Mentalization: Ontogeny, Assessment, and Application in the Treatment of Borderline Personality Disorder

Bateman and Fonagy (3) define mentalization as “the mental process by which an individual implicitly and explicitly interprets the actions of himself or herself and others as meaningful on the basis of intentional mental states such as personal desires, needs, feelings, beliefs, and reasons. Within this definition, Bateman and Fonagy identify three dimensions of mentalization: the first related to two modes of functioning (i.e., implicit and explicit), the second related to two objects (i.e., self and other), and the third related to two aspects (i.e., cognitive and affective) of both the content and process of mentalizing.

Implicit mentalization refers to unconscious, automatic, or procedural operations of an individual’s ability to imagine his own and others’ mental states. Jon Allen (17) refers to conversational turn-taking as an example of implicit mentalization in operation. Without deliberate reflection, individuals naturally and instinctually hold the mind of their conversation partners in mind, anticipating when the other might want to respond in turn. In contrast, explicit mentalization involves deliberately exercised and conscious uses. The activities of psychotherapy provide an example of explicit mentalization. The therapist works to consciously and deliberately imagine the mental states of the patient and also encourages his patient to consciously and deliberately focus on her mental states. While the implicit and explicit modes define the two poles of the process of mentalizing, they are not mutually exclusive and completely discrete (17) . Individuals can alternate between these two modes and use them simultaneously. For example, a therapist can be consciously reflecting on the patient’s mental states but can also be attuned to the patient in more unconscious, intuitive ways.

With respect to objects, i.e., the self and the other, in the mentalization framework, each has a set of mental states, including feelings, thoughts, motives, intentions, beliefs, desires, and needs, to name a few. Fundamentally, the two objects in this framework mentalize interactively. The process of imagining one’s own thoughts and feelings potentially determines one’s idea of what is in the other’s mind and vice versa. For example, a person may become aware that he or she feels angry at a friend and then assumes the friend is also angry at him or her. Through talking and interacting, both parties can develop a more complex, enriched, and realistic idea of what is going on in their minds. In short, through mentalizing, the two friends can come to understand a misunderstanding. As illustrated by this example, the mental contents of both objects are dynamic: feelings, thoughts, and intentions constantly shift in response to changes in the interpersonal milieu.

A third dimension of the mentalization concept, in addition to the implicit/explicit dimension and the self/other dimension, relates to its cognitive and affective aspects. The content of mentalizing activity, that is, the “intentional mental states” in oneself and others, can be cognitively focused and affectively laden to varying degrees. Additionally, there are cognitive and affective aspects of the process of mentalizing. Mentalization requires a panoply of intact cognitive skills that enable individuals to imagine mental states with plausibility, flexibility, and complexity, but it optimally integrates this cognitive realm concerning reason and insight with emotion. The integration of cognitive and affective aspects of both the process and content of understanding mental states allows individuals to “feel clearly” and enhances “emotional knowing” (17) .

The domain of mentalization appears “all-encompassing,” so we will use these three dimensions of the concept (i.e., implicit/explicit modes, self/other objects, and cognitive/affective aspects) to anchor this review of the mentalization concept and its relationship to its conceptual sources and conceptual overlaps.

In a 1991 article titled “Thinking About Thinking,” Fonagy introduced his concept of mentalization, simply defined as “the capacity to conceive of conscious and unconscious mental states in oneself and others” (1) . Fonagy’s new use of the term combined the psychoanalytic idea of symbolization with the scientific and philosophical concept of theory of mind (18) . Frequently and without clear definition, the term mentalization had been intermittently scattered throughout psychoanalytic writings since the 1960s (19) . In their review of this literature, Lecours and Bouchard (20) report that the concept of mentalization signifies the basic intrapsychic transformation of one’s own inchoate somatic experiences into increasingly organized images, ideas, and words that could be modified, linked, and communicated. Other terms used interchangeably with mentalization in this literature include “metabolization” and “representation” (21), “symbolization” or “symbol formation” (22), “secondary mental processes” (23), and “alpha-function” (24) . An individual’s felt experience can be perceived in several forms ranging from physical (somatic and motor) to internally visualized (images and dreams) to interpersonally communicated, verbally articulated forms.

To illustrate the distinctions among these forms of mentalization just described, we can consider the experience of anxiety on these three levels of representation. Starting with somatic and motor representations, one can experience stomachache, sweaty palms, and pacing as markers of anxiety. At this level of representation, anxiety is experienced more viscerally than mentally. To move to an increasingly conscious level of mental representation, one can imagine anxiety-laden images or dreams, like showing up at school without one’s homework. Finally, at the most mentalized or self-reflective level, one can conceive of a mental representation of anxiety in an idea or thought such as “I am anxious because I am getting close to my boyfriend, and I am afraid of intimacy.” This verbally articulated representation of one’s affect state is not only the most easily and unambiguously communicated form but also the easiest to potentially link to other representations of felt experience. Of importance, all of these representations of anxiety facilitate awareness of one’s own internal state. Greater awareness of internal experience can be achieved through linking multiple representations of experience together.

Theory of mind refers to mental faculties that allow an individual to first appreciate the existence of different mental states in others and then to accurately identify others’ mental states (e.g., intentions, motives, beliefs, desires, and feelings) in order to interpret their behavior. The term theory of mind was introduced into the scientific literature by primatologists (18) who observed a chimpanzee’s ability to understand the intentions of an actor in film clips, which enabled her to predict the actor’s next move. Research on this particular social cognitive capacity expanded after developmental psychologists (25) introduced the “false-belief task,” the first experimental paradigm for studying theory of mind. In this experimental paradigm, children are told a character named Maxi puts a candy bar in a cupboard in the kitchen and then leaves the kitchen, after which time his mother comes in and moves the candy bar to a drawer. The child is then asked where Maxi will look for the candy. A child who has developed theory of mind will understand that Maxi falsely believes the chocolate is in the cupboard. Using this and other versions of the false-belief task, researchers have shown that a majority of children master this task between 4 and 6 years of age. A number of other experimental paradigms targeting theory of mind functions have been developed for use in both children and adults. The operationalization of theory of mind into experimental protocols has enabled researchers to localize its function to specific regions in the brain (26) and also to differentiate individuals with high-functioning autism and Asperger’s syndrome from normal comparison subjects (27) . Baron-Cohen et al.’s conception of the interpersonal functioning difficulties in autism as theory of mind deficits (28) influenced Fonagy to apply the theory of mind concept to borderline personality disorder. In his appropriation of the theoretically derived and clinically applied broad psychoanalytic term mentalization, Fonagy brought its internally or self-oriented and affectively rich dimensions to bear on the more empirically derived, externally or other-oriented, cognitively focused construct of theory of mind.

Fonagy outlined several salient propositions relevant to this new chimeric term. First and foremost, Fonagy emphasized a developmental model, contextualizing the formation of mentalization in the setting of secure early attachment relationships. More specifically, Fonagy argued that the primary caretaker’s marked and contingent mirroring of a child’s internal states within a secure attachment facilitates that child’s development of a capacity to mentalize. To clarify through an example, consider an infant who is distressed and crying. His caretaker responds, not by crying and looking distressed herself but rather with an exaggerated frown and furrowed brow, expressing both concern and loving. This constitutes an expression that not only mirrors the imagined emotional state of the baby but modifies it as a “re-presentation” (29) so that it is clearly congruent with the infant’s state but also differentiated as the caregiver’s response to it (2) . By mirroring the internal states of the child in a way that is both marked (i.e., the caretaker’s metabolized “representation” of it) and contingent (i.e., accurate and responsive), the caretaker helps the child solidify an understanding of an internal experience that he is only dimly aware of initially (2, 30) . The caretaker’s mirroring helps the child convert a felt, physical, sensory experience into a contained mental, conscious awareness. This mirroring process facilitates the development of a child’s capacity to mentalize his internal experience, which in turn enables him to regulate his affect and distress. As was earlier identified by Kohut (31), this mirroring process also theoretically contributes to a sense of self .

This developmental theory states that when the capacity to mentalize insufficiently develops, as is the case in borderline personality disorder, capacities for self-awareness and self-regulation remain impaired (2) . Here, while Fonagy suggested that the borderline personality disorder subject’s failures in mentalization represent a developmental deficit, born in the context of insecure attachment with insufficient mirroring, he also argued that subjects with borderline personality disorder who have backgrounds of trauma employ a defensive inhibition of mentalization, as a self-protective way to avoid considering the malicious intents of an abusive or neglectful figure. This original model of the mentalization deficits in borderline personality disorder combined theories of deficit and defense, which some authors have framed hard to reconcile (32), while others have claimed as quite compatible (33) . In Fonagy’s trauma model, the defensive inhibition of mentalization presumably arose in children who had an intact capacity to mentalize. But in his developmental model, he suggests that in the absence of secure early attachments, the capacity to mentalize inadequately develops, leading to a deficit.

More recently, Fonagy and Bateman (34) proposed a more complex relationship between early attachment, trauma, and borderline personality disorder that incorporates three mechanism by which mentalization becomes destabilized or impaired in borderline personality disorder: first as a deficit, second as a defense, and third as a derailment due to dysregulated affect. Their theory now notes that a variety of factors other than trauma, such as genetic contributions and temperament, may contribute to a suboptimal fit between infant and caregiver that interferes with the establishment of a secure attachment as well as the process of marked contingent mirroring. Incorporating research on family environment that suggests that factors such as neglect, lack of support, excessive control, and emotional maltreatment are predictive of borderline personality disorder, Fonagy and Bateman suggest that the family context may impair the development of mentalization both in combination with and apart from any incidents of trauma. This kind of family context, trauma, and even innate biological factors may cause a dysfunction and hypersensitivity in the stress-response system, leading to a cascade of hyperarousal, affective dysregulation, and inhibition of the orbitofrontal cortex, a brain region associated with mentalizing activity (34) .

The broad nature of Fonagy’s concept of mentalization contributes to its appeal as well as its potential to be confusing. The territory of the concept spans a number of other “conceptual cousins” (17), including mind blindness, emotional intelligence, insight, rationality, and imagination (17) ; theory of mind (18) ; and a number of psychoanalytic concepts including alpha function (24) and potential space (35, 36) . Reviews of the overlap between mentalization and a number of these concepts have been published elsewhere (17, 20, 36) . This review of conceptual overlaps is restricted to the related concepts of mindfulness, psychological mindedness, empathy, and affect consciousness, all of which have been operationalized into empirical measures and studied in relation to borderline personality disorder or integrated into psychotherapeutic treatments. In order to elucidate the domain of the mentalization concept, we will examine each of these four “conceptual cousins” and the ways in which they overlap with each other. A graphical depiction of the conceptual overlaps is summarized in Table 1 and Figure 1 .

Fonagy and collaborators have formulated borderline personality disorder as a syndrome organized around an unstable capacity for mentalization. The capacity for mentalization normally develops in the context of secure early attachments; in later attachment contexts, that capacity is vulnerable (2, 3) . When an individual is unable to mentalize, three “prementalistic” or “nonmentalistic” modes of thinking become evident, which Fonagy refers to as psychic equivalence mode, pretend mode, and teleological stance . In psychic equivalence, individuals equate what is in their mind with what is in the world, such that no alternative perspectives on reality exist and the contents of the mind become unbearably real (e.g., posttraumatic stress disorder flashbacks, transient paranoid ideation). Pretend mode, the converse of psychic equivalence, involves “decoupling” (2) or “cutting loose” (17) what is in the mind from reality. For example, an individual with borderline personality disorder might find himself or herself lost in a self-critical pessimistic train of thought that is unanchored to and unmodulated by reality. Developmentally, children acquire the ability to mentalize when they integrate these two modes and understand internal experience and external reality as “linked, but separate” (3) . Finally, in the teleological stance, an individual only accepts the existence of mental states when they are observed in concrete and physically evident ways. In the teleological stance, communication of internal mental states only occurs through action; ideas and even words mean very little. For example, cutting provides an index of internal pain and sexual intercourse or physical affection serves as an index for love or caring.

In the mentalization-based treatment model, the borderline personality disorder symptom profile is organized around these primitive nonmentalizing modes of thinking. The mentalization-based treatment model instructs clinicians to identify moments in which mentalizing fails and these prementalistic modes of functioning become manifest. Then the clinician and patient can work together to revive the patient’s capacity to mentalize instead of reverting to psychic equivalence mode, pretend mode, or teleological stance. Fully functioning mentalization capacities are established when patients can integrate psychic equivalence and pretend modes of functioning, thereby linking what is in their mind with what is observable in reality in a way that is separate but related.

In revising the mentalization concept, Fonagy’s ambition was to build a conceptual framework based on psychoanalytic theory, validated by scientific evidence, and applied effectively in clinical programs. Of importance, while he was developing the concept of mentalization, Fonagy was also involved in developing its measure, originally referred to as reflective self-function (55) and subsequently as reflective function. Initially designed as an adjunct to the Adult Attachment Interview (unpublished coding manual by Kaplan, Goldwyn, and Main, 1985), reflective function is a coding scheme that can be applied to a variety of clinical interviews, including the Parent Development Interview (unpublished protocol by Aber et al., 1985). The original scale expanded a subscale of the Adult Attachment Interview called the Metacognitive Monitoring Scale, which regarded a subject’s ability to reflect on his own thought process. Reflective function incorporated the core features of the Metacognitive Monitoring Scale, extending its scope to rate a subject’s ability to reflect on his own and others’ mental states.

The Reflective Function Scale’s validity is primarily derived from research demonstrating the ability of mothers’ prenatal reflective functioning to predict infant attachment security (55), a finding that has been replicated (29) . Reflective function is also shown to differentiate “difficult to treat” personality disordered inpatients at Cassel Hospital from matched normal comparison subjects recruited from an ambulatory general medical clinic (56) . While Fonagy did demonstrate statistically significant differences in Reflective Function Scale scores in subjects with borderline personality disorder compared to subjects without personality disorder, these findings supporting the scale’s validation were completed before the manualization of the Reflective Function Scale in 1998 and have not been replicated since then.

The Reflective Function Scale has been applied as an outcome measure in the Borderline Personality Disorder Research Foundation/Personality Disorder Institute randomized, controlled trial comparing dialectical behavioral therapy, transference-focused psychotherapy, and supportive psychotherapy (9) . However, its validity and stability as an outcome measure are uncertain because adequate reports of its test-retest reliability and convergent and divergent validity are still needed. While interrater reliability is well established in a number of studies (intraclass correlation coefficient ranging from 0.70 to 0.91) (9, 29, 55, 56), test-retest reliability of the scale is not established. Scores for reflective function may vary within the same therapist regarding different patients and can be below average for a therapist regarding their patients after 1 year of treatment with transference-focused psychotherapy (57) . In short, the reflective function of an individual can vary depending on who he or she is mentalizing about, and even therapists can have below-average scores for reflective functioning.

Aside from the need for further empirical elaboration of the validity and reliability of the Reflective Function Scale, there are limitations in interpreting the meaning of a given reflective functioning score. The capacity that is assessed by the Reflective Functioning Scale is multidimensional, with factors such as plausibility, consistency, complexity, and originality. However, the grading is done using a unidimensional score that cannot be submitted to factor analysis. In two different transcripts coded for reflective functioning with a score of 3, one transcript may reflect a consistently superficial, clichéd, and general understanding of mental states, while the other transcript reflects a highly variable capacity to understand mental states with some moments of antireflectiveness and other moments of marked reflectiveness.

Adequate evidence demonstrating that deficits in reflective functioning distinguish individuals with borderline personality disorder from individuals with other forms of psychopathology is lacking. At present, reflective functioning has been shown to be predictive of borderline personality disorder diagnosis only in subjects with histories of abuse (56) . Additionally, Fonagy and colleagues readily admit that mentalization capacities of individuals with borderline personality disorder are variable. The reflective functioning of an individual with borderline personality disorder is expected to vary depending on the relationship context in question, the level of distress, or the intensity of affect. Reflective functioning scores in individuals with borderline personality disorder might be sensitive to the context or timing of the assessment. Therefore, the assessment of mentalization as a marker of borderline personality disorder remains highly problematic.

Further research using the Reflective Functioning Scale is limited by the time-consuming and costly nature of the instrument. Adjacent concepts such as theory of mind, mindfulness, psychological mindedness, empathy, and affect consciousness have been operationalized into less cumbersome, albeit partial self-report measures. An exploration of findings related to these concepts may indicate simpler ways to evaluate the different dimensions of the mentalization concept in relation to borderline personality disorder psychopathology. As mentioned in the introduction, enhancing reflective function is thought to be a common feature of all psychotherapeutic approaches. Fonagy and Bateman propose that without changes in mentalizing, no improvements in functioning will be seen as a result of such treatments. Using self-report measures of overlapping constructs that isolate different dimensions of mentalizing relevant to different psychotherapeutic modalities would provide empirical evidence of their particular mechanisms.

A review of the concept of mentalization reveals the breadth of its territory, spanning from the psychoanalytic domain of understanding and transforming internal experience into mentally contained forms to the social cognitive and developmental research domain of imagining the mental states of others in interpersonal interactions. An analysis of the overlap between mentalization and other related concepts, such as mindfulness, psychological mindedness, empathy, and affect consciousness, helps to clarify similarities and distinctions that are outlined in Table 1 and Figure 1 . Research investigating the correlation of measures that operationalize these concepts may elucidate these relationships and also potentially lead to localization of mentalization deficits, in particular its implicit, explicit, self-oriented, other-oriented, cognitive, and affective components.

Ultimately, Fonagy’s introduction of the concept of mentalization has catalyzed the development of novel and effective treatments as well as promising avenues of inquiry in psychopathology, psychotherapy, developmental psychology, and neuroscientific realms of research. However, one obstacle to its empirical utility is that it remains difficult to assess. While the assessment tool of reflective functioning importantly measures cognitive processes related to mentalization in an attachment context, the validity of this measure is underdeveloped, and it remains difficult to employ in large-scale research and routinely in clinical settings. Research to further develop the validity of reflective functioning, its relationship with related scales, and its relationship to borderline personality disorder and other diagnostic groups is needed to elucidate the boundaries of the concept and its usefulness in understanding and treating borderline personality disorder.

Presented in part at the 160th annual meeting of the American Psychiatric Association, San Diego, May 19–24, 2007. Received Aug. 28, 2007; revision received Feb. 24, 2008; accepted May 1, 2008 (doi: 10.1176/appi.ajp.2008.07081360). From the Department of Psychiatry, Harvard Medical School, Belmont, Mass. Address correspondence to Dr. Choi-Kain, Department of Psychiatry, McLean Hospital, 115 Mill St., Belmont, MA 02478; [email protected].

Supported by grants from the Program for Minority Research Training in Psychiatry and the McLean Hospital Psychosocial Fellowship (LWC).

The authors report no competing interests.

The authors thank Glen Gabbard, M.D., for his critical review and revisions on this manuscript.