Howard Gardner only has to hear a piece of music once to remember it. But meet him for the second time and he will look at you blankly, unable to recognise a single detail of your face. He was 25 before he realised that other people could recall faces more easily than music, he says. Grasping that their minds worked differently from his own was one factor which prompted him to develop his theory of multiple intelligences, which has dominated his work and reputation as professor of cognition and education at Harvard for the past 20 years.

The child of refugees from Nazi Germany, Mr Gardner, 59, began his career in psychological research by studying the development of gifted children and the brain processes of people with head injuries. The unusual thinking patterns of both groups told him that it was absurd to assume there was a single characteristic which could be labelled “intelligence”. Two children with equally stratospheric IQs might learn things in two completely different ways. Brain-damaged adults might achieve average intelligence scores in parts of IQ tests, yet be unable to remember their own names.

The way he and fellow psychologists discussed intelligence was, he began to argue in the mid-1980s, heavily influenced by the culture they inhabited:

“Our culture has valued the language-logical mode of intelligence, the law professor mode, and the more you resemble the law professor, the smarter you will be seen to be. If Bobby Fischer hadn’t lived in the 20th century, in a place where chess was valued, he wouldn’t have been called smart.”

Of course, it is only by possessing law professor-type credentials that Mr Gardner has been able to question them. On one of his rare visits to England, to speak about thinking at a recent conference in Harrogate, his style is Ivy League avuncular: redefining terms, preferring questions to answers, steely in response to criticism.

For there have been critics of his theory that people possess a whole set of “intelligences”: abilities associated with learning, related to particular areas in the brain, and valued by at least one culture. Initially he identified seven: linguistic, logical-mathematical, spatial (used by pilots, surgeons), bodily-kinaesthetic (used by dancers, footballers), musical, interpersonal (awareness of others’ feelings) and intrapersonal (awareness of one’s own feelings).

He has since upped the seven to eight and a half, adding naturalistic intelligence (an ability to classify the environment, as possessed by Darwin) and, more hesitantly, spiritual intelligence, which he defines as “raising the big questions: what’s going to happen to us? Why are we here?”

Each of us possesses all these intelligences to some extent, he argues. One particular intelligence - different ones for different people - probably acts as an entry point to engage the others. In his case, it is music. But that is as egalitarian as he is prepared to be. “I do believe that if people have different profiles of intelligence, you can’t simply ignore that fact. You can’t have a uniform school system in which everybody is taught the same thing in the same way. When people say multiple intelligences mean that everybody is giftedI I don’t believe that. It worries me when people say it, because I don’t think life is that fair.”

Mr Gardner is wary of the multiple intelligences industry that his research and books - 18 to date - have spawned: “It’s a growth industry, even though I never had anything to do with promoting it. I developed this theory as a psychologist writing for other psychologists. I was very surprised that the most interest in it has come from educators. I’ve even been sent a multiple intelligences board game from Taiwan.”

But his critics are warier still. If seven intelligences can be upped to eight and a half, they argue, then why not 10, or even 20? The sub-divisions are arbitrary. Mr Gardner’s list over-plays the arts and undervalues practical intelligence - the one that gets things done. And, most damagingly, it contains no hard psychological evidence to prove these different intelligences exist.

Producing that evidence would mean designing psychometric tests for each intelligence and finding thousands of people to undertake them in order to create “norms” against which others could then be measured. It would, Mr Gardner says, be 10 years’ work: “I think it’s worth doing, but I don’t want to spend my life doing it. It’s hard to assess how people’s minds work. I think if people are doing fine in life we shouldn’t interfere by making them do clinical tests; let’s save those for the kids with problems. What we can do is observe what children are doing; any teacher who’s really awake watching kids in a classroom designed to encourage the use of multiple intelligences will learn a lot about how they learn.”

So does that mean that he believes multiple intelligence theory is making a real difference in today’s classrooms, either in the United States or elsewhere? Unfortunately not, he says. “There are hundreds and hundreds of schools which say they are doing multiple intelligence work. Policy-makers in the US, at least, have learned that it costs them nothing to talk about individual differences in learning style - because it’s politically correct. But that’s quite different from the kind of policies which they legislate.”

Howard Gardner has worked closely with two schools, one in St Louis, the other in Indianapolis, in trying to create classrooms which stimulate every possible intelligence. He does advisory work with groups of schools and policy-makers in the US, the Far East and Europe - in Holland, Sweden, and Italy, but not in the UK.

His Harvard colleagues have set up research projects to look at schools - private and public, deprived and wealthy, elementary and senior - which try to apply multiple intelligences in the classroom. One in particular, Project Sumit (schools using multiple intelligence theory) - has identified six features that characterise such schools.

These include children working in depth on single topics. In one school, for example, students spent five to six weeks perfecting topographical maps of Africa; there was lots of collaboration between teachers, and plenty of choices for children in the ways they work - model-making, computer drawings, creating surveys, for example - and a significant role for the arts.

But even this is still a long way from Mr Gardner’s more radical vision, outlined most fully in a speech he gave in Holland last year. The logical outcome of believing that everyone has a different set of intelligences is an individualised education system, he argues.

Such a system would find out as much as possible about the way children learn and match them to curriculums designed to suit individuals. It would also match them with a series of role models, inside and outside school, who possess the same mix of intelligences as their own.

New technology should make this kind of individualised education possible, he argues. It would be based on what he calls “disciplines”: the capacity to think intelligently in scientific, humanistic, historical, artistic and mathematical ways. It would also radically reduce the number of facts children learn, and replace them with “an intimate knowledge of a limited number of really important issues”: evolution, for example; the Holocaust; the music of Mozart.

As he lays out his vision, it becomes clear how very different its direction is from the fact-laden, centralised education policies current in the US and the UK. To study evolution properly, he argues, requires not only definitions and stories and works of art and numerical puzzles, but also “the raising of the most profound existential questions: why are we here? What will happen to us and our species in the future?”

These kinds of moral issues have dominated his research in recent years: his latest books are not about education at all, but about political leadership and what he calls “good work”: the need for professionals to put ethics above personal success, and the tension which that produces in a market-driven economy.

“Good work” is as crucial in education as in any other profession, he argues. Even eight and a half intelligences are not enough; schools must also imbue their pupils with humanist, not market-driven, values. “We must help students to find meaning in daily life, to feel connected to other individuals and their community - past, present and future - and to feel responsible for the consequences of their actions.”

But this is not a vision he sees taking imminent shape in American schools. “The US couldn’t be going in a worse direction in its education policy at present,” he says. In the UK, too, he suggests that policy-makers have done little more than scratch at the surface of his research. “The think-tank Demos has shown interest in my work,” he says. “Tony Blair has given a speech on multiple intelligences. Gordon Brown did once ask to see me. But the hardest thing is to get a difference in the classroom.”

Howard Gardner was visiting the Learning Conference at Harrogate in July. His books include: Frames of Mind: the theory of multiple intelligences (Basic Books, 1983); Multiple Intelligences: the theory in practice (Basic Books, 1993); The Disciplined Mind: beyond facts and standardized tests, the K-12 education that every child deserves (Penguin Putnam, 2000); Intelligence Reframed (Basic Books, 2000); Good Work: when excellence and ethics meet (Basic Books, 2001). Details of Project Sumit can be found at www.pz.harvard.edusumit.