Pictures and an exposition

The Oxford History of Western Art edited by Martin Kemp (Oxford University Press, £40, 546pp)

Art history has been one of the big academic growth areas of the last half century. In 1950, you could only study it in three British universities: the figure now stands at around 50. For many years Oxford and Cambridge refused to take it seriously, only offering the subject as the second part of a degree. Both have now jumped on the bandwagon and offer full courses. Although the number of university applicants recently dipped by around 10 per cent, this year's opening of Lottery-funded museums is likely to fuel further expansion.

So the provision of books to students and art lovers is increasingly big business. Tate Modern, when it opens on Bankside in May, will boast the world's biggest art bookshop. Potentially the most lucrative area is the introductory survey. The best-selling art book of all time is Ernst Gombrich's The Story of Art (Phaidon, £19.95). Since 1950, it has sold over 6 million copies.

Oxford University Press is now trying to muscle in on this market. The Oxford History of Western Art is edited by Martin Kemp, Professor of Art History at Oxford, and includes contributions by 51 scholars. OUP have come out with all guns blazing. A polemical press release, which slighted the authors of rival works, has already prompted a scathing rebuke in Private Eye.

Kemp himself makes no bones about the iconoclastic nature of his magnum opus. Instead of a hagiographic plotting of individual careers, the emphasis is on the contexts within which artists worked. There are sections devoted to the settings for art, from churches and piazzas to museums, and to interpreters of art - historians and critics. Discrete sections discuss "lesser" forms such as prints, photographs and the decorative arts.

While some of the essays - especially on the non-canonical artforms - are impressive, it feels like a work in progress rather than the finished article. OUP hasn't heard the old adage "too many profs spoil the broth". This is a babel of different tongues with far too much repetition, inconsistency, and some extraordinary omissions. The design is poor, with illustrations bunched in chaotic groups rather than keyed into the text.

It is also full of slack writing. which should have been edited. Thus the Great Altar at Pergamon (c.160 BC) is said to be "in an operatizing, 'baroque' mode"; a Bronzino panel painting is included in a section on painting on canvas; we hear that in 1906-7 Picasso "sought to tap more primal reservoirs"; Cornelia Parker's installations are "somewhat discretely and sometimes insistently spectacular." Such sloppiness is worrying in any book but, in an introductory survey, it is fatal.

With the more innovative aspects, it is a case of "what the Lord giveth, the Lord taketh away". Kemp's introduction states that the separate sections on the print stress that "one of the 'lesser arts' played a hugely significant role in conveying imagery to an over-widening audience". This point appears to find confirmation when we are told in a later essay that Raphael "is said" to have hung Dürer's prints around his studio.

But drawings played an equally important role in the diffusion of styles. Raphael and Dürer exchanged drawings, and we know that Raphael was most impressed of all by a "marvellous" Dürer self-portrait in gouache and watercolour. Yet not a single drawing is reproduced anywhere in the Oxford History - a first for a book of its kind. This is surprising, not least because Kemp made his name as a scholar of Leonardo: he recently catalogued a collection of Leonardo drawings acquired by Bill Gates.

What of the competition? Gombrich's Story of Art is showing its age, but still stands as the most accessible concise introduction to western figurative art (it can't cope with modern art). The US market leader, H W Janson's History of Art (Thames & Hudson, £40), is an amiable if plodding behemoth with almost every paragraph headed by an artist's name.

If we could only have one book, it would have to be Hugh Honour and John Fleming's A World History of Art (Lawrence King, £28). First published in 1982, it has been regularly revised. The book is beautifully written and organised, full of lucid distillations of complex ideas. It is in no way monolithic, and the text is enlivened by quotation of contemporary opinion. Works in a huge variety of media are discussed, and non-western art given unprecedented attention.

I ought to declare an interest: I have admired Honour and Fleming's work ever since coming across it as a student, and am one of several people to whom they talked when revising the contemporary art section. But I don't think anyone can seriously doubt the unrivalled usefulness of their book.

Kemp only gives non-western art a belated, walk-on part. Among many barmy editorial decisions, this one stands out. His final section, "Modernism and After 1914-2000", suddenly changes focus, and almost a quarter of the space is devoted to "alternative centres" - the Soviet Union, Latin America, India, Africa, the Caribbean, Canada and Australia. Thus we have the bizarre situation whereby Soviet Realism gets more attention that Russian Constructivism; and a derivative bronze sculpture by the Caribbean artist Edna Manley, from 1938, is discussed and reproduced on the front cover while "classic" tribal art, such as Benin bronzes, is ignored. This tokenism may have something to do with the fact that many of these places are former colonies in which OUP has a strong presence.

But perhaps we should be wary of dismissing this as a folie de grandeur. After all, the final chapter - by Profesor Paul Crowther - suggests that this is the shape of art history to come. In a resounding last paragraph, we are told that thanks to new technologies, a "revolution" has taken place in the way images are created, manipulated and distributed. "Issues of this kind are currently being explored by the British art historian Martin Kemp (1942-) at Oxford University as the basis of a 'new visual history'. It may be that a cross-disciplinary approach on these lines will become important for art history in the next century."

We shouldn't let a minor error (the text should presumably have read "in this century") detract from the momentousness of this announcement: art books of the future will look and read like computer manuals.

James Hall's 'The World as Sculpture' is now a Pimlico paperback