Why certain people reach places in us others don't. Why the same fight keeps finding you in different relationships, wearing a different face. And what changes when you recognize who's sometimes in the room with you.
Would you hand a five-year-old the keys and let him drive the bus? Would you let a five-year-old pick your career, sign your mortgage, choose who you marry? Of course not. Nobody would. And yet, more often than any of us would like to admit, that's exactly who's behind the wheel — which is the entire reason this piece exists.
Someone I respect deeply once said something to me that I couldn't shake: when two adults meet for dating and romance, it isn't really the two adults meeting. It's their two five-year-olds — each one looking, underneath all the adult language of attraction and compatibility, for healing. It instantly had an air of truth to it, and the more I thought about it, the more true it became. It turns out the research supports more of the idea underneath it than I expected.
So here is the invitation before the research: think about who you're drawn to, and who you're not. Think about the fight that keeps recurring, in different relationships, with different people, wearing a different face each time — the one where you can predict your own reaction before it happens. Think about what you actually needed in the moment you went quiet, or raised your voice, or reached for your phone to check if they'd written back. Somewhere underneath the adult account of what happened, there is very often a smaller, older account running the show. This piece is for reading with your own relationships open on the table, not at arm's length — because the point isn't to learn a theory. It's to recognize the five-year-old the next time they show up, know what they're actually asking for, and understand the growth work that recognition opens.
None of this is a metaphor dressed up to make a hard moment sound more forgivable. It's a real, identifiable, researchable mechanism: adult relationships can activate emotional states, needs, and reactions laid down in early childhood, and doing so predictably enough that it's been studied from four separate directions for decades. What follows is what that research actually says — and, more usefully, what to do with it once you see your own pattern in it.
The foundation is attachment theory, built by the psychiatrist John Bowlby in the 1950s and 60s and tested empirically by Mary Ainsworth in her "Strange Situation" studies of infants and caregivers.1,2 Bowlby's proposal was that a child forms an internal working model — an unconscious template for what closeness is, whether it is safe to need someone, and what to do when that person becomes unavailable. Ainsworth's lab work sorted infants into patterns still in use today: secure, anxious, avoidant, depending on how a child responded when a caregiver left the room and came back.
The leap to adulthood came two decades later, and it is the real hinge of this whole inquiry. In 1987, Cindy Hazan and Phillip Shaver published a study treating romantic love as an attachment process in its own right, and found the same three patterns showing up in how adults described their marriages and relationships.3 That paper is the first time "the child you were" and "the partner you are" were tested as one continuous system rather than compared as a metaphor.
The third figure matters more than it looks. An effect size in that range means early attachment is a real, replicated influence — not a footnote — but one input among several, not a sentence handed down at five and served for life. The research does not say your childhood decided your marriage. It says your childhood wrote a first draft, and the draft is still open.
What's striking is that four schools of thought, working decades apart from different starting premises, landed on roughly the same picture: under the right kind of threat — a raised voice, a long silence, the feeling of not being seen — an adult body can drop into a younger pattern of response before the thinking mind has time to object.
Attachment theory formalized this as what Mario Mikulincer and Phillip Shaver call the attachment behavioral system4 — a switch that flips automatically once a bond feels threatened, regardless of the person's age or competence elsewhere in life. Their research shows the switch is measurable: reaction time, word choice, even brain activation shift in predictable ways the moment it trips.
Schema therapy, developed by Jeffrey Young through the 1990s and 2000s specifically for patterns that outlast ordinary talk therapy, names the mechanism almost exactly the way the folk phrase does. One of its core concepts is, literally, the Vulnerable Child mode5 — a state Young describes an adult flipping into, complete with a younger emotional register, when an old wound gets touched. It is one of the better-tested frameworks here, with randomized controlled trials behind it and evidence that it outperforms standard treatment for some longstanding personality-level patterns.
Imago Relationship Therapy, popularized by Harville Hendrix in the late 1980s, is built almost entirely on this premise: that we unconsciously seek partners who recreate the emotional terrain of our early caregivers, and that conflict is often the unfinished business of that terrain replaying itself.6 It is the most widely known version of the idea and the least rigorously tested — its evidence is mostly clinical and observational rather than experimental, which doesn't make it wrong, only a lens to hold a little more loosely than the ones above it.
Polyvagal theory, and the "window of tolerance" work built on it, offers one biological account for what may be happening underneath these psychological accounts.7,8 Stephen Porges's proposal is that perceived relational threat can shift the nervous system out of a calm, socially engaged state and into a more primitive one — fight, flight, or shutdown — the same limited repertoire available to a much younger person with far fewer tools. In this framing, "acting like a five-year-old" is not a figure of speech. It is a description of which part of the nervous system currently has the wheel.
Note — the two-brain problem
Explicit knowledge ("this reaction is out of proportion") and the implicit, limbic sense of threat are processed in largely separate systems, and the second does not take orders from the first.9 This is why intelligent, self-aware adults keep having the same fight: the knowing and the reacting live in different places, and only one of them is listening in the moment.
Underneath every framework in the last section sits one more piece of research that belongs in this inquiry on its own: the study of attunement and presence, specifically — what happens in a person, child or adult, when it's offered, and what happens when it isn't. This is the part of the literature that explains why "just be there for me" is not a small ask. It's the ask.
In 1975, the developmental psychologist Edward Tronick ran what became one of the most replicated experiments in the field.11 A mother plays normally with her infant, then, on cue, goes still — flat expression, no response, present in the room but absent in every way that matters. Within three minutes the infant sobers, grows wary, and mounts a series of increasingly urgent bids to bring the mother back: reaching, smiling, vocalizing, escalating. When none of it works, the infant gives up — withdraws, turns the body and face away, and settles into what Tronick's team described as a hopeless expression. That collapse, inside three minutes, in a child too young to speak, is the clearest picture the field has of what happens when presence is withdrawn: first protest, then despair, then a shutdown that looks very much like the one an adult reaches for decades later in a different kind of silence.
What's easy to miss in that finding is the twin discovery underneath it. Tronick's later work showed that attuned, perfectly synchronized interaction — mother and infant matching each other's state at the same moment — happens far less often than people assume in healthy relationships: by his own measurements, somewhere between 10 and 21 percent of the time even in normal, well-functioning pairs.11 Mismatch is not the danger. Mismatch is the baseline. What matters, and what predicts good outcomes, is repair — the return to attunement after a rupture, over and over, reliably enough that the child learns disconnection is temporary and survivable rather than permanent and catastrophic. This is the finding that should recalibrate the whole subject: nobody is asking you to be perfectly attuned to another person's inner five-year-old. The research asks something much more achievable — that you come back.
Note — rupture is not the failure
Every relationship, parent-child or adult-adult, runs on a constant cycle of attunement, mismatch, and repair. The child who grows up secure is not the one whose caregiver never went still. It's the one whose caregiver reliably came back.11 The same appears to hold for adults, which is the argument for staying in the room rather than aiming for flawless presence.
Adult research asks close to the identical question and gets a strikingly similar answer. John Gottman's Love Lab followed newlywed couples for six years, coding how each partner responded to the other's small daily "bids" for connection — a comment, a glance, a request to look at something out the window.12 Couples who were still together at the six-year mark had turned toward those bids 86 percent of the time. Couples who had divorced had turned toward them only 33 percent of the time. The mechanism is the same one Tronick found in the crib, translated into adult vocabulary: what predicts the outcome isn't grand gesture or the absence of conflict, it's the accumulated answer to a small, constant, largely wordless question.
Sue Johnson, the developer of Emotionally Focused Therapy, names that question directly. Whether the relationship is a marriage or a therapy room, she argues the query underneath almost every conflict is the same one an infant asks a caregiver: are you there for me?10 Her clinical model builds a bond around three qualities — accessible, responsive, engaged, sometimes shorthanded as A.R.E. — and the research behind it treats those three as close to the adult translation of what an attuned caregiver offers a child: available, attentive, and not performing availability from somewhere else.
A separate line of adult research approaches presence from the practice side rather than the outcome side. A 2004 randomized controlled trial introduced couples to a structured mindfulness practice built specifically for relationships — shared meditation, mindful touch, exercises in simply being present with a partner without managing or fixing anything about them.13 The couples who practiced more showed measurable gains in relationship satisfaction, closeness, and acceptance of each other, with the benefit tracking day to day: more presence on a given day predicted a better relationship day, not just a better month. Presence, in other words, is not only a childhood outcome. It is a trainable adult skill with its own dose-response curve.
| Condition | What the research shows | What it tends to produce |
|---|---|---|
| Childhood, received | Rupture happens constantly; the caregiver reliably repairs it | Secure attachment; the felt sense that need is survivable and connection returns |
| Childhood, withheld | Still-face and chronic unresponsiveness; protest, then despair, then withdrawal | The template this whole inquiry has been describing — the five-year-old who learned to expect absence |
| Adulthood, received | Turning toward bids 86% of the time; accessible, responsive, engaged partners | Relationship stability, trust, and — per the mindfulness data — day-to-day satisfaction |
| Adulthood, withheld | Turning toward bids only 33% of the time; chronic unavailability, distraction, defensiveness | The same protest-then-despair arc, in adult form — pursuit, then shutdown, then quiet resignation |
Read side by side, the four rows describe one continuous story rather than two separate ones. What a nervous system needed at five is close to what it still needs at forty-five: not flawless attention, but reliable return. The adult who learned early that bids go unanswered will read an ordinary Tuesday-night distraction as confirmation of something much older. The adult who learned early that someone comes back will have more slack to extend when a partner is, for one evening, somewhere else.
| Framework | Core claim | Evidence base | What it adds |
|---|---|---|---|
| Attachment theory | Early bonds build a template that adult relationships activate | Strong — decades of longitudinal and experimental work | The mechanism itself, and language to name your own pattern |
| Schema therapy | A Vulnerable Child mode gets triggered by matching cues | Strong — randomized controlled trials, clinical outcomes | A structured way to catch the mode while it's happening |
| Polyvagal / nervous system | Threat shifts the body into a younger, more reactive state | Moderate — well-established physiology, more contested at the level of specific claims | Why it feels physical, not only emotional |
| Imago Relationship Therapy | Partners are chosen to replay early caregiver dynamics | Weaker — mostly clinical and observational | The most usable, human-scale story of the four |
It is worth separating what's well established from what's simply a satisfying story, because the subject tends to get flattened into one or the other.
Early attachment experience measurably shapes adult relational behavior. The attachment system activates under perceived threat, producing predictable shifts in behavior, language, and physiology — replicated across hundreds of studies and multiple countries.
Naming and interrupting the triggered state — whichever framework's language you reach for — is associated with better relationship outcomes in controlled trials, particularly in schema therapy and in Sue Johnson's Emotionally Focused Therapy, a related model with strong outcome data of its own.10
"Inner child" work specifically — reparenting exercises, letters to your younger self — is widely used and widely liked by participants, but has comparatively little controlled outcome research behind it. Useful as a practice. Not yet a well-validated construct.
The honest summary: the mechanism is real and well documented. "Two five-year-olds in the room" is a simplification of that mechanism — a teaching image, not a citation.
For years the instinct is to hide it — the vulnerability, the woundedness in your own five-year-old, kept out of sight the way you'd keep any old injury covered so no one could press on it. The more useful move, once you can make it, is the opposite: to look at that five-year-old with a different kind of attention. Not as a flaw to manage or a weakness to outgrow, but as a child who deserves the compassion and empathy any real child deserves — because in the moment he's running the show, he is one.
The same holds for whoever you're in relationship with. When a reaction seems much larger than the moment — the withdrawal, the sharpness, the need that arrives dressed as a demand — it can be useful to wonder whether the adult argument has touched something older. The person in front of you is still an adult, still accountable for what they do next; but sometimes the intensity belongs partly to an earlier experience, reaching for something that was missing the first time and asking, however badly, in the only language a five-year-old has. And a five-year-old, anyone's, deserves the same tenderness you'd offer an actual child standing in front of you: not indulgence, not a blank check, but the recognition that a scared child is asking to be seen, not managed.
This is the real shift the research points toward, underneath all four frameworks: you are not there to repair your partner's childhood, and you cannot. Nobody can go back and re-parent a five-year-old from the outside; the work is theirs to integrate, the way any of us has to integrate our own. What you can do — what changes the whole texture of a relationship once you start doing it — is stop treating the reaction as an indictment of the adult and start treating it as a child asking to be witnessed. Not fixed. Witnessed. Stayed with. Given the room to be scared without being told the fear is a problem to solve.
That distinction matters because witnessing is not the same as absorbing. Sitting with someone's five-year-old is not a standing invitation to be shouted at, managed, or made responsible for a wound you didn't create — compassion for the child does not require you to let the child run the relationship unchecked. It means meeting the fear with steadiness instead of matching it, and it is a very different practice than either fixing the person or tolerating whatever the fear produces.
I know this one from the inside, not the outside. For years, my instinct in a highly charged relationship — married or dating — was to shut down: go quiet, turn away, put distance between myself and whatever emotional state was standing in front of me, because it was more than I could hold. It took a long time to understand how little of that response belonged to the moment I was actually in. It was the exact coping strategy I built at five, when withdrawal was the only real protection a small boy had. As a grown, physically strong, reasonably intelligent man, that same strategy stopped protecting anyone and started doing harm — because what the women in front of me were actually asking for, underneath the charge, was for their own five-year-old to be witnessed and heard. I didn't know how to do that. My system kept reading the intensity as a threat, the way it correctly had been at five. It no longer was one.
There is a distinction worth holding onto here, because compassion for a five-year-old is not the same as tolerance for whatever that five-year-old does. Presence alone isn't always enough, and there are moments when the dysregulation in front of you crosses into something flatly abusive rather than simply activated. That is the moment to step outside the old reflex — neither shutting down nor absorbing it — and name plainly what you're seeing: that the other person is acting out their childhood in this moment, and that naming it is an invitation to their accountability, not an accusation. Staying present enough to say that, instead of retreating into silence, is the harder move. It's also the one that doesn't abandon them while still refusing to disappear into the pattern with them.
That distinction deserves more than a sentence, because getting it wrong in either direction costs something. Read every hard moment through the five-year-old lens and you end up excusing what shouldn't be excused. Read none of them that way and you lose the compassion that makes a relationship survivable. The two questions worth asking, in the moment, are not "is this their childhood" — it almost always is, at the root — but "is this still reaching toward me, or has it turned into something aimed at hurting me," and "when I name what I'm seeing, does it soften, or does it escalate."
| Signal | Signals of dysregulation | Signals of a harmful pattern |
|---|---|---|
| Pattern | Tied to a specific trigger; not every conflict looks like this | Chronic, recurring regardless of the trigger — a standing tactic, not a reaction |
| Direction | Reaching for connection even while dysregulated — fear, not control | Aimed at control, isolation, humiliation, or punishment |
| When named | Softens, gets sheepish, becomes more reachable | Escalates, denies outright, or turns the accusation back on you |
| Afterward | Remorse, some account of what happened, an attempt at repair | No account taken; blame lands elsewhere; the same thing recurs unchanged |
| What it calls for | Steady presence, naming, patience while the nervous system settles | A boundary, distance, and often professional support — not more presence |
These two columns are not mutually exclusive, and treating the table as a clean diagnostic would be a mistake. A person doing real harm can still show remorse, soften when confronted, or be genuinely frightened underneath it. A person who is only dysregulated, not abusive, can still escalate in the moment or reach for blame before they've settled. Childhood activation can explain harmful behavior without making it not-harmful — which is the whole point of the paragraph that follows.
The five-year-old lens still explains the origin of the second column, not only the first — nearly all abusive behavior is also a wound operating without an adult's brakes on it. But an origin is not a license, and understanding where a pattern came from has never been the same question as whether it's acceptable now. The actual practice is holding both truths at once: this very likely began somewhere, in a person too small to do anything but survive it, and the adult standing in front of you today is the one responsible for what they do with it. Compassion for the child underneath does not obligate you to keep absorbing what the adult chooses to do on top of it.
The five-year-old is real, and the research backs the mechanism more solidly than the phrase that carries it around. Yours deserves the same compassion you'd give any frightened child. So does theirs. You cannot fix it. You can witness it. You can sit with it. And, when it is safe and right to do so, you can stay.