Theory of mind is the ability to attribute mental states to other people and to yourself: beliefs, desires, intentions, emotions, knowledge. It is what lets you realize that someone else can hold a belief you know to be wrong, or that a friend’s smile might mask disappointment. The concept sits at the center of developmental psychology, neuroscience, and philosophy of mind, and it touches nearly every social interaction you have, from telling a white lie to judging whether someone harmed you on purpose.
How Children Learn That Other People Have Minds
The classic way researchers test theory of mind in young children is the false-belief task. In the best-known version, called the Sally-Anne test, a child watches a short scenario: Sally puts a marble in a basket and leaves the room. While she is gone, Anne moves the marble to a box. The child is then asked where Sally will look for her marble when she comes back. To answer correctly, the child has to understand that Sally holds a false belief about the marble’s location. She did not see it move, so she will look in the basket even though the marble is actually in the box.
For decades, the standard finding was that children could not pass this test until around age four. Before that, they would point to the box, answering based on what they themselves knew rather than what Sally believed. But more recent research has complicated that timeline. Studies using eye-tracking and anticipatory looking rather than verbal questions have found signs of false-belief understanding in much younger children. One study using an anticipatory looking method found meaningful differences between 12-month-olds and older infants, though even some of the 12-month-olds showed evidence of an implicit grasp of others’ beliefs.1Korean Journal of Child Studies. The Developmental Differences of Implicit Theory of Mind in Infants Using Anticipatory Looking Paradigm A review of this evidence concluded that a robust, flexible understanding of false belief appears to emerge in infancy, and that older children who fail verbal tasks may be tripped up by the demands of the task itself rather than by a genuine lack of understanding.2Current Directions in Psychological Science. The Developmental Origins of False-Belief Understanding
Computer-based versions of the Sally-Anne test have also been developed for specific populations. One pilot study adapted the scenario into an interactive tablet application with filmed characters, designed for children with autism, where children watched a short clip and then pointed to the location they thought the character would search.3PubMed Central. Assessing False-Belief Understanding in Children with Autism Using a Computer Application: A Pilot Study The shift toward non-verbal and technology-assisted testing has been important because it allows researchers to separate the skill they actually want to measure from other cognitive demands like language ability and impulse control.
Cognitive Versus Affective Theory of Mind
Theory of mind is not a single, monolithic skill. Researchers now distinguish between at least two components. Cognitive theory of mind is the ability to infer what someone else thinks or believes. Affective theory of mind is the ability to infer what someone else feels. The distinction matters because these two abilities do not always travel together: a person can be sharp at reading beliefs but poor at reading emotions, or the reverse.
Brain-imaging studies show that these components rely on partly separate neural circuits. Damage to certain prefrontal areas tends to impair cognitive theory of mind while leaving the affective side relatively intact, and vice versa for damage to other regions.4PubMed. Dissociable prefrontal networks for cognitive and affective theory of mind: a lesion study An imaging study found that affective theory of mind recruited the basal ganglia, a set of deep brain structures not activated during purely cognitive mind-reading tasks.5PubMed Central. Comparing the neural correlates of affective and cognitive theory of mind using fMRI: Involvement of the basal ganglia in affective theory of mind And when researchers temporarily disrupted a specific part of the right prefrontal cortex using magnetic stimulation, participants’ performance on cognitive theory of mind tasks changed while affective tasks were unaffected.6PubMed. Dissociating cognitive from affective theory of mind: a TMS study
This split has practical consequences. Clinical conditions do not impair theory of mind in a blanket way. People with schizophrenia, for instance, tend to show broad impairments across both dimensions, while people with borderline personality disorder show a different pattern: they tend to over-read intentions into ambiguous social situations rather than miss them.7PubMed. Factors contributing to social cognition impairment in borderline personality disorder and schizophrenia Understanding which component is affected changes the therapeutic approach.
The Brain Network That Reads Minds
There is no single “theory of mind area” in the brain. Instead, neuroimaging research has identified a network of regions that activate together when you reason about other people’s mental states. A review of the imaging literature described several core regions in this network, including parts of the prefrontal cortex and the superior temporal sulcus, along with more peripheral regions that join in depending on the specific task.8PubMed Central. Are there theory of mind regions in the brain? A review of the neuroimaging literature
One long-running debate is how this mentalizing network relates to the mirror neuron system, which activates when you observe someone else performing an action. For a while, these were treated as competing explanations: either you understand others by simulating their actions in your own motor system (the mirror account) or by reasoning about their mental states using a separate, higher-level system (the mentalizing account). The evidence now points to both systems working together. During imitation of communicative gestures, brain areas associated with both the mirror system and the mentalizing system activated, with the specific pattern depending on whether the gesture carried social meaning.9PubMed. Differential role of the Mentalizing and the Mirror Neuron system in the imitation of communicative gestures During live social interactions involving imitation, the two systems showed strong coupling, with the mirror system engaging in action preparation and simulation while the mentalizing system anticipated the other person’s intentions.10PubMed. Mirror Neuron System and Mentalizing System connect during online social interaction
A more recent study using brain scanning of two people simultaneously during cooperative tasks found that this coupling extended across brains, with synchronized activity between the motor, mirror, and mentalizing systems of the two interacting partners.11PubMed. Synergy of the mirror neuron system and the mentalizing system in a single brain and between brains during joint actions The picture, then, is one of integration: reading minds is not purely about abstract reasoning or purely about motor mirroring. Your brain recruits whichever tools the situation calls for and coordinates them on the fly.
How Siblings, Adversity, and Culture Shape the Skill
Theory of mind does not develop in a vacuum. The social environment a child grows up in can accelerate or slow its emergence. One well-studied factor is siblings. Children who had at least one sibling close in age scored higher on a battery of theory of mind tasks than only children or those whose siblings were much older or much younger, even after controlling for language ability.12British Journal of Developmental Psychology. Mental playmates: Siblings, executive functioning and theory of mind The likely mechanism is straightforward: children with same-age siblings spend more time negotiating, pretending, arguing, and cooperating, all of which require tracking someone else’s point of view.
Socioeconomic adversity also plays a role. A study of children in Colombia found that socioeconomic hardship predicted lower theory of mind scores, while exposure to armed conflict predicted lower executive function but, interestingly, higher empathy.13PubMed. The Role of Socioeconomic Adversity and Armed Conflict in Executive Function, Theory of Mind and Empathy in Children These findings suggest that different forms of stress do not uniformly help or harm social cognition; the type of adversity matters. Chronic poverty may limit the cognitive resources or social exposure needed for mind-reading, while living in a dangerous environment may heighten sensitivity to others’ emotions as a survival strategy.
What about childhood trauma specifically? A large study of adults found that neither interpersonal loss nor child maltreatment was significantly associated with lower theory of mind in adulthood, and the age at which trauma occurred did not make a difference either.14PubMed Central. Associations Between Childhood Trauma Characteristics and Theory of Mind in Adults: Results From a Large, Diverse Sample This does not mean trauma has no social consequences, but it suggests that theory of mind specifically may be more resilient to early adversity than researchers once assumed.
Culture shapes how children perform on different kinds of theory of mind tests, though perhaps not in the ways you might expect. A cross-cultural comparison of Chinese and Australian children found no cultural differences on tasks measuring the ability to interpret social situations or to recognize faux pas. However, Chinese children performed lower on the Reading the Mind in the Eyes test and on Strange Stories, tasks that lean more heavily on reading subtle emotional cues from faces and interpreting ambiguous social narratives.15PubMed. Age-Related Changes in Theory of Mind in Middle Childhood: A Cross-Cultural Comparison Between Australian and Chinese Children Whether this reflects differences in socialization norms, familiarity with the task stimuli, or genuine variation in social cognitive development is still debated.
Second-Order Reasoning and Why Irony Is Hard
First-order theory of mind is understanding that someone else can hold a belief. Second-order theory of mind is understanding that someone can hold a belief about another person’s belief: “She thinks that he thinks the ball is in the basket.” This nested reasoning is considerably harder and develops later in childhood. Research shows that children find second-order reasoning easier when the scenario involves positive desires and true beliefs, and harder when it involves negative desires or false beliefs.16PubMed Central. Desires and beliefs: the development of second-order Theory of Mind reasoning in preschoolers and in school-age children
This matters because many sophisticated forms of communication depend on second-order mind-reading. Irony is a case in point. When someone says “What lovely weather” during a downpour, understanding the irony requires grasping not just what the speaker believes (the weather is terrible) but what the speaker intends the listener to understand about the gap between their words and their belief. Several studies have found a link between second-order theory of mind and irony comprehension, though the relationship is not airtight: some children who pass second-order tasks still struggle with irony, and some studies have failed to find the connection at all.17Philosophical Transactions of the Royal Society B. Affective Theory of Mind and linguistic intimacy in irony production: a developmental perspective Children with cerebral palsy, who often show difficulties with second-order theory of mind, also tend to struggle with irony comprehension, consistent with the idea that the two abilities share cognitive architecture even if one does not strictly require the other.18PubMed. Theory of mind and irony comprehension in children with cerebral palsy
Theory of Mind and Moral Judgment
Your ability to read someone else’s intentions is deeply entangled with how you judge their behavior morally. Most adults consider an intentional harm worse than an accidental one, even when the outcome is identical. Brain imaging research has shown that the right temporoparietal junction, a key node in the theory of mind network, plays a central role in this distinction. It activates when you are weighing someone’s intentions against the outcome of their actions, and it responds most strongly in cases of attempted harm, where someone tried to hurt another person but failed.19PubMed Central. The neural basis of the interaction between theory of mind and moral judgment Neural patterns in this region can even be used to predict how a person will judge a moral scenario.20PubMed Central. Decoding moral judgments from neural representations of intentions
Children develop this intent-based moral reasoning between about age four and eight, and the shift maps closely onto their developing theory of mind. Younger children tend to judge actions by their outcomes: if something bad happened, the person who caused it is bad. Older children increasingly weigh the actor’s intent, forgiving accidental harm and condemning attempted harm even when it fails.21PubMed. The development of intent-based moral judgment This trajectory suggests that moral development is not purely about learning rules; it is partly about gaining the cognitive tools to peer inside other people’s heads.
What Happens to Theory of Mind as You Age
Theory of mind does not plateau after childhood. It continues to develop through adolescence and then, in older adulthood, begins to decline, at least on standard lab tasks. Both cognitive and affective components show age-related drops, with younger adults consistently outperforming older groups on theory of mind batteries.22PubMed. Affective and Cognitive Theory of Mind in Aging: What’s the Role of Executive Functions and Cognitive Reserve on the Decline?
But there is a twist. When researchers tested older adults using implicit eye-tracking measures rather than explicit verbal tasks, the age difference vanished. Both younger and older participants showed the same looking patterns, automatically tracking another person’s false belief.23PLoS ONE. Explaining age-related decline in theory of mind: Evidence for intact competence but compromised executive function The same study found statistical evidence in favor of there being no real age difference in this implicit measure. The implication is that older adults may still possess the core competence to read minds but struggle with the executive demands of traditional tasks: holding multiple pieces of information in memory, inhibiting their own perspective, and formulating a verbal response. This echoes the infant research, where younger children also appear to understand false beliefs implicitly before they can demonstrate it explicitly. The pattern suggests that theory of mind, once developed, may be more durable than test scores imply.
The Double Empathy Problem and Neurodivergence
For decades, difficulties with theory of mind were treated as a defining feature of autism. The original framing was straightforward: autistic people struggle to model other minds, and this explains their social difficulties. That account is not wrong exactly, but it is incomplete in a way that has real consequences.
The double empathy problem, a concept introduced by autistic researcher Damian Milton, argues that the communication gap between autistic and non-autistic people runs in both directions. A study tested this by having non-autistic participants watch autobiographical accounts from both autistic and non-autistic narrators. Participants were significantly less accurate at reading the emotions of autistic narrators, especially for happy and sad emotions. At the same time, viewing autistic narrators produced stronger bodily emotional responses in participants, particularly for anger and fear.24PubMed Central. Do you feel me? Autism, empathic accuracy and the double empathy problem In other words, non-autistic people also struggle to read autistic minds. The difficulty is mutual, not one-sided.
This reframing matters practically. Therapies built on the assumption that autistic people simply have a broken theory of mind module may miss the point. If the challenge is partly about differing social styles and expressive norms rather than a straightforward deficit, then interventions need to work on bridging the gap from both sides, not just training autistic people to perform non-autistic social conventions.
Overmentalizing and Clinical Presentations
Theory of mind difficulties in clinical populations are not always about under-reading others. People with borderline personality disorder, for instance, tend toward a pattern called overmentalizing: attributing complex, often hostile intentions to other people in ambiguous situations. In contrast, people with schizophrenia tend to undermentalize, missing social cues altogether and struggling with basic inference about others’ beliefs and feelings.25PubMed. Factors contributing to social cognition impairment in borderline personality disorder and schizophrenia One condition involves reading too much into the social world, the other too little, and the subjective experience of each is dramatically different.
Oxytocin, often casually called the “bonding hormone,” appears to play a role in these differences. In schizophrenia, patients showed lower plasma oxytocin levels compared to healthy controls, and those levels were positively correlated with affective but not cognitive theory of mind. Oxytocin levels also partly explained the link between impaired mind-reading and poor social functioning in these patients.26PubMed. Relationship between the domains of theory of mind, social dysfunction, and oxytocin in schizophrenia A pilot trial in children with ADHD, another condition associated with social cognition difficulties, found that intranasal oxytocin improved reaction times and response consistency on theory of mind and emotion recognition tasks.27PubMed. The effect of intranasal oxytocin on theory of mind in children with attention deficit/hyperactive disorder These are preliminary findings, but they point toward the possibility that social cognition is not purely a matter of learning and experience. The neurochemical environment your brain operates in also affects how well you read other people.
Do Other Animals Have Theory of Mind
The question of whether theory of mind is uniquely human has been debated since 1978, when researchers first asked whether chimpanzees understand what others know. Three decades of follow-up work has produced a nuanced answer. Chimpanzees clearly understand goals, intentions, and perceptions: they know what another individual can see, they adjust their behavior based on what a competitor knows, and they anticipate what another chimpanzee is trying to do. But despite multiple well-designed attempts, there is currently no convincing evidence that chimpanzees understand false beliefs the way human children do by age four or five.28PubMed. Does the chimpanzee have a theory of mind? 30 years later The conclusion, at least for now, is that chimps operate with something like a perception-goal psychology: they understand what others see and want, but not what others mistakenly believe.
Some of the most intriguing evidence for non-primate mind-reading comes from corvids. Western scrub-jays that have been observed while caching food will later return to re-hide it, but only if they were watched by another bird and only if they themselves have experience pilfering other birds’ caches. They also keep track of which specific bird was watching and adjust their protective strategies accordingly. These behaviors raise the possibility that the jays simulate what another bird knows about the cache location, a form of mental attribution.29PubMed Central. Social cognition by food-caching corvids. The western scrub-jay as a natural psychologist Whether this qualifies as “true” theory of mind or a simpler rule-based strategy remains open, but it suggests that the pressure to predict what others know may not require a primate brain.
Can Reading Fiction Exercise Your Mind-Reading Muscles
A widely publicized set of experiments reported that reading literary fiction, as opposed to popular fiction or nonfiction, temporarily improved performance on tests of both affective and cognitive theory of mind.30PubMed. Reading literary fiction improves theory of mind The proposed mechanism was that literary fiction, with its ambiguous characters and unresolved motivations, forces readers to actively construct mental models of others in a way that genre fiction, with its more predictable characters, does not.
The finding generated excitement across the humanities, where it seemed to offer hard evidence for the value of literary education. But subsequent replication attempts have been mixed, with preregistered studies yielding inconsistent results.31Social Psychological and Personality Science. Reading Literary Fiction and Theory of Mind: Three Preregistered Replications and Extensions of Kidd and Castano (2013) The effect, if real, may be small or may depend on factors like how absorbed the reader is, what specific text they read, and how theory of mind is measured. The broader question of whether habitual fiction reading over years shapes social cognition is harder to test experimentally and remains genuinely open.
Can AI Pass Theory of Mind Tests
The rise of large language models has prompted a new and slightly surreal question: can AI systems that produce human-like text also demonstrate human-like mind-reading? The answer depends heavily on which model and which test you are looking at. In one rigorous evaluation using forty custom-designed false-belief tasks where the model had to solve all eight scenarios within a task to get credit, older language models solved none. GPT-3.5 solved about a fifth. GPT-4 solved three-quarters, matching the performance of six-year-old children.32PubMed Central. Evaluating large language models in theory of mind tasks
Another study administered a range of established theory of mind tests to multiple models, including false-belief tasks, faux pas recognition, irony comprehension, and the hinting task.33Nature Human Behaviour. Testing theory of mind in large language models and humans The results were striking but also hard to interpret. These models have been trained on enormous quantities of human text that describe mental states, intentions, and beliefs. Whether they are genuinely modeling others’ perspectives or pattern-matching on the statistical structure of how humans describe such scenarios is a question that current tests cannot definitively answer. A systematic review of this research area noted that existing benchmarks may not adequately capture the flexible, context-sensitive nature of real human theory of mind.34arXiv. A Systematic Review on the Evaluation of Large Language Models in Theory of Mind Tasks
The Debate That Will Not Resolve
Since the 1980s, two rival theories have tried to explain how theory of mind actually works in the brain. Theory-Theory says we build folk-psychological models of other minds the way we build theories about the physical world: through inference, evidence, and updating. Simulation Theory says we understand others by running a kind of mental simulation, using our own mind as a model to predict what someone else will do or feel. Both accounts have intuitive appeal, and both can explain many of the same findings. A detailed critique argued that neuroimaging methods, despite early claims, have not provided data that clearly discriminates between these two accounts for the kind of mental-state reasoning that matters most, and it is uncertain they will in the future.35PubMed. Beyond Simulation-Theory and Theory-Theory: why social cognitive neuroscience should use its own concepts to study “theory of mind” The field increasingly treats both frameworks as partial descriptions rather than competing explanations, and many researchers have moved toward studying the neural and cognitive mechanisms directly rather than trying to shoehorn findings into one philosophical camp.