A note on scale: what follows begins with one evening, one team and one task. We are not presenting it as research. Two published studies appear later in the piece; the piece itself begins in the field, not in the literature.
Early May, evening. The tents had just gone up at an overnight camp on Taiwan’s north coast. On a patch of grass lay a few coloured footprint mats and a length of rope. The task brief was a single sentence:
“Everyone has to get from this end to that end, and nobody’s feet may touch the ground.”
No right answer. No steps to follow. No demonstration. What remained was for that team of children to work it out themselves.
Over the next dozen or so minutes, roughly this happened. One child grabbed the lead at once and started assigning who would stand where. Another stood quietly to one side and said nothing. One child seemed to have already spotted the problem, but did not speak.
The first attempt: the line jammed in the middle and someone’s foot touched the grass. The second: the children at the front moved too fast and the ones behind were left without a mat. The third: they stopped and talked — the first time they had actually stopped to talk rather than rushing straight into another go.
Then the child who had said nothing spoke, not loudly:
“We need to decide who goes first and who goes last, or we’ll get stuck.”
After that, the team got across.
What the leader did during those minutes
The leader did not step in with the answer. That was deliberate: the next attempt was still theirs to make.
What had been designed was not the leader’s silence but the frame around it. The task had a clear condition for success (the whole team across, no feet on the ground), clear resources (a few mats, one rope) and a clear boundary (the leader stood beside them but did not intervene in the method). Inside that frame, the cost of failing was clear and limited: what they lost was one attempt, not their safety.
If the leader had supplied that answer after the first failure — “You could try deciding on an order first” — the child would not have had the same opportunity to arrive at it and voice it himself. And voicing it mattered far more than solving the task.
This is something we keep seeing in the field. What children need is not to avoid failure, but a structure in which failure still leaves room for another attempt. Arriving at that sentence after three failures is not the same experience as receiving the hint from an adult at the start.
A classroom deals with the systematic side of knowledge: teach the method clearly first, then let children practise. For much of learning that is efficient and necessary. The field deals with something else: what you do when nobody has taught you first. The two complement each other; neither replaces the other.
One well-held failure is not enough
At a camp we can design a single failure so that it can be tried again. Across those three failures, nobody was told off, compared with anyone, or hurried into a correction.
But the places a child actually meets setbacks in a year are mostly not on a mountain. For the other three hundred-odd days, failure happens at home, at the after-school class, on the evening a bad test paper lies on the dinner table.
The structure in the field and the everyday reactions of adults together shape how a child comes to understand failure. Getting only one of them right is not enough.
Children cannot read what you believe
Here is a finding that many parents will find a little surprising.
Haimovitz and Dweck, in a series of studies published in Psychological Science in 2016 — including 73 parent–child pairs and experimental conditions with more than a hundred parents — set out to test an intuitive assumption: that parents who believe ability can grow raise children who believe ability can grow.
They did not find it. Parents’ own beliefs about intelligence were unrelated to their children’s.
What did predict the children’s beliefs was something else: the parents’ beliefs about failure — whether you see failure as something that diminishes a person, or as something that helps them grow. That significantly predicted the child’s beliefs about intelligence.
The mechanism is not complicated. Parents who see failure as harmful tend, in the moment their child fails and almost without thinking, to focus on whether the child lacks ability: reassuring them (“never mind, you were never good at this”), moving the hard problem out of the way, changing the subject. And the child can see that reaction.
What you believe in your head, a child has no way to read. Your face on the evening of the bad test, and whether your first question is “How did this happen?” or “Where did you get stuck?” — those a child reads, and reads accurately.
It starts at fifteen months
If that finding feels heavy, this next one suggests there is something to be done.
Leonard, Lee and Schulz, in a study published in Science in 2017, worked with 262 infants aged fifteen months, in three groups. The first group watched an adult pick up a toy, try repeatedly, switch between several methods, and only then succeed. The second watched an adult succeed effortlessly on the first go. The third was a control group.
Each infant was then given a new and difficult task.
The infants who had watched the adult “try many times” made significantly more attempts than either of the other groups. The authors concluded that adult models causally affected infants’ persistence, and that the infants generalised the value of persistence to a task they had never seen before.
Fifteen months old. Not yet talking.
Even at fifteen months, watching an adult persist can affect how long a child keeps trying at something new. Watching you take three tries to pitch the tent is not wasted on a child.
Four things that happened on that grass
Back to the grass. Looking back at those dozen or so minutes, four things are visible: seeing what happened, naming where it got stuck, having another try, and leaving the answer with the child. We use these words to describe what we saw, not as a sequence every situation has to follow. We have seen the same four things in many different settings, and they can happen on a mountainside just as easily as at a dinner table.
See what happened. Failure has to be seen, not skipped over. Adults are sometimes too quick with “never mind, go again”, and the failure is already behind the child before they have understood what just happened. Before trying again, “where did that get stuck” needs a chance to be seen. That evening, after the first failure, the leader paused and let “that got stuck” hang in the air for a few seconds.
Name where it got stuck. Feelings can be received, but reflection has to go one step further, towards what specifically happened. “What did you make of that?” usually gets “It was hard.” The more useful question is “Which step did it get stuck at?” The team’s third discussion worked because they were finally talking about something specific: the order.
Try again. The threshold for another go has to be low. If a second attempt means lining up again, hearing the rules again and waiting for an adult’s permission again, children tend not to try. The mats and the rope were exactly where they had been; whenever the team wanted another go, they could take it. The cost of trying again decides whether a child will try a third time.
Leave the answer with the child. “We need to decide who goes first” was not the leader’s sentence. And when the group gathered afterwards, it did not turn into “Good thing the teacher reminded you.” The child who walked off that grass took with him a solution he had thought of himself.
We put it that way because the solution was his to begin with. The adults simply did not take it from him.
We cannot say that one task changed him. But over the following two days he was noticeably more willing to take part than on the first: pitching tents, lighting the fire, the next day’s route — he kept moving towards all of it. At least for that weekend, he seemed to have confirmed one thing: an idea he put forward could be taken seriously by a whole group.
The same structure grows into a different shape somewhere else. In April, on a walk up Tianshang Shan — a hill on the edge of the Taipei basin — with a group of ten- and eleven-year-olds, one child said on the way down: “I thought I couldn’t go on, but I still got there.” He said it himself. An adult can say “You can do it” a hundred times and not buy that sentence once.
An honest limitation
Neither of the studies above was conducted in Taiwan. We are not aware of comparable quantitative research from Taiwan on how parents respond when their children fail — and that gap is itself worth pointing out. What we can offer is fifteen years of field observation with several thousand children. That is a kind of evidence, but it is not that kind.
The four things described above are drawn from practice we have seen recur in the field; they are not an experimental result. We have not run a control group, and we have not yet systematically tested how it differs across ages.
Nor will we claim that one hill walk or one overnight camp changes how a child faces failure. That is not something a few days can do. What the child who spoke after three failures did experience that day was one complete pass through “I messed up, and then I worked something out myself.” It takes many such experiences before that becomes part of who a person is.
Finally
Children cannot read what you believe about failure. They read your face in the moment they fail.
And in the field, there is one more thing we can do: leave failure a next time.
See it, name it, let them try again — and leave the answer with the child.