Identity parade

Sander's work can also be seen at the Museum of Modern Art in Oxford, as part of an exhibition mouthfully entitled "In Visible Light: Photography & Classification in Art, Science and the Everyday". The theme is specimenisation, the dubious ways the camera has been used to provide specimens for various human and natural sciences. An extensive and fascinating archive has been assembled.

The show divides broadly into two strands. There are pictures from photography's first hundred years whose ostensible empirical purpose betrays dodgy agendas - ethnography, criminology, psychiatry, sexology, sociology. Here we have Charcot's pictures of women in hysterical convulsions, Muybridge's studies of human motion, demonstrations of the insane, the degenerate and the exotic, ancient police scene-of-murder photos. Then, as corrective, there are photo-projects by contemporary artists which question this sort of procedure: works by Andy Warhol, Cindy Sherman, Andres Serrano, Jo Spence, where a taxonomic approach is suggested only to be refused.

Here's one such play-off from the crime section. One of Francis Galton's composite photos from around 1880 superimposes the faces of several criminals so as to reveal a supposed basic criminal face-type. Galton invented the term eugenics, and used this form of photographic breeding to construct essential physiognomies. Hard by, there's a Christian Boltanski booklet from 1987 in which the artist has juxtaposed snapshots of both murderers and murderees. But there are no identifying captions, so your impulse to detect facial signs of viciousness or victimhood is aroused and then utterly thwarted.

That's a neat but representative example of the moral polarity involved here. A bad past is criticised by a better present. Of course I agree with that particular moral. Galton's work was wrong and its influence pernicious; Boltanski's is salutary. But I think the show offers a little too readily the pleasures of superior enlightenment, and it does tend to suggest that all empirical photography is suspect, by including very little that we might still consider valid and useful.

But there's no question that this moralism, which you can hardly resist, is exciting - and here the old photos, which we're to see through, are more exciting than the contemporary ones whose scepticism demands assent. You get a sort of chamber of horrors effect at one remove. Look at those mid 19th-century pictures of the mad by Hugh Diamond and Henry Hering. It's not what they're of that provides the frisson, but the troubling psychiatric culture of which they're evidence - troubling just because it is itself so clearly troubled and excited by its subject. You see the ghastly theatricality of these attempts to make objective case studies of delusional paranoia.

In a way, our own approach to these images duplicates that of the original photographers to their original subjects. They become case studies for us too, specimens now of a historical sickness, to be viewed at a fascinated distance. They must have haunted their contemporaries; in a different way they haunt us still. There's no escaping the fact that, one way or another, photography creates these distances - between camera and subject, between present viewer and the past act of photography. Or perhaps there is one escape route, what August Sander does: to imply a distance so extreme that there's no way that anyone can't be on the receiving end of itn

To 6 July. Museum of Modern Art, Oxford (01865 728608) Tom Lubbock