It doesn’t take a child long, by such steps, to grasp the basic idea of a cello, the relationship of the bow, the string, and the left hand. But while he has been figuring this out, he has been ceaselessly active. One could say that he is having too much fun – a weak word, really – playing the cello to want to take time to figure it out. A scientist might say that, along with his useful data, the child has collected an enormous quantity of random, useless data. A trained scientist wants to cut all irrelevant data out of his experiment. He is asking nature a question, he wants to cut down the noise, the static, the random information, to a minimum, so he can hear the answer. But a child doesn’t work that way. He is used to getting his answers out of the noise. He has, after all, grown up in a strange world where everything is noise, where he can only understand and make sense of a tiny part of what he experiences. His way of attacking the cello problem is to produce the maximum amount of data possible, to do as many things as he can, to use his hands and the bow in as many ways as possible. Then, as he goes along, he begins to notice regularities and patterns. He begins to ask questions – that is, to make deliberate experiments. But it is vital to note that until he has a great deal of data, he has no idea what questions to ask, or what questions there are to be asked.

But the greatest difference between children and adults is that most of the children to whom I offer a turn on the cello accept it, while most adults, particularly if they have never played any other instrument, refuse it.

A child’s understanding of the world is uncertain and tentative. If we question him too much or too sharply, we are more likely to weaken that understanding than to strengthen it. His understanding will grow faster if we can make ourselves have faith in it and leave it alone.

He has to get the correct hunch many times, and test it, and see it proved right, before he can feel sure of it. Each time he is right, his hunch becomes stronger and surer; but it takes a long time – longer for some children than others – before it becomes what we think of as certain knowledge. […] Children’s first hunches about anything are extremely faint and tentative, the merest wisps of intuition that a certain thing may be so. Each time children test one of these faint hunches and have it confirmed by experience, the hunch becomes a bit stronger. What we might call a 5 percent hunch becomes a 10 percent, the 10 percent a 20 percent, and so, slowly, all the way to the point where they will say with conviction that they know that such-and-such is true – something, as I showed in How Children Fail, that even the “brightest” children in the “best” schools will rarely ever say.
[…]
Knowing this about children’s hunches makes me understand more clearly than every why, and how, our constant checking up on children’s learning so often prevents and destroys learning, and even in time most of the capacity to learn. […]
The first reason has to do with this matter of hunches. When we constantly ask children questions to find out whether they know something (or prove to ourselves that they don’t), we almost always cut short the slow process by which, testing their hunches against experience, they turn them into secure knowledge. Asking children questions about things they are only just beginning to learn is like sitting in a chair which has only just been glued. The structure collapses. Under pressure, children stop trying to confirm and strengthen their faint hunches. Instead, they just give them up. More times than I can remember, I have heard children being tested say of their hunches, “This must be wrong,” or “I know it’s wrong”.