# From Attribution of Mental States to Mentalisation and Empathy

## 2.1. From Attribution of Mental States to Mentalisation and Empathy

Early work on mindreading treated the attribution of beliefs and desires
as a largely intellectual exercise, but subsequent attachment-based
research demonstrated that the same capacity is fundamentally affective
and relational. Fonagy and Target (1997) showed that infants internalise
an understanding of mind not through abstract inference but through
caregivers’ emotionally attuned mirroring of their inner states. This
process, termed reflective functioning, is the developmental engine that
transforms simple mental-state attribution into full mentalisation,
providing the scaffolding for later empathic concern and prosocial
behaviour (Fonagy, et al., 1991; Fonagy et al., 2002).

### 2.1.1. Mentalisation: A Multidimensional Social-Cognitive Capacity

With the foundational work of Premack and Woodruff (1978), who asked the
now-famous question, "does the chimpanzee have a theory of mind?", ToM
became a formal field of research within psychology. This question
marked the start of decades of research into how humans, particularly
children, come to understand that others have beliefs, desires, and
intentions different from their own. The development of ToM was most
notably studied through false-belief tasks (Wimmer & Perner, 1983).
Usually demonstrated through the "Sally–Anne" scenario, this task
evaluates a child's capacity to understand that another person could
have a belief different from either reality or the child's own knowledge
(Wimmer & Perner, 1983). Along the course of normal development,
children succeed at this task around the age of four, marking the first
strong evidence of ToM acquisition (Wellman, et al., 2001).

Throughout the late 1980s and the 1990s, research broadened the
conceptualisation of ToM beyond the grasp of false beliefs to include
the comprehension of deceit, irony, sarcasm, metaphor, and other forms
of non-literal communication (Happé, 1993). Subsequent studies have
underlined that ToM is not a monolithic faculty but a constellation of
processes that operate either explicitly (that is, consciously and with
effort) or implicitly, automatically and without reflection (Apperly &
Butterfill, 2009). Moreover, evidence indicates that cognitive elements
of ToM (inferring beliefs, intentions, and knowledge) and affective
elements (understanding or sharing emotions) follow partially separate
developmental timelines from preschool through adolescence (Peterson &
Wellman, 2019) and rely on partly distinct neural networks (Frith &
Frith, 2012; Shamay-Tsoory, 2011).

Today, ToM is recognised as a multifaceted construct shaped by
developmental, cultural, and environmental factors, combining biological
predispositions with social learning (Schaafsma et al., 2015). While it
remains a cornerstone of social-cognition research, scholars
increasingly emphasise the need for broader frameworks that incorporate
not only cognitive inference but also emotional and relational
sensitivity. One such framework is mentalisation, which builds on and
extends ToM by stressing the capacity to reflect on one’s own and
others’ mental states within emotionally meaningful contexts.

Mentalisation emerges in early attachment relationships and continues to
develop across the lifespan, moving along the implicit-to-explicit
continuum described by Lieberman (2007) as it progresses from preverbal
caregiver–infant exchanges to sophisticated self–other reflections in
adulthood. Mentalisation operates at both conscious and pre-reflective
levels, integrating what Lieberman (2007) terms explicit and implicit
processes. In infancy, caregivers’ accurate mirroring of the child’s
internal states fosters implicit mentalising: the baby comes to “feel
felt” and begins to expect that mental states govern behaviour (Fonagy
et al., 1991). During the preschool years, secure attachment predicts
mastery of false-belief tasks, marking the transition to explicit
recognition that others can hold perspectives different from one’s own
(Fonagy & Target, 1997). Middle childhood and adolescence bring further
differentiation, as children refine the ability to toggle between self-
and other-focused viewpoints and to integrate cognitive (beliefs,
intentions) with affective (emotions) components of social understanding
(Dumontheil et al., 2010; Sebastian et al., 2012). In adulthood,
mentalisation is shaped continuously by new relational contexts (e.g.
romantic partnerships, parenthood, and professional roles) while
remaining sensitive to stress that can “down-shift” functioning to
earlier, more automatic modes (Fonagy & Luyten, 2009). Thus the capacity
first cultivated within parent–infant attachment is elaborated,
reorganised, and sometimes compromised throughout the lifespan,
reflecting both ongoing neural maturation and the quality of later
interpersonal experiences (Fonagy & Bateman, 2019).

### 2.1.2. The Function of Mentalisation in Social Cognitive Process

Mentalisation plays a central role in enabling individuals to navigate
the complexities of social life by making sense of their own and others’
internal experiences, a capacity indexed empirically by reflective
functioning (RF). Acquired in infancy through caregivers’ attuned
mirroring, high parental RF predicts secure attachment and fosters
children’s early ToM success, setting the trajectory for adaptive
emotion regulation and prosocial behaviour in later development (Fonagy
et al., 1991; Fonagy & Target, 1997). Across childhood and adolescence,
growing RF supports increasingly sophisticated perspective-taking,
negotiation of peer conflict, and academic engagement, while in
adulthood it underpins self-soothing under stress, empathic attunement
in intimate relationships, and flexible shifts between rapid, implicit
readings of social cues and slower, explicit reflection when automatic
assumptions fail (Fonagy & Luyten, 2009). Deficits or imbalances in RF
constitute a transdiagnostic vulnerability: under high affective load,
some individuals “hyper-mentalise” (over-interpret yet misconstrue
others’ motives), whereas others collapse into non-mentalising modes
characterised by impulsive action or concrete thinking, which are
patterns observed in borderline personality disorder, trauma-related
disorders, and certain externalising pathways (Allen et al., 2008;
Fonagy & Bateman, 2019). Thus, RF functions as a developmental bridge
linking early attachment quality to life-span outcomes in
self-regulation, relationship quality, and mental health.

Mental state attribution and reasoning not only underpins everyday
social interactions but also fuels creative engagement with stories:
readers continuously track characters’ motives, infer hidden intentions,
and predict outcomes (Mar & Oatley, 2008; Oatley, 1999). Evidence shows
that this capacity is context-sensitive: education, cultural background,
anxiety, and even brief exposure to narrative fiction can modulate
performance on empathy and ToM tasks (Kidd & Castano, 2013; Mar et al.,
2006). For example, reading literary fiction, which is rich in complex,
ambiguously motivated characters, has been found to increase subsequent
scores on mental-state attribution and empathy measures, presumably
because such texts demand intensive perspective-taking (Dodell-Feder &
Tamir, 2018; Kidd & Castano, 2013; Mar et al., 2006). Although most
studies frame these effects in ToM or empathy terms, the present thesis
interprets them within the broader construct of mentalisation, which
encompasses not only making correct inferences about others’ mental
states, but also emotional insight, self-reflection, and relational
context (Fonagy & Bateman, 2019).

### 2.1.3. Mentalisation in Relation to Self–Other Understanding and Empathy

Mentalisation integrates two complementary capacities (self-related
mentalisation and other-related mentalisation) that together provide the
cognitive architecture of mental state attribution. Self-related
mentalisation refers to recognising and regulating one’s own thoughts
and feelings, a prerequisite for reflective self-control and affect
regulation; other-related mentalisation involves inferring the beliefs,
intentions, and emotions of those around us, allowing accurate prosocial
responsiveness, which together provide the cognitive architecture of
mental state attribution, supported by evidence for distinct neural
system for self- and other-related social cognition (Amodio & Frith,
2006).

Empathy itself comprises two partially dissociable components: an
affective facet involving automatic emotional resonance with others and
a cognitive facet entailing deliberate perspective-taking (Decety &
Jackson, 2004). When these capacities operate in tandem, they generate
the double helix of empathy: affective attunement grounded in an
internal sense of “what it feels like,” and cognitive perspective-taking
that appreciates “how it must be for you” (Shamay-Tsoory, 2011).
Mentalisation is most strongly tied to the cognitive aspect of empathy;
higher self-related and other-related mentalisation scores predict
better performance on perspective-taking tasks and elevated scores on
cognitive-empathy subscales of multidimensional empathy inventories
(Mazza et al., 2014). At the same time, balanced self-related
mentalisation tempers affective sharing, preventing emotional contagion
from overwhelming self-boundaries and thereby sustaining pro-social
concern (Fonagy & Bateman, 2019).

Longitudinal attachment studies support this dual-facet model. High
parental reflective functioning, the capacity to discuss attachment
experiences in mental-state terms, predicts balanced growth of
children’s self-related and other-related mentalisation and, in turn,
higher cognitive-empathy scores in middle childhood (Ensink & Mayes,
2010; Fonagy & Target, 1997). Conversely, disrupted attachment or low
caregiver RF is associated with asymmetric profiles: some children
over-read others while remaining unaware of their own affect, whereas
others show the opposite pattern (Sharp et al., 2011). These findings
indicate that both implicit, pre-reflective attunement and explicit,
conscious perspective-taking must be integrated early for empathy to
flourish across development (Fonagy, et al., 2002).

Balanced development of self-related and other-related mentalisation is
linked to favourable social outcomes throughout life, including
cooperative problem-solving, relationship satisfaction, and resilience
under stress (Allen et al., 2008). By contrast, an imbalance invites
difficulty. Individuals who over-focus on others’ minds can become
hyper-mentalising (i.e. over-interpreting motives and generating
inaccurate, often paranoid, social narratives). Those who
under-represent their own inner states may drift into alexithymia,
struggling to label feelings and to draw on them for decision-making.
Both patterns are observed trans-diagnostically, from externalising
trajectories to personality pathology, and illustrate Luyten and
colleagues’ (2020) contention that psychological wellbeing depends on
flexible movement along the self–other axis of mentalisation.

Crucially, mentalisation remains open to experiential refinement.
Engagement with literary fiction, for instance, exposes readers to
richly layered characters whose motives must be inferred from subtle
cues; even brief exposure has been shown to boost performance on empathy
and ToM tasks (Dodell-Feder & Tamir, 2018; Kidd & Castano, 2013). Such
findings highlight the plasticity of the system: social learning,
cultural practices (Mar et al., 2006), and targeted mentalisation-based
interventions (Luyten et al., 2020) can strengthen both poles of
mentalisation, fostering a more integrated empathic stance that supports
adaptive functioning across diverse relational contexts.