The Royal Academy’s exhibition of ‘American Art in the 20th Century’ at Burlington House and the Saatchi Gallery is a honeymoon with a marvellous girl who has been stitched into her nightie. No less than one in three of the 230 works arouse a desire to have them in a permanent collection here, but no more than three of the rooms in the show give a feeling of satisfaction.

The first is Gallery Nine at Burlington House, a square room where superb Frank Steallas of his black period confront us from the back wall while in the middle of the floor, humped in silence, is Robert Morris’s big low cage of a steel sculpture of 1967 and to either side whitish paintings by Robert Ryman and Agnes Martin. No self-respecting museum would present a combination which was so insouciant art-historically, but it does look very good.

The second room is Gallery 12 nearby, diagonally traversed by Dan Flavin’s 1968 piece, knee high and fifty-odd feet long, coolly dazzling and hotly blazing, An Artificial Barrier of Blue, Red and Blue Fluorescent Light (to Flavin Starbuck Judd), and with word pieces by Lawrence Weiner on two of the walls. The ghostly lunar light fills the coved ceiling of the room and caresses the gilded ornament, while in Gallery Nine the shape of the coved ceiling is perfectly echoed by that of the Morris cage. Both rooms provide an exhilarating harmony of rampantly Modernist art with emphatically traditional architecture.

The third efficacious room is the long narrow gallery at the west end of the Saatchi once we’re a quarter of the way in (so it’s not quite a room, because the chosen Basquiats are dullish and Haring is redundant) and the works before us are those by Jeff Koons, Mike Kelley, Robert Gober, a series of Cindy Sherman photographs which include a black-suited blonde-wigged creature whose one visible eye is also a mouth and a nostril and a wound, and a stunningly handsome and inventive example of Jenny Holzer’s electronic signboards. Women artists are at last beginning to be gold medallists rather than bronze medallists among their generation, like Jewish artists fifty years ago.

Those moments of pleasure and excitement, with their play between the works and two nicely contrasting contexts, one very much of, one very much not of, their period, underline the loss of an opportunity to do a great bipartite exhibition, had it been given greater thought and care. I cannot further delay declaring an interest. Early in 1990 the Royal Academy invited me as a student of American art of the 20th century, to co-curate the present exhibition with the old firm of Norman Rosenthal and Christos Joachimides. After three meetings – which were very enjoyable though they threatened squalls ahead – and some subsequent interchanges with Rosenthal, I realised that I had to resign rather than go on taking up a great deal of their time and my own to no useful purpose. I was not replaced. Evidently no other student of the subject could bear to work with them either or they couldn’t bear to work with any student of the subject or else they didn’t feel they needed to.

Several matters were making me feel I should leave, but I see from the correspondence that the insurmountable problem, because I took it to be symptomatic, was my inability to convince Rosenthal that you simply cannot do an exhibition called ‘American Art in the 20th Century’ without including at least one of Arshile Gorky’s portraits or double portraits of the Thirties (and, if there were more space, a related de Kooning and ideally a John Graham too, but at all events one such Gorky). Rosenthal’s view was that ‘in no way’ could such a painting be ‘described as one of the most central works of 20th-century American art. Gorky’s significance was to come later, in the last few years of his life.’ For me the absence of such a Gorky seemed like a British social history of the period that made no mention of the Jarrow March. It turned out to be a venial sin of omission by comparison with some of the total exclusions of significant artists in the finished exhibition.

It goes without saying that no survey exhibition – unless it either consists of works on a miniature scale or has licence to expand into the entire British Museum, say – can treat its subject as comprehensively as a well-illustrated book can do. Normally a survey exhibition of a century’s art in a large country is a contradiction in terms. It is bound to have crucial omissions; its selection is as sure to inspire anger as that of a Test team. Furthermore, as with Test teams, the selectors will have had objective problems to contend with that the critics don’t know about. For example, selectors are often severely criticised for neglecting to include an example of a certain crucial type of work when in fact they had been working frantically for two or three years trying to borrow one and getting turned down every time. Again, it is part of the job of selectors not to find themselves with more works than they can show to advantage, but often there is so much delay in response to loan requests that distracted selectors go for a substitute, get it, then find that they can have their first choice after all don’t want to leave it out because of its quality, don’t want to offend the owners of the substitute, and are landed with two works where they wanted one.

But problems of this sort can’t have impinged greatly on the present exhibition’s apparently cavalier omission of notable artists. This has provoked so many cries of outrage in the press – in letters to editors as well as reviews – that writing in their wake I have to resist the temptation to register more than a few complaints, not necessarily those I care about most but rather those that do not seem to have been the most widely mentioned.

The exclusion of Thomas Hart Benton is tantamount to leaving Steinbeck out of a survey of the modern American novel. The exclusion of Mark Tobey has to be the result of amnesia (or megalomania). The exclusion of Dorothea Tanning as of every other exponent of post-de Chirico Surrealism suppresses a key part of the story. The exclusion of Josef Albers shows bias. The exclusion of Mark di Suvero means the omission of the one artist (David Smith is something else) who has created a sculptural equivalent of Abstract Expressionism, the movement which forms the nucleus of the exhibition. The exclusion of Chuck Close, accompanied by the inclusion of three large works by Jonathan Borofsky, can only inspire the traditional football crowd’s advice to the referee to get himself a dog and a white stick. As for the exclusion of all signs of Land Art, the failure to put a photograph or two of Walter de Maria’s Lightning Field into an exhibition of the American art of our time is equivalent to not putting an engraving or two after the Sistine Chapel into an exhibition of the Italian art of its time. The presence of an illustrated essay on Land Art in the expensive catalogue* does not answer that objection. (Exhibitions are made for man, not man for exhibitions.)

R. & J. have surely not been surprised by all the criticism of this kind, for, right at the start of the catalogue, Joachimides writes that this is ‘an exhibition devised from a European viewpoint’ and that its ‘perception is concentrated on essentials, and the result is not a broad panorama but a critical focus’, that, rather than aiming ‘to achieve encyclopedic completeness, this exhibition seeks to open a debate’.

The main debating point has been, inevitably, the wholesale exclusion of colour-field painting, notably Louis and Noland. Joachimidcs writes that America’s ‘essential contribution to the art of our time’ was made in the third quarter of the century. Why, then, do R. & J. fill so much of their space with artists who have emerged in the fourth quarter while omitting a whole group of artists who were very conspicuous in the third? The predecessors who have mainly influenced American painting in our time have been Matisse, Mondrian, Picasso and Duchamp. The omission of the colour-field painters is only part of R. & J.’s strategy for diminishing the influence of Matisse and of Mondrian and thereby the importance of that physicality which for many of us is the supreme quality in modern American painting, from Pollock and Newman to Twombly and Stella – infinitely weightier than those iconographic matters, such as responsiveness to popular culture, which seem to excite R. & J.

Even within the limited space available, there was no real need for the subject surveyed to be so heavily filtered through the sensibilities of the selectors. In determining your strategy for selecting any mixed exhibition the primary choice is whether you go for a restricted number of artists generally represented by several works or whether you go for a much larger number of artists most of whom are represented by one or two works only. The recent fashion has been to take the former course (for example, at the Whitney Annual, which some time ago abandoned its traditional policy, which made it so delightfully banal, of one man, one work). And the present exhibition, like its predecessors in the series, has chosen that course.

I personally do not think that it is the best approach for an exhibition of this kind, a survey whose audience is not generally going to be familiar with the field. Think of what it was like in the days when you were going round all sorts of museums for the first time, getting to know the repertory of art. Think of how many times you saw a single work by some artist you had never heard of and were not only moved by it but took away a clear impression of a particular artistic personality. A single Albers, a single Tobey could have had that effect. Equally, having a single Demuth and a single Sheeler, instead of four works by each (to take two of many possible instances), could have made room for a number of extra artists.

Furthermore, several important artists who are included seem to me to be inadequately represented. The display begins with beautiful muted bangs and showers of coloured lights in the form of three radiant structures of bright emblems painted by Marsden Hartley in 1913-14, but Hartley, who died in 1943, was also notable for his dark threatening seascapes, of which there is no sign here. Elsewhere, there is a group of three exquisite paintings of labelled objects by Stuart Davis, dating from 1921-4, but Davis, who died in 1964, was mainly notable for the later grand jazzy pictures, none of which are in the exhibition (the feel of popular culture rather than its iconography and language).

The central question isn’t how many works represent each artist, but how many phases or aspects of an artist’s work need to be exemplified. With a few exceptions, artists are represented in this exhibition by several works of a single period. The policy is enunciated thus by Joachimides: ‘lt is one of the purposes of this exhibition to tell the story of American art as a history of innovations. The selection concentrates on the moment at which the artist invents something essentially new, or at which he succeeds in a specific act of concentration that endows a work with exemplary character ... In fact, the guiding principle throughout has been to trace the artistic process to its inception.’

Think of what would happen if this principle were applied to another story which could reasonably be treated as ‘a history of innovations’ – that of Italian art in the 16th century. We would be given Titian’s Assumption of the Virgin but not the Rape of Europa, let alone the Flaying of Marsyas. Could that choice seriously claim to represent a slice of the history of Italian art in the 16th century? The ‘guiding principle’, then, is facile and shallow. And R. & J. tacitly admit as much when they include powerful examples of two separate phases in Stella’s art, which might be described as his translation of Reinhardt into low relief and his translation of Pollock into high relief.

But why was this good sense not extended to cover such other major figures as de Kooning, Johns and Serra? The Saatchi section of the exhibition, which covers the years after 1970, could have been vastly strengthened by the inclusion of a big Serra, a bigger Serra than has been or could have been fitted into Burlington House; Serra belongs in any case in the Saatchi section, though this is a historical fact of which Rosenthal is evidently unaware. In a letter in the Independent of 22 September, in which he put in their place two distinguished correspondents who had criticised the selection, he represented himself as knowing all about developments beyond the ken of those old fogeys and proceeded to show his credentials by providing them with a potted history of American art since their time, in the course of which he described Serra as a Minimalist. Yes, and Matisse was an Impressionist.

To construct an exhibition from groups of works each of which is homogeneous does, of course, make it much easier to achieve a decent hang than if every work is clearly different from the next, being by a different artist or by the same artist but of a different period And there is a lot to be said in favour of making exhibitions in which a sense of responsibility to history is a lower priority than the visual impact of the work; we can go to books for instruction. Exhibitions are for catharsis or, if not that, at least pleasure.

Given, furthermore, that the point about homogeneous groups of works is especially true in the case of contemporary American art – artists today work with the ensemble of their next one-man show in mind – I really ought to tear up everything I’ve written here and say that the trouble with this exhibition is that it didn’t confine itself to a score of artists, each with a whole space to himself, a space containing examples of one type of work, the whole thing entitled ‘Some American Art of the 20th Century’. Certainly, R. & J.’s strategy has been on the side of achieving good results in terms of aesthetic experience rather than education. This is why the deepest disappointment of the exhibition is its dismal failure to make the works look as good or as significant as they are.

Mostly this happens through overcrowding, an overcrowding which (as my friend Max Coombes put it) makes looking at the pictures feel like flicking through the plates in the catalogue. Sometimes it happens through an insensitive or uneducated juxtaposition of artists, as when silkscreen paintings by Rauschenberg, instead of being shown along with his own combine paintings, which are doing the same thing with a different technique, are hung (repeating a mistake made in the Pop Art exhibition) along with silkscreen paintings by Warhol, which are doing something profoundly different. Both of these defects happen in an especially disastrous way at the heart of the exhibition, the two linked spaces that are given over to the Abstract Expressionists – the usual hall of honour and an octagonal space beyond which becomes the inner sanctum of the temple. Uniting Newman and Rothko, it turns out to be an Unholy of Holies. Their work looks rather alike as chalk and cheese look rather alike. Really they are as different – and as alike – as Ingres and Delacroix. This doesn’t mean that they shouldn’t be in the same space, but they shouldn’t be in a space that crams them together. There are seven equal facets in this space and each of them is occupied where three of them should have been left empty. As it is, the canvases hang cheek by jowl, each dissolving sweatily into the next like members of a rugger scrum, when their need is to be as separate as sentries.

What a shame that R. & J. didn’t get Bryan Robertson to give them a hand! Henry Moore used to say that Robertson’s Rothko show at the Whitechapel was the most moving exhibition of a modern artist he had ever seen. Robertson also did a beautiful Pollock show there. The hang of the Pollocks here is execrable. It was bound to be difficult because R. & J. had managed to borrow the biggest picture Pollock ever painted (eight feet high by 20 long) – a splendid coup but a rash one, rather like bringing an elephant home as a pet when one only has a horsebox to house it in. But did their taking that risk really mean setting up the hopeless juxtaposition of Number II (1951) with Ocean Greyness (1953)?

Answering criticism of the overcrowding which I made directly to him, Rosenthal, who is, of course, not only co-curator of this show but Exhibitions Secretary of the Royal Academy, told me that I apply the wrong standards to his way of hanging shows at the Academy. He believes that the sort of spaced-out hanging which I favour would be quite inappropriate, maintains that his crowded hanging is in the best traditions of the Academy, and argues that, while he has not seen installation shots of the legendary exhibition of Italian art held there in 1930, the number of items in the catalogue suggests that the hang must have been crowded. The first Royal Academy Winter Exhibition that I myself remember vividly was the Dutch exhibition of 1952-3. In what was always called Gallery Three, daylit in those days, Rembrandts were alternated with Cuyps in one of the most serene and Iuminous rooms I have ever seen. That certainly wasn’t croeded.

But I’m quibbling, because Rosenthal is certainly right that Old Master paintings can be hung almost frame to frame and still look great; after all, in fresco cycles there’s little division between the images. But big Modernist paintings are not like that. With an Old Master there’s an illusion of space within the picture, whose energy can exercise itself within that space. In a Modernist painting there is not that space; the energy can only expand outwards into a space surrounding the canvas. In hanging two such pictures side by side, you have to judge where their fields of energy end and thus neither hang the pictures so close that they start to fight nor so far apart that the wall between them goes dead.

Again and again in the Royal Academy’s modern exhibitions there’s this terrible failure to make the pictures work within the given spaces, whether the exhibitions are blockbusters or modest in scale. With ‘Monet in the Nineties’ there was an utterly crazy hang in Gallery Three – everything very symmetrical in this grand symmetrical space except that on one side wall the pictures were double-banked and in a totally cack-handed way. The exhibition of early Rouault, based on the selection of works which had looked so impressive at Pompidou, looked dead, the works stifled by their hanging. The current Pissarro exhibition also is too full. Why was it not foreseen that so many pictures of the same size in that space would look like frames in a strip cartoon? Why, in short, doesn’t Rosenthal stop and think?

A failure to stop and think may be one explanation, however partial, of a staggering statement made by Rosenthal in the course of a recent catalogue essay for an exhibition of Gilbert and George in Peking and Shanghai. It is a passage which could have been treated as the expression of a considered and coherent point of view had its author been John Berger, say, or Magritte.

Western art of this century seems to have lost any moral significance on account of its fruitless search for formal purity. Meaning and ornament, closely related qualities, have been marginalised by the mainstream European aesthetic of our century. The black square painting is a goal that can appeal only to very few aesthetes. Not only the black square but equally the crushed automobile, the Coca-Cola can and other examples of Western cultural detritus, all threaten to take over the world.

But what on earth are we to make of this when it comes from the co-curator of the present exhibition? Is it that Rosenthal thinks that this demolition job is needed to create a launching-pad for a defence of Gilbert and George? Is it that Rosenthal is trying to ingratiate himself with Chinese official circles.? Is it that Rosenthal doesn’t really like the work of Newman or Reinhardt or Chamberlain or Warhol? Are they types of modern art with which he is basically uncomfortable – this would help to explain how badly he instals Newman and Reinhardt – and handles only because it is part of the job?

Are there kinds of modern art with which he feels totally at home? He certainly loves German Expressionism. He has written eloquently about Baselitz, and in the catalogue of the present exhibition he writes passionately about aspects of Pollock which relate to Expressionism while tending to plod dutifully through most of the rest of the story. And when you come to think of it, Rosenthal’s own style as a curator could be described as expressionist – energetic, sweeping, unstructured and charged with a sense of omnipotence. Okay for making art; dangerous for exhibitions.