Social Cognition

Abstract

Social cognition is the set of mental processes by which people perceive, interpret, store, and act on information about other people and about themselves as social agents. It spans the rapid reading of faces, gaze, and gestures; the inference of others' beliefs, desires, and intentions; and the regulation of one's own behaviour within a group. Two theoretical traditions organise the field: theory theory, which holds that people reason about minds using a tacit body of causal knowledge, and simulation theory, which holds that people model other minds by covertly running their own. Functional imaging has localised a mentalizing network, centred on medial prefrontal cortex and the temporo-parietal junction, that is partly separable from the action-matching mirror system. Deficits in social cognition are central to autism and schizophrenia, making it a clinical target as well as a basic science.

Keywords: social cognition, theory of mind, mentalizing, social brain, mirror neurons

Social cognition denotes the information processing that supports social behaviour: how observers encode, remember, and reason about the mental states and dispositions of other people, and how they position themselves within a social group (Frith, 2008). It is distinguished from social psychology by its level of analysis, treating the perception of persons with the same representational and neural rigour that cognitive psychology brings to the perception of objects, and from general cognition by evidence that reasoning about minds recruits partly dedicated processes (Frith & Frith, 2012). The field matters because social inference is ubiquitous, developmentally early, and selectively vulnerable: the same machinery that lets a toddler predict where a playmate will search is disrupted in autism and in schizophrenia, so a precise account of it carries directly into the clinic (Adolphs, 2009).

What Social Cognition Is

Social cognition covers every process that takes another person, or the self as a social object, as its content. At the fast end it includes the near-automatic extraction of social primitives: the direction of another's gaze, the emotion on a face, the goal implied by a reaching hand. At the deliberate end it includes explicit reasoning about what someone knows, wants, or intends, and the strategic management of one's own reputation and alliances. What unifies these is not a single mechanism but a shared subject matter, agents with minds, which imposes computational demands that inanimate perception does not: minds are hidden, they hold representations that can be false, and they respond to being modelled (Frith & Frith, 2012).

The fast end has a well-characterised perceptual front-end. Another's eyes are a privileged social stimulus: seeing where a face is directed triggers a reflexive shift of the observer's own attention toward the gazed-at location, a gaze-cueing effect that operates rapidly and largely outside voluntary control and supplies the joint-attention signal on which later mind-reading is built (Frischen, Bayliss, & Tipper, 2007). This automaticity is one pole of a distinction that cuts across the whole field by processing mode rather than by content. Satpute and Lieberman marshalled neuroimaging evidence for two broad systems: a reflexive, automatic system — which they labelled the X-system, after the x in reflexive — that reads social primitives quickly and effortlessly, and a reflective, controlled system, the C-system, that supports slow, deliberate, and revisable inference (Satpute & Lieberman, 2006). The division recurs throughout social cognition and maps roughly onto the fast and deliberate ends sketched above.

The construct is deliberately broad, and MeSH places it within social perception, which in turn sits under perception in the tree of mental processes. It overlaps with, but is not identical to, emotional intelligence and with attention to social stimuli; the distinction that has structured the most research is between the perception of transient states, such as an expression, and the inference of enduring or hidden states, such as a belief (Lieberman, 2007).

Types of Social Cognition

The Medical Subject Headings classification treats social cognition as a descriptor with one narrower descriptor directly beneath it. Its parent descriptor in MeSH is Social Perception, and the single direct child is listed in Table 1. Because MeSH is an indexing vocabulary rather than a theory of the mind, this hierarchy records how the biomedical literature is catalogued, not a claim that the field carves social cognition at exactly these joints; the subtype below is not mutually exclusive with the broader construct but a specialised region of it.

Table 1. Direct subtypes of Social Cognition in the MeSH classification (tree F02.463.593.752.500).

Theory of Mind and Mentalizing

The best-studied component of social cognition is theory of mind, the capacity to attribute mental states such as beliefs, desires, and intentions to oneself and others and to use those attributions to predict behaviour. The term entered psychology when Premack and Woodruff asked whether a chimpanzee could impute mental states to a human actor, and framed the ability as a theory because the states inferred are unobservable (Premack & Woodruff, 1978). The operational test that followed was the false-belief task: a child watches a character leave an object in one place and, in the character's absence, sees it moved. A child who understands that the character holds a belief about the object's location, distinct from its true location, predicts that the character will search where the object was left. Passing this task requires representing a representation, and typically developing children begin to pass it around the age of four (Baron-Cohen et al., 1985).

This ability is now understood to be mentalizing rather than a monolithic module. It fractionates: a cognitive component, reasoning about beliefs and knowledge, is partly separable from an affective component, sharing and inferring feelings, and the two dissociate across disorders (Happe et al., 2017). The measurement is fragile, however. Many tasks that carry the theory-of-mind label share little variance and can be passed by non-mentalizing strategies, so a low score is not by itself evidence of a mentalizing deficit (Quesque & Rossetti, 2020).

How the Inference Is Made: Theory Versus Simulation

Two accounts compete over the mechanism of mental-state inference. Theory theory holds that people carry a tacit, law-like body of knowledge about how mental states cause behaviour and apply it as an intuitive science of mind. Simulation theory holds that people instead use their own cognitive apparatus as a model, imagining themselves in the other's situation and reading off the states that result, so that understanding another mind is a covert enactment rather than a deduction (Gallese & Goldman, 1998). The discovery of mirror neurons, cells that discharge both when an action is performed and when the same action is observed, gave simulation a candidate neural substrate and made the debate tractable (Rizzolatti & Craighero, 2004).

The evidence does not crown a winner. Simulation plausibly underlies the fast, automatic understanding of observed actions and expressions, whereas explicit reasoning about false or counterfactual beliefs seems to demand theory-like representation that pure simulation struggles to supply. Contemporary accounts therefore treat the two as complementary routes operating at different levels, an arrangement that also maps onto partly distinct brain systems (Frith & Frith, 2012).

The Social Brain

Lesion and imaging work converge on a distributed social brain rather than a single centre (Adolphs, 2009). Reasoning about other minds reliably engages a mentalizing network: medial prefrontal cortex, the bilateral temporo-parietal junction with a right-hemisphere weighting, the precuneus, and the temporal poles. Within this network the medial prefrontal cortex supports the representation of mental states as such, integrating information about the self and others (Amodio & Frith, 2006), while the right temporo-parietal junction is recruited selectively when the content of the inference is another person's belief rather than merely a fact about them (Saxe & Kanwisher, 2003).

A quantitative synthesis sharpens the picture. Meta-analysis of neuroimaging places theory of mind and empathy on partly separate but hierarchically related substrates, with abstract belief reasoning drawing on the mentalizing network and affect sharing drawing on regions closer to sensory and interoceptive processing (Schurz et al., 2021). The action-observation, or mirror, system, centred on inferior frontal and inferior parietal cortex, is engaged by understanding what others do and is anatomically distinct from the mentalizing network that supports understanding what others think (Molenberghs et al., 2016). That much of the mentalizing network coincides with the brain's default mode, active when a person is not engaged in an external task, has been read as evidence that the resting brain is biased toward social thought (Lieberman, 2007).

Development and Clinical Significance

Social cognition has a protracted developmental trajectory whose most-charted landmark is the passing of explicit false-belief tasks near the fourth birthday, with precursors such as gaze following and joint attention appearing much earlier (Baron-Cohen et al., 1985). The developmental curve is steep and roughly sigmoidal: below three years most children fail, by five most pass, and the transition concentrates around age four. That graded, S-shaped rise, rather than a sudden switch, is the picture a meta-analysis of scores of false-belief studies recovered from beneath the scatter of individual results, and it holds across cultures and task variants (Wellman et al., 2001); it is the shape captured by the logistic model in the Worked Example and the demonstration below.

The tidy four-year story is itself complicated by evidence from nonverbal measures. Using a violation-of-expectation procedure that demands no verbal report, Onishi and Baillargeon found that 15-month-old infants look reliably longer when an agent reaches for an object in a location inconsistent with the false belief it ought to hold, as if the infants expected the agent to act on its belief rather than on reality (Onishi & Baillargeon, 2005). Whether this early sensitivity reflects genuine belief representation or a leaner tracking of behavioural regularities remains debated, but it has shifted the question from when theory of mind switches on to how an implicit competence present in infancy relates to the explicit understanding that surfaces years later.

Its clinical weight is considerable. A specific difficulty in mentalizing was the first cognitive account of autism to make experimental predictions, and social-cognitive atypicality remains central to the condition (Baron-Cohen et al., 1985). In schizophrenia, impairments in emotion perception and theory of mind are robust, partly separable from general neurocognitive deficit, and among the strongest predictors of how well a patient functions in daily life, which is why social cognition is now a target for remediation (Green et al., 2015).

Worked Example

Consider a stylised model of when children pass an explicit false-belief task. Let the probability of passing be a logistic function of age, P(age) = 1 / (1 + e^(-k(age - a50))), with the crossover age a50 = 4.0 years and a slope k = 1.6 per year. The model is illustrative, chosen to reproduce the well-replicated shape of the developmental transition (Wellman et al., 2001) rather than to fit a particular dataset.

At age 3.0 the exponent is 1.6 x (3.0 - 4.0) = -1.6, giving P = 1 / (1 + e^(1.6)) = 0.168, so about one child in six passes. At age 4.0 the exponent is zero and P = 0.500, the crossover. At age 5.0 the exponent is 1.6 x (5.0 - 4.0) = 1.6, giving P = 1 / (1 + e^(-1.6)) = 0.832, so roughly five children in six pass; the odds of passing at five are 0.832 / 0.168 = 4.95 to one. The steepest change occurs at the crossover, where the slope equals k / 4 = 0.40 per year: near the fourth birthday the pass rate climbs by about ten percentage points every three months. This concentration of change around a single age is exactly what makes false-belief performance such a sharp developmental marker, and it is why measurement noise there, of the kind that inflates or deflates individual scores, translates into large apparent differences in competence (Quesque & Rossetti, 2020).

Discussion

Social cognition has matured from a single flagship task into a structured, partly fractionated domain with identifiable neural systems. The theory-versus-simulation debate that once split the field has largely resolved into a division of labour: simulation for the fast reading of actions and expressions, theory-like representation for explicit reasoning about beliefs, each with a distinguishable substrate (Frith & Frith, 2012). Meta-analytic work has replaced the search for a social-cognition centre with a map of interacting networks, and has shown that empathy and theory of mind, often lumped together, occupy related but separable territory (Schurz et al., 2021).

Two cautions temper the progress. First, measurement: the proliferation of theory-of-mind tasks with poor convergent validity means that claims about group differences can reflect task artefacts rather than mentalizing itself, and the field is now revisiting its instruments (Quesque & Rossetti, 2020). Second, translation: although social-cognitive deficits predict functional outcome in schizophrenia and define much of the phenotype of autism, the path from a laboratory dissociation to an effective intervention remains long (Green et al., 2015). The domain's value to cognitive psychology is that it forces the discipline to model a content, other minds, that is hidden, recursive, and reactive, and in doing so it has produced some of the clearest examples of dedicated cognitive machinery (Adolphs, 2009).

Common Misconceptions

Theory of mind is a single ability that a person either has or lacks.

Theory of mind fractionates into cognitive and affective components that dissociate across conditions, and the tasks meant to measure it often fail to correlate, so a single pass-or-fail verdict misrepresents the construct (Happe et al., 2017).

Mirror neurons explain how people understand other minds.

Mirror neurons plausibly support understanding of observed actions, but the action-observation system is anatomically distinct from the mentalizing network engaged when reasoning about beliefs, so mirroring is at most part of the story (Molenberghs et al., 2016).

Social cognition is just social psychology by another name.

The two ask different questions: social psychology studies how situations shape behaviour, whereas social cognition studies the information-processing and neural mechanisms that represent other agents, with the same rigour applied to object perception (Frith, 2008).

Glossary

Affective theory of mind.

The inference and sharing of another person's feelings, as distinct from reasoning about their beliefs.

Automatic and controlled processing.

The distinction between a fast, reflexive, effortless mode of social processing (Lieberman's X-system) and a slow, reflective, deliberate mode (the C-system), cutting across social cognition by processing mode rather than content.

Cognitive theory of mind.

Reasoning about another person's beliefs, knowledge, and intentions, independent of emotional content.

Default mode network.

A set of brain regions active during rest and internally directed thought that overlaps substantially with the mentalizing network.

Empathy.

The capacity to share and vicariously experience another person's emotional state, related to but separable from theory of mind.

False-belief task.

A test of theory of mind in which success requires attributing to another a belief the participant knows to be false.

Gaze cueing.

The reflexive shift of an observer's attention toward the location another person is looking at; a rapid, largely involuntary front-end that feeds joint attention and later mind-reading.

Implicit theory of mind.

An early, nonverbal sensitivity to others' beliefs, evidenced by infants' looking behaviour, that precedes the explicit false-belief understanding measured around age four.

Medial prefrontal cortex.

A frontal midline region central to representing mental states of the self and others during mentalizing.

Mentalizing.

The process of representing and reasoning about mental states; a near synonym for theory of mind in its explicit form.

Mirror neuron.

A neuron that responds both when an action is executed and when the same action is observed, proposed as a substrate for simulation.

Sense of agency.

The felt experience of controlling one's own actions and their external effects; the sole MeSH subtype of social cognition.

Simulation theory.

The view that people understand others by covertly running their own mental processes as a model of the other's.

Social brain.

The distributed set of brain regions that support processing information about other agents.

Social cognition.

The mental processes by which people perceive, interpret, and act on information about other people and themselves as social agents.

Temporo-parietal junction.

A cortical region, right-lateralised for mentalizing, engaged selectively when reasoning about another person's beliefs.

Theory of mind.

The attribution of mental states to oneself and others in order to explain and predict behaviour.

Theory theory.

The view that people understand others by applying a tacit, causal body of knowledge about how mental states produce behaviour.

Key Researchers

Ralph Adolphs (Caltech). Bren Professor of Psychology, Neuroscience, and Biology; mapped the neural basis of social knowledge and argued for a distributed social brain rather than a single module. ORCID - Faculty Page - Google Scholar

Simon Baron-Cohen (University of Cambridge). Director of the Autism Research Centre; with Alan Leslie and Uta Frith provided the first experimental evidence of a specific false-belief difficulty in autism. ORCID - Faculty Page - Wikipedia

Christopher D. Frith (University College London). Emeritus Professor of Neuropsychology; helped found the cognitive neuroscience of social cognition and, with David Amodio, set out the role of medial prefrontal cortex in mentalizing. ORCID - Royal Society - Wikipedia

Uta Frith (University College London). Emeritus Professor of Cognitive Development; co-authored the founding false-belief study of autism and developed the mentalizing-deficit account. ORCID - Faculty Page - Wikipedia

Matthew D. Lieberman (UCLA). Professor of Psychology and a founder of social cognitive neuroscience; catalogued its core processes and linked social thought to the brain's default network. ORCID - Faculty Page - Wikipedia

Rebecca Saxe (MIT). John W. Jarve Professor in Brain and Cognitive Sciences; identified the right temporo-parietal junction as selectively engaged in reasoning about others' beliefs. ORCID - Faculty Page - Wikipedia

Frequently Asked Questions

What is social cognition?
Social cognition is the set of mental processes by which people perceive, interpret, store, and act on information about other people and about themselves as social agents (Frith, 2008).

How does social cognition differ from social psychology?
Social psychology studies how situations shape behaviour, whereas social cognition studies the information-processing and neural mechanisms that represent other agents, applying the rigour of cognitive psychology to the perception of persons (Lieberman, 2007).

What is theory of mind?
Theory of mind is the capacity to attribute mental states such as beliefs, desires, and intentions to oneself and others and to use them to predict behaviour, a term introduced in a study of chimpanzees (Premack & Woodruff, 1978).

When do children develop theory of mind?
Typically developing children begin to pass explicit false-belief tasks around the age of four, with earlier precursors such as gaze following and joint attention (Baron-Cohen et al., 1985).

What is the difference between theory theory and simulation theory?
Theory theory holds that people reason about minds using tacit causal knowledge, whereas simulation theory holds that people model other minds by covertly running their own; the evidence supports a mixture of the two (Gallese & Goldman, 1998).

Which brain regions support social cognition?
A mentalizing network of medial prefrontal cortex, the temporo-parietal junction, the precuneus, and the temporal poles supports reasoning about minds, partly separable from the action-observation mirror system (Schurz et al., 2021).

Do mirror neurons explain empathy and mind-reading?
Mirror neurons plausibly support understanding of observed actions, but the mirror system is anatomically distinct from the mentalizing network, so mirroring accounts for only part of social understanding (Molenberghs et al., 2016).

Why does social cognition matter clinically?
Impaired social cognition is central to autism and is among the strongest predictors of everyday functioning in schizophrenia, making it a target for assessment and remediation (Green et al., 2015).

References