample evidence that mentalizing, as indicated by the activation (or inhibition) of the network of neurons involved in Theory of Mind, is critical for evaluating the morality of others (including those who are harmful), for predicting how competitors will behave, and for learning from those who have special information about a shared environment (e.g., stock markets) [Reference Park, Kim, Young, Gilead and Ochsner5].

We have argued that vulnerability to mental health disorders is the flipside of the advantages that mentalizing brings [Reference Fonagy, Campbell and Constantinou6]. On the basis of data from epidemiological studies, only one in five people will go through life without experiencing a diagnosable mental health condition [Reference Schaefer, Caspi and Belsky7]. Looking at such prevalence figures from the perspective of natural selection, it is clear that whatever the neural systems are that underpin mental disorder, they must have other functions that are critical for survival. A defining feature of mental disorder is the experience of “wild imagination,” and we consider that mentalizing difficulties—the tendency to get caught up in unhelpful ways of imagining what is going on both for ourselves and for other people—are the price we as a species pay for the immense benefits of the human imagination [Reference Fonagy, Campbell and Constantinou6].

The Non-Mentalizing Modes

Mentalizing is a complicated and multifaceted process. We have pinpointed three non-mentalizing modes, which describe different states of mind that are encountered when mentalizing is ineffective, to help clinicians to understand and recognize when a client’s mentalizing is poor. These modes are called psychic equivalence, teleological mode, and pretend mode.

The Mentalizing Dimensions

To help us understand the how and why of the ways in which mentalizing and non-mentalizing can take these different forms, we need to appreciate that mentalizing draws on several different kinds of social-cognitive activities, underpinned by different neurobiological processes. We have categorized these into four mentalizing dimensions (which are sometimes called polarities, as they are not dimensional in the everyday sense):

• automatic/controlled

• self/other

• cognitive/affective

• internal/external.

We all move backward and forward along these dimensions in response to changes in our environment and adjustments in what we are thinking. At times, for example, we might be focused on our own mental states, while at other moments we might be more caught up in what we think is in the minds of other people. We often find that individuals who are experiencing emotional distress or behavioral difficulties tend to be more “stuck” at one pole of one or more of the dimensions. We shall briefly explain these dimensions here, but we shall also refer to them in our account of clinical processes, to illustrate how thinking about them can help us in our work.

Developmental Aspects of Mentalizing: Child/Adolescent/Adult

Mentalizing is a developmental achievement, and the full repertoire of mentalizing skills emerges over the course of childhood and adolescence. There are two reasons why it is helpful to understand how mentalizing develops. The first reason is that, when we are working with children and adolescents, it is useful to know what level of mental-state understanding we can expect from them, and what signs there might be of problems in the development of their individual functioning. The second reason why it is useful to understand the unfurling of ever more complex mentalizing skills across development is that, from the perspective of developmental psychopathology, it helps us to make sense of a patient’s vulnerabilities and how these affect their functioning.

The non-mentalizing modes, discussed earlier in this chapter, were partly inspired by careful observations of how children operate—recognizing the literality of the toddler’s horror of “poisonous” cabbage, or understanding the precursor of pretend mode in young children’s play. Indeed, in some of our earlier work we described these modes as pre-mentalizing rather than non-mentalizing modes. Moving forward through development from toddlers to adolescents, our growing understanding of the heightened and unstable experience of mentalizing in adolescence helps to explain why this life stage is a period of particular vulnerability in terms of the emergence of mental health difficulties.

Bertram Malle has provided a helpful summary of the developmental emergence of mentalizing in what he describes as the “tree of social cognition” [Reference Malle, Gilead and Ochsner18]. At the base of the “tree,” in the first 6 months of life, infants show capacities that are shared with non-human primates, such as differentiating between biological and non-biological agents. They develop a special interest in reading faces, and can detect that biological agents have goals, in contrast to movement by objects that is not goal directed. By the end of the first year, they can follow another person’s gaze (i.e., they show an interest in the direction of other people’s attention), they socially reference with others (e.g., they might look at a caregiver as if to ask “Should I be frightened of this stranger or new object?”), and, critically, they can engage in shared social attention—they enjoy the feeling of looking at things together. After another 6 months, infants are able to detect intent behind an action and can mimic others, usually to establish social links. Between the second and third years of life, a capacity for genuine empathy emerges, along with deliberate imitation and the capacity to infer wishes and desires. During the third and fourth years, children start inferring knowledge (which can be tested in the well-known false-belief test of Theory of Mind) and start to show self-knowledge and self-awareness. Non-human primates can also make inferences about what others want, but the evidence for their ability to infer false beliefs is debated and, if this ability is present, it is certainly not robust.

From middle childhood onward, the child develops an ever-increasing sophistication in ascribing mental states to others. Arguably, this ability continues to improve gradually right through to young adulthood. For example, by the age of 7–9 years, children master third- and fourth-order false beliefs (e.g., “If you think that I think that you think …” [Reference Osterhaus, Koerber and Sodian19]), and develop increasingly complex explanations of actions [Reference Atance, Metcalf, Martin-Ordas and Walker20]. At the top of the “tree” is explaining action in terms of the stable characteristics of an agent (so-called trait inference). This capacity reflects how the child has come to view the world as being occupied by people with different behavioral tendencies, based on a sense of “what so-and-so is like,” and therefore, by sampling another person’s actions, we can create a general picture of them that will help us to predict their behavior in a new situation. Of course, infants do this implicitly (behaviorally) via non-mentalizing processes from the earliest days of attachment to their caregivers, but at this point attachment moves to the level of representation and is experienced in terms of the anticipated mental states of attachment figures.

A defining feature of adolescence is the increased importance of social relationships, marked by an increase in social sensitivity [Reference Somerville21]. There is evidence that adolescents are more likely to report bad mood and anxiety following an experience of rejection [Reference Sebastian, Viding, Williams and Blakemore22], and that they expect less favorable positive feedback from their peers [Reference Moor, Guroglu and Op de Macks23]. The network of brain regions involved in mentalizing (especially the medial prefrontal cortex) is more responsive to social rejection in adolescents than in adults [Reference Sebastian, Tan and Roiser24], perhaps as a result of greater functional connectivity between the regions of the mentalizing network [Reference Burnett and Blakemore25].

This brief review of the development of mentalizing across childhood and adolescence has highlighted two things. The first is that mentalizing capacities are based on the emergence of multiple interlinked components, with explicit, content-led mental state ascriptions—what we think of as fully fledged mentalizing—developing quite late. A second repeated theme concerns the role of the relationship between the self and the other in the emergence of these capacities. The abilities that we describe may be conceptualized in purely cognitive terms, but the content of mental-state attributions, especially in the early years when a child is totally dependent on their caregivers, can be of the utmost seriousness, implying the availability of food, protection, and warmth, or indeed the opposite—risk, danger, and death. Self-awareness and self-regulation play a critical role in differentiating and modifying numerous mentalizing capacities, thus providing important functions that make social living possible, and this self-awareness and self-regulation is supported by the relationships and social environment that construct the sense of self. In the next section we shall describe the interpersonal nature of these developmental processes in more detail.

The Impact of the Environment on the Emergence of Mentalizing

How do we acquire our understanding of mind across development? Findings from neuroscience research show that the ability to mentalize is a prewired evolutionary adaptation—the process of acquiring increased social cognition, described in the preceding section, is a normative developmental trajectory. However, the achievement of full and robust mentalizing is dependent on environmental input. There are good reasons for this—in the same way that we are prewired for language acquisition, but not for the acquisition of any particular language, we need to be open to learning a different “language” for thinking and talking about mental states. We learn to understand mental states from how others around us frame them. Infants and young children have wants and feelings, and caregivers “read” their behaviors—their need for food, or assistance, or comfort—and offer evidence of intuitive understanding through actions or verbal responses [Reference de Villiers, Gilead and Ochsner26]. Carers are predisposed to guess, with typical utterances such as “You are tired, you have had enough of this game” or “Oh, it was the shiny silver foil that you were after!” Different cultures vary in the interpretations that they give to a child’s behavioral reactions [Reference Aival-Naveh, Rothschild‐Yakar and Kurman27], and even within the same culture caregivers will not give the same labels to expressions of supposed internal states. However it is expressed, toddlers rapidly acquire internal-state language from those around them. They like to explore and often talk confidently about others’ feelings, preferences, desires, and perceptions, allowing rich communication within families that includes increased empathy but also, for example, teasing [Reference Dunn, Brophy, Astington and Baird28]. The constant asking of questions by young children—the notoriously persistent “why” questions—is, we argue, partly about obtaining information, but more fundamentally about reveling in the process of joint attention on a shared object. The more children hear mental-state words and are exposed to mentalizing talk, the more their ability to successfully interpret false-belief tasks is likely to improve [Reference Dunn, Brophy, Astington and Baird28,Reference Meins, Fernyhough and Arnott29]. Some families engage in a lot of mental-state talk, and their children’s ability to mentalize is advanced, whereas others do much less—parents’ self-reported use of elaborated mentalizing conversation appears to predict higher scores on tasks that assess Theory of Mind [Reference Ebert, Peterson, Slaughter and Weinert30].

A young child develops a sense of their own mind, and then of other people’s minds, through early experiences of having how they are feeling being accurately recognized, mirrored, and responded to by a caregiver. These interpretations of and responses to what is going on for children help them to build a sense of who they are and to develop a sense of agency—that is, the ability to carry an idea of themselves as an active presence in the world, whose experiences matter, are of interest to those around them, and can make things happen. The caregiver’s responses that reflect back to the child what they are feeling are called “secondary representations.” It is as a result of having adults around us to decipher and delineate “who we are” for us that we can develop a coherent sense of self (sometimes called self-representation); in that sense selfhood, rather than being all about individuality, is an intrinsically social construct. Individuals who have not had the benefit of such secondary representations may have a diminished or less coherent self-representation (see this chapter: The Alien Self); these may include individuals who have been exposed to such responsive care but, for a complex range of possible reasons, have not been able to benefit from it.

There is plenty of evidence of the importance of being mentalized for a child’s development. The significance of parental mentalizing was first revealed in a large empirical study exploring parents’ capacity for reflective functioning, which found that a parent’s ability to think about and understand their childhood relationship to their own parents, measured during pregnancy, strongly predicted their infant’s attachment security [Reference Fonagy, Steele and Moran31]. A follow-up study found that antenatal parental mentalizing continued to be a predictor of the child’s reflective functioning 17 years later, when the children had grown into young adults [Reference Steele, Perez, Segal and Steele32]. Since this initial research, there has been a great deal of work on the effects of caregivers’ mentalizing on outcomes for their children. In the area of child attachment outcomes, a meta-analysis by Zeegers et al. examined the relationships between parental mentalizing, parental sensitivity, and infant attachment outcomes. This meta-analysis found that both mentalizing and sensitivity had significant direct effects on infant–parent attachment after controlling for the effects of each other, which suggests that parental mentalizing is directly related to infant–parent attachment security [Reference Zeegers, Colonnesi, Stams and Meins33].

As well as evidence for the impact of parental mentalizing on attachment security, there is also a growing body of evidence which indicates that better parental mentalizing fosters mentalizing in children [Reference Meins, Fernyhough and Wainwright34] and in adolescents [Reference Rosso and Airaldi35,Reference Rosso, Viterbori and Scopesi36]. Whereas associations between parental mentalizing and infant attachment typically represent small effect sizes (defined as Cohen’s $d=0.20$), the association between parental and infant mentalizing is typically stronger, representing medium to large effect sizes (Cohen’s $d=0.50-0.80$). For instance, Rosso and Airaldi found a particularly strong association between mothers’ ability to mentalize negative and mixed-ambivalent mental states, but not positive mental states, and the corresponding ability in their adolescent children ($r=0.40-0.50$) [Reference Rosso and Airaldi35]. Findings such as these suggest that the capacity of caregivers to reflect on difficult and affect-charged mental states is particularly important in the context of the transmission of mentalizing between generations, from parents to their children.

Studies on mentalizing and adversity during childhood have provided some of the strongest evidence for the potential role of caregivers’ mentalizing capacities in the development of their children’s mentalizing capacity. Early adversity and complex trauma (i.e., negative experiences in early life involving neglect or abuse, typically within an attachment/caregiving context) in particular have been shown to have the potential to severely impair mentalizing. This impairment is apparent in strongly biased mentalizing, hypersensitivity to the mental states of others, and a defensive inhibition of mentalizing, or a combination of these features (for reviews, see Borelli et al. [Reference Borelli, Cohen and Pettit37] and Luyten and Fonagy [Reference Luyten, Fonagy, Bateman and Fonagy38]). At the same time, there is increasing evidence that high levels of caregivers’ reflective functioning, particularly reflective functioning with regard to their own traumatic experiences (see Chapter 11), may be an important buffer in the relationship between early adversity and child outcomes (reviewed in Borelli et al. [Reference Borelli, Cohen and Pettit37]). For instance, higher trauma-related reflective functioning in parents with a history of sexual abuse and neglect has been shown to be related to lower risk of infant attachment disorganization [Reference Berthelot, Ensink and Bernazzani39], and to a substantially lower risk of exposure to childhood sexual abuse in their own infants [Reference Borelli, Cohen and Pettit37]. These findings are of particular clinical interest as they suggest the value of both prevention and intervention work, making use of mentalizing, with vulnerable parents. Finally, mentalizing difficulties reported in childhood are associated with social-emotional and cognitive problems, such as difficulties in emotion regulation and interpersonal problems, internalizing and externalizing problems, and difficulties in attentional control, effortful control, and academic achievement (for reviews, see Fonagy and Luyten [Reference Fonagy, Luyten and Cicchetti40–Reference Luyten and Fonagy42]).

Whereas our earlier work tended to heavily emphasize the role of attachment in supporting or inhibiting the development of mentalizing [Reference Fonagy43,Reference Fonagy, Steele and Steele44], more recently our views have evolved into a more comprehensive set of considerations about the role of family members, peers, and broader sociocultural factors, such as socioeconomic deprivation, social isolation, and school climate, in the development of mentalizing [Reference Fonagy, Luyten, Allison and Campbell45,Reference Fonagy, Luyten, Allison and Campbell46]. In keeping with this view, there is growing evidence for mentalization-based treatment (MBT)-related interventions that focus on the family and the broader social context that surrounds an individual. Several studies have provided evidence of the effectiveness of MBT in various groups, such as substance-abusing mothers and their infants Reference Suchman, DeCoste, Borelli and McMahon[47], fostered and adopted children Reference Redfern, Wood and Lassri[48], mothers living in underserved poor urban communities with children at high risk of maltreatment [Reference Slade, Holland and Ordway49,Reference Byrne, Sleed and Midgley50], individuals supporting a family member with borderline personality disorder (BPD) Reference Bateman and Fonagy[51], and school-based prevention and intervention programs Reference Fonagy, Twemlow and Vernberg[52]. These types of family and “system-level” interventions might be most effective in addressing the problems that non-mentalizing social environments (e.g., neighborhoods with high levels of crime and violence, or schools with a culture of bullying) tend to generate, by creating a mentalizing climate as a counterweight against competitive, hostile, and aggressive wishes and tendencies. We are of the view that the social network around an individual serves to support the recovery of mentalizing when it is inevitably lost in the course of the daily difficulties encountered in life—the “micro-traumas” to which we are all continuously exposed. Connecting to others at moments of stress affords powerful protection against these problems because it ensures some balanced mentalizing, and thus resilience, in relation to the experience [Reference Fonagy, Allison, Campbell, Bateman and Fonagy53,Reference Fonagy, Steele and Steele54]. We consider that creating a mentalizing climate around the individual or family is particularly important for children and young people, and for providing support to their families/carers (for further discussion of mentalizing-based interventions that aim to do this, see Chapters 15 and 17). An important additional aspect of this view is that it acknowledges the need to provide a supportive mentalizing system around mental health professionals, given the many internal and external pressures and anxieties that they face in their work.

The Alien Self

The idea of the alien self is something we use in MBT to capture the experience of individuals who, in a high state of stress or distress, find their sense of self to be so frightened/frightening and incoherent that an “alien self” steps in. As we shall see in Chapters 14 and 18, the alien self has been particularly useful in understanding some cases of suicidality, serious acts of self-harm, interpersonal aggression/violence, and other mental health crises. When we speak of the alien self, we mean that disturbances in core self, sensory, and cognitive psychological processes lead the individual to experience various forms of loss in their personal agency. Their experience, which could be related to bodily, cognitive, or affective (dys)control, is of a “not-me,” and there is a failure to integrate the experience of self coherently—essentially, parts of selfhood, experienced in the here and now, are alienated. We consider this representation to originate from the internalization of an insensitive or overwhelming caregiver, as a consequence of the child’s exposure to failed mirroring by the caregiver. By “mirroring,” we mean interactions where the caregiver recognizes the infant’s mental states and is able to reflect those states back to the infant in a “marked,” or regulated, way which shows that they understand and empathize with the infant’s feelings but are not overwhelmed by them. Later in development, hostile or abusive caregiving is managed by the activation of a split-off part of the self-representation—the alien self—which presents the child with the possibility of obtaining a degree of control over an uncontrollable social environment by identifying with the aggressor (the abusive caregiver) using this alien self; the individual then becomes host to a torturing figure within the self-structure. In a sense, alien self-representations can help individuals to cope with disturbances of the self; the emergence of the alien self is an attempt at an emergency rescue from the intense emotional experiences that the individual is unable to manage or regulate. Although this defensive process affords something of a sense of temporary control, it also causes an acute experience of fragmentation and a feeling of being tortured from within. The alien self is a threat to psychological integrity, and can often lead to behaviors such as self-harm and suicide attempts that are linked to the affective pressure coming from the alien self-experience. Importantly, psychological contact with alien self-representations may also lead to dissociation, which is a hallmark of a mental process coming in as an attempt to protect psychological self-integrity. The importance of the alien self in clinical work is considered further in Chapters 10 and 12.

As we described in the previous section, the sense of self, and the ongoing coherence of the sense of self, is partly constructed on the basis of the secondary representations of who we are that are communicated to us, in the first instance, by our caregivers in early life. An individual who has missed out on sufficiently benign, coherent, or accurate secondary representations, or who has difficulty in recognizing and internalizing these secondary representations (possibly because of a lack of exposure to such representations), may be particularly vulnerable to developing a fragmented self-structure (i.e., an alien self), which is then called upon in response to a chaotic or empty sense of self-experience. The latter is sometimes called a “black hole” by individuals with BPD, who unsurprisingly find it terrifying, as it threatens their very existence.

Attachment and MBT

MBT is strongly shaped by attachment theory—the idea that the development of mentalizing emerges in the context of attachment relationships has always been central to our thinking, and evidence from research supports the view that there is a “loose coupling” between attachment, emotional sensitivity, and parental mentalizing capacities [Reference Luyten, Campbell, Allison and Fonagy55]. In MBT work, a patient’s attachment experiences and representations will play a key part in the therapist’s understanding of the challenges that the patient faces when thinking about other people’s mental states. However, in the practice of MBT, it is also vital that the therapist understands the impact of attachment arousal in the here and now of treatment, and how the experience of such strong emotions can hinder effective mentalizing.

Brain-imaging studies have shown that the activation of the attachment system inhibits the neural systems that support mentalizing [Reference Bartels and Zeki56–Reference Nolte, Bolling and Hudac58]. Trauma can activate the attachment system, and attachment trauma may cause chronic hyperactivation. This explains how some patients may be able to mentalize when they are under little or no stress; however, arousal related to attachment may cause a collapse of mentalizing at moments of stress in individuals who are prone to this form of hyperactivation. This has important implications for clinical work. Any intervention that calls for reflection—for example, by asking the patient to clarify or elaborate on a thought—is by its very nature asking the patient to engage in controlled mentalizing. If the topic that the patient is being asked to reflect on triggers arousal related to attachment, it may be very difficult for them to engage in such reflective activity. This may be one of the reasons why, historically, dynamic therapy for BPD was often found to be ineffective—traditional psychotherapeutic methods trigger the attachment system, both through the nature of the relationship between therapist and patient and through the subject matter that is explored in therapy. In effect, the patient is being asked to carry out a potentially impossible task—to think about mental states in a regulated way while being placed under circumstances that make such reflection almost impossible.

For all that MBT is about mentalizing, the work of MBT should not involve a relentless onslaught of reflective mentalizing on the part of the therapist. A patient who is in a state of high emotion, whose attachment system is ringing alarm bells, and whose mentalizing has gone “offline” should not be met with further demands to mentalize—these will either be ignored or be experienced as insensitive and unhelpful. Rather, the appropriate response from the therapist is recognition and validation of the patient’s feelings at that moment. The reasons for this will unfold in the next sections of this chapter, in which we shall explain how the capacity for shared attention and thinking together is made possible by a sense of personal recognition. First, we shall explain how our clinical experiences directed our thinking in this regard.

Borderline Personality Disorder and MBT

As we mentioned in Chapter 1, MBT was first developed as a response to an acute awareness of how few patients with a diagnosis of BPD were receiving the therapeutic help and support that they needed. MBT was deliberately built on the awareness that a form of treatment was required that (a) really addressed the clinical needs and real-life experiences and difficulties endured by these patients, and (b) was sufficiently usable for a broad range of mental health workers to implement reliably without undergoing years of intensive training.

Given how much thinking about BPD has influenced the development of MBT, we believe it will be useful to give an overview of the nature of the difficulties associated with BPD, to help the reader to understand some key elements of the general mentalizing approach and the rationale for this approach. One of the most common features of BPD is emotional dysregulation, which involves a hypersensitivity to emotional stimuli that can spark intense and unstable emotional responses. Individuals with BPD may struggle to find coping strategies that help to reduce their distress, and the emotional dysregulation is often coupled with impulsive symptoms such as self-injurious behavior (including impulsive self-harming behaviors), cognitive disturbances (including dissociation), and interpersonal problems. Individuals with BPD often experience intense anguish, and they are 50 times as likely to attempt suicide as members of the general population [Reference Pompili, Girardi, Ruberto and Tatarelli59].

Stress-related dissociation is common in individuals with BPD Reference Korzekwa, Dell and Links[60], and can affect the sense of identity, episodic memory, perception, and consciousness [Reference Miller, Townsend and Grenyer61]. Disruptions of such fundamental social-cognitive processes can be understood as defensive (i.e., self-protective) in nature, by which we mean they are adaptations to stressful internal experiences that function to limit the impact of these experiences on the person (or their body, as in derealization or emotional numbing). We see the symptoms of BPD as understandable mental adjustments made by the individual to cope with what would otherwise be overwhelming subjective experiences of stress or distress Reference Vermetten and Spiegel[62]. Sometimes the impulsivity observed in BPD plays out as impulsive aggression [Reference Ende, Cackowski and Van Eijk63,Reference Sebastian, Jung and Krause-Utz64]. Aggression shares with emotional dysregulation an association with high levels of stress and distress Reference Krause-Utz, Keibel-Mauchnik and Ebner-Priemer[65], but also features a lack of consideration for the mental state of the person toward whom the anger is directed Reference Bateman, Bolton and Fonagy[66]. A range of negative behaviors, such as aggressive outbursts, binge eating, risky sexual behavior, substance abuse, or suddenly ending a relationship, can be triggered in negative emotional contexts where the individual perceives these actions as uncontrollable and inevitable reactions to an impossible (social) situation [Reference Cackowski, Reitz and Ende67,Reference Jacob, Zvonik and Kamphausen68].

Interpersonal disturbances are widely thought to be central in the work of helping people with BPD. There is strong evidence that individuals with BPD have difficulty in developing trust in others, a heightened sensitivity to social exclusion and rejection, and lengthy histories of conflicted unstable relationships with unresolved conflict about separating [Reference Lis and Bohus69,Reference King-Casas, Sharp and Lomax-Bream70]. Often the histories of individuals who present for treatment amply explain their reluctance to place trust in social relationships [Reference Reinhard, Dewald-Kaufmann and Wustenberg71,Reference Jowett, Karatzias and Albert72]. Whether interpersonal disturbances should be regarded as attempts to manage earlier developmental problems (e.g., as in transference-focused psychotherapy) or as caused by other difficulties (as conceptualized in dialectical behavior therapy) remains a source of controversy, but from the pragmatic standpoint of a clinician, both the emotions themselves and their impact on relationships need to be tackled.

Mentalizing and BPD

A growing literature suggests that mentalizing difficulties are an intrinsic aspect of BPD and may drive the interpersonal disturbance, affect dysregulation, and impulsivity that are characteristically shown by individuals with this diagnosis. The mentalizing difficulties found in people with BPD can cover all four of the mentalizing dimensions (which were outlined in detail earlier in this chapter).

In relation to automatic/controlled mentalizing, the dysregulation and impulsivity associated with BPD reflect a tendency toward automatic mentalizing. In what was perhaps the first demonstration of the failure of reflective mentalizing in BPD, Fonagy et al. showed a dramatically reduced ability of BPD patients to reflect on their own and others’ mental states in the context of attachment relationships [Reference Fonagy, Leigh and Steele73]; this deficit in reflective functioning has been shown to be reversible by psychotherapy [Reference Levy, Meehan and Kelly74]. The findings on reflective functioning have been replicated by two relatively large studies [Reference Fischer-Kern, Schuster and Kapusta75,Reference Gullestad, Johansen and Høglend76] and confirmed in a recent meta-analysis [Reference Bora77].

A number of studies have shown that BPD patients show mentalizing impairments when tested with the Movie for the Assessment of Social Cognition (MASC), which is currently probably the best validated and most reliable measure of this mentalizing domain [Reference Preissler, Dziobek and Ritter78–Reference Sharp, Pane and Ha80]. The MASC is a video-based test of mentalizing that requires participants to recognize the mental states of characters as they interact in an everyday-life scenario involving relationships. The misinterpretation of interpersonal situations because of a limited reflective capacity may generate extreme or inappropriate emotions in social interactions [Reference Dziobek, Preissler and Grozdanovic81], particularly if there is also an overemphasis on externally oriented social cognition.

The self/other mentalizing dimension can often appear in BPD as an excessive focus on the mental states of others. Within the MBT model, knowledge of self and knowledge of other are complementary but interdependent. We learn about ourselves from others, and this self-knowledge helps us to place ourselves in another’s shoes. We would expect mentalizing problems to manifest as a reduction in self-knowledge and unstable self-representations. Another area of mentalizing the self is its role in generating a sense of ownership over one’s own actions. Several studies suggest that the experience of agency may be reduced in people with BPD relative to those without BPD [Reference Bekrater-Bodmann, Chung and Foell82–Reference Pavony and Lenzenweger84]. There is emerging evidence of difficulties with self–other differentiation in BPD [Reference De Meulemeester, Lowyck and Luyten85], with, for example, observations of greater proneness to the Rubber Hand Illusion [Reference Bekrater-Bodmann, Chung and Foell82,Reference Neustadter, Fineberg and Leavitt86], and confusion in identifying the self from graded morphed self-images [Reference De Meulemeester, Lowyck and Panagiotopoulou87]. Several investigations have shown that people with BPD are more likely to recognize negative than positive self-attributes, suggesting distortions in their self-concept toward greater negativity [Reference Auerbach, Tarlow and Bondy88–Reference Kanske, Schulze and Dziobek91]. However, these studies gave a less clear picture in relation to the structural complexity of the self-concept. One study of adolescents and young adults reported that patients with BPD endorsed a greater number of attributes as being self-related, indicating a more diffuse sense of self [Reference Auerbach, Tarlow and Bondy88], whereas another study of adults did not mirror this observation [Reference Vater, Schroder-Abe and Weissgerber90]. The differences in the findings of these studies could reflect the tendency of adolescents with emerging BPD to hypermentalize, which is likely to decrease with age [Reference Sharp, Ha and Carbone92]. In line with the greater prominence of the self in BPD, memory studies suggest that individuals with BPD have enhanced recall of self-relevant social events uninfluenced by emotional valence [Reference Winter, Koplin and Schmahl93]. A “mind-wandering” paradigm suggested that self-related thoughts fluctuate more, and are more extreme, in individuals with BPD [Reference Kanske, Schulze and Dziobek91].

The evidence in relation to the internal/external dimension of mentalizing in BPD is mixed. Accuracy of emotion recognition appears to be reduced for static facial stimuli in individuals with BPD in some studies [Reference Ritzl, Csukly, Balazs and Egerhazi94,Reference Lowyck, Luyten and Vanwalleghem95], but not in others [Reference Bertsch, Krauch and Stopfer96–Reference Bertsch, Gamer and Schmidt98]. Similarly, whereas some studies reported that patients with BPD take longer than individuals without BPD to decode faces with emotional expressions [Reference Bertsch, Gamer and Schmidt98], other studies did not find this difference [Reference Kobeleva, Seidel and Kohler99,Reference Dyck, Habel and Slodczyk100], and a meta-analysis confirmed the absence of a reliably identifiable deficit in the speed of emotional face processing in people with BPD [Reference Hanegraaf, van Baal, Hohwy and Verdejo-Garcia101]. There are also mixed findings regarding facial reactivity to brief emotion-induction videos. One study reported reduced reactivity in individuals with BPD (effect size = –0.67) [Reference Renneberg, Heyn, Gebhard and Bachmann102], whereas another found reduced reactivity only to positive emotions, and also reported that BPD patients displayed more negative and mixed facial emotions in response to social exclusion than individuals without BPD [Reference Staebler, Renneberg and Stopsack103]. There is evidence that facial mimicry is enhanced in BPD. In a task in which participants’ facial muscle activity in response to dynamic facial expressions was recorded, BPD patients showed enhanced responses to angry, sad, and disgusted facial expressions, and attenuated responses to happy and surprised faces [Reference Matzke, Herpertz and Berger104]. This finding suggests a specific rather than general hypersensitivity to the emotional state of others, expressed by exaggerated responses to negative social signals and attenuated responses to positive signals. Some studies have also identified a strong negative attribution bias, with BPD patients attending more quickly to negative faces and spending more time looking at them [Reference Schulze, Schmahl and Niedtfeld105].

A large number of studies have explored mental-state discrimination in BPD. Most of these studies reported a medium-sized effect across a large number of measures. The most prominent of these measures is the Reading the Mind in the Eyes test, which has yielded both the largest reported advantage for mind reading [Reference Fertuck, Jekal and Song106,Reference Frick, Lang and Kotchoubey107] and the greatest deficits in those with BPD [Reference Berenson, Dochat and Martin108,Reference Anupama, Bhola, Thirthalli and Mehta109]. A meta-analysis by Hanegraaf et al. revealed a substantial and significant heterogeneity between the studies ($Q=151.55$, $p<0.0001$) [Reference Hanegraaf, van Baal, Hohwy and Verdejo-Garcia101]. However, this comprehensive meta-analysis provides extensive and robust evidence that people with BPD tend to be poorer at mental-state discrimination than people from other clinical populations, including those with other personality disorders (cluster C, avoidant, and narcissistic personality disorders) and those with major depression. Obviously this adds empirical support to the MBT approach and justifies our focus on mentalizing in our therapeutic work with patients with a diagnosis of BPD. BPD tends to be associated with difficulties in both selecting and displaying optimal social signals during interpersonal interactions, through facial expressions, language, and caregiver–infant behavior, and when interpreting the social signals of others.

Not surprisingly, given that the diagnostic criteria for BPD include rejection sensitivity, in experimental studies of social exclusion it has been found that individuals with this diagnosis report more severe experiences of ostracism than control participants without BPD; in fact, BPD patients are more likely to report feelings of exclusion even in conditions of social inclusion (e.g., [Reference Brown, Plener and Groen110]). There is evidence that rejection sensitivity is mediated by social-cognitive factors (tolerance of ambiguity, effortful control) and attachment-related factors (attachment anxiety, feeling of belonging, self-criticism) [Reference Sato, Fonagy and Luyten111,Reference Sato, Fonagy and Luyten112]. In the meta-analysis by Hanegraaf et al., BPD patients were also found to tend to feel more excluded than individuals with social anxiety disorder, major depressive disorder, or non-suicidal self-injury [Reference Hanegraaf, van Baal, Hohwy and Verdejo-Garcia101]. A failure of mentalizing combined with the expectation of negative outcomes from social interactions leaves individuals with BPD vulnerable to misinterpreting such interactions as instances of rejection.

Widening the Lens: Mentalizing in Psychopathology

We hope the description of the mentalizing challenges that individuals with BPD have to face has helped to provide a good sense of the rationale for the MBT approach. We shall now explain how and why this thinking can also inform the psychological treatment of other mental health disorders, ranging from other personality disorders to anxiety and depression, and to psychosis. This widening application of MBT is based on the view that mentalizing difficulties are common to many forms of mental health problems.

There is a lively and growing body of research around what is called the general psychopathology factor, or “p factor,” in mental disorder. This is the idea that it might be more useful to understand the many overlapping mental health diagnoses that exist as sharing a common vulnerability factor, which then interacts with individual circumstances and genetic tendencies to manifest as a particular set of symptoms. To give quite a crude example, adolescent boys on the whole are statistically more likely to show externalizing symptoms, such as conduct disorder, whereas adolescent girls are more likely to develop internalizing symptoms, such as anxiety or depression. What is more, an adolescent boy with conduct disorder may well go on to experience depression and anxiety as an adult, with or without continuing conduct problems. So, across different populations, or across one individual’s life course, there will be a range of experiences and behaviors that appear quite different in terms of both outward behavior and emotional experience, but, according to the thinking around the p factor, the underpinning vulnerability that leads to all of these different symptoms remains the same—it is just that they play out differently for different people as a result of a combination of their different social circumstances and genetic factors. There are various reasons for giving credence to this idea—for example, how do we make sense of the numerous concurrent and often overlapping difficulties that people experience? Findings from behavioral genetics and molecular biology are also consistent with the argument that there is a common underlying factor. Studies of families and twins tend to show that genetic risk is not specific to particular disorders, but is largely a transdiagnostic vulnerability factor [Reference Franic, Dolan and Borsboom113–Reference Pettersson, Larsson and Lichtenstein116]. Similarly, a genome study of 17 mental disorders with over one million participants found that, in terms of genomic markers, the mental disorders shared common variance risks [Reference Anttila, Bulik-Sullivan and Finucane117].

The argument for the possibility that a single common factor underpins vulnerability also arises from population studies that use bifactor modeling of symptoms and diagnoses. Bifactor modeling involves exploring a possible general factor that a broad range of different variables—in this case symptoms and diagnoses—may share in common. Work in this area has found that psychiatric disorders are more convincingly explained if a general psychopathology factor is assumed to exist alongside clusters of symptoms (internalizing, externalizing, and psychosis) and individual mental disorders—for example, schizophrenia, generalized anxiety disorder, and depression [Reference Caspi and Moffitt118–Reference Lahey, Applegate and Hakes120]. Numerous studies with samples of children, adolescents, and adults indicate that the p factor appears to capture an underlying propensity for any kind of psychopathology as indicated through the apparent symptoms [Reference Lahey, Krueger and Rathouz121–Reference Patalay, Fonagy and Deighton123].

The emphasis on the p factor leaves us with the question of what this factor actually is, other than a useful statistical construct. Many studies suggest that all mental disorders involve a dysfunction of just one or two parts of the brain—the prefrontal cortex and perhaps the limbic regions [Reference Macdonald, Goines, Novacek and Walker124,Reference Wise, Radua and Via125]. There may be, and indeed there are very likely to be, many other areas of the brain that place an individual at risk of having a mental disorder. In fact, it is increasingly likely that it is not so much the structure of particular parts of the brain but the connections between different areas that cause these problems [Reference Hinton, Lahey and Villalta-Gil126]. So general psychopathology, or the p factor, or perhaps most accurately the risk of mental disorder, at the level of the brain, may be to do with irregular connectivity in the cortex of the brain. This hypothesis is consistent with the growing consensus that psychopathology is best seen as a dysfunction of brain networks rather than of localized brain regions [Reference Protzner, An, Jirsa, Diwadkar and Eickhoff127]. A study in which high-quality resting-state brain scans were obtained in 6593 children aged 9–10 years found that a high p factor was associated with reduced connectivity within the default mode network (DMN), and increased connectivity between the DMN and multiple control networks [Reference Sripada, Angstadt and Taxali128]. The DMN is active when the mind is not engaged in any specific cognitive task—that is, when a person is “mind-wandering”. The DMN overlaps with and extends beyond the network of brain regions involved in mentalizing. One study showed that a set of tasks concerned with social cognition (i.e., reasoning about the mental states of other persons) activates the DMN [Reference Jack, Dawson and Begany129]. The DMN is activated in thought processes such as daydreaming, imagining the perspectives of others, imagining the future, and recalling autobiographical memories [Reference Ekhtiari and Paulus130], and is generally considered to subserve mentalizing. In other words, vulnerability to psychopathology, as measured by a high p factor, is associated with anomalies in connectivity in the neural network that is implicated in mentalizing processes.

Neuroscience research has consistently implicated two related major domains of functioning across most mental health disorders. The first is emotional dysregulation. Poorly managed emotions interfere with purposeful activity [Reference Beauchaine131]. People who regulate their emotions well have an accurate idea about the risks of a given situation that triggered emotion, and can direct their attention to what they need to do to cope with it, use their attention to focus but also to distract themselves if focusing is unlikely to help, and reassess the risks as the results of their actions unfold [Reference Gross and Gross132]. Emotional dysregulation has been shown to be a feature of almost every psychiatric diagnostic condition [Reference Beauchaine and Cicchetti133]. Researchers believe that a weakness in emotion regulation can amplify emotional experience that is going on in the background, cause distortions in the way children see and experience social situations, and lead to intense emotional reactions just because the person is anticipating and experiencing an intense feeling. It can generate inappropriate feelings and lead to sometimes dramatic actions to avoid unpleasant intense emotion.

The second general domain that is consistently associated with mental disorder is executive function [Reference McTeague, Huemer and Carreon134]. Executive function is the capacity to control thought and generally direct information in the appropriate direction relevant to any specific task; it includes self-regulation, decision making, sequencing of actions, planning, prioritizing, and navigating new tasks [Reference Banich135]. These cognitive processes are deemed to be essential for healthy functioning. The quality of executive function is thought to affect the p factor. Individuals with a high p factor score, signifying what we might term high persistent psychological distress, are oversensitive to difficult social interactions, and they find it hard to reliably interpret the reasons for others’ actions and put out of their mind potentially upsetting memories of experiences. This state leaves them vulnerable to emotional storms.

To return to the role of mentalizing, we can see that because mentalizing supports executive function and affect regulation, both emotional dysregulation and poor executive function—the two areas reliably associated with psychopathology—are associated with poor mentalizing. We do not know where cause and effect lie in relation to mentalizing here; from a pragmatic clinical perspective, whether mentalizing problems are the consequence or the cause of emotional dysregulation and failures in executive function may be something of a “chicken-and-egg” question. The issue, clinically, is that all of these processes are closely linked to one another. The restoration of the capacity to think about mental states, both in oneself and in others, may be critical to making affect regulation and more protective executive function possible; this provides the rationale for the application of MBT approaches to a range of mental health disorders. In the next section, we shall set out why we consider mentalizing to have such great significance for healthy functioning.

Mentalizing, Social Learning, and Epistemic Trust

We have defined epistemic trust, which is a key part of mentalizing theory, as “trust in the authenticity and personal relevance of interpersonally transmitted knowledge about how the social environment works and how best to navigate it” (Fonagy et al. [Reference Fonagy, Campbell and Bateman136], p. 177). We have argued that mentalizing has a specific function in enabling individuals to take a position in relation to having epistemic trust in others—something that is vital if they are to benefit fully from opportunities to learn from others about the world in general and the social world, including themselves, in particular. We suggest that this is one of the reasons why disruptions in mentalizing appear to be so detrimental to social functioning [Reference Fonagy, Campbell and Constantinou6] and so closely linked to vulnerability to any long-term mental disorder, and why they appear to be associated with the p factor [Reference Sripada, Angstadt and Taxali128]. To survive in the socially complex networks that humans inhabit, we have to be able to learn from other people [Reference Masten, Morelli and Eisenberger137]. In learning the skills that we need to survive, we observe and listen to experts, we avoid punishments and gain rewards by imitating others, we make complex strategic decisions after taking others’ experience and views into account, and throughout all of this we learn about ourselves through social interactions. Learning efficiently from others is a central part of human existence. The transmission of information from parents to children, as well as between peers, constitutes a core mechanism of adaptive cultural learning, but can also cause maladaptive behaviors, such as antisocial actions, exaggerated avoidance, and anxiety [Reference Espinosa, Golkar, Olsson, Gilead and Ochsner138].

The development of mentalizing is intricately related to epistemic trust and openness to social learning—that is, the ability to learn from others by trying to understand what they have to teach us. Human infants quickly learn to infer a model’s goal and copy successful (but not failed) actions, but non-human primates acquire this capacity earlier than humans. Humans take a slightly different approach to learning. By the age of 3, children start showing what is called over-imitation, by which we mean that young children readily learn and copy behaviors that are unnecessarily complicated and do not seem to have a more functional purpose, if they seem to be the “right” way of doing things (a simple example would be learning to eat with a spoon rather than the faster and more efficient method of using one’s hands) [Reference Hoehl, Keupp and Schleihauf139]. For a period, young children are prone to unselectively and unreflectively copy the behavior and acquire the beliefs of the majority of people with whom they interact [Reference Lyons, Damrosch and Lin140], and will copy a person’s failed as well as successful attempts to complete a task [Reference Huang, Heyes and Charman141]—something that non-human primates have not been observed to do [Reference Clay and Tennie142]. The reasons for this are that what the child is learning when, for example, they learn to use a spoon is complex. They are not only learning how to break up food and transport it to their mouth [Reference Gergely, Csibra, Enfield and Levinson143]. In fact, most small children find it easier to use their hands to eat—but learning to use a spoon correctly is one tiny example of the huge ocean of cultural and social knowledge that infants are born into. Trusting the demonstration by reliable elders that spoons are the correct way to eat, and copying how to use them, is an example of the small steps the young child starts to make in learning to navigate the world. Relatively inefficient routines are frequently socially transmitted as part of the cultural knowledge shared by members of a child’s community [Reference Clay and Tennie142]. Acquiring such routines is therefore an important developmental task that allows children to become members of their cultural group. The term “over-imitation” does not signify that children imitate knowledgeable adults when they should not. Rather, the tendency to over-imitate actions that are cognitively opaque is a cultural learning strategy, presumably identified through natural selection, based on the assumption that learning to do things that one does not fully understand is beneficial because they are culturally relevant, despite their opacity to the learner [Reference Altınok, Hernik, Király and Gergely144]. Preschool children’s detailed mimicking of new behaviors may represent an openness to learning novel skills, unusual social norms, and rituals, and thus to affiliating with members of their community [Reference Nielsen145]. Over-imitating declines as mentalizing emerges, but it remains part of culture and of how humans learn, because a comprehensive understanding of causation is beyond humans even in relation to the physical world. The predisposition to over-imitate is sustained into adulthood [Reference Flynn and Smith146], and is enhanced by the social context [Reference Gruber, Deschenaux, Frick and Clement147,Reference Marsh, Ropar and Hamilton148].

This account of social learning is relevant to our thinking about mentalizing and MBT because we suggest that learning and trusting others so that we can learn from them—that is, epistemic trust—is enabled by mentalizing. We have said elsewhere, “If I feel that I am understood, I will be disposed to learn from the person who understood me, who I feel is a trustworthy potential collaborator. This will include learning about myself but also learning about others and about the world I live in” (Fonagy et al. [Reference Fonagy, Campbell and Constantinou6], p. 7). The outcome of this process is what we have called an epistemic match. This sounds like quite a simple proposition, but there are several mentalizing stages involved in this process. It requires the listener to (a) have a coherent enough sense of self to be able to recognize the match in the first place, (b) recognize the image of themselves that the communicator is describing to them (in a situation where the communicator is somehow expressing their own perception of the listener), and (c) judge the similarity or difference between their own self-perception and how the communicator is describing them. In addition, the process requires the communicator to have mentalized the listener well enough to have a good understanding of them. Thus epistemic trust as a social process depends on mentalizing on the part of both the listener and the communicator.

When co-mentalizing as occurs in an epistemic match is achieved, a particular subjective experience of social cognition is generated, which has been labelled the “we-mode.” The we-mode is a mental state that describes an interpersonal experience of being enjoined with someone else, where intentional states are shared with a common purpose [Reference Gallotti and Frith149,Reference Higgins150]. The we-mode triggers trust, which in turn triggers the potential for learning from the trusted source. The task of restoring mentalizing in MBT is thus undertaken to (re)open the possibility of experiencing the we-mode and an epistemic match that is crucial for enabling the social learning that underlies social functioning. In addition, there is a relational aspect to the we-mode, which generates the feelings of both agency and belonging that arise from a sense of “jointness,” or the epistemic connectedness that shared intentionality creates. It is important to note that the we-mode is not some kind of merging of minds. The self, with its full sense of agency, remains important; in fact, in the absence of a robust self that temporarily subverts itself for a joint purpose, access to the we-mode may be only illusory and achieved by self-distortion. The we-mode is not an abdication of the awareness of separate mental states, nor is it some sort of psychic fusion. Rather, it involves holding in mind the idea that others are separate “agents or persons just as real as oneself” (Tomasello [Reference Tomasello151], p. 56), with whom we momentarily share a view of the world, an ambition, a plan, or an understanding that supersedes our individual perspective. Indeed, the benefits of the we-mode accrue from separate minds joining together to focus on a shared object; joint attention brings new knowledge and perspectives that a single mind cannot [Reference Tomasello4]. We shall describe in more detail in the next sections how disruptions in epistemic trust and in experiencing the we-mode are implicated in the development of psychopathology.

Trust in BPD

Trauma and Mentalizing

The mentalizing model originated from understanding of the psychological consequences of trauma (see also Chapter 11), with the assumption that one possible outcome of childhood adversity was a limited capacity for mentalizing, driven by anxiety. A child’s fear of understanding the mental states of an individual who poses a genuine threat to them is understandable, and generalization of this fear to other minds might be expected [Reference Allen, Fonagy, Bateman, Vermetten and Lanius186]. Considerable evidence has accumulated to suggest that limitations in mentalizing are commonly associated with the experience of trauma, particularly post-traumatic stress disorder. Differences in mentalizing have been demonstrated in tests of emotional intelligence [Reference Janke, Driessen and Behnia187], tests of empathy and compassion [Reference Palgi, Klein and Shamay-Tsoory188,Reference Mazza, Tempesta and Pino189], and tests developed specifically to measure cognitive mentalizing, such as the Faux Pas test and the Strange Stories task [Reference Nietlisbach, Maercker, Rossler and Haker190,Reference Mazza, Giusti and Albanese191]. Our theoretical approach predicts a bias against internal cues to mental states (mental-state understanding) but a potential hypersensitivity to external indicators of mental states (observations) to balance the defensive avoidance of “looking behind” the external. In line with this prediction, patients with a history of trauma appear to show no or only small deficits in emotion-recognition tests [Reference Nazarov, Frewen and Oremus192–Reference Bell, Colhoun and Frampton194].

Our assumption in relation to childhood trauma is rooted in the psychoanalytic history of MBT, but a rare aspect of this history is shared by modern psychiatry. Anna Freud and Dorothy Burlingham, who observed children’s responses to the bombing of London during the Second World War, noted that the objective danger a child was in appeared to be less predictive of a phobic reaction than the mother’s level of anxiety [Reference Freud and Burlingham195]. Their observations antedated the substantial interest in the vicarious learning of children in the context of social referencing [Reference Klinnert, Campos, Sorce, Plutchhik and Kellerman196,Reference Sorce, Emde, Campos and Klinnert197] and other related, particularly facially transmitted, ways of deriving emotional information about risk and danger [Reference Askew and Field198–Reference Olsson, Knapska and Lindstrom200].

In our developmental approach to trauma, we have suggested that adversity becomes traumatic when it is compounded by the sense that one’s mind is alone. Normally, an accessible other mind provides the social referencing that enables one to frame a frightening and otherwise overwhelming experience [Reference Allen, Fonagy, Bateman, Vermetten and Lanius186]. Recent experimental studies have provided strong support for this view. A study of conditioning found that vicarious safety learning (when participants watch a calm-looking demonstrator modeling safety) leads to better attenuation of a conditioned threat response than traditional direct safety learning training when a model is not present [Reference Golkar, Haaker, Selbing and Olsson201]. This observation suggests that the mind is programmed to be attuned to socially accessible agents so as to judge the threat response through the availability of another person, such as an attachment figure.

Three Communication Systems

In drawing out some of the clinical implications of the theory we have covered so far in this chapter, we have developed the idea that there are three “communication systems” which are associated with effective therapeutic help [Reference Bateman, Campbell, Luyten and Fonagy202]. These communication systems do not just apply to MBT; rather, we suggest that any form of meaningful psychotherapeutic help tends to involve the communication, internalization, and (re-)application of new forms of learning about oneself, and about oneself in relation to other people.

Summary of Empirical Research on MBT

In 2020, a Cochrane Review identified MBT as one of two therapies (the other being dialectical behavior therapy) with a reasonable evidence base for the treatment of BPD [Reference Storebø, Stoffers-Winterling and Völlm203]. A meta-analysis that included 33 randomized controlled trials (RCTs) of specialized psychotherapies for BPD compared with non-specialized psychotherapies for adult patients diagnosed with BPD supported the efficacy of MBT in patients with BPD [Reference Cristea, Gentili and Cotet204]. Another meta-analysis of psychological treatments for BPD, which covered 87 studies, found that MBT (along with schema therapy and reduced dialectical behavior therapy) was associated with higher than average effect sizes compared with treatment as usual (TAU), which was associated with lower than average effect sizes [Reference Rameckers, Verhoef and Grasman205]. A review of RCTs found that MBT for BPD tended to be associated with medium to large or very large effect sizes on a range of outcome measures, including BPD symptoms, educational attainment, and interpersonal functioning [Reference Volkert, Hauschild and Taubner206]. A systematic review similarly concluded that MBT is associated with significant improvements in BPD symptom severity, comorbid symptom severity, and quality of life [Reference Vogt and Norman207].

Bateman and Fonagy tested the effectiveness of MBT as a treatment for BPD in a series of RCTs. A day-hospital program of MBT was investigated in outcome studies culminating in an 8-year follow-up study—the longest follow-up of treatment for BPD conducted to date [Reference Bateman and Fonagy208]. Compared with TAU, MBT achieved decreases in suicide attempts, emergency-room visits, inpatient admissions, medication and outpatient treatment use, and impulsivity. Far fewer patients in the MBT group than in the TAU group met the criteria for BPD at follow-up (13% versus 87%). Patients in the MBT group also showed greater improvements in interpersonal and occupational functioning.

An intensive outpatient MBT program was found to be more effective than structured clinical management for BPD at the end of the 18-month treatment period [Reference Bateman and Fonagy209], particularly for patients with more than two personality disorder diagnoses [Reference Bateman and Fonagy210]. Compared with TAU, the outpatient MBT treatment resulted in lower rates of suicidal behavior and non-suicidal self-injury, and fewer hospitalizations. The primary outcome measure of 6-month periods free of suicidal behaviors, severe self-injurious behaviors, and hospitalization showed improvements from 0% to 43% in the structured clinical management group and 73% in the MBT group. In addition, the MBT group showed improved social adjustment and reduced depression, symptom distress, and interpersonal distress [Reference Bateman and Fonagy209]. An 8-year follow-up of this study found that, of the 73% of the sample who agreed to participate, a significantly higher proportion of patients in the MBT group (74%, compared with 51% in the structured clinical management control group) still met the primary criteria for recovery [Reference Bateman, Constantinou, Fonagy and Holzer211]. Finally, an RCT of MBT for comorbid BPD and antisocial personality disorder found that MBT was effective in reducing symptoms such as hostility, paranoia, anger, and frequency of self-injurious behavior and suicide attempts, and achieved improvements in negative mood, general psychiatric symptoms, interpersonal problems, and social adjustment [Reference Bateman, O’Connell and Lorenzini212].

An RCT in Denmark that investigated the efficacy of MBT compared with a less intensive manualized supportive group therapy in patients with BPD found that MBT achieved better results in terms of clinician-rated Global Assessment of Functioning [Reference Jørgensen, Freund and Boye213]. These results were sustained 18 months later [Reference Jørgensen, Bøye and Andersen214]. In a second study in Denmark, patients who were treated with partial hospitalization followed by group MBT showed significant improvements after treatment (average duration 2 years) on a range of measures, including Global Assessment of Functioning, hospitalizations, and vocational status, with further improvements at 2-year follow-up [Reference Petersen, Toft and Christensen215]. A multi-site RCT in the Netherlands that compared day-hospital MBT with specialist TAU in patients with BPD found that both forms of treatment were effective, although the MBT program was associated with higher patient acceptability, indicated by significantly lower early dropout (9% for MBT versus 34% for TAU) [Reference Laurenssen, Luyten and Kikkert216].

Another study in the Netherlands investigated the effectiveness of an 18-month manualized program of MBT in 45 patients with severe BPD [Reference Bales, Bateman, Bateman and Fonagy217]. Patients showed significant positive changes in symptom distress, social and interpersonal functioning, and personality pathology and functioning, and effect sizes were moderate to large; however, it should be noted that this study did not use a control group. A naturalistic longitudinal study in Norway that compared psychodynamic group-based treatment and MBT found that greater clinical severity of BPD was associated with poorer outcomes in the group-based treatment, whereas it had no significant effect on treatment outcome in the MBT group [Reference Kvarstein, Pedersen and Folmo218]. This observation supports Bateman and Fonagy’s finding that MBT may be particularly indicated for patients with more severe BPD [Reference Bateman and Fonagy210].

An RCT compared a form of MBT for eating disorders and specialist supportive clinical management for patients with eating disorders and symptoms of BPD. This study had a high dropout rate, with only 15 of the 68 participants who were eligible for randomization (22%) completing the 18-month follow-up, making the results difficult to interpret. However, MBT was associated with a greater reduction in patients’ concerns about body shape and weight as assessed by the Eating Disorder Examination [Reference Robinson, Hellier and Barrett219].

In an RCT of MBT for adolescents (MBT-A) (see also Chapter 14), 80 adolescents who presented to mental health services and had engaged in self-harm in the preceding month were randomized to MBT or TAU; 97% were diagnosed with depression, and 73% met the criteria for BPD. At the end of 12 months of treatment, MBT-A was found to be more successful than TAU in reducing self-harm and symptoms of depression. On the basis of self-report, the recovery rate was 44% for MBT-A and 17% for TAU, whereas on the basis of interview assessment the recovery rates were 57% and 32%, respectively. There was also a greater reduction in depressive symptoms and BPD diagnoses and traits in the MBT-A group. Both groups showed significant reductions in self-harm and risk-taking behavior, based on Risk-Taking and Self-Harm Inventory scores, following both a linear and a quadratic pattern. The group × time interaction term was also significant for both self-harm and risk taking, indicating that the linear decrease was significantly greater for the MBT-A group on both variables. At 12 months, self-harm scores were significantly lower for the MBT-A group [Reference Rossouw and Fonagy220]. However, a recent Cochrane Review of treatments for self-harm in adolescents has suggested that more evidence for effectiveness is needed for MBT-A [Reference Witt, Hetrick and Rajaram221].

An uncontrolled pilot study in Denmark of 34 female adolescents with BPD who participated in 12 months of structured mentalization-based group therapy found that of the 25 adolescents who completed the study, 23 displayed improvements in BPD symptoms, depression, self-harm, peer and parent attachment, mentalizing, and general psychopathology. Enhanced trust in peers and parents in combination with improved mentalizing capacity was associated with a greater reduction in borderline symptoms, pointing to a candidate mechanism for the efficacy of the treatment [Reference Bo, Sharp and Beck222]. On the basis of this pilot study, an RCT was developed to compare group MBT with TAU (which involved individual supportive care) for 112 adolescents with BPD in an outpatient setting; this study found that both treatments had similar effectiveness. The authors suggest that group MBT might be better indicated as an early-stage intervention, and that a group MBT program also including individual sessions should be tested [Reference Beck, Bo and Jorgensen223]. Finally, a naturalistic multi-informant study of MBT in 118 adolescent inpatients with personality pathology symptoms found that the participants improved on general and personality pathology measures, and in terms of health-related and generic quality of life. These improvements were clinically important as well as statistically significant, particularly in internalizing domains [Reference Jørgensen, Storebø and Bo224].

Concluding Remarks

In Part II of this book we shall discuss the translation of this developmental, evidence-based theory into clinical practice. To summarize, the central question is how to help patients to generate robust and effective mentalizing that will be conducive to constructive personal and social functioning in their life. Better mentalizing, particularly when attachment processes are activated, fuels epistemic learning within the social world and allows personal change over time. MBT is designed to stimulate this pathway of change, and is organized as a coherent set of interventions that are systematically applied over time to help the patient to learn about and practice mentalizing skills in their everyday life, while taking into account how attachment anxiety and other stressors can undermine the stability and effective use of mentalizing.