The big theories - Newton’s laws of motion, Darwin’s theory of natural selection, Einstein’s theories of relativity - are demonstrably true and are never likely to be overturned. The last great discoveries - quantum mechanics, DNA - emerged several decades ago. Though something like 95 per cent of all the scientists who ever existed are alive now, their misfortune is to be born at the wrong end of time. All they can do, writes Horgan, is to refine and apply “the brilliant, pioneering discoveries of their predecessors . . . to measure the mass of quarks more precisely, or to determine how a given stretch of DNA guides the growth of an embryonic brain”.

It is true that scientists still do not know exactly how the universe was created or whether there are infinite numbers of parallel universes or how life got started on Earth. But, Horgan argues, these are questions to which it may be impossible to get definitive answers. All scientists can do is speculate, and Horgan dismisses that as “ironic science”: in the same way as the literary critic (according to present thinking) cannot hope to divine the “true” meaning of a text, so a scientist cannot verify empirically the ” truth” about life, the universe and everything.

I must say that I found this thesis unconvincing - if there had been such things as science journalists in the 17th century, they could plausibly have argued that Newton had already wrapped up the physical universe. But the important point, which even Horgan is compelled to acknowledge, is that there is one final frontier, where no one remotely approaches the authority of Darwin or Einstein. The human mind, human memory, human behaviour, human learning -these remain as mysterious as ever.

Scientists, to put it bluntly, don’t have a clue. Consider their attempts to build something that imitates the human mind. Dozens of scientists have laboured to create a computer that eventually beat Kasparov at chess. But since chess is after all played in bounded space with a finite number of options, that is one thing that a computer ought to be good at.

Now, I don’t want to discourage professors Michael Barber and David Reynolds, Tim Brighouse and Baroness Blackstone, and all those other New Labour luminaries whose hour has come round at last. I merely wish to counsel caution and humility, and to warn that they should prepare themselves for disappointment.

Newton said that he saw further because he stood “on the shoulders of giants”. Alas, Professor Barber et al won’t be able to see very far because most of education’s giants - Rousseau, Dewey, Piaget, Cyril Burt, to name but a few - have been at least partially discredited.

We are no clearer, for example, about the role of IQ, or even how to define it, than we were 50 years ago. We are apt to think that, because our understanding of biology and physics is so far in advance of, say, the 15th century’s, our understanding of other subjects, such as education, must also be superior. I don’t think it is. Education, bound up with all those unanswered questions about learning and behaviour, remains the most mysterious of all human activities.

Even the things that we would expect intuitively to be true are stubbornly resistant to proof. Children must surely learn more in smaller classes, we think. Yet the best that research can do is to show some small gains if we reduce class sizes for five to seven year-olds.

We naturally expect that more resources generally will help. Yet research keeps coming up with perverse results. As The TES reported last month, one of those giant American surveys has found that extra spending helps above-average children but not the disadvantaged. To be fair, research does now tell us that schools make a difference (though exactly how is still very uncertain) and that nursery education for children from poor homes shows long-term advantages. But it took an unconscionably long time to do so. It is rather as if Newton had established the laws of gravity only after several experiments in which apples failed to fall to the ground.

The lesson for David Blunkett and his advisers is that all research and all big ideas should be taken with a pinch of salt. Education doesn’t have a second law of thermodynamics, so why pretend it has? The new administration is entering dangerous territory.

Up to 10 years ago, governments had very little power over what went on in classrooms. The Conservatives then began to prescribe, sometimes in unnecessary detail, the content of the curriculum. New Labour will be tempted to go further, telling schools how much homework to set, how to stream or set classes, how to teach maths Taiwanese-style. Official inquiries have always commented on these matters (think of Plowden, Bullock, Cockroft, for instance) but, until now, governments lacked the machinery to implement their recommendations. If the fashion was for “look-and-say”, a few brave souls (more than a few, in fact) could remain faithful to phonics in the privacy of their classrooms. They might damage their promotion prospects; and they would certainly be nagged by any passing local authority advisers, but they didn’t face the prospect of being denounced as incompetents and driven from the profession.

Perhaps we shall never crack the secrets of human learning; perhaps the study of education is, of its nature, an “ironic” activity, in which one theory is as good as another, and the only criterion for the value of a proposition is whether it leads anywhere interesting.

Alternatively, perhaps education is simply in a pre-Newtonian phase, waiting for the apple to drop on somebody’s head. Either way, an excess of dogma, certainty, missionary zeal and - to put it at its simplest - bossiness, is the last thing we need.

Tony Blair has said that the three priorities for his government are “education, education, education”. I hope that, in pursuing them, his three favourite theories are common sense, common sense, common sense.