Francis Spufford has been there, done that and found the books to prove it
It's a world of forecasts. Today is littered with fragments of hypothetical tomorrows which we view with mostly automatic scepticism. The business of measuring out our credence is something we do almost unconsciously, operating a set of calculations about probability without noticing it. Weather over one day: yes. Interest rates over six months: maybe. Political promises over five years: no.

Often we're wrong. Studies show we persistently overestimate some dangers and underestimate others. But we know what kind of activity it is, and it doesn't provide any feeling of contact with a future. For that, we have an appetite that our pragmatic judgement can't satisfy. We want not so much to know the future as to be able to examine it as a temporarily stable object. We want someone to make us a proposal, properly glued together, like the past is in a historical novel.

It doesn't matter that futures change all the time. Remember the Space Age? The actual contest between Apollo and Soyuz in the 1960s and early 1970s towed along a shadowy extrapolation of itself into lunar holidays by the year 2000. Then the Age of Cyberdreck arrived to replace it, with one stage-set giving way to another.

Suddenly, the present to be extrapolated was the anti-social paradise of Reagan and Thatcher and the images in the crystal ball were from Blade Runner. Efforts to dislodge this paradigm haven't worked yet, because our present is still dedicated to low taxes and shafting the poor. The future fuzzes and focuses, fuzzes and focuses.

What's permanent is the wish of mortal creatures for a picture ahead of them. But we ask for incompatible qualities. Any fool can make a prediction. We need prophets with some claim to authority. At the same time we want the detail, the pleasurable specifics of dates and places and people.

It isn't an easy combination. Phoenix House has hired 12 experts, with 12 more to follow in the spring, on a range of subjects (Predictions, pounds 2 each) nicely blended to tickle our curiosities about apocalypse, weird machines, and what our street will look like in 2050. Their blurbs promise a pack of cross-time Polaroids. The experts, however, have had other ideas. They have not taken up the format's licence to play.

Instead they have stayed with their subjects' knotty issues. Only one of the little books, Francois Heisbourg's Warfare, gives us future events - and his suite of wars, in the old World War Three mock-doc mode, are really there to illustrate the ways his set of military technologies might work out in practice.

At the other end of the spectrum, Conor Gearty's Terrorism, Stephen Tumim's Crime and Punishment and John Gribbin's very elegant Cosmology have nothing to do with the future at all. The authors politely ignore the commission to offer, respectively, a polemic against the hypocrisies of "anti-terrorism", a report on the British prison system, and a masterly summary of how we know what we do about the universe. There are two paragraphs on the next- to-last of Gribbin's pages about "the way things may develop further in the 21st century". Yet he says that "I don't want to leave you with speculation"; and his finale is the Big Bang, in all its grand, known choreography.

In the middle stand eight pamphlets that in different ways are about the future without often making the series-title literal. The political and cultural batch lean toward discussion, that civilised form in which possibilities are stirred around. Hugh Thomas's Europe reads likes a brief prepared by a highly skilled civil servant for a minister who needs the options laid out (federation/ confederation/ free trade area). Bernard Lewis's equally handy Middle East sounds a more urgent note, pointing out that superpower withdrawal will let good things happen in the region if they are allowed to. His recommendations of Arab-Israeli co-operation and greater democracy will not surprise anyone.

In Men, Dave Hill makes a witty and likeable plea for the burlier half of the human race to reinvent themselves. The most vividly realised future here is the one he wants us to avoid: a vision of men in self-inflicted redundant, lifelong boyhood. Felipe Fernndez-Armesto manages to be incredibly fluent by virtue of instant value-judgements on Religion, which he prefers in the form of unapologetic mainstream monotheism on speaking terms with liberal tolerance. He is absolutely certain that when a religious impulse takes a tacky shape, it can itself only be trivial or funny.

Last, there is the scientific core to the project. Genetic Manipulation is the simple one; Robert Winston gives a neat, reassuring commentary on the next turns in the medical road, all of which are already envisaged by doctors and only need their practical i's dotting and ethical t's crossing. With Population (John J Clarke), Disease (Matt Ridley) and Climate (Andrew Goudie), we arrive at topics where numerical data exist; where firm predictions are frequently demanded and the refusal to deliver them becomes more interesting.

The styles are different - you won't find anything in the others that quite resembles the brutal humour of Ridley's plague's-eye view of human behaviour - but they share an impressive commitment to explaining the toolkits used by medical, demographic and meteorological forecasters.

Critics make much of the inability of all three to get things right decisively more often than a "naive forecast" - the guess that tomorrow will repeat today's temperature, and so on. But that doesn't mean the specialists know no more than a guesser. The ever-more complex structural understanding of populations is not imaginary. The refinement of the maths of climate isn't a mirage. Things are now understood about the evolution of pathogens that used not to be. It just means that the things they know cannot be reduced from a matrix of variables and assumptions to something simpler.

So Andrew Goudie cannot say whether global warming is happening, or how much. He can only suggest we prepare for surprises. "Even though the risk posed by such possibilities is impossible to assess and cannot therefore easily be quantified, it would not perhaps be prudent to ignore entirely unexpected possibilities in weighing the action that may be necessary." Not the voice of a prophet.