Mentalization

Abstract

Mentalization, a type of mental process, is the capacity to read behaviour, one's own and other people's, as arising from mental states such as beliefs, desires, and feelings rather than surface appearances alone. The construct has two independent origins that converged: the developmental science of theory of mind, which asked when children come to attribute beliefs, and the clinical tradition begun by Peter Fonagy, which recast the same capacity as reflective function and made its failure central to personality disorder. This article follows the false-belief task from Wimmer and Perner through the autism findings of Baron-Cohen, Leslie, and Frith, the Wellman meta-analysis, the implicit-versus-explicit debate and two-systems account, the mentalizing brain network, and the attachment-based model behind mentalization-based treatment. Three interactive demonstrations let the reader run a false-belief task, trace the developmental curve, and drive the arousal switch that shuts mentalizing down under stress.

Keywords: mentalization, theory of mind, false belief, reflective function, mentalization-based treatment

Mentalization is the mental process by which a person interprets action, their own and that of others, as flowing from underlying mental states, and so is able to treat other people, and themselves, as psychological agents with minds rather than as objects that merely move. In the Medical Subject Headings vocabulary it is catalogued as descriptor D000076563, defined as the mental process by which an individual implicitly and explicitly interprets the actions of self and others as meaningful on the basis of intentional mental states such as personal desires, needs, feelings, beliefs, and reasons. The term is closely bound to theory of mind, the ability to attribute mental states, but it reaches wider: it covers the automatic and the deliberate, the cognitive and the emotional, the reading of the self as much as the reading of others, and it carries a clinical charge that theory of mind, born in comparative and developmental laboratories, does not. The two traditions grew up apart, one measuring when a child first understands that a belief can be false, the other asking why some adults chronically misread the minds around them, and this article follows both as they converged on a single capacity. The sections below trace the capacity from the theory-of-mind idea and the false-belief task that operationalised it, through its developmental course, its implicit precursors, and its dependence on executive control and a specific brain network, to the attachment-based account of reflective function and the clinical treatment built on it.

What Mentalization Is

Mentalization names a capacity that most people exercise continuously and without effort: seeing a colleague hurry past without a greeting and inferring preoccupation rather than rudeness, or noticing one's own irritation and recognising it as tiredness rather than as a fact about the person who happens to be nearby. In each case behaviour is interpreted through an inferred inner state, a belief, a wish, a feeling, an intention, and it is this interpretive move, holding a mind in mind, that the term picks out. The capacity is imaginative in a specific sense: mental states are not observable, so their attribution is always an inference, a best guess about what is going on behind an action, and it can be done well or badly.

Several features mark the construct out and explain why it is broader than any single laboratory test. It is organised along four dimensions, set out in full below: it runs from the automatic to the controlled, from fast intuitive reading to slow deliberate reasoning; it is directed at the self as much as at others, and the two develop together; it has a cognitive side, reasoning about beliefs and intentions, and an affective side, resonating with feeling, which can dissociate; and it draws on external cues, the face and voice, as much as on inner felt experience. Cutting across all four, it is not all-or-none but a capacity that varies across people, across situations, and within the same person from moment to moment, and that can be temporarily lost.

That last point separates mentalization from the more static notion of theory of mind. In developmental psychology the question was typically whether a child possesses a theory of mind, a competence that is either present or absent at a given age. Mentalization, arising in a clinical setting, is treated instead as a graded and fragile performance: an adult who is perfectly capable of reading minds when calm may lose the capacity in the grip of strong emotion, and the interest of the clinical tradition lies precisely in those failures and their repair. This is metacognition turned outward and inward at once, and it borders the study of empathy, emotional intelligence, and social understanding without being reducible to any of them.

The Four Polarities of Mentalizing

The clinical tradition sharpened these features into four dimensions, or polarities, along which any act of mentalizing can be placed, and the modern account reads a person's mentalizing profile as their characteristic balance across all four (Fonagy & Luyten, 2009; Luyten, Campbell, Allison, & Fonagy, 2020). The first polarity runs from the automatic and implicit to the controlled and explicit: fast, intuitive, unreflective reading of a mind at one pole, slow and effortful reasoning about it at the other. The second runs between self and other, the mentalizing of one's own states and the mentalizing of someone else's, capacities that usually grow together but can come apart. The third runs between cognitive and affective mentalizing, the reasoning about beliefs and intentions on one side and the resonant understanding of feeling on the other. The fourth runs between internal and external: reading a mind from its outward signs, the face, the voice, the posture, as against reading it from inner felt thoughts and feelings.

Framed this way the polarities become a clinical map, because different presentations sit at different imbalances. A person locked at the external, automatic pole may read faces and tones with quick, sometimes uncanny accuracy yet be unable to reflect on the inner states behind them; another, collapsed onto the self pole, loses sight of the other; a third, stranded in cognitive mentalizing, reasons fluently about mental states without feeling their weight. Mature mentalizing is not a fixed point on any axis but the flexibility to move along all four as the situation demands, and the loss of that flexibility, a rigid over-reliance on one pole, is much of what the clinical work sets out to restore (Luyten, Campbell, Allison, & Fonagy, 2020).

Theory of Mind and Its Origins

The intellectual root of the developmental strand is a single question asked about a chimpanzee. David Premack and Guy Woodruff, studying whether an ape could infer the goals of a human actor, coined the phrase theory of mind and defined it precisely: an individual has a theory of mind if it imputes mental states to itself and others, a theory in the sense that such states are not directly observable and are used to predict behaviour (Premack & Woodruff, 1978). Their evidence for the chimpanzee was contested, but the framing outlasted the data, because the commentators who responded to the paper supplied the tool that would define the field for the next forty years.

Chief among them was the philosopher Daniel Dennett, who pointed out that most apparent evidence of mind-reading can be explained without crediting the reader with any grasp of belief, since predicting that an agent will pursue what it wants, where it is, requires no notion that the agent might be wrong about the world. The decisive test, Dennett argued, is a situation in which the agent holds a false belief: only someone who can predict that another will act on a belief that does not match reality has shown that they represent the belief as a belief, distinct from the state of the world it is about (Dennett, 1978). This is the intentional stance made operational. A person who tracks only the world will predict action from the world; a person who tracks minds will, in the crucial case, predict action from the false belief and against the facts. The false-belief prediction thus became the litmus test for genuine mentalizing, and the search for a paradigm that could put it to children began at once.

The False-Belief Task

Heinz Wimmer and Josef Perner built that paradigm. In their change-of-location task a child watches a story in which a character, Maxi, puts an object in one place and leaves; in his absence the object is moved to a second place; Maxi then returns, and the child is asked where Maxi will look for it (Wimmer & Perner, 1983). The correct answer is the original location, where Maxi wrongly believes the object still is. A child who answers with the object's true, current location has failed to separate Maxi's belief from reality and has, in effect, credited Maxi with knowledge the child possesses but Maxi does not. Wimmer and Perner found a sharp developmental shift: most children below about four years pointed to the object's real location, most above it to Maxi's belief. The structure of the task is laid out in Figure 1, in the familiar Sally-Anne form that Baron-Cohen and colleagues later made standard.

The task acquired its lasting importance when it was turned on autism. Simon Baron-Cohen, Alan Leslie, and Uta Frith gave the Sally-Anne task to children with autism, children with Down syndrome, and typically developing children, matched so that the autistic group had the highest mental age of the three. The great majority of the autistic children failed the false-belief question while passing control questions about what had actually happened, whereas most of the children in the two comparison groups passed, even though several had lower mental ages (Baron-Cohen, Leslie, & Frith, 1985). The pattern could not be attributed to general intellectual delay, and the authors proposed that autism involves a specific impairment of the mechanism that represents mental states, a mindblindness at the level of belief attribution. The claim reshaped the understanding of autism and made the false-belief task the most influential single measure in the field. The demonstration below runs a false-belief task directly, letting the reader set whether the character witnessed the object being moved and see how that single fact changes what the character believes, and so where they will look.

The Development of Mentalizing

No single experiment fixes the age at which children come to understand false belief, because the answer depends on the exact wording, the number of control questions, and dozens of other procedural choices. What settles the developmental picture is aggregation across the whole literature. Henry Wellman, David Cross, and Julanne Watson pooled results from 178 separate studies and several hundred conditions, and found a highly consistent trajectory beneath the procedural noise: children below about three and a half years perform below chance, systematically pointing to the object's real location, and performance climbs steadily to well above chance by about four and a half to five years (Wellman, Cross, & Watson, 2001). The transition is genuine and not an artifact of any one method, because manipulations that made the task easier or harder shifted the whole curve up or down together without erasing the developmental change itself, and the same ordering held across countries and languages. The meta-analysis turned a scattered set of findings into a robust developmental fact.

The shape of that trajectory is worth examining in its own right, because it is smooth rather than a sudden switch. The proportion of children who answer the false-belief question correctly rises gradually with age, passing through the fifty-percent point somewhere just before the fourth birthday, and this graded, S-shaped rise is what the Worked Example below computes and the demonstration draws. Table 1 places the false-belief transition among the other milestones of mentalizing development that the literature has charted, from the early precursors that precede it to the more demanding forms that follow.

Table 1

Milestones in the Development of Mentalizing

Note. Representative ages for landmarks in the growth of mentalizing; individual children vary widely, and the ages mark central tendencies rather than fixed thresholds. The explicit first-order transition is the one quantified by the meta-analysis of false-belief studies (Wellman, Cross, & Watson, 2001); the implicit precursor is treated in the next section.

Implicit and Explicit Mentalizing

The tidy story of a competence acquired around four years was complicated by evidence that infants far too young to pass any verbal task are nonetheless sensitive to what others believe. Kristine Onishi and Renée Baillargeon adapted the change-of-location scenario for fifteen-month-olds using looking time as the measure. Infants watched an actor hide a toy and then saw it moved, or not moved, in the actor's presence or absence, and then watched the actor reach. The infants looked reliably longer when the actor reached in a way that violated the belief she should hold given what she had witnessed, as if they expected her to act on her belief rather than on the true state of the world (Onishi & Baillargeon, 2005). Because the infants could neither speak nor be asked anything, the finding suggested a form of belief tracking that is implicit, automatic, and present years before the explicit task is passed.

Ágnes Kovács, Ernő Téglás, and Ansgar Endress pushed the point further and into adulthood. In their task the mere presence of another agent who held a belief about where a ball was located altered how quickly participants, infants and adults alike, detected the ball, even when the other agent's belief was entirely irrelevant to the participant's own task and there was no reason to compute it (Kovács, Téglás, & Endress, 2010). Adults could not help but represent the other's belief; it intruded on their reaction times automatically. This automatic, hard-to-suppress belief tracking looks quite unlike the effortful reasoning a four-year-old must marshal to pass the Sally-Anne task.

Ian Apperly and Stephen Butterfill drew the two bodies of evidence together in a two-systems account, by analogy with dual-process theories elsewhere in cognition. On their proposal humans have two ways of tracking beliefs: an early-developing, fast, automatic, and cognitively cheap system that handles a limited range of belief-like states with little demand on other resources, and a later-developing, flexible, and effortful system that can represent beliefs in their full generality but draws heavily on language and executive control (Apperly & Butterfill, 2009). The first system explains the infant and reaction-time findings; the second explains the explicit task and its late, resource-hungry arrival. The division maps closely onto the automatic-versus-controlled distinction that the clinical tradition independently drew within mentalization, and it reframes the developmental puzzle: the question is not simply when mind-reading appears but how many systems are doing it.

Mentalizing and Executive Function

If explicit false-belief understanding is the effortful, resource-hungry system, one would expect it to depend on the general machinery of cognitive control, and it does. Passing the false-belief task requires more than representing a belief; it requires holding that belief in mind while inhibiting the pull of one's own knowledge of where the object really is, and this is a demand on executive function. Josef Perner and Birgit Lang reviewed the accumulating evidence that theory-of-mind development and executive control develop in lockstep, correlate strongly even after age and language are controlled, and are jointly impaired in autism (Perner & Lang, 1999). The tight coupling admits more than one reading: executive control may be a prerequisite that lets an already-formed concept of belief be expressed in action, or the very act of constructing a concept of mind may build the capacity for self-control, and the review laid out both possibilities without forcing a choice.

Stephanie Carlson and Louis Moses supplied the sharpest empirical anchor. Testing preschoolers on a battery of theory-of-mind tasks and a battery of inhibitory-control tasks, they found a strong association between the two that survived controls for age, verbal ability, and general intelligence, with inhibitory control, the ability to suppress a prepotent response, the executive component most closely tied to mentalizing performance (Carlson & Moses, 2001). The link makes sense of the false-belief task's characteristic error: the child who blurts out the object's true location is failing to inhibit their own privileged knowledge, not necessarily failing to understand belief in principle. It also connects the development of mentalizing to working memory and inhibition, and it foreshadows the clinical observation that mentalizing fails precisely when arousal overwhelms the control needed to sustain it.

The Neural Basis of Mentalizing

Functional neuroimaging has mapped mentalizing onto a consistent and now well-replicated network. When people read stories or watch animations that require attributing mental states, and these are compared with closely matched materials that do not, a characteristic set of regions activates: the medial prefrontal cortex, the temporo-parietal junction bilaterally but especially on the right, and the temporal poles. Helen Gallagher and Christopher Frith reviewed the early imaging studies and identified this network, arguing that the medial prefrontal cortex in particular supports the decoupling of mental states from reality that mentalizing requires, the same representational feat that the false-belief task demands of a four-year-old (Gallagher & Frith, 2003). Chris and Uta Frith argued that this capacity to represent other minds has a dedicated biological basis rather than being a by-product of general reasoning (Frith & Frith, 1999), and later consolidated the accumulated imaging into the same three-region network, assigning each node a role in the mentalizing computation (Frith & Frith, 2006).

The strongest claim for functional specialisation came from Rebecca Saxe and Nancy Kanwisher, who isolated a region of the right temporo-parietal junction that responded far more when participants read about a person's thoughts and beliefs than when they read about that same person's physical appearance, their bodily sensations, or other true facts about them (Saxe & Kanwisher, 2003). The region was not a general-purpose social or self-referential area; it was engaged specifically by the attribution of beliefs. This selectivity gave the belief-attribution system a candidate cortical home and sharpened the whole enterprise, moving it from a diffuse claim that social thinking recruits frontal and temporal cortex to a testable proposal that a particular patch of cortex is recruited by a particular representational task. The convergence of the developmental, the individual-difference, and the neural evidence on the representation of belief is one of the tighter cross-level stories in cognitive science.

Reflective Function and Attachment

While the developmental laboratories were charting theory of mind, a separate tradition was building around the same capacity from the clinic. Peter Fonagy took the term mentalization from French psychoanalysis and gave it an empirical form that could be measured and studied. His central move was to operationalise the capacity as reflective function, the demonstrated ability to understand behaviour in terms of mental states, and to score it from the way a person talks about their own and others' minds in the Adult Attachment Interview. In a study of expectant parents, Fonagy, Miriam Steele, Howard Steele, George Moran, and Anna Higgitt found that a parent's reflective function, assessed before the birth, predicted the security of the infant's later attachment to that parent, and did so most strikingly among parents whose own histories were adverse: high reflective function appeared to buffer the transmission of insecurity across generations (Fonagy, Steele, Steele, Moran, & Higgitt, 1991).

Fonagy and Mary Target developed this into a developmental theory in which mentalizing is not merely a cognitive attainment but an achievement of relationships. On their account the capacity grows out of secure attachment: an infant comes to understand its own mind by being accurately understood, by a caregiver who reflects the infant's states back in a marked and manageable form, and a failure of that early mirroring leaves the self poorly mentalized and vulnerable (Fonagy & Target, 1997). Before mentalizing matures, they argued, the child experiences mind in more primitive modes, and these same modes resurface in adults whenever mentalizing fails. In psychic equivalence, inner states are felt to be literally true of the outer world, with no sense that a thought is only a representation that might be mistaken: a patient who feels worthless experiences the worthlessness as an established fact about themselves rather than as a feeling, and a terrifying mental image is met as an actual threat. In the pretend mode, thoughts are sealed off from reality altogether and cannot be linked back to it, so that a person may talk fluently and at length about their feelings in a circular, affectless way that touches nothing felt. In the teleological mode, states of mind are credited only when they issue in physical proof: care is believed only if it takes a tangible form, and distress is expressed, or reassurance demanded, through concrete acts such as self-harm or insistence on some visible demonstration. Mature mentalizing integrates these three, holding mental states as representations that are about reality without being identical to it. This account gave the clinical tradition a bridge to attachment theory and a developmental story of its own, parallel to the theory-of-mind trajectory but framed around emotional relationships rather than cold cognition.

Mentalization in Clinical Practice

The clinical payoff came in the treatment of borderline personality disorder, which Fonagy and Anthony Bateman reconceived as, in large part, a disorder of mentalizing. On their model the emotional storms, unstable relationships, and impulsivity of the condition follow from a fragile capacity to mentalize that collapses under the very attachment-related arousal that close relationships provoke: as emotional intensity rises past a threshold, controlled, reflective mentalizing shuts down and the person reverts to the prementalizing modes described above, psychic equivalence, pretend, and teleological, experiencing thoughts as literal realities or demanding physical proof of others' intentions, and losing the ability to read minds accurately (Fonagy & Luyten, 2009). This biobehavioural switch, in which high arousal deactivates the neural systems for controlled mentalizing and hands control to faster automatic responses, is the target of the arousal demonstration below.

Bateman and Fonagy built a structured therapy on this analysis, mentalization-based treatment, whose single overarching aim is to restore and strengthen mentalizing rather than to challenge distorted thoughts or teach coping skills directly, allying it with but distinguishing it from cognitive behavioural therapy. They tested it against ordinary psychiatric care in a randomised controlled trial and found substantial reductions in self-harm, suicide attempts, hospital use, and symptoms in the treated group (Bateman & Fonagy, 1999), and a later trial against a well-specified control condition confirmed the advantage on the most serious outcomes (Bateman & Fonagy, 2009). Lois Choi-Kain and John Gunderson reviewed the construct as it stood, mapping its overlap with related ideas and pressing for sharper measurement (Choi-Kain & Gunderson, 2008), and Patrick Luyten, Chloe Campbell, Elizabeth Allison, and Fonagy later set out the mature mentalizing approach across the range of disorders, arguing that a compromised capacity to mentalize, and a related failure to trust and learn from others, is a common thread in many forms of psychopathology rather than a feature of one diagnosis (Luyten, Campbell, Allison, & Fonagy, 2020). The demonstration below models the arousal switch at the centre of this account, letting the reader raise emotional arousal and watch controlled mentalizing give way to automatic responding.

Epistemic Trust and Social Learning

Fonagy's more recent work reframes what mentalization-based treatment ultimately repairs, placing at its centre a concept drawn from social epistemology: epistemic trust, the readiness to treat what another person communicates as trustworthy, relevant, and worth taking in as general knowledge about the world rather than a one-off (Fonagy & Allison, 2014). Human beings depend on such trust to learn from one another at all, since most of what anyone knows is acquired not from direct experience but on the word of others, and that channel stays open only where the listener judges the speaker to be a reliable source who has the listener's interests in mind.

Against this background, early adversity does more than weaken mentalizing. A child whose caregivers are frightening, neglectful, or chronically unattuned learns, adaptively, not to take in what they say, and this epistemic vigilance can harden into a pervasive hypervigilance or a defensive freeze, a state Fonagy and colleagues call epistemic petrification, in which no communication is allowed through as genuinely meant (Fonagy, Luyten, & Allison, 2015). Sealed off in this way, the person cannot readily update from new social experience, which is why difficulties prove so stubborn and why relationships, including the therapeutic one, fail to correct them: the corrective information is simply not admitted.

This recasts how the therapy is thought to work. On the newer model the value of mentalization-based treatment lies not only in steadying affect and rebuilding reflective capacity but in reopening epistemic trust, because a clinician who reliably shows that the patient's mind has been accurately understood offers exactly the marker of trustworthiness that lowers epistemic vigilance, and once the channel reopens the patient can begin again to learn from the people around them, inside the consulting room and beyond it (Fonagy & Allison, 2014; Luyten, Campbell, Allison, & Fonagy, 2020). The therapeutic relationship becomes less a treatment in itself than a gateway back into ordinary social learning.

Worked Example

The developmental curve of false-belief understanding is worth computing by hand, because it makes the meta-analytic result concrete and it is exactly what the development demonstration draws. Model the probability that a child of age a years passes the false-belief task as a logistic function rising from near zero to near one:

p(a) = 1 / (1 + exp(−k × (a − a₀)))

with the midpoint a₀ = 3.92 years, the age at which half of children pass, and the slope k = 2.0 per year setting how sharply the transition occurs. Take a two-year-old, a = 2. The exponent is −2.0 × (2 − 3.92) = −2.0 × (−1.92) = 3.84, so p(2) = 1 / (1 + exp(3.84)) = 1 / (1 + 46.5) = 0.021, about 2 percent: two-year-olds essentially never pass. For a three-year-old, the exponent is −2.0 × (−0.92) = 1.84, giving p(3) = 1 / (1 + 6.30) = 0.137, about 14 percent. At the midpoint itself, a = 3.92, the exponent is zero and p = 1 / (1 + 1) = 0.50 by construction, the half-way point of the transition.

Now step across the fourth birthday. For a = 4 the exponent is −2.0 × (0.08) = −0.16, so p(4) = 1 / (1 + exp(−0.16)) = 1 / (1 + 0.852) = 0.540, about 54 percent, just past chance. For a five-year-old the exponent is −2.0 × (1.08) = −2.16, giving p(5) = 1 / (1 + 0.115) = 0.897, about 90 percent, and for a six-year-old p(6) = 1 / (1 + exp(−4.16)) = 1 / (1 + 0.0156) = 0.985, about 98 percent. The curve therefore climbs from 2 percent at age two, through 14 percent at three and 54 percent at four, to 90 percent at five and near-ceiling by six, crossing the fifty-percent line just before the fourth birthday. This is the graded, S-shaped transition the meta-analysis found beneath the scatter of individual studies, and setting the development demonstration to these parameters reproduces the same values point for point.

Discussion

Mentalization is the point at which two independent research traditions turned out to be studying one capacity. From the developmental side came the concept of theory of mind, the false-belief task that operationalised it, the autism findings that made it clinically urgent, the meta-analysis that fixed its developmental course, and the imaging that gave belief attribution a place in the brain (Premack & Woodruff, 1978; Wimmer & Perner, 1983; Baron-Cohen, Leslie, & Frith, 1985; Wellman, Cross, & Watson, 2001; Saxe & Kanwisher, 2003). From the clinical side came reflective function, its roots in attachment, and a treatment that works by rebuilding the capacity when arousal has stripped it away (Fonagy & Target, 1997; Bateman & Fonagy, 2009). The two accounts illuminate each other. The developmental finding that explicit mentalizing depends on executive control (Carlson & Moses, 2001) and the clinical observation that mentalizing collapses under emotional arousal (Fonagy & Luyten, 2009) are the same fact seen from two directions: a capacity that is effortful and controlled is, for that very reason, fragile, easily lost when the resources it depends on are consumed by inhibition failure in a child or by emotional flooding in an adult. The clinical account has since widened: the same collapse that strips away mentalizing also closes off epistemic trust, so that a person in crisis both misreads the minds around them and stops taking in what those minds try to tell them, two faces of a single withdrawal from the social world (Fonagy & Allison, 2014).

Open questions remain, and they are substantial. The two-systems account leaves unsettled how the early automatic system and the late explicit one relate, and whether they are truly distinct or two ends of one developing competence; the automatic belief-tracking that motivates the early system is real in the paradigms that reveal it, but its scope, and whether it amounts to representing belief at all, remains contested (Apperly & Butterfill, 2009; Kovács, Téglás, & Endress, 2010). The clinical construct, for all its therapeutic success, is broad enough that sharpening its measurement and separating it from neighbouring ideas such as empathy and emotional intelligence is ongoing work (Choi-Kain & Gunderson, 2008). What is not in doubt is the centrality of the capacity itself. Reading behaviour as the expression of a mind is so basic to human social life that its absence, whether the developmental absence of autism or the situational absence of an adult in emotional crisis, is among the most consequential things that can go wrong in the mind, and the interactive figures above are illustrative renderings of the phenomena, the false-belief task, its developmental curve, and the arousal switch, that the two traditions independently uncovered.

Common Misconceptions

Mentalization and theory of mind are just two names for the same thing.

They overlap but are not identical. Theory of mind, from developmental and comparative work, is typically treated as a competence that is present or absent; mentalization, from the clinic, is a broader and graded capacity that includes the self as well as others, feeling as well as belief, and the automatic as well as the deliberate, and that can be transiently lost (Fonagy & Target, 1997; Luyten, Campbell, Allison, & Fonagy, 2020).

Passing the false-belief task shows mind-reading first appears at about four years.

The four-year transition is for the explicit, verbal task. Infants of fifteen months already show implicit sensitivity to others' false beliefs in looking-time measures, and adults track others' beliefs automatically, so belief sensitivity long predates the explicit milestone (Onishi & Baillargeon, 2005; Kovács, Téglás, & Endress, 2010).

Mentalizing is a purely cognitive, unemotional kind of reasoning.

It has a cognitive side, reasoning about beliefs and intentions, and an affective side, understanding and resonating with feelings, and the two can dissociate. The clinical tradition stresses that mentalizing depends on emotional regulation and fails under high arousal, making it anything but cold computation (Fonagy & Luyten, 2009).

A capable adult mentalizes reliably in every situation.

Mentalizing is a fragile performance, not a fixed possession. Because it draws on controlled processing, it is lost when emotional arousal in a close relationship overwhelms the systems that sustain it, which is why an otherwise insightful person can badly misread others in the grip of strong feeling (Fonagy & Luyten, 2009; Bateman & Fonagy, 2009).

Glossary

Affective mentalizing.

The understanding and resonant appreciation of emotional states, one's own and others', as distinct from the more cognitive attribution of beliefs and intentions; the two can dissociate.

Attachment.

The emotional bond between an infant and a primary caregiver; in Fonagy's account the relational matrix within which the capacity to mentalize develops, with secure attachment fostering it.

Automatic mentalizing.

Fast, intuitive, unconscious reading of mental states, requiring little effort or cognitive resource; the mode captured by infant looking-time and adult reaction-time findings and by the early system of the two-systems account.

Cognitive mentalizing.

The reasoning-based side of mentalizing, the attribution of beliefs, intentions, and reasons, as distinct from the resonant understanding of feeling; one pole of the cognitive-affective polarity, from which it can dissociate.

Controlled mentalizing.

Slow, effortful, explicit reasoning about mental states, dependent on language and executive control; the mode assessed by the verbal false-belief task and the one that fails under high arousal.

Epistemic trust.

The readiness to treat what another person communicates as trustworthy, relevant, and generalisable to oneself; its opening is, on Fonagy's later account, what allows learning from others and what mentalization-based treatment sets out to restore.

Epistemic vigilance.

A wariness towards communicated information that treats it as potentially unreliable; adaptive in moderation, it can harden after early adversity into a hypervigilance or defensive freeze that seals the person off from social learning.

External mentalizing.

The reading of mental states from outward, observable cues such as facial expression, tone of voice, and posture, rather than from inner felt experience; one pole of the internal-external polarity.

False-belief task.

A test in which a character holds a belief the child knows to be false; predicting that the character will act on the false belief rather than on reality is the criterion for understanding belief as a representation.

Intentional stance.

Dennett's term for the strategy of predicting an agent's behaviour by treating it as a rational holder of beliefs and desires; the false-belief case is where this stance shows itself as genuine mind-reading.

Internal mentalizing.

The reading of mental states from inner, felt thoughts and feelings rather than from outward signs; one pole of the internal-external polarity, complementary to external mentalizing.

Mentalization-based treatment.

A structured psychotherapy developed by Bateman and Fonagy whose central aim is to restore and strengthen the patient's capacity to mentalize, with demonstrated efficacy for borderline personality disorder.

Mentalization.

The capacity to interpret behaviour, one's own and others', in terms of underlying mental states such as beliefs, desires, feelings, and intentions; holding a mind in mind.

Pretend mode.

A prementalizing mode in which thoughts are sealed off from reality altogether, treated as having no bearing on the world, the counterpart to the over-literal collapse of psychic equivalence.

Psychic equivalence.

A prementalizing mode in which inner states are felt to be literally true of the outer world, so that a thought or fear is experienced as a fact rather than as a representation that might be mistaken.

Reflective function.

The operationalised, measurable form of mentalizing, scored from how a person represents mental states in the Adult Attachment Interview; a parent's reflective function predicts the infant's attachment security.

Social learning.

The acquisition of knowledge from what others communicate rather than from direct experience; on Fonagy's account it depends on epistemic trust and is blocked when epistemic vigilance hardens into petrification.

Teleological mode.

A prementalizing mode in which mental states are acknowledged only when expressed in physical action or outcome, so that care or intention is credited only if it produces a tangible result.

Temporo-parietal junction.

A cortical region, especially on the right, that responds selectively when people attribute beliefs to others, a candidate neural substrate for belief attribution within the wider mentalizing network.

Theory of mind.

The ability to impute unobservable mental states to oneself and others in order to predict behaviour; the developmental and comparative construct at the core of, and narrower than, mentalization.

Two-systems account.

Apperly and Butterfill's proposal that belief tracking is served by two systems, an early, fast, automatic, limited one and a later, flexible, effortful one, reconciling the infant and explicit-task findings.

Key Researchers

Simon Baron-Cohen. Professor of developmental psychopathology at the University of Cambridge; with Alan Leslie and Uta Frith he showed that autistic children are selectively impaired on the false-belief task, arguing autism involves a specific deficit in the mechanism for representing mental states, and later devised the Reading the Mind in the Eyes test of adult mental-state attribution. Faculty Page - ORCID - Google Scholar - Wikipedia

Anthony Bateman. Consultant psychiatrist and, with Peter Fonagy, the co-developer of mentalization-based treatment; he led the randomised controlled trials that established its efficacy for borderline personality disorder, including sustained gains at long-term follow-up. Faculty Page - ORCID

Peter Fonagy. Professor of contemporary psychoanalysis and developmental science at University College London and head of the Anna Freud Centre; he brought mentalization into empirical psychology, operationalised it as reflective function scored from the Adult Attachment Interview, tied its development to secure attachment, and co-created mentalization-based treatment. Faculty Page - ORCID - Google Scholar - Wikipedia

Christopher D. Frith. Emeritus professor of neuropsychology at University College London; he helped establish the social-cognitive neuroscience of mentalizing, mapping the brain network, the medial prefrontal cortex, temporo-parietal junction, and temporal poles, recruited when people attribute mental states. Faculty Page - ORCID - Google Scholar - Wikipedia

Uta Frith. Emeritus professor of cognitive development at University College London; a co-author of the landmark study linking autism to a theory-of-mind deficit and, with Chris Frith, of the reviews charting the neural basis of mentalizing and its development. Faculty Page - ORCID - Google Scholar - Wikipedia

Josef Perner. Professor of psychology at the University of Salzburg; with Heinz Wimmer he devised the change-of-location false-belief task that became the field's standard test of belief attribution, and later argued that theory-of-mind development is bound up with the growth of executive control. Faculty Page - ORCID - Google Scholar - Wikipedia (German)

Rebecca Saxe. Professor of cognitive neuroscience at the Massachusetts Institute of Technology; she isolated a region of the right temporo-parietal junction that responds selectively when people reason about another person's beliefs rather than other facts about them, sharpening the neural localisation of belief attribution. Faculty Page - ORCID - Google Scholar - Wikipedia

Frequently Asked Questions

What is mentalization?
Mentalization is the capacity to interpret behaviour, one's own and other people's, in terms of underlying mental states such as beliefs, desires, feelings, and intentions (Fonagy & Target, 1997). It is the everyday act of holding a mind in mind, treating people as psychological agents whose actions flow from what they think and feel rather than from surface appearances alone.

What is the difference between mentalization and theory of mind?
They overlap but are not the same. Theory of mind, from developmental and comparative research, is usually treated as a competence that is present or absent; mentalization, from clinical psychology, is broader and graded, covering the self as well as others, feeling as well as belief, and automatic as well as deliberate reading, and it can be temporarily lost (Premack & Woodruff, 1978; Luyten, Campbell, Allison, & Fonagy, 2020).

What is the false-belief task?
It is a test in which a character comes to hold a belief the child knows to be false, for example about where an object is hidden after it has been moved in the character's absence (Wimmer & Perner, 1983). A child who predicts that the character will act on the false belief rather than on reality has shown they represent belief as a representation, separate from the world it is about.

At what age do children pass the false-belief task?
A meta-analysis of 178 studies found that children below about three and a half years typically fail, and performance rises to well above chance by about four and a half to five years, crossing the half-way point just before the fourth birthday (Wellman, Cross, & Watson, 2001). The transition is gradual and holds across countries and languages.

Is mentalizing a single ability?
Probably not. Infants and adults track others' beliefs automatically and early, while the explicit false-belief task arrives late and demands effort, which led to a two-systems account of an early, fast, automatic system and a later, flexible, effortful one (Onishi & Baillargeon, 2005; Apperly & Butterfill, 2009). Mentalizing also has separable cognitive and affective sides.

Which brain regions support mentalizing?
A consistent network activates when people attribute mental states: the medial prefrontal cortex, the temporo-parietal junction on both sides but especially the right, and the temporal poles (Gallagher & Frith, 2003). The right temporo-parietal junction responds selectively to others' beliefs rather than to other facts about them (Saxe & Kanwisher, 2003).

What is reflective function?
Reflective function is the measurable form of mentalizing that Fonagy operationalised, scored from how a person represents mental states when talking about attachment relationships in the Adult Attachment Interview (Fonagy, Steele, Steele, Moran, & Higgitt, 1991). A parent's reflective function predicts the security of their infant's attachment, especially where the parent's own history was adverse.

What is mentalization-based treatment?
It is a structured psychotherapy developed by Bateman and Fonagy whose central aim is to restore and strengthen the capacity to mentalize rather than to challenge thoughts or teach skills directly (Bateman & Fonagy, 2009). Randomised controlled trials have shown it reduces self-harm, suicide attempts, and hospital use in borderline personality disorder (Bateman & Fonagy, 1999).

References