Your four year old sets an extra place at the table. There is a rabbit called Marcel who lives in the hallway, hates carrots and has strong opinions about bedtime. You have started stepping around him. And somewhere in the back of your mind sits a question you may not have said out loud: is this normal, or is something going on?
It is normal. It is more normal than most parents realise, and decades of developmental research suggest it is a sign of capability rather than a sign of trouble. Here is what the evidence actually shows about imaginary companions between ages three and seven, what your child is quietly practising when they talk to someone who is not there, how to respond without either squashing it or moving into the fantasy yourself, and the narrow set of situations that are worth mentioning to a professional.
Is it normal for a child to have an imaginary friend?
Yes, and it is close to typical. In a study by psychologists at the University of Washington and the University of Oregon, 65 percent of children reported having had an imaginary companion at some point by the age of seven. That is roughly two children in three, which makes the experience more common than not.
The same research also broke a long-standing stereotype. Imaginary companions are not the preserve of only children or firstborns, and their appearance is not in itself a marker of loneliness or distress. You can read the original announcement of the findings and ScienceDaily's summary.
What age do imaginary friends appear, and when do they fade?
Most imaginary companions appear between ages three and five, and many persist well into the school years. The same study found that having an imaginary companion was at least as common among school-age children as among preschoolers: 31 percent of school-age children were playing with one at the time of the interview, compared with 28 percent of preschoolers.
That detail matters, because plenty of parents assume the companion should have disappeared by the time school starts and worry when it has not. It also means the fade-out is rarely dramatic. Companions usually shrink in importance over months rather than vanishing on a particular day, and it is common for a child to simply stop mentioning one they were inseparable from a year earlier.
What is my child practising when they talk to someone who is not there?
They are rehearsing the hardest social skill there is: holding another mind in their head. An imaginary companion has to be given a personality, preferences, a mood, and reactions that stay consistent from one day to the next. That is theory of mind under construction, and it does not come free.
A classic study by Marjorie Taylor and Stephanie Carlson examined 152 three and four year olds and found that pretend play with imagined characters was strongly correlated with theory of mind performance, even after controlling for age and verbal intelligence. The study is indexed on PubMed, and Taylor's later review in Learning & Behavior sets out the psychological significance of imaginary companion play in more depth.
The knock-on effects show up in language too. Children who narrate for a companion produce longer, more complex sentences than they would in ordinary conversation, because they are voicing two speakers instead of one. It is the same mechanism at work when children think out loud, which we covered separately in our piece on what self-talk means at ages three to seven.
Does an imaginary friend mean my child is lonely?
Usually not. The research consistently finds that children with imaginary companions are, if anything, more socially able than their peers, not less. A 2021 review in Imagination, Cognition and Personality gathered the evidence and reported associations with creativity, richer vocabulary and stronger social understanding rather than social withdrawal.
There is also a control element that has nothing to do with loneliness. A relationship with an imaginary companion is one the child fully directs, which makes it a safe place to test out arguments, apologies, bossiness and generosity before trying them on a real four year old who might not cooperate. The review is available from SAGE, and a piece in The Conversation summarises the developmental case for parents. Research has even found social benefits among autistic children who create imaginary companions, which is the opposite of the isolation story people expect.
Should I play along, and how far?
Play along at your child's invitation, and stop short of running the fantasy yourself. The companion belongs to the child, and the developmental work happens when the child is the author. A parent who takes over the plot is, gently, taking the exercise away.
A quick guide to responding
- Do set the extra place if asked. It costs nothing and it signals that you take your child's inner world seriously.
- Do ask open questions: what does Marcel think about that, what is he scared of, what does he like to eat.
- Do let your child correct you when you get the companion wrong. That correction is the skill working.
- Do not invent major plot points or new companions on your child's behalf.
- Do not let the companion carry blame. "Marcel drew on the wall" is a moment for warm, non-negotiable clarity about who is responsible.
- Do not tell your child the companion is not real. There is no benefit, and children this age generally know already.
That last point surprises people. Studies of imaginary companion play find that children usually maintain a clear line between pretending and believing. They know Marcel is pretend. They would simply prefer you did not say so out loud.
When is an imaginary friend worth mentioning to a professional?
In the overwhelming majority of cases there is nothing to raise with anyone. The situations worth a conversation with your paediatrician are specific, and they are about distress or displacement rather than about the companion existing at all.
Flag it if the companion is a consistent source of fear rather than comfort, if your child seems unable to stop interacting with it in situations where they want to join other children, if it is systematically replacing real friendships your child otherwise wants, or if it arrives alongside a marked change in mood, sleep or behaviour after a difficult event. A companion that persists into the pre-teen years and is treated as genuinely real is also worth a mention. None of these is an emergency; they are simply reasons to ask.
Worth separating from all of this: a child who invents a companion after a house move, a new sibling or a bereavement is not showing a symptom. They are using a normal tool to process something hard. Our guide to helping a three to six year old handle big emotions covers what else helps in those weeks.
Where does a talking toy fit into this?
Somewhere between a companion the child invents alone and a friend who talks back. Plenty of children hand their imaginary companion a body: a soft toy becomes the vehicle, and the child does both voices. That is pretend play in its purest form, and researchers have described the same dynamic playing out with invisible companions in the classroom. It is also the reason a plush toy tends to survive every clear-out.
A plush that actually answers changes the exercise rather than replacing it. With Ted, an interactive plush companion, the child is not voicing both sides; they are having a real back-and-forth, in any of more than sixty languages, with the topics and tone set by the parent in the app. That is closer to conversation practice than to imaginative play, and the two are worth keeping distinct. The best case is a child who does both: invents Marcel in the hallway, and separately has someone to ask why the moon follows the car. There is no camera and no screen involved, and what is and is not collected is set out on our FAQ page. And if you want the wider case for unstructured make-believe, we made it in why pretend play matters so much.
Frequently asked questions
At what age do children get imaginary friends?
Most imaginary companions appear between ages three and five, though some arrive as early as around thirty months. They commonly continue into the school years, and research found that 31 percent of school-age children were actively playing with one, slightly more than the 28 percent of preschoolers.
Is having an imaginary friend a sign of a problem?
Almost never. Around 65 percent of children have had one by age seven, and studies link imaginary companions to stronger theory of mind, richer language and better social understanding. Concern is warranted only if the companion causes fear, replaces friendships the child wants, or accompanies a marked change in mood or behaviour.
Should I tell my child their imaginary friend is not real?
No. There is no developmental benefit, and children of this age generally already know the difference between pretending and believing. Correcting them tends to shut down the conversation rather than clarify anything.
Do imaginary friends mean my child is lonely?
Research does not support that reading. Children with imaginary companions are typically at least as socially skilled as peers without them, and the pattern is not limited to only children or firstborns. The companion functions as low-risk practice, not as a substitute for company.
When do imaginary friends usually go away?
There is rarely a clear end date. Most companions fade gradually across the primary school years as other relationships and interests take up more space. A companion that is still treated as genuinely real into the pre-teen years is worth mentioning to your paediatrician.