Visual Arts: The project: to reconstruct the lost works of a dead and u njustly neglected artist. The result: a case of misattribution or just mistaken identity?

Mondrian, you might well think, looking at the image at the top left of this page. It looks like a Mondrian. But it isn't. And in a way, writes Tom Lubbock, it couldn't be, because it does something that Mondrian never does.

In Mondrian's pictures that look like this, there's always at least one vertical line and one horizontal stretching right across the canvas, uninterrupted. Here the only vertical is broken, divided by two horizontals. That kind of rupture, Mondrian always ruled out. It's not a Mondrian. It's a Marlow Moss. A who?

Marlow Moss was a gay, British, female artist - what's the right order for those adjectives? - and a pupil-follower of Mondrian. She is, to put it mildly, little known. She lived from 1890 to 1958. She was christened Marjorie, and became Marlow, and adopted the clipped, mannish dress and haircut that was popular with more or less "out" lesbians in the early 20th century. She spent a good deal of time in Paris, exhibited alongside other De Styl artists, and ended up in Cornwall, where she kept very much to herself.

Moss made several other radical departures from the Mondrian way, one of them significant, because Mondrian borrowed it back from her. She introduced the Double Line, and you see it in this picture too, with its close parallels. It sounds funny to say that - "Oh, the Double Line, big deal!" - but when you reckon that much of the force and fineness of Mondrian's paintings from the late Thirties (his best, I think) depend on the use of double lines, it's a vital development. But Moss's name hasn't flourished, and partly because most of her work pre-1940 was lost when German bombs destroyed her French studio. Much is known only from photos. In fact the picture above isn't strictly a Marlow Moss. It's a Florette Dijkstra.

Florette Dijkstra is a Dutch woman artist in her early thirties who has an exhibition about Marlow Moss at the Tate Gallery in St Ives. Dijkstra has thoroughly researched the artist, and her centrepiece is a painted reconstruction of the works of Moss, not just the lost ones, but her complete known works. It's a chance, then, to know this unknown art - except that knowledge isn't exactly what the show offers.

Dijkstra doesn't attempt a literal reconstruction. All her "Moss" versions are miniaturised, done half-size. They're painted very perfunctorily and thinly, not with the sturdy and intense brushwork of Neo-Plasticism. What's more, their colours are standardised: throughout, all the reds, yellows and blues are the same unvarying red, yellow and blue. (Obviously some of the colours had to be guessed from black-and-white photos.) And then, a few of these canvases are blank, to represent pictures that are known from unillustrated catalogues. This isn't the recovering of a body of work; rather, something to stand in for it, the tokens of an uvre. Accordingly, they're not hung in the normal way, but all close together in a block, about 60 items five rows high, up one wall - as if to say: that's the lot, a life's art, all that survives of it.

So while Dijkstra says that her project's "ultimate goal is to open up a new place in art history for Marlow Moss's work", the show doesn't do a straight reclamation job on a lost and neglected woman artist. It's more about loss and neglect, and making you feel the artist and her art at a remove. Photos of Moss and her Dutch lover exist, but none are in the show; there are only Dijkstra's vague drawings after those photos. Moss's sculptural constructions are represented only by another set of paintings, no 3-D replicas. There are images derived from a TV programme about Moss, a hand leafing through a portfolio (obliquely showing some of her later work, which is mostly in storage). Everything is kept distant, indirect.

A canny but unknowing viewer might even suspect that Marlow Moss never existed at all, is actually one of those fictional artists that artists invent - and that doubt would be to the exhibition's point. Moss might as well not have existed for all she's remembered, and re-establishing her now is a quasi-fictional enterprise, given how little of her and her work is known; and then thinking about how modern art history would look if Moss had a more central place in it, which arguably she deserves, makes you think this history itself is a kind of fiction. Who decides? Who tells the story? All those questions.

Not bad questions, but there's a danger of turning mischance into mystification. Moss, after all, is almost too good an example of a missing artist - she has everything not going for her. Much work destroyed, much squirrelled away from view; not just a woman, whose influence male artists may have been reluctant to admit, but a very retiring and reclusive operator who didn't really try to get into history. But still, she did exist and some of her work still does, and we can have news about it, rather than making it into an emblem of loss.

There is actually a real Moss on show in the next door gallery at the Tate, among real Nicholsons and Hepworths, while its half-size token hangs among the rest round the corner - a curious near-juxtaposition, but it's telling that the genuine article isn't in the same room as the replicas. It would rather break the cloud of unknowing that Dijkstra cultivates. It would let in the possibility of direct knowledge and response, and that seems the important thing, however late in the day, however little we have to go on.

Take the Double Line business: Moss started it, and should get credit for that, but it may still be true that the Double Line is only valuable because of what Mondrian did with it, not what Moss did with it. Originality isn't everything. To judge from the replicas, I'd say that was so, and that Moss's best work is when she departs further from Mondrian, with much more fractured scatterings of lines and oblongs. But from these small- scale, notionalised replicas, you can't really tell, and aren't really meant to.

The point is, it could be done another way. Dijkstra's treatment stresses our helplessness to know the absent work. But she or another artist might have been more helpful, might have taken on a more literal reconstruction of the Moss uvre. The real issue this show raises is about originality in the other sense: the possibility of usefully transmitting an artist's work at second hand. Both modernism and post-modernism are against this. Modernism says that only the artist's authentic hand will do, that or nothing; there can be no valid substitutes for the unique handiwork. Post- modernism favours imitation as a way of casting doubt on the very idea of authenticity and uniqueness.

But it seems to me there's a middle way here, which allows for unique values in art, but allows too that some of that value can be usefully communicated by substitutes - eg by someone copying a picture very faithfully, or by somebody trying to repaint a lost picture, known from reproduction, as it might have been, drawing on surviving work of the artist (something like Deryck Cooke's completion of Mahler's 10th Symphony).

A gallery of such copies and reconstructions, seriously done, wouldn't be invalid. It wouldn't just be fakes and pastiches. Something of the originals could come through to us. Many old masters are known only through others' copies, but when the originals turn up, you find the copy often wasn't far off the mark. And, frankly, it can't be that hard to do a passable version of a Moss or even a Mondrian - not as hard as it would be to do a Vermeer. Besides, the De Styl artists set no store by individuality, so it would be quite in the spirit, too. Of course, I'm imagining a different show from the one in St Ives, and a different approach to art from any that now prevails: one not fussed about patent law, but very interested in the distribution of goods.

The Tate Gallery, St Ives (01736 796226), until 12 March.