How do you make school as alluring as James Bond?

Recently I had been away for a few days, and when I got back Gabriel, my 11-year-old son, was in a state of high excitement. He'd solved a tricky section in the best-selling James Bond computer game: he'd had to work out which flagstones to safely step on to get to a casket holding the Golden Gun. Take a wrong step and automatic machine guns blow you to bits. It had taken him three hours, he said, but he'd cracked it and proudly showed me the sequence - left two, forward one, right one and so on.

What amazed me was his perseverance. It certainly wasn't a characteristic he'd shown with his homework - the suggestion that he spend more than half an hour doing maths problems or finding out historical details were met with groans and highly imaginative evasions. And it wasn't just him, several school friends - equally casual about school work - also had devoted hours to cracking the Golden Gun code. There was intense competition among them as to who had done it the fastest.

It is a scenario that must be familiar to almost every parent in the land. The child whose grip on the most elementary facts of geography or literature is very shaky, yet who has an impressively encyclopaedic knowledge of every football club, British and European. Generally we respond with resignation and we use the interesting stuff as a carrot - so it's "finish your homework and you can go and play on the computer".

But why should there be this yawning chasm between the, mostly, grind of school and effortless absorbing of huge amounts of information about popular culture?

Normally answers to such question are framed in terms of endless tug- of-war between the progressives and the traditionalists over the curriculum. "Let them decide for themselves about team games, make schools more fun," say the progressives. "Learn the basics, more tests, more competition," reply the traditionalists.

What both sides are resolutely ignoring is several decades of scientific research into how children actually learn - research which provides a convincing explanation for what is going on with Gabriel and his mates, and suggests how to make school as alluring as James Bond or The Simpsons.

It turns out that both sides in the curriculum debate consistently underestimate what children can do. What's more, they make the mistake of treating children as empty vessels ready to be filled up with knowledge. And the result is disastrous.

"It's been clear for 20 years that children are capable of more abstract and effective learning than we used to believe," says Professor Alison Gopnik, a psychologist with the University of California at Berkeley. "Teachers still have the idea that children remain egocentric and aren't really capable of proper reasoning until they are about seven. But research in our lab and others shows that isn't true.

"For instance, children as young as two or three can understand that someone else can see things differently from the way they do. Five and six-year-olds can understand ideas about growth and life and death. They can get to grips with ideas of cause in biology, and have a much more sophisticated psychology than we assume."

The picture that emerges from Gopnik's work is of even very young children as small scientists, who are trying to make sense of the world long before they arrive at primary school. "It seems that kids have been hard-wired by evolution to make hypotheses about the world and test them," she says.

"Our hominid ancestors survived because that was what they did. If you give kids information relevant to a problem they are trying to solve, they can make changes to their ideas - just like a scientist who is working with a hypothesis."

Ignoring this ability is at the heart of the Golden Gun problem. Schools take virtually no notice of the definite ideas about the world children already have when they first arrive at school.

"Teachers still assume that children have a rather superficial idea about causes, and focus on appearances," says Gopnick, whose book, How Babies Think: the Science of Childhood, is being published by Weidenfeld & Nicolson early next year. "Even progressive schools concentrate on immediate sensitivities and manipulating objects, rather than encouraging an understanding of more abstract ideas and concepts."

The result is that children are first presented with a diet of facts, on the assumption that they will develop their own ideas later. "This is fatal," says Gopnik, "because, by failing to give young children information that is relevant to them, they don't get the chance to see how the new facts fit with what they already know."

The idea that schools need to be far better integrated with the lives of the children and the culture that surrounds them is also central to The Making of Intelligence by Ken Richardson of the Open University, also published by Weidenfeld and Nicolson, in July. However, he arrives there from a different starting point. For Richardson, it's our mistaken idea about intelligence that is the key. Seduced by the false promise of the IQ test, - that intelligence is something individual and measurable, like height or weight - we have created schools geared to a particular sort of fragmented learning.

After demolishing the case that IQ tests measure anything useful at all - no agreement on what it measures, failure to predict any intellectual ability except passing secondary school exams - he then lays into the notion that it is coded in our genes. Instead he describes it as dynamic process between the ability that all mammals have to make representations of the world inside their head, and the particular society a human grows up in.

"Human intelligence arose in the context of human culture, and is not viable outside it," he says. "The optimal conditions for learning occur when it is tied into meaningful cultural activity." Just as a jigsaw is much easier when you can see the full picture, so the "scaffolding of a social purpose" is what children need to learn best.

So what might a curriculum look like that attempted to heal the split between school and real life? What new teaching techniques might emerge, if we taught children according to how they actually think? Richardson is rather more concerned with eliminating IQ tests, and all that goes with them, than redesigning the curriculum, but he does suggest that schools might organise some of their teaching around real-life problems facing local businesses, which, if it was a football team, could be a winner. Gopnik, however is much more detailed.

One surprising hero of her analysis is the sports coach. "Humans seem to be designed to learn from adults in a very specific way that you get with the old apprentice system or sports coaches," she says. "What's important is that you watch an adult perform a skill, then you do it yourself, then you get feedback. This is what happens when you learn music or ballet, but it is rare in the classroom."

One of the crucial elements here is the notion of performance, of actually doing something. "Just suppose children were taught sport the way we teach them science. Almost no one would do it" she points out. "Until they were 12 children would read about baseball technique and occasionally hear inspirational stories about great players. They would answer quizzes about the rules.

"Conservative coaches would make children practice fundamental baseball skills, throwing the ball to second base 20 times in a row. Undergraduates might be allowed under strict supervision to reproduce famous historic baseball plays. But only in graduate school would they, at last, actually get to play a game."

The reason why computer games are so attractive is that they allow children just the kind of hands-on experience Gopnik seeks. Children do something, like stepping on a flagstone on the way to the Golden Gun, and get immediate results. There is absolutely no reason why that kind of learning shouldn't be incorporated in schools, even if the cost of personalised apprentice schemes for all make them wildly impractical.

A style of teaching that, from the moment children arrive at school, builds on what they already know, could be as rigorous as the traditionalist demand, and as fun as the progressives keep pushing for. Once that connection had been made, the division between boring school and fun screens would vanish.

With the child's interests and actively doing, rather than passively absorbing at the heart of a curriculum, one way forward would be to study a few subjects very deeply.

This is a suggestion put forward in another new book - The Disciplined Mind: what all children should understand by Harvard psychologist Howard Gardner, published in the States.

"Essentially it's a political decision," says Gopnik. "Do we want to produce children with a real understanding of what they learning or to turn them out with a rag-bag of scraps of information that has no connection to their own experience of the world? Having 30 kids in a room with one person lecturing them is a 19th-century invention. It doesn't have to be like that."