When people today say "I have a dream" they are not necessarily expressing an ambition to change the world for the better, as Martin Luther King did. They sometimes want to sell it - or at least to offer it - to a sleep laboratory. The important thing is it should be worth probing for neurological analysis.

It should be the kind of dream that most of us have now and then - a sequence of images and sensations that are jumbled, episodic, unlinked, possibly on the crazy side. At its grotesquely pictorial best, the ideal dream is depicted in Inception, the new blockbuster movie in which Leonardo DiCaprio and Ellen Page can bend streets, walk up walls, and obliterate a cafe by the force of thought. The actors are able to manipulate the scenario to their will. They know they are dreaming. They can, and do, change the dream's scenario.

Can ordinary folk do that? The departments of cognitive science at several western universities believe many of us can. At Swansea, for instance, Professor Mark Blagrove runs a sleep laboratory. He is an expert in lucid dreams, that is to say, dreams over which the slumberer has more or less complete control. Always, that sort of dreamer is aware that he is safely in bed experiencing a cinema of the mind. He can re-arrange the content.

Professor Blagrove says the number of people who can manipulate their dreams has grown by up to 40 per cent since 1980. How he can possibly know that is not entirely clear to me. It suggests too swift and gigantic a leap in evolutionary consciousness. And how can he, for instance, be entirely sure that his human guinea pigs are telling him the truth when they describe their pillow dramas?

The other day, I heard of an experiment - not at Swansea - in which sleepers were attached to a monitor that transcribed their mind-imaginings onto a screen in measurable squiggles. It is thought that one day these squiggles may be translated into cinematic pictures. The guys in white coats will be able to peer into our unwoken minds.

Scarifying that, I think - more horrible than Orwell's sinister imaginings in 1984, but Prof Blagrove is an eminent psychologist. He doesn't, I should guess, have ambitions in that sinister direction.

At Leeds, the chief dream analyst Prof Martin Conway says most dreams are fuelled by memories that are fixedly recent. I find this hard to accept: my own nocturnal fantasies make no chronological sense, the events of 40 years ago mingling with utterly unrelated things I have done the night before. Nor can I change them. If I am beginning to enjoy myself in dream, something makes me wake up. If I am having a nightmare, I have to put up with it before the clock radio goes off.

Tread not on my dreams, professor. I intend to keep them to myself.