Philip Hensher: That's a Hirst? It'll never sell...

The Institute of Contemporary Arts has had an interesting idea. In its new show, Surprise, Surprise, it has asked 40 well-known artists to supply work in an unexpected style. Many artists have a particular, distinct manner or characteristic subject that they can be identified by. This exhibition has asked them for work that is different from their characteristic manner.

Damien Hirst has sent in a drab little assemblage of book and driftwood, made before he discovered pickled animals and spot paintings. Anish Kapoor, who specialises in disorientating effects produced by (often massive) organic sculptures, has sent in a single dripping red line on paper, called Vein. Wolfgang Tillmans, the photographer, and Martin Creed, the most austere of conceptualists, have sent in paintings.

Interestingly, most artists have supplied art they made before they were famous, or had established a manner. Peter Doig has an effective painting from 1982, though his complicated, dreamlike style is already evident. But Hirst's assembly, from 1985, would never be suspected as being by him. And, in supplying juvenilia, no one goes as far as the Chapman brothers, who put in a charming papier mâché pig (Dinos) and a penguin (Jake) made when they were eight and five.

What is rare, apart from Kapoor's elegant piece, is a sense of something made as a diversion from the artist's "fingerprint" pieces. The show has to fall back on juvenilia, because most successful artists work within defined, recognisable forms, which could in most cases be copyrighted. There is no particular gain to be had from diverting from the brand.

So Rachel Whiteread does casts of empty spaces. Antony Gormley casts his own body in various media. Chris Ofili does glittery cross-ethnic paintings with elephant dung. Till-mans does low-impact casual photography of groovy and grubby urban lives. Sarah Lucas turns pub detritus (often cigarettes) into sculpture.

Artists can move from one thing to another, but their work tends to be successful when they take the trouble to establish "lines", rather in the way fashion houses have haute couture and "diffusion" styles, or establish different but still characteristic collections in subsequent seasons. Gary Hume made his reputation with a series of very similar gloss paintings of hospital doors; followed with lurid gloss-paint Pop Art shapes; went on to overlapping outlines of human figures, and so on.

Hirst had his pickled animals and the vitrine/tableaux for museums to buy; then the "diffusion range" of the spot paintings, usefully transferable to mugs, T-shirts etc. For a while now, he's struggled to establish a new "line"; his giant anatomical models, such as the one now in the courtyard of the Royal Academy, haven't caught the public imagination in the same way as his early Nineties pieces.

There are painters who only do black paintings, and who only do white ones. How would one recognise a Barnett Newman if he'd suddenly decided to produce a painting with one horizontal line, rather a vertical one? And contemporary artists have taken the point. To establish yourself, produce something clear and simple; do it again and again until someone notices. Once the market is saturated, move on to something else, and start doing that repeatedly.

In this situation, there isn't much space for an artist who just feels like doing something a bit different. The market and the artist collude not to confuse matters by offering unidentifiable examples of a named artist's work for sale.

And, strangely, the market doesn't value the uncharacteristic piece of work, but the highly characteristic, even the commonplace one. You could probably plot a graph of the phenomenon. Our artist, X Peri-Mental, paints cows' bottoms. If he paints five such paintings, they will probably be less feted and valuable than if he paints 20, and the value will go on rising to a large number; probably only when he has painted 500 or 1,000 will the value start to drop. However many Mr Peri-Mental paints, however, they will still be more valuable than a watercolour of Exmoor he painted when bored, which does not look like a Peri-Mental at all. Should he decide to start painting goats' bottoms instead, he will probably have to paint 30 or 40 before the market starts to value them at all highly, or, indeed, to notice them.

It's a curious situation, which forces artists to be less adventurous, surely, than they might like to be. Let's rediscover that curious pleasure of looking at a work that doesn't look like its author's work at all.