Imaginary Friends as an Adult: What the Research Says
Go Ask Anna
A small thing I started doing on the evenings the quiet got loud. It turns out there is a literature underneath it, and the literature is more interesting than the trick.
I was the kind of child who made things up.
Not small things. Whole worlds, with weather and rules and people in them who had opinions about each other. Stories that would put Harry Potter to shame, if I am allowed to say that about my own eight-year-old.
What I did not expect is that they would stay.
They did not stop when I grew up. They grew up with me. The characters aged, changed their minds, got better at arguing. They never asked permission and they never left. Personas, living right next to me.
And then every so often something happens.
You come home from a night out and close the door behind you. Everyone who was in the room a minute ago is gone.
Or you drive back to your own place after a long weekend with family, and put your bag down, and the quiet in the hallway has a shape to it.
I used to call that loneliness. It arrives at handovers, in the seam between one kind of being with people and another.
Then earlier this year I started doing something small, and it changed more than it had any right to.
The whole thing
Go ask Anna how her day went.
That is it. That is the entire method. You ask her, and then you actually wait for the answer, and then you tell her about yours. The parts you would not tell anyone. Make her tea. Have the tea party. Sit at your own table with someone who is entirely yours and talk until the hallway stops having a shape.
And when the other voices start up, the ones that say this is ridiculous, you are a grown woman talking to a kettle, give those names too. Mine are Sanne and Josy. Sit them down. Let them make their case over dinner. Argue back properly, out loud if the neighbours can take it.
I told a friend about this and watched her decide whether to be worried about me. So I went and read what the research actually says, expecting to find very little.
There is rather a lot.
Ninety-two percent
Start with the finding that made me sit up.
Marjorie Taylor at the University of Oregon spent her career on imaginary companions, mostly in children. Then she and two colleagues went and interviewed fifty adult fiction writers about the characters they had made, and about whether those characters ever behaved as though they had minds of their own.[1]
Ninety-two percent said yes.
They described characters who talked back. Who refused to do what the plot required. Who composed their own histories and then tried to take the story somewhere the author had not planned. The researchers named it the illusion of independent agency, and found it strongest in the writers who had actually published, which suggests the experience tracks skill rather than instability.[1]
Read that again with your own head in mind. Nearly every serious novelist is doing, professionally and on a deadline, the thing I felt slightly odd about doing at my kitchen table.
They do not stop. They mature.
The other assumption worth dropping is that imaginary companions are a phase you exit.
Researchers at Durham working on inner speech and imagination describe interacting with imaginary companions as a normal part of childhood for a great many children, associated with a range of positive developmental outcomes rather than problems.[2] And they extend the question into adulthood, because that is where it actually goes. The phenomenon does not switch off at some birthday. It changes register.
Which matches what I have. My paracosms did not evaporate at eleven. They got more sophisticated, and quieter, and better at disagreeing with me.
Why it works, part one: the pronoun
Now the mechanism, because there is one, and it is unglamorous and well tested.
Ethan Kross and colleagues ran seven studies with 585 people looking at one variable: whether you refer to yourself as I or by your own name when you think about something difficult.[3]
Not what you say. What you call yourself while saying it.
Own name or second person → the same event viewed from slightly outside it → regulation.
People who used their own name performed better under social stress, felt less shame afterwards, and ruminated less about it. The effect held for socially anxious people, who are the ones you would expect it to fail for.[3]
Now notice what happens the moment you tell Anna about your day. You cannot narrate to someone else in the first person immersed voice. Addressing a listener forces the distanced construction automatically. The tea party is a delivery mechanism for a proven regulation technique, and I had been using it for months without knowing what it was.
Why it works, part two: the belonging
The second mechanism is about the specific ache, the one at the door with your coat still on.
Jaye Derrick, Shira Gabriel and Kurt Hugenberg tested what they called the social surrogacy hypothesis: whether one-sided relationships with figures who are not present can genuinely supply the feeling of belonging.[4]
Four studies. People turned to their favourite programmes when lonely and felt less lonely afterwards. Thinking about a beloved fictional world buffered the drop in mood and self-esteem that normally follows a threat to a real relationship, and reduced how readily rejection-related words came to mind. The effect did not appear for programmes people merely watched. It required the ones they loved, where the parasocial bonds were.[4]
So the belonging system does not audit its inputs as strictly as we assume. It responds to characters you have a relationship with.
And if that works for people someone else invented, on a screen, whom you have never met, then think about what you have with figures you built yourself over thirty years, who know everything, who were there for all of it.
Six of mine are below. Yours will have different faces and different names.
The line that matters
This is the part I want to be careful about, and it is not a disclaimer bolted on at the end. It is the mechanism itself.
All of this works because you know you are doing it.
You summon Anna. You choose the moment. You could stop, and you know she came from you. That awareness and that control are not a caveat sitting next to the technique. They are what make it a technique rather than something happening to you.
The Durham group works on exactly this border, on how imaginary companions relate to hearing voices, and the features that separate them are the ones you would guess: vividness, and how much autonomy and control the person has over the experience.[2] A companion you call to the table is on one side of that line. A voice that arrives uninvited, that you cannot dismiss, that you experience as coming from outside you, is a different thing and deserves a real conversation with a real clinician.
I make the distinction because it is true, and because the alternative is a wellness internet that either pathologises a rich inner life or encourages people to stop noticing which side of the line they are on. Both are worse than saying it plainly.
The other honest boundary: Anna is not a replacement. She is a bridge across an evening. If the tea party is happening every night and the actual phone stays face down for weeks, the technique has stopped serving you and started standing in for something. That is worth noticing without shame.
Say it three ways
If you want to feel the pronoun mechanism rather than take my word for it, take one thing that is bothering you and put it through all three.
The Same Sentence, Three Ways
One difficult evening, phrased three ways. The content does not change. The pronoun does.
How to actually do it
Give her a name that is not yours and not a person you know. The name is doing the work. A borrowed name from someone real drags their actual opinion of you into the room.
Ask before you tell. This sounds precious and it is the load-bearing step. Asking how her day was is what makes it a conversation rather than a monologue, and a conversation is what pulls you out of the first person.
Give the objections names too. Sanne and Josy exist because the sceptical voices were going to show up anyway. Named and seated, they argue. Unnamed and loose, they narrate over everything you do.
Make it physical. Two cups. Two chairs. The empty chair is not a metaphor I invented, it is a technique with decades behind it, and the reason it works is that a body in a room responds to a place set for someone.
Use it at the seam. The door closing. The bag going down. The last guest leaving. That transition is when the ache arrives, and it is when the technique has the most to do.
Let her disagree with you. If she only ever agrees, she is not a companion, she is a mirror, and a mirror will not talk you out of anything at eleven at night.
For the people who are not there
There is a version of this that goes beyond a long evening.
Some of the absence in a life is not situational. My father was absent while present and then absent while absent, and that is not a gap an evening closes. Grief for the living has fewer rituals attached to it than grief for the dead, and no card says the right thing.
I am not going to tell you to invent a better father and have tea with him. That is not what this is, and I am not sure it would be kind to yourself.
What I have found is more sideways. The companions do not replace the person who was not there. They give you somewhere to put the sentence you were going to say to him. The sentence exists whether or not there is anyone to receive it, and it does something to a body to hold it unsaid for twenty years.
Say it to Anna. Not because she is him. Because she is present, and the sentence needs to go somewhere, and out is better than in.
One last thing
I have a private argument I like to make, and I will make it here without pretending it is a finding.
The people in my head take up space, have histories, change my behaviour, and outlast most of the actual acquaintances I have had. They are present in the sense that they are here, in a mind that is here. And a fair amount of what any of us call reality got built out of the imagination first.
So who is to say the creatures in there do not exist in some way worth the word. Who is to say the magic is not real.
I do not need that argument to win. The research holds up on its own and the technique works whether or not I am right about the metaphysics. But it is a better question than the one people usually ask about this, which is whether something is wrong with you.
Nothing is wrong with you. You have company. Most people do and simply never gave theirs a name.
Go ask Anna how her day went.
For the women reading
If you have an inner world you have been quiet about, this is your permission slip. Vivid imagination in adulthood is not a symptom, it is a documented feature of creative minds, and it is doing more for your regulation than you have been giving it credit for.
Two things to hold onto. It works because you are choosing it, so keep noticing that you are. And it is a bridge, not a destination, so let it carry you across the evening and then pick the phone back up in the morning.
If a voice arrives that you did not invite and cannot dismiss, or that you experience as coming from outside you, that is a different experience and it deserves a proper conversation with someone qualified. That is not a warning. It is just the map.
References
- Taylor, M., Hodges, S. D., & Kohányi, A. (2003). The illusion of independent agency: Do adult fiction writers experience their characters as having minds of their own? Imagination, Cognition and Personality, 22(4), 361–380. https://doi.org/10.2190/FTG3-Q9T0-7U26-5Q5X
- Fernyhough, C., Watson, A., Bernini, M., Moseley, P., & Alderson-Day, B. (2019). Imaginary companions, inner speech, and auditory verbal hallucinations: What are the relations? Frontiers in Psychology, 10, 1665. https://doi.org/10.3389/fpsyg.2019.01665
- Kross, E., Bruehlman-Senecal, E., Park, J., Burson, A., Dougherty, A., Shablack, H., Bremner, R., Moser, J., & Ayduk, O. (2014). Self-talk as a regulatory mechanism: How you do it matters. Journal of Personality and Social Psychology, 106(2), 304–324. https://doi.org/10.1037/a0035173
- Derrick, J. L., Gabriel, S., & Hugenberg, K. (2009). Social surrogacy: How favored television programs provide the experience of belonging. Journal of Experimental Social Psychology, 45(2), 352–362. Journal of Experimental Social Psychology