Lead
"She's feeding her stuffed animals." "He's giving injections to the dolls and playing doctor." From around age two, most children begin to engage in pretend play. To an adult, it is charming. To a developmental psychologist, the question of why a child can treat something unreal as real has been a central research problem for nearly four decades.
That pretend play is deeply connected to the development of Theory of Mind — the ability to understand that other people hold beliefs and desires different from one's own — has been shown by multiple studies. But how they are connected, and specifically whether pretend play causes Theory of Mind to develop or whether both are expressions of the same underlying cognitive capacity, remains actively debated. This article lays out the theory and evidence honestly, without smoothing over what is still uncertain.
The Cognitive Foundation of Pretend Play
A two-year-old who holds a banana to their ear as a telephone, or pretends to drink from an empty cup, is doing something cognitively demanding. Alan Leslie, in a 1987 paper in Psychological Review, argued that pretend play requires the child to hold two representations simultaneously: the real one (a banana is food) and the pretend one (the banana is a telephone) — and to avoid confusing them [1]. Leslie called this capacity metarepresentation: the cognitive ability to hold a mental state that is decoupled from reality — representing a representation rather than the thing itself: the ability to represent a representation, to hold a mental state at one remove from the real.
The further claim in Leslie's argument, and the one that makes it foundational for subsequent research, is that this same metarepresentational capacity underlies Theory of Mind [1]. Understanding that another person holds a false belief — that they believe something about the world that is not true — also requires representing a mental state that is decoupled from reality. The developmental coupling between pretend play and false-belief understanding, in Leslie's account, follows directly from their shared cognitive mechanism.
The Developmental Co-occurrence
Leslie's theoretical prediction has received partial empirical support.
Jerome Belsky and Robert Most's 1981 longitudinal study observed 40 infants from 7.5 to 21 months and documented 12 distinct stages in the development of pretend play [2]. The sequence begins with simple object manipulation (mouthing, banging), moves to self-directed pretense (pretending to sleep), then to object-directed pretense (feeding a doll), and eventually to the combination of multiple pretend actions into sustained, complex scenarios [2]. The progression is orderly and gradual.
Marjorie Taylor and Stephanie Carlson's 1997 study brought theory and measurement together more directly [3]. They assessed 152 children aged three and four on measures of fantasy engagement — including whether the child had an imaginary companion and the richness of their pretend play — and on a battery of Theory of Mind tasks (false-belief tasks, appearance-reality distinctions). Among four-year-olds, fantasy engagement and Theory of Mind performance were significantly positively correlated, and the association held after controlling for language ability [3].
The study is important. It is also, at bottom, a correlation study — and correlation is not causation. Whether pretend play produces Theory of Mind, or whether a shared developmental maturation produces both, cannot be settled by a correlation alone.
A Critical Look at "Pretend Play Develops Cognition"
The claim that pretend play confers specific developmental advantages has circulated in parenting culture for decades. It is worth examining the evidence base more carefully.
Angeline Lillard and colleagues undertook exactly that exercise. Their 2013 review in Psychological Bulletin examined the empirical case for pretend play's benefits across cognitive, social, linguistic, and creative development [4]. The conclusions were measured. Evidence that the absence of pretend play impairs development is thin. Evidence that pretend play is the cause of observed developmental gains, rather than a co-occurring product of the same underlying growth, is generally insufficient — because most relevant research is observational rather than experimental [4]. Establishing causation would require experimental manipulation of pretend play, which faces obvious ethical and methodological constraints.
Lillard's position is not that pretend play is unimportant. It is that the evidence for pretend play being superior to other developmental pathways is currently weak [4]. That is a more modest and more honest claim.
This distinction matters because it changes the stakes. If pretend play were uniquely necessary, a child who didn't engage in it would be at risk of specific developmental deficits. The evidence does not support that reading. Children develop Theory of Mind through a variety of social and conversational experiences — being read to, overhearing adult conversations, negotiating with siblings — and pretend play appears to be one route among several, not a privileged one [4]. Parents who worry that their child's pretend play is "not rich enough" can probably set that worry down; the same underlying cognitive maturation is being exercised in other contexts as well.
How Parents Can Engage
If the causal story is more complicated than often assumed, what does that mean for how a parent relates to their child's pretend play?
Paul Harris, in The Work of the Imagination (2000), argued that pretend play functions as a rehearsal space for imagining [5]. Imagining another's perspective, imagining a past or future state, imagining a counterfactual — all of these are practiced repeatedly in pretend play scenarios. The child who assigns the doll a name, a history, and a set of needs is exercising exactly the imaginative capacity that, in other contexts, supports understanding other minds.
The implication for parents is not "structure pretend play as a teaching tool." It is closer to the opposite: join the play that children initiate rather than scripting it from the outside. Agreeing to be the patient in the doctor game. Going along with the doll's name and circumstances. This kind of participation extends and enriches the imaginative world the child is constructing, without appropriating it.
The reverse concern is also worth naming. When adults take over the narrative — correcting the plot, redirecting the characters, supplying the "right" pretend scenario — they can reduce the child's own degree of representational freedom. The beneficial adult role is to participate without over-directing.
A practical note on "my two-year-old doesn't do pretend play yet": as Belsky and Most's stage model shows, pretend play develops from 7.5 months onward, in stages [2]. By age two, a child may be at the self-directed pretense stage — pretending to eat or sleep — rather than at the complex, other-directed stage. Asking whether any pretense behaviors are present is more informative than asking whether complex pretend play is there. The developmental distance between a toddler who pretends to drink from an empty cup and a four-year-old running a multi-character drama is not a gap in engagement; it is simply the same trajectory observed at different points.
Summary
Pretend play and the development of metarepresentation and Theory of Mind are theoretically and empirically connected [1,3]. Pretend play develops in stages from infancy, reaching complex forms between ages two and four [2].
That said, the causal claim — that more pretend play directly produces greater Theory of Mind competence — is not well supported by current evidence [4]. What matters is whether children have rich opportunities to exercise imagination through play [5]. Pretend play is not something to engineer; it is something children initiate, and something adults can meaningfully accompany without steering.
References
- Leslie AM. Pretense and representation: the origins of "theory of mind." Psychol Rev. 1987;94(4):412–426. doi:10.1037/0033-295X.94.4.412.
- Belsky J, Most RK. From exploration to play: a cross-sectional study of infant free play behavior. Dev Psychol. 1981;17(5):630–639. doi:10.1037/0012-1649.17.5.630.
- Taylor M, Carlson SM. The relation between individual differences in fantasy and theory of mind. Child Dev. 1997;68(3):436–455. doi:10.1111/j.1467-8624.1997.tb01950.x. PMID: 9249959.
- Lillard AS, Lerner MD, Hopkins EJ, Dore RA, Smith ED, Palmquist CM. The impact of pretend play on children's development: a review of the evidence. Psychol Bull. 2013;139(1):1–34. doi:10.1037/a0029321. PMID: 22905949.
- Harris PL. The Work of the Imagination. Oxford: Blackwell; 2000.