At some point between two and four, most children start asking why, and then do not stop. Adults experience this as a loop with no exit. Answer the question and the answer becomes the next question. Answer that and so does the next.
The folk theory is that this is a game, or a bid for attention, or a bedtime stalling tactic that happens to be extremely effective. Some of it certainly is. But the overall pattern, looked at properly, does not fit the attention theory well, and fits an information seeking theory rather better.
What the pattern of follow up shows
The useful evidence here does not come from asking children what they mean. It comes from watching what they do with the answer.
Studies of children's conversational exchanges have looked at what happens after an adult responds. When the response contains an actual explanation, children tend to accept it and move on, often to a related but new question. When the response is a non answer, a deflection, a repetition of the fact, or a because, children tend to re ask, rephrase or press harder.
That is not what attention seeking looks like. If attention were the goal, any response would do, and the child would be equally satisfied by a warm non answer as by a cold informative one. They are not.
What counts as an explanation to a three year old
Adults often assume that a real explanation must be complete and technically correct, which makes the task feel impossible and encourages the deflection that provokes the next why.
It does not have to be complete. Children appear to be satisfied by explanations that supply a mechanism or a purpose at roughly the level of the question. Why is the grass wet? Because water fell on it in the night. That is a mechanism. It is not meteorology, and it is enough.
What tends not to satisfy is a restatement. Why is the grass wet? Because it is. Why do we go to bed? Because it is bedtime. These give no new structure, and children can tell.
Children are building theories, and testing them badly on purpose
A productive way to think about early childhood, and one with a good deal of support, is that children behave like people building and revising informal theories of how things work. They form expectations, notice violations, and go looking for the missing piece.
Why questions cluster in exactly the places a theory builder would expect: around causes, around purposes, around the behaviour of people, and around anomalies. Children ask why the moon follows the car far more often than they ask why the car has four wheels, because the first violates something and the second does not.
The endlessness is a feature of the method. Every explanation opens the next level down. Anyone who has tried to explain something properly to an adult knows that a good question generates more questions; children simply have not yet learned to stop asking out of politeness.
The purpose bias, and why it is not a mistake
Young children show a strong tendency to explain things in terms of what they are for. Rocks are pointy so animals do not sit on them. Rain is for the plants. Researchers call this a teleological bias, and it is robust across cultures.
It is easy to treat this as an error to be corrected. It is more useful to notice that a purpose based explanation is often the correct kind of explanation for the things a small child deals with most: objects made by people, and the behaviour of people. Chairs really are for sitting on. The bias is a good default applied slightly too broadly, which is what most useful heuristics are.
It fades, unevenly, with knowledge. It does not fade entirely in anybody. Adults reach for purpose explanations under time pressure too.
Answering well, without turning it into a lesson
This magazine does not hand out instructions, and there is no evidence base that would justify a technique here. But there are some observations worth having.
A real answer at the level of the question ends more loops than a deflection does. This is the practical upshot of the follow up research, and it has the pleasant property of being less work.
Not knowing is an answer. Saying that you do not know, and wondering aloud how you might find out, models the thing the question is reaching for. It is not a failure to say it.
Turning the question back is not always generous. What do you think is a useful move sometimes and an evasion at others, and children who are genuinely after information notice when it is the second.
Not every why deserves the full treatment. Some are stalling. Some arrive at the seventh minute of a bedtime that was meant to take four. Recognising the difference is a domestic skill and not a research finding.
What happens when the questions stop
Parents often notice that the phase ends, and assume the curiosity has ended with it. It usually has not. What ends is the specific verbal form.
Several things change at once. Children acquire more background knowledge and therefore need fewer explanations of the basics. They become able to seek information in other ways, including reading. And they become socially aware enough to notice that constant questioning has costs, which is the beginning of the long process by which adults learn to sit through meetings without asking the obvious question.
That last one is worth pausing on. Some of the decline in asking is a gain in knowledge and some of it is a loss of nerve. Both are real. If you are interested in what happens to curiosity when it is measured formally, we have written about how curiosity gets measured in research at all.
The wider point about explanation
Explanation seeking is one of the more distinctively human things a small child does. It is not obvious that it should exist. Learning could in principle be done entirely by association, and a great deal of it is. Asking for the reason is something else: it treats the world as having structure that can be stated.
Which is why the why phase is worth more patience than it usually gets. It looks like noise from the outside and it is a search from the inside. The adult experiences repetition; the child experiences progress.
For the connected question of what unstructured time does for this kind of thinking, see what play actually does and boredom and what it is for.