The headlineTheory of mind is the achievement of this age
Piaget framed preschoolers by what they lack — they can't decenter, can't conserve, are "egocentric." The modern story is what they're gaining: a working model of other people's minds. This is arguably the central cognitive milestone of early childhood.
To have a theory of mind is to treat other people as intentional agents whose behavior flows from inner states — beliefs, desires, intentions, and knowledge — rather than from the observable situation alone. The decisive test is whether a child can represent a state of the world that is false yet still governs how someone acts. That criterion was made rigorous in the study that launched the field, in which children heard about a boy who left his chocolate in one cupboard and, while he was away, had it moved to another; only children who could hold the boy's outdated belief separate from the current reality predicted he would search where he had left it (Wimmer & Perner, 1983). Passing that task means representing a representation — thinking about someone else's thought as a thing that can be accurate or mistaken — which is why theory of mind is often described as metarepresentational.
The stakes of this achievement reach well beyond a laboratory puzzle. A child who grasps that beliefs can be false can also lie strategically, keep a secret, understand that a playmate is mistaken rather than dishonest, follow a story built on dramatic irony, and repair a misunderstanding by imagining what the other person does or does not know. Understanding of desires and simple emotions comes earlier and more easily; the belief–reality distinction is the harder, later-emerging piece, and its arrival reorganizes a preschooler's social world. Framing early childhood around this gain, rather than around Piaget's catalog of deficits, captures what is genuinely new about the preschool mind.
Theory of mind is understanding that other people have beliefs, desires, and knowledge that differ from your own — and can even be false. The classic test is the false-belief task: Sally puts her marble in a basket and leaves; Anne moves it to a box; where will Sally look when she returns? Children under ~4 say "the box" — they can't separate what they know from what Sally believes. Around age 4–5 they pass, answering "the basket." Wellman, Cross, and Watson's (2001) meta-analysis of 178 studies confirmed this shift is robust and cross-cultural in its timing, though language and having siblings speed it along.
Piaget's "egocentrism" gets taught as a flat inability — the child trapped inside their own viewpoint. But perspective-taking is graded and emerging, not all-or-nothing. Even 2-year-olds adjust what they show a person based on what that person can see; 3-year-olds understand desires differ before they master beliefs; and by 4–5 the full false-belief understanding is in place. The honest picture is a developmental gradient of increasingly sophisticated mind-reading — not a wall that suddenly falls at age 7. Egocentrism describes an early difficulty, not a total blindness to other minds.
Theory of mind develops in steps. Tap one to see what emerges when.
NuanceHow solid is the false-belief milestone?
The single most reassuring fact about theory of mind is how orderly its development is. A meta-analysis of 178 conditions found that performance on standard false-belief tasks climbs from well below chance in three-year-olds to reliably above chance by about age four or five, with the same trajectory appearing whether the task involved an unexpected location, a deceptive container, or an appearance–reality contrast (Wellman, Cross, & Watson, 2001). The developmental order that precedes it is equally stable: understanding that people want different things comes first, then that they believe different things, then that seeing leads to knowing, and only then that a belief can be outright false — a sequence captured by a psychometric scale that holds across children (Wellman & Liu, 2004). That same ordered progression, on the same rough timetable, has been documented in children from Canada, India, Peru, Samoa, and Thailand, which argues that the shift reflects a maturing cognitive capacity rather than a peculiarly Western schooling effect (Callaghan et al., 2005).
Two forces reliably move the timetable. Language is the strongest: the richer a child's syntax and vocabulary, and especially their command of sentence complements — the grammar of "she thinks that the marble is in the basket" — the earlier they pass, an association robust enough to survive meta-analytic scrutiny (Milligan, Astington, & Dack, 2007). Social experience is the other; children with more siblings, who spend their days negotiating conflicting wants and rival versions of events, tend to pass sooner. Both fit the picture of theory of mind as something built through talk and interaction rather than switched on by a single maturational event.
Whether anything mind-reading exists before children can talk is the field's liveliest controversy.
The standard verbal task may underestimate infants badly. When false belief is probed nonverbally — by measuring how long a baby looks at an event that violates what a mistaken agent should expect — infants as young as 15 months already look longer when an actor reaches to the location she has no way of knowing about, as if surprised that she "knows" what she should not (Onishi & Baillargeon, 2005). Interpretations of these looking-time findings are contested: some read them as evidence for an early, implicit tracking of others' beliefs that later matures into the explicit understanding tapped by the Sally–Anne task, while others argue that simpler attentional rules could produce the same looking patterns without any genuine belief attribution. The unresolved gap between early implicit sensitivity and later explicit reasoning is now one of the central puzzles in the study of social cognition, and it is a useful reminder that a failed verbal task marks the limit of a method as much as the limit of a mind.
The classic accountPiaget's preoperational stage — what it got right and wrong
For Piaget, the years from roughly two to seven make up the preoperational stage, defined by the arrival of symbolic thought and by the child's continuing inability to perform mental "operations" — reversible, logical transformations carried out in the head. The gains are real and dramatic. A preschooler can let a banana stand in for a telephone, act out elaborate pretend scenarios, use words and drawings to refer to absent things, and reason about objects that are not in front of them. This flowering of symbols, most visible in the explosion of pretend play, is exactly the capacity that theory of mind and language build upon (Piaget & Inhelder, 1969).
Piaget defined the stage, though, mostly by its limits. Egocentrism named the child's difficulty representing a viewpoint other than their own, illustrated by the three-mountains task, in which young children asked what a doll seated across the table can see tend to choose their own view. Centration named the tendency to lock onto a single salient dimension of a situation while ignoring the rest, and it is the engine behind the stage's most famous failure: the conservation error. Shown two identical glasses of juice, then watching one poured into a taller, thinner glass, a preoperational child insists the tall glass now holds "more," centering on height while neglecting the compensating change in width and forgetting that the pour could simply be reversed. Failures of conservation of number, mass, and volume were, for Piaget, the signature that logical operations had not yet arrived. He also catalogued animism, the tendency to grant thoughts and feelings to the sun, the moon, or a broken toy.
The modern verdict is that Piaget was a superb observer of what preschoolers do and an unreliable guide to why. His tasks routinely buried competence under heavy demands — confusing wording, unfamiliar settings, and questions that seemed to ask children to contradict an adult. Strip those demands away and hidden ability appears. When number is tested with tiny sets a child can take in at a glance, three-year-olds track numerical identity and react with genuine surprise when a hidden item is secretly added or removed, revealing an appreciation of number invariance that the classic conservation task masks (Gelman, 1972). Perspective-taking shows the same pattern: even two- and three-year-olds adjust what they show or tell someone based on what that person can see, so the "egocentric" child is better described as one whose perspective-taking is fragile and effortful than as one blind to other viewpoints. The stage names a set of characteristic difficulties, not a wall of incapacity, and its boundaries are far softer and more task-dependent than the textbook staircase implies.
The reappraisal worth knowingThe marshmallow test, thirty years on
The delay-of-gratification study is the most famous thing in developmental psychology, and it's usually taught wrong. The self-control story is real but far smaller than the legend.
Mischel's original finding — preschoolers who waited for two marshmallows fared better decades later — got inflated into "willpower at age 4 determines your life." A large preregistered replication with a more diverse, tenfold-bigger sample found the long-run association much weaker, and it largely disappeared once you controlled for the child's family background, home environment, and early cognitive skills. The lesson isn't that self-control is fake — it's that the ability to wait is partly a readout of a child's circumstances (Is the world reliable? Are resources predictable?), not a fixed inner trait. A hungry kid in an unpredictable environment is being rational to grab the treat now.
The original studies are worth stating precisely, because the legend has drifted from them. Preschoolers at a university lab were left alone with a treat and told they could ring a bell and eat it at once or wait — often around fifteen minutes — for a larger reward; the seconds a child held out predicted, years later, higher rated adolescent competence and, in follow-ups, stronger academic performance (Mischel, Shoda, & Rodriguez, 1989). What made the finding theoretically interesting was never raw willpower but strategy: the children who waited longest were the ones who looked away, distracted themselves, or mentally recast the marshmallow as a puffy cloud rather than a delicious food. Delay of gratification, in this account, is less an inner reserve of grit than a set of learnable attention-management skills — turning the "hot," appetitive pull of the reward into a "cool," abstract representation.
The predictive story is the part that has been cut down to size. A conceptual replication with a sample roughly ten times larger and far more socioeconomically diverse found the association between preschool wait time and later achievement to be less than half the size the early work implied, and to shrink toward nothing once family background, home environment, and early cognitive ability were held constant (Watts, Duncan, & Quan, 2018). Delay behavior, on this reading, is partly a readout of the conditions a child grew up in rather than a fixed trait that causes later success. A complementary experiment makes the logic vivid: children who had just watched an adult break a promise — an unreliable environment — waited far less time for a second marshmallow than children who had just experienced the same adult as trustworthy (Kidd, Palmeri, & Aslin, 2013). Waiting only pays when the promised future actually arrives, so a child who grabs the sure treat in an unpredictable world is not failing a test of character; they are making a sensible bet. The honest lesson is that self-control is real and trainable, but it is embedded in circumstance, and no four-year-old's fate is sealed by a marshmallow.
Two clever findingsSymbols and imitation
Two studies that illuminate how preschool minds handle representation — both memorable.
Dual representation (DeLoache, 2000)
Show a 2.5-year-old where a toy is hidden in a scale model of a room, then send them to find it in the real room. They fail — they can't hold the model as both an object and a symbol standing for something else. Three-year-olds succeed. This dual representation — treating one thing as a sign for another — is the cognitive root of understanding maps, pictures, and eventually written words. A tidy window into symbolic development that goes beyond the usual "animism" discussion.
Overimitation
Show a preschooler how to open a puzzle box using steps that are obviously causally irrelevant (tap the lid twice, then slide the latch). Children faithfully copy the useless steps too — and do so more than chimpanzees, who skip the pointless bits. Far from being a bug, this overimitation is thought to be a powerful engine of cultural learning: copying high-fidelity, even when you don't see why, lets kids absorb practices whose logic they can't yet reconstruct.
Both findings speak to how a preschooler handles a symbol, and both have been sharpened by later work. The scale-model result is a clean demonstration that using something as a sign requires holding two incompatible construals at once — the model as a real, interesting object and the model as a stand-in for the room — and that this dual representation is genuinely hard for young two-year-olds; reducing the model's salience as an object, for instance by placing it behind a window so children treat it as a mere picture, actually improves their performance (DeLoache, 2000). The same demand underlies why toddlers struggle with maps, photographs, and scale replicas long after they can name the things depicted.
Overimitation points the other way — toward the surprising reach of preschool learning. When an adult opens a puzzle box with a sequence that visibly includes useless steps, children reproduce the pointless actions along with the effective ones, and they do so even when explicitly warned to skip anything silly; wild chimpanzees, by contrast, edit the causally irrelevant steps out (Lyons, Young, & Keil, 2007). What looks like a failure of causal reasoning is better understood as a strong bias to treat a knowledgeable demonstrator's actions as intentional and worth copying in full. That high-fidelity, "copy first, understand later" strategy is costly in a single puzzle box but powerful across a childhood, because it lets a novice absorb culturally accumulated practices — cooking, ritual, tool use — whose purposes are opaque and could never be reinvented from scratch. Overimitation, in short, may be a design feature of a species that survives by culture.
Vygotsky vindicatedPrivate speech is a tool, not a symptom
Private speech shows up under Vygotsky's theory. Worth making explicit: this is a place where Vygotsky was right and Piaget was wrong, and the disagreement is worth understanding.
Piaget dismissed the way preschoolers talk to themselves as "egocentric speech" — a sign of immaturity that simply fades. Vygotsky countered that private speech is a cognitive tool: children externally narrate to guide and regulate their own behavior, and over time this self-talk goes underground to become silent inner speech — the running verbal thought adults use to plan and self-control. The evidence backs Vygotsky: kids use more private speech on harder tasks and when working alone toward a goal, exactly what you'd expect if it's self-regulation, not leftover egocentrism. Private speech doesn't disappear — it internalizes.
The empirical case for Vygotsky's reading has held up well. Careful observation of preschoolers shows that self-directed speech rises precisely when a task gets difficult or when a child hits an obstacle, falls off as the task is mastered, and predicts better subsequent attention and performance — the profile of a self-regulatory tool, not of leftover immaturity (Berk, 1994). Over the preschool and early school years the talk does not so much vanish as change form, moving from full out-loud commentary to muttering and inaudible lip movement before going fully silent as inner speech, the verbal channel adults still use to plan, count, and talk themselves through a hard moment. Piaget's own late position conceded much of this ground.
Vygotsky's engineThe zone of proximal development and scaffolding
Private speech is one expression of Vygotsky's larger claim that higher mental functions are social before they are individual: a child first regulates a task with a more capable partner and only later regulates it alone, having internalized the shared activity. The concept that carries this idea is the zone of proximal development — the gap between what a child can do unaided and what the same child can do with skilled help. Development, on this view, happens at the leading edge of that zone, where instruction pitched just beyond a child's solo reach pulls competence forward (Vygotsky, 1978).
The support a partner supplies in that zone was later named scaffolding: a good tutor does not solve the problem but structures it — highlighting the important features, breaking the task into reachable steps, keeping the child oriented toward the goal, and controlling frustration — then withdraws the help as the child takes over, exactly as a construction scaffold comes down once the building stands (Wood, Bruner, & Ross, 1976). The contrast with Piaget is the whole point. Where Piaget pictured the child as a lone scientist constructing knowledge through solitary action on the world, Vygotsky pictured a child apprenticed into the intellectual tools of a culture — its language, number systems, memory strategies, and habits of problem-solving — through guided participation with others. Neither account is complete, but the Vygotskian half explains something Piaget's cannot: why the shape of cognitive development varies so much with the tools, talk, and tutoring a particular culture makes available.
The synthesisDeficits reframed as construction
The through-line of this age is a shift in how we ask the question. Piaget's catalog of what preschoolers cannot do — decenter, conserve, take a viewpoint, resist a treat — is not wrong so much as incomplete and often unfair to the child, because his tasks loaded the answer with demands that hid real competence (Gelman, 1972; Wellman & Liu, 2004). Read forward instead of backward, the preschool years are a burst of construction: a mind assembling a working model of other minds, learning to hold one thing as a symbol for another, copying its culture with a fidelity that outruns its own understanding, and taking over the regulatory talk it first heard from adults. Even the famous marshmallow is better read as a child reasoning about a world than as a child failing a trait test (Kidd, Palmeri, & Aslin, 2013; Watts, Duncan, & Quan, 2018). The honest chapter title for early childhood is not what these children lack, but what, at remarkable speed, they are building.
SourcesCited in APA 7
Berk, L. E. (1994). Why children talk to themselves. Scientific American, 271(5), 78–83.
Callaghan, T., Rochat, P., Lillard, A., Claux, M. L., Odden, H., Itakura, S., Tapanya, S., & Singh, S. (2005). Synchrony in the onset of mental-state reasoning: Evidence from five cultures. Psychological Science, 16(5), 378–384.
DeLoache, J. S. (2000). Dual representation and young children's use of scale models. Child Development, 71(2), 329–338.
Gelman, R. (1972). Logical capacity of very young children: Number invariance rules. Child Development, 43(1), 75–90.
Kidd, C., Palmeri, H., & Aslin, R. N. (2013). Rational snacking: Young children's decision-making on the marshmallow task is moderated by beliefs about environmental reliability. Cognition, 126(1), 109–114.
Lyons, D. E., Young, A. G., & Keil, F. C. (2007). The hidden structure of overimitation. Proceedings of the National Academy of Sciences, 104(50), 19751–19756.
Milligan, K., Astington, J. W., & Dack, L. A. (2007). Language and theory of mind: Meta-analysis of the relation between language ability and false-belief understanding. Child Development, 78(2), 622–646.
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children. Science, 244(4907), 933–938.
Onishi, K. H., & Baillargeon, R. (2005). Do 15-month-old infants understand false beliefs? Science, 308(5719), 255–258.
Piaget, J., & Inhelder, B. (1969). The psychology of the child. Basic Books.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard University Press.
Watts, T. W., Duncan, G. J., & Quan, H. (2018). Revisiting the marshmallow test: A conceptual replication investigating links between early delay of gratification and later outcomes. Psychological Science, 29(7), 1159–1177.
Wellman, H. M., Cross, D., & Watson, J. (2001). Meta-analysis of theory-of-mind development: The truth about false belief. Child Development, 72(3), 655–684.
Wellman, H. M., & Liu, D. (2004). Scaling of theory-of-mind tasks. Child Development, 75(2), 523–541.
Wimmer, H., & Perner, J. (1983). Beliefs about beliefs: Representation and constraining function of wrong beliefs in young children's understanding of deception. Cognition, 13(1), 103–128.
Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving. Journal of Child Psychology and Psychiatry, 17(2), 89–100.