Toddler tantrums
A Socratic walk-through of toddler tantrums — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why do tantrums peak in the third year and then fade without anyone teaching a child to stop?
Tantrums arrive on a schedule. They are rare in the first year, common through the second and third, and largely gone by five. What is strange is the exit. Nobody runs a course in tantrum cessation. Parents who are strict and parents who are indulgent both watch them fade at roughly the same age, in roughly the same order, and mostly cannot say what they did that worked.
A behaviour that appears without being taught and disappears without being untaught is not really being learned or unlearned. It is being permitted by something and then stopped by something. The useful question is what that something is — and why it has a window.
Reasoning it through
REASONING #Ask what a tantrum minimally requires. It needs a goal: a specific thing wanted, held in mind, and not obtained. A six-month-old is distressed by states — hunger, cold, fatigue — but does not want the blue cup rather than the red one. It also needs the goal to be blocked, and the child to have no better move available than the tantrum. And it needs the frustration of the block to exceed what the child can absorb.
So a tantrum sits at the meeting point of one capacity that has arrived and several that have not. Goal formation and will develop fast in the second year. What lags is the ability to say the goal — to name it, argue for it, propose a substitute, accept "after lunch" — and the ability to hold back a response and recover from disappointment, which rides on slow-maturing frontal control.
Now the shape of the curve falls out. Before the second year there is no specific goal to be thwarted, so nothing peaks. Through the second and third years, wanting outruns both saying and self-soothing by the widest margin, and the gap is the tantrum. After that, language and regulation catch up — not because anyone drilled them, but because they are developing anyway, on their own timetable and with ordinary conversation as scaffolding. The tantrum does not need to be extinguished. It runs out of room.
There is a detail here that argues strongly for this reading, and against the obvious rival. Careful analysis of recorded tantrums finds two overlapping processes rather than one. Anger — screaming, kicking, stamping — rises sharply and decays. Sadness — crying, whining, seeking comfort — runs at a lower level throughout and is what remains at the end. They are not stages the child passes through in sequence, and the anger component in particular is not paced like something aimed at an audience. This is Potegal and Davidson's finding, and I am recalling it rather than reading it, so take the detail lightly and the shape seriously.
Why does that matter? Because it makes an unusually specific prediction about what helps: questions and negotiation during the anger phase re-engage the goal and prolong it, whereas saying little and waiting lets the anger decay to the sadness underneath, at which point comfort works. Parents converge on this by trial and error, which is part of why the fading looks untaught.
The analogy
THE ANALOGY #Think of a river that has cut faster than its banks have been built. The water is the child's wanting — rising quickly in the second year. The banks are the ways of channelling it: words to ask, patience to wait, the capacity to be told no and stay in one piece. For a stretch, the water rises faster than the banks, and it goes over the top regularly. Nobody stops the flooding by scolding the river. The banks come up, and the same volume of water stays inside them.
Banks are built by deliberate work and the child's are not — they mature partly on their own schedule — and the water does not fall afterwards, which is exactly the point: a five-year-old wants things just as fiercely and simply has somewhere to put it.
Clarifying the model
THE MODEL #The main rival is that tantrums are strategic: the child has learned they work, and stops when they stop working. That account gets something right — reinforcement does shift how often tantrums occur, and consistency does help. But it predicts the wrong curve. If tantrums were a strategy, they should peak where the child is the better strategist, and instead they fade across exactly the years when children become far more capable of manipulating adults. They also happen with no audience, and over things no adult could grant. And the anger phase's rise-and-decay shape is not what goal-directed pressure looks like.
Two things worth separating from this, because they answer neighbouring questions and can be mistaken for it. Why the tantrums land on the parent who cares most is a question about safety and audience, not about timing. Why a child regresses when a sibling arrives is a question about a changed situation, not a changed capacity. Both can raise the rate of tantrums in a given child. Neither explains why the window opens at eighteen months and closes by five, which is what a capacity account is for.
Nor is "the terrible twos" an explanation. It names the window without saying what sets it, and so predicts nothing about which children stay in it longer.
That gives the test. If the gap between wanting and being able to say or absorb is what drives tantrums, then children whose expressive language is delayed — with wanting intact — should tantrum more, and for longer into childhood. They do. And children whose regulation matures atypically should not exit on schedule, which is also observed. If tantrums faded on the calendar regardless of a child's language and self-control, this account would be finished.
A picture of it
THE PICTURE #How to readThese values are an ordinal sketch of the argument, not measurements — the radar is repurposed here to show the shape of a profile rather than data. Compare the three outlines. At twelve months the whole figure is small, so there is little wanting to overflow. At forty-eight months it is large but roughly even. The tantrum window is the thirty-month outline, which is stretched far out on the "wants a thing" spoke and pulled in on the four that would let the child do something about it. The lopsidedness is the phenomenon.
What became clearer
WHAT CLEARED #Tantrums are not a behaviour a child acquires and then gives up. They are what a specific mismatch looks like from outside: a want that has become sharp and particular, in a person who cannot yet state it, defer it, or ride out being refused. That mismatch has a window because its two sides mature at different rates, and it closes when the slower side catches up. Which is also why nobody had to teach the ending — the ending is what happens when there is no longer a gap for the tantrum to occupy.
Where to go next
ONWARD #- Why some children exit the window years late, and what distinguishes ordinary tantrums from clinically significant ones.
- How the same capacity gap shows up in older children under fatigue or illness, when regulation is temporarily withdrawn.
Key terms
TERMS #| Term | What it means |
|---|---|
| Emotion regulation | the capacity to modulate the intensity and duration of an emotional response rather than acting it out whole. |
| Expressive language | the ability to produce speech, which lags comprehension in the second and third years. |
| Inhibitory control | the ability to withhold a prepared response, one of the executive functions that matures slowly through the preschool years. |
Every term the collection defines is gathered in the glossary.