Mentalizing

Mentalizing is the capacity to attribute mental states to oneself and others — to understand and interpret behavior in terms of underlying thoughts, feelings, intentions, beliefs, and desires. It is the ability behind everyday acts of social understanding: reading a friend’s mood from their voice, inferring why a stranger acted a certain way, anticipating how a partner will respond to news, recognizing one’s own emotional state and naming it. The capacity develops through childhood and continues to be refined throughout life.

Two largely independent literatures use the term, sometimes interchangeably and sometimes with subtle distinctions. The cognitive-science Theory of Mind (ToM) tradition, originating in the 1978 Premack and Woodruff chimpanzee paper and developed through false-belief tasks (Wimmer & Perner 1983; Baron-Cohen, Leslie & Frith 1985), focuses on the cognitive ability to attribute beliefs and infer mental representations. The clinical-psychological mentalization framework developed by Peter Fonagy and colleagues treats mentalizing as a context-dependent, often-affective capacity central to attachment, emotional regulation, and psychotherapy. In autism research, mentalizing is operationalized in screening tools including the RAADS-14 Mentalizing Deficits subscale (7 items, max score 21, the largest of the three RAADS-14 subscales).

Mentalizing matters at several levels.

For everyday social functioning: nearly every act of human cooperation depends on mentalizing. Conversation, friendship, romantic partnership, parenting, workplace collaboration, and conflict resolution all rest on participants’ ability to model what others think, feel, and want. When mentalizing breaks down — under stress, in unfamiliar contexts, in conditions like autism or borderline personality disorder — the social difficulty that follows is often what brings people to clinical attention rather than the mentalizing difficulty itself.

For autism diagnosis and self-recognition: mentalizing differences have been one of the most-studied features of autism since the 1985 Baron-Cohen, Leslie & Frith paper that applied the false-belief task to autistic children. The framing has shifted substantially over four decades — from the strong "ToM deficit" hypothesis of the 1980s and 1990s to the more cautious 2025 critical evaluation by Long, Catmur & Bird, which argues that no existing theoretical account adequately explains ToM in autism and that existing tests have validity problems. The construct remains central to autism screening tools (the RAADS-14 has 7 of its 14 items in the Mentalizing Deficits subscale) but its theoretical interpretation is more contested today than ten years ago.

For psychotherapy: Mentalization-Based Treatment (MBT), developed by Fonagy and Bateman, has become a standard evidence-based treatment for borderline personality disorder and is increasingly applied to other conditions. Mentalizing is treated in MBT as a capacity that can be supported and trained, which differs from the more strongly trait-like framings sometimes implicit in autism ToM research.

For research methodology: Mentalizing is measured by an unusually wide range of instruments — false-belief tasks, the Reading the Mind in the Eyes Test, Strange Stories, the Movie for the Assessment of Social Cognition (MASC), the Reflective Functioning Scale, the Reflective Functioning Questionnaire, the Adult Theory of Mind Questionnaire, the Th.o.m.a.s. clinical interview, and recent additions including the Interview Task developed alongside Long et al.’s 2025 Mind-space framework. Different instruments tap different facets of the construct, which is one reason results across studies sometimes appear inconsistent.

One reason "mentalizing" is a slippery term is that it sits at the intersection of two large research literatures that developed largely in parallel and use the word with subtly different emphases.

The two literatures have begun to converge over the past decade. Recent work in autism increasingly cites Fonagy-tradition concepts (Reflective Functioning, contextual mentalizing), and recent work in clinical mentalization increasingly engages with cognitive-science measurement (Reading the Mind in the Eyes, computational modeling of belief attribution). The 2025 Long, Catmur & Bird paper proposing the Mind-space framework explicitly attempts to integrate insights from both traditions. But the conceptual emphasis differs enough that "mentalizing" in a cognitive-neuroscience paper often means something subtly different than "mentalizing" in a Fonagy-tradition clinical paper.

Origins (1978–1985)

The phrase "theory of mind" was coined by David Premack and Guy Woodruff in their 1978 paper "Does the chimpanzee have a theory of mind?", which asked whether non-human primates can represent mental states in others. The paper itself answered cautiously, but the question was taken up in human developmental psychology.

The most consequential operationalization was the false-belief task, introduced by Heinz Wimmer and Josef Perner in 1983. The task — in versions including the Sally-Anne test — assesses whether a child can attribute a false belief to another character: Sally puts a marble in a basket and leaves the room; Anne moves it to a box; when Sally returns, where will she look? The correct answer requires understanding that Sally holds a belief that no longer matches reality. Typically-developing children pass false-belief tasks around age four; children with intellectual disability or other developmental conditions follow developmental rather than chronological trajectories.

Baron-Cohen, Leslie & Frith (1985, Cognition) applied the false-belief task to autistic children and found that most failed at ages where typically-developing children pass. The paper proposed that autism is characterized by a specific impairment in Theory of Mind, sometimes informally called "mindblindness". This framing dominated autism cognitive research for the following two decades.

Theoretical accounts

Several theoretical accounts of how Theory of Mind works have been proposed:

• Modular ToM mechanism (Leslie 1987, 1994) — ToM is a specific cognitive module that develops on a relatively fixed timetable; impairment in this module produces autism-like difficulty.
• Simulation theory (Goldman 2006) — we understand others by simulating their mental states using our own minds as model.
• Theory-theory (Gopnik & Wellman 1992) — children develop ToM by constructing implicit theories about how minds work, much like scientists construct theories about the natural world.
• Two-systems accounts (Apperly & Butterfill 2009) — humans have an early-developing, fast, automatic, signature-limit system and a later-developing, slower, flexible, full-belief system; both contribute to mentalizing.

Beyond false belief: contemporary tasks

Modern ToM measurement extends well beyond the original false-belief task. Common adult tasks include:

• Reading the Mind in the Eyes Test (RMET) — participants identify mental states from photographs of the eye region; Baron-Cohen et al. 2001.
• Strange Stories (Happé 1994) — vignettes requiring inference about why characters say what they say.
• Movie for the Assessment of Social Cognition (MASC) — video-based scoring of mental-state attribution in social interactions.
• Spontaneous-mentalizing tasks — assessed via anticipatory eye gaze, reaction times, or pupillometry rather than explicit answers; the 2024 Onda et al. paper in Frontiers in Psychiatry examined whether these measures cohere in autistic adults.
• Adult Theory of Mind Questionnaire (A-ToM-Q) — recent forced-choice online instrument measuring decision accuracy and confidence.
• Th.o.m.a.s. clinical interview (Bosco et al.) — structured interview operationalizing five dimensions of ToM (first-order, second-order, allocentric, egocentric, meta-representation), used in the 2024 Fadda et al. Frontiers in Psychology study of autistic adolescents.

The clinical mentalization framework was developed primarily by Peter Fonagy and colleagues over roughly three decades, beginning with the foundational Reflective Functioning (RF) scale (Fonagy, Target, Steele & Steele 1991, 1998) applied to the Adult Attachment Interview. The framework grew out of attachment research and psychoanalytic clinical practice and has subsequently become the theoretical foundation for Mentalization-Based Treatment (MBT), developed for borderline personality disorder by Bateman and Fonagy (2004 onward) and now applied across multiple conditions.

The four polarities

Fonagy and colleagues (2011, formalized in the 2019 framework update) describe mentalizing as multidimensional, organized along four polarities. Most healthy mentalizing involves balanced movement between poles in each dimension; difficulty at one pole defines characteristic patterns of dysfunction.

The three pre-mentalizing modes

When mentalizing capacity is overwhelmed — by stress, attachment activation, intoxication, exhaustion — Fonagy describes three pre-mentalizing modes that emerge in its place. These are not pathological per se; they are typical childhood modes that adults revert to when the more sophisticated mentalizing capacity fails.

• Psychic equivalence mode — mental states are experienced as equivalent to external reality; if I think it, it is true. The thought that "no one likes me" is felt as fact rather than as a thought.
• Pretend mode — mental states are decoupled from reality, often appearing as overly intellectualized or dissociated talk about feelings without genuine connection to them.
• Teleological mode — mental states are recognized only when they are expressed in concrete action; "she loves me only if she does X." Common in interpersonal demands escalating toward self-harm or relationship rupture.

Reflective Functioning

The most-cited operationalization of Fonagy mentalizing is Reflective Functioning (RF), scored from -1 (negative or absent reflective functioning) to 9 (exceptional) when applied to Adult Attachment Interview transcripts. The scoring assesses the quality of an interviewee’s spontaneous reasoning about mental states in self and others during emotionally-charged narrative recall. RF correlates with parental sensitivity, attachment security in offspring, and outcomes across multiple psychiatric populations.

The full RF scale is labor-intensive (transcripts must be coded by trained raters) which limited large-scale research. The Reflective Functioning Questionnaire (RFQ-8 and RFQ-54), developed by Fonagy, Luyten and colleagues and validated in 2016 (PLOS ONE), is a self-report alternative measuring two factors: certainty about mental states (RFQ-c) and uncertainty about mental states (RFQ-u). The RFQ has limitations — some authors note that simple polar-scored items may not fully capture mentalizing capacity, since hyper-certainty and hyper-uncertainty can both indicate mentalizing failure — but it has enabled much larger-sample research than the original RF scale supports.

Mentalizing differences in autism are among the most-studied features in cognitive neuroscience, but the field has become substantially more cautious about strong "ToM deficit" framings since the 2010s.

The classical account (1985–2010s)

For roughly thirty years following Baron-Cohen, Leslie & Frith 1985, the dominant framing in autism cognitive research was that autism involves a specific impairment in Theory of Mind. Evidence supporting this framing included: failure of false-belief tasks at ages where typically-developing peers pass; difficulty with second-order false belief; lower scores on Strange Stories and Reading the Mind in the Eyes; and neuroimaging differences in brain regions associated with mentalizing (medial prefrontal cortex, temporo-parietal junction).

The 2025 critical re-evaluation

The current direction

Contemporary research on mentalizing in autism emphasizes:

• Spontaneous vs explicit mentalizing — the 2024 Onda et al. study examined whether anticipatory gaze (a spontaneous measure) and reaction time biases (also spontaneous) cohere in autistic adults. The finding suggested that spontaneous and explicit mentalizing may dissociate in autism in ways one-shot tasks miss.
• Context-dependent and reciprocal mentalizing — recognizing that mentalizing is not a fixed trait but varies with social context, emotional pressure, attachment activation, and interpersonal stakes. Many autistic adults perform well on standalone ToM tasks while still describing real-life social-cognitive difficulty in dynamic interactions.
• Cognitive vs affective dissociation — some autism research finds that cognitive mentalizing (inferring beliefs) can be substantially preserved while affective mentalizing (sharing emotional states) is more impaired, mapping to the cognitive-affective polarity in the Fonagy framework.
• Double empathy problem — an alternative framing developed by Damian Milton (2012, autistic researcher) which proposes that "social-cognitive difficulty" in autism is partly a mismatch between autistic and non-autistic communication styles rather than a deficit located within autistic individuals. The double empathy framing has gained traction in autism research over the past decade as a complement (not replacement) to traditional ToM accounts.
• Population-level humility — recognition that mentalizing variability within the autistic population is large and that the average difference between autistic and non-autistic groups, while real on most tasks, masks substantial individual variation. Many autistic adults, particularly women and those with later-life diagnosis, perform indistinguishably from non-autistic peers on standard tasks.

The RAADS-14 Screen for adult autism (Eriksson, Andersen & Bejerot, 2013, Molecular Autism, CC BY 2.0) operationalizes mentalizing as one of three factor-analytic subscales. The Mentalizing Deficits subscale is the largest of the three: 7 of the 14 items, maximum score 21. The construct it measures is the social-communicative dimension of mentalizing relevant to autism diagnosis — difficulty with theory of mind, reading nonverbal cues, conversational reciprocity, and adapting to social expectations.

The seven verbatim items

What the items capture

The seven items map roughly onto four mentalizing-relevant features:

• Affective mentalizing (items 1, 4) — difficulty inferring emotional states from verbal and nonverbal cues during interaction
• Social-pragmatic mentalizing (items 2, 3, 6) — difficulty inferring expectations, conversational norms, and implicit communication
• Cognitive style (item 5) — detail-focused processing; this item taps a feature that is mentalizing-adjacent rather than mentalizing-direct
• Cognitive flexibility (item 7) — distress at unexpected change; another mentalizing-adjacent feature, often grouped with restricted/repetitive behaviors in DSM-5

The subscale’s factor-analytic basis means it captures these features as they co-vary in the validation sample, which is appropriate for adult autism screening but does not correspond to either the cognitive-science ToM tradition or the Fonagy framework one-to-one. A high Mentalizing-Deficits score on RAADS-14 is one signal among several that comprehensive autism evaluation may be warranted; it does not by itself diagnose autism, and it does not capture the full breadth of mentalizing as either of the two literatures defines it.

The Mentalizing-Focused archetype

The LBL Adult Autism Test additionally classifies respondents into one of five archetypes based on their sub-dimensional profile across the three RAADS-14 subscales. The Mentalizing-Focused archetype indicates that the respondent’s pattern is dominated by Mentalizing-Deficits responses, with Sensory-Reactivity and Social-Anxiety items contributing less. The archetype is informational rather than diagnostic; it provides language for the respondent’s sub-dimensional profile and is paired with cross-tool referrals (the Cognitive Reserve Estimator and the Adult ADHD Test, since AuDHD overlap is common in this profile). Full archetype thresholds are documented on the tool’s methodology page.

The wide range of mentalizing instruments reflects that the construct is multidimensional and context-dependent. Different instruments tap different facets and are not interchangeable. The table below summarizes major instruments organized by type.

Cross-instrument correlations are often modest, reflecting that "mentalizing" is not a single thing measured by all of them but a family of related abilities that overlap imperfectly. Researchers selecting instruments increasingly attend to which facet they want to measure, the population they are studying, and the context in which mentalizing will need to be deployed.

Mentalizing differences are not specific to autism. The Fonagy clinical literature and recent transdiagnostic research has documented mentalizing patterns associated with several conditions:

Borderline personality disorder

BPD is the condition for which the Fonagy mentalization framework was originally developed and where the evidence base is largest. Characteristic mentalizing patterns in BPD include hyper-sensitivity to external cues with poor internal access (external>>internal pole), rapid breakdown of mentalizing under attachment-relevant emotional pressure, and frequent reversion to teleological or psychic-equivalence modes. Mentalization-Based Treatment for BPD has demonstrated efficacy in multiple randomized controlled trials.

Depression and anxiety

Depression is associated with negative mentalizing biases — over-interpreting others’ mental states as critical or rejecting — and with reduced self-mentalizing during depressive episodes. Generalized anxiety often involves over-mentalizing (excessive rumination about others’ mental states) rather than under-mentalizing. The 2026 rapid scoping review by Petersen et al. noted that the "better mentalization is always better" assumption is empirically complicated; eight of twenty-two studies in their review found null or reverse associations between RF and internalizing symptoms.

Eating disorders

Anorexia nervosa research has found particular patterns of mentalizing impairment, especially around bodily and affective self-mentalizing (recognizing one’s own emotional and bodily states). Mentalizing-based treatment for eating disorders is being developed and tested.

Trauma and complex PTSD

Childhood trauma is associated with altered mentalizing development — sometimes hypervigilant other-mentalizing as adaptive response, sometimes blunted mentalizing during dissociative episodes. Mentalization-based treatment is increasingly applied to complex trauma.

Schizophrenia and the psychoses

Schizophrenia involves substantial social-cognitive impairment, including mentalizing impairment that may be more severe than in autism on standard tasks (Pinkham et al. and others). The patterns differ qualitatively: schizophrenia mentalizing impairment often includes paranoid over-attribution (assuming malicious intent), whereas autism mentalizing tends toward under-attribution or literal interpretation.

The transdiagnostic view

Luyten et al. 2024 and others have proposed that mentalizing impairment may function as a transdiagnostic factor — partially explaining symptoms across multiple conditions and serving as a moderator of psychotherapy outcomes. The transdiagnostic framing remains an active research question rather than established consensus, but it has shaped how mentalizing is now studied: increasingly across conditions rather than within single diagnostic categories.

The LBL Adult Autism Test reports your Mentalizing Deficits subscale score (0–21) alongside Social Anxiety (0–12) and Sensory Reactivity (0–9), plus a presentation archetype based on your sub-dimensional profile. If mentalizing is part of your reason for taking the test, the subscale breakdown gives you that signal directly.

Note: the RAADS-14 Mentalizing-Deficits subscale is one operationalization of mentalizing, not a comprehensive measurement. It captures the social-communicative dimension relevant to autism screening but does not assess the affective, attachment-related, or context-dependent dimensions emphasized in the Fonagy framework. A high subscale score indicates a pattern worth discussing with a clinician familiar with adult autism; it does not by itself diagnose autism or fully characterize a person’s mentalizing capacity.