Why a three year old talking to a banana is doing some of the most sophisticated cognitive work of early childhood, what the evidence says it builds, and what adults should and should not do about it.
Pretend play is the ability to hold two realities in mind at once. When a child lifts a banana to their ear and says hello, they know perfectly well it is a banana. They are not confused. They are deliberately treating one thing as another while keeping the truth in view, and that double bookkeeping, the banana is a phone AND a banana, is called symbolic representation. It is the same mental move that underpins language, where a sound stands for a thing, and later literacy, where a squiggle stands for a sound. A child who can make a banana stand for a phone is rehearsing the foundation of reading.
This is why developmental psychologists treat pretend play not as a pleasant distraction from learning but as one of the most advanced things a young child does. The benefits are concrete and well evidenced: stronger language, better self-control, an earlier grasp of other minds, and a safe rehearsal space for difficult feelings. The rest of this page walks through each of those, and through what the research says about your role in it.
The Soviet psychologist Lev Vygotsky made the most famous claim about pretend play, and it remains the frame researchers use a century later. In play, he argued, a child behaves "as though he were a head taller than himself". A three year old who cannot sit still at the table will stand motionless for minutes because she is being a guard, and guards do not fidget. The role supplies a rule, and the child holds herself to the rule voluntarily, because breaking it would break the game.
Vygotsky called the gap between what a child can do alone and what they can do with support the zone of proximal development, and he saw pretend play as a place where children create that zone for themselves. In play, the child performs above their everyday ability: more patience, more vocabulary, longer attention, better self-control than they show anywhere else. That makes pretend play less like recreation and more like a training ground the child builds and runs without being asked.
Key citation
Vygotsky, L. (1933/1967). Play and Its Role in the Mental Development of the Child · The source of the "head taller" observation and the zone of proximal development as applied to pretend play.
To play a character, a child must model a mind that is not their own. The doll is hungry, the dragon is cross, the shopkeeper wants paying. Each of those requires the child to hold beliefs and wants that differ from their actual beliefs and wants, which is precisely the skill psychologists call theory of mind: the understanding that other people have inner lives unlike yours. Children who do a lot of role play are getting daily, self-directed practice at it, and it is the bedrock of empathy, friendship and getting along in a classroom.
Staying in role is inhibitory control. The child playing the patient must wait while the doctor fusses with the bandage; the child playing the dog must not talk. Pretend play asks children to suppress their first impulse in favour of what the role requires, to hold a storyline in working memory, and to switch flexibly when a playmate changes the plot. Those three capacities, inhibition, working memory and flexibility, are the core of executive function, and they predict school success more strongly than early academic knowledge does.
The talk that happens inside pretend play is measurably richer than ordinary toddler talk. Children narrate, negotiate, assign roles, voice characters and explain invisible scenarios, all of which pulls vocabulary and sentence structure upward. Researchers Kathy Hirsh-Pasek and Roberta Golinkoff have spent careers showing that language grows fastest in exactly this kind of meaningful, back-and-forth, child-led exchange, and pretend play generates more of it per minute than almost any other activity in a young child's day.
Key idea from the research
Hirsh-Pasek, K. & Golinkoff, R. · decades of work on playful learning finds that the conversational, child-directed talk produced during pretend play is among the richest language input and output in early childhood.
Children replay what happens to them. The trip to the doctor becomes a week of teddy bear injections; the new baby becomes a doll that must be fed and, occasionally, posted behind the sofa. This is not avoidance, it is digestion. In pretend play the child controls the script, so frightening or overwhelming experiences can be run again at a safe size, with the child as the doctor rather than the patient, until the feeling is mastered. Play is where the small process the big.
Pretend play arrives in a fairly predictable sequence, with wide normal variation around every milestone.
The evidence on adult involvement is consistent and slightly humbling: the best thing an adult can do in pretend play is follow. When a grown-up takes over, directing the plot, correcting the facts, steering the game towards something educational, children's engagement, language and persistence all drop. The play becomes the adult's, and the child becomes the audience.
Three working rules. First, follow, do not direct: respond to what the child sets up rather than proposing improvements. Second, accept the assigned role: if you are told you are the baby, you are the baby, and you should commit to it. Third, resist improving the plot. The fiftieth identical tea party feels pointless to you and is doing exactly what it should for the child. Your job is to be a warm, reliable supporting actor who occasionally adds a line of rich language ("Oh no, the kettle's gone cold, whatever shall we do?") and otherwise lets the director direct.
Pretend play needs less equipment than the toy industry suggests, but the equipment it does use matters. The useful distinction is between props that are open-ended and props that arrive with a script. A plain wooden animal can be anything: today a pet, tomorrow a patient, next week a passenger on a block-built bus. A franchise character cannot, or at least not easily, because the child already knows who it is, how it talks and what it does. The film has written the script, and the child's job shrinks to performing it.
Key citation
Trawick-Smith, J. et al., TIMPANI Toy Study, Eastern Connecticut State University · Year after year, simple, open-ended, non-branded toys produced the highest-quality preschool play: longer engagement, richer language and more creative use than elaborate or media-linked toys.
This is why the TIMPANI study's annual winners are so consistently plain: blocks, simple figures, basic vehicles. The less a toy specifies, the more the child must supply, and the supplying is the development. A small shelf of simple wooden toys →, some animals, some blocks, a few fabric squares, will sustain richer pretend play than a bedroom full of characters whose stories have already been told.
A child watching a story is receiving one. A child pretending is authoring one. Both involve imagination in the loose sense, but only one of them exercises it, because only one requires the child to generate the characters, hold the plot, solve the problems and supply the dialogue. This is not an argument that television is poison; it is an arithmetic point about who is doing the work. Every hour of a young child's day spent receiving stories is an hour not spent authoring them, and the preschool years are the once-only window in which authoring is learned. Screen time crowds out pretend play at exactly the age pretend play matters most, which is reason enough to guard the unstructured, slightly boring stretches of the day in which pretend play reliably appears.
Usually not. The timeline has wide normal variation: some children show barely any pretend play before two and a half and then produce elaborate scenarios within months, and some temperaments simply prefer building, sorting and physical play. It is worth mentioning to your health visitor if a child shows no pretend acts at all by around age two, especially alongside limited pointing, gesturing or shared attention, because absent pretend play is one of several things professionals look at together. On its own, a quieter imagination is a style, not a symptom.
It is not a competition, but if it were, pretend play would hold its own against anything. Role play exercises language, self-control and social understanding simultaneously, which no single-purpose toy does. The best educational toys → for this age are really pretend-play props in disguise: simple, open-ended objects that leave the thinking to the child. A toy that drills letters teaches letters; a hospital made of cushions teaches the child who will one day use the letters.
Because repetition is how mastery works. Each run of the identical tea party lets the child rehearse the sequence, the language and the feelings until they are fully owned, and only then does the script get varied. Adults experience the fiftieth repetition as tedium; the child experiences it as consolidation. The repetition will end on its own, precisely when it has done its job.
Boys pretend play differently more than they pretend play less. Group averages show girls gravitating earlier towards domestic and relationship scenarios while boys skew towards adventure, vehicles and physical role play, but the underlying capacity and the benefits are the same, and individual children vary far more than the averages do. A boy being a digger driver and a girl running a café are doing identical cognitive work. Offer both kinds of scenario to both, follow whatever they take up, and ignore the rest.
Britain's oldest toy maker. Every toy open-ended, wooden, and built for exactly the kind of play the research recommends.