About a hundred years ago, a group of artists embarked on that unprecedented adventure known as modernism. Critics and philosophers were soon on hand with explanations and justifications. Realist painting was denounced as "a pale reflex of actual appearance" requiring no more than "technical capacity in the imitation of nature". The true aim of art, it was declared, is not the depiction of actual forms, but the expression of feeling, or the creation of new forms. Some went even further, claiming that we ascribe "reality" to certain works of art merely on the basis of our familiarity with their stylistic conventions. These ideas went on to dominate the 20th century. Critics and historians now use the term "realism" only in scare quotes, and although some artists still aim at verisimilitude, it is usually in a spirit of postmodern irony and not because they regard it as an intrinsically worthwhile pursuit.

The Objective Eye applies a keen philosophical scalpel to this mêée of ideas. It scrupulously dissects the various myths and confusions surrounding the concept of depiction, with the aim of rehabilitating realism as "one kind of excellence in art". Against the misgivings of sophisticates, it champions what it sees as the natural, pre-theoretical stance of artists themselves. "Whereas philosophers and psychologists are fascinated by illusions and by images in the mind, artists have more often said that they are interested in nature, reality and truth."

John Hyman's main target is the theory, which he traces back to Descartes, that pictures rep resent the world not by imitating it directly, but by simulating its effects on the viewer's senses. Pictures, so this argument goes, are instruments of illusion, and the history of pictorial art is one of the refinement of illusionistic techniques. This set of ideas has often served to denigrate pictorial representation as a tricksy, mechanical enterprise, which the pioneers of abstraction were quite right to abandon. Illusions may be amusing or instructive, but only incidentally are they of any aesthetic interest. Indeed, the perfect illusion could not be of aesthetic interest, for it would not be visible as an illusion. Art's paradoxical quest, according to this view, is to make itself disappear. "Naturalistic art dissembled the me dium," wrote the influential champion of abstract expressionism Clement Greenberg, "using art to conceal art."

In a tone of weary patience, Hyman points out that illusionism, far from being the soul of pic torial art, is a rare anomaly. Trompe 'oeil "works" only when seen from a particular position, in particular lighting conditions, whereas most pictures, including the most realistic, can happily be seen from many positions, in many lighting conditions. Trompe 'oeil tries to conceal the painted surface, whereas the old masters (think of Titian or Rembrandt) often announce it with swirling brush strokes and thick impasto. Illusionism, in short, is not the same as realism, although it adapts the techniques of realism to its own ends.

If some critics denigrate realism as illusionism, others deny that it has any objective meaning at all. Pictorial styles, they argue, are like scripts - systems of conventional symbols, with no inherent resemblance to reality. If some seem more realistic than others, this is only because long acquaintance has dulled us to their conventionality. Especially guilty of this imposture is the western system of perspective painting. Like the market economy, whose progress around the globe it has often shadowed, it has a proclaimed universality that is simply a mask for cultural chauvinism. It had to wait for cubism, with its crude, iconic pipes and guitars, to reveal the true nature of pictorial symbolism.

Hyman wants to insist that some styles are objectively more realistic than others. But by what right, we might ask? What objective measure of realism remains, now that it has been dissociated from illusionism? It is in response to this dilemma that Hyman introduces his concept of "modality". The measure of a picture's realism, he explains, is not its tendency to deceive the eye, but rather the range of questions we can ask about its content. To ask if an Egyptian statue is standing stiffly or at ease is senseless; it is, in Ernst Gombrich's nice analogy, like asking for the age of the king on the chessboard. Egyptian sculpture is therefore deemed less realistic than Greek sculpture, about which such questions can be raised. But this does not imply that realism progresses along a single trajectory, culminating in the virtual reality machine, for the range of questions we can ask about a picture's content can be expanded in many different directions. Nor does it turn realism into a passive, mechanical affair; on the contrary, it shows it to be a feat of imagination and intellect. The concept of modality is Hyman's most important innovation. It opens up the fascinating prospect of a cross-cultural history of pictorial representation, freed from dependence upon the psychology of illusion. Historians of art - take note.

Hyman writes fastidious, elegant prose, without the technicism that so often mars analytical philosophy. He has a wide knowledge of European art and a discriminating eye. Yet I fear his hope of finding an audience beyond professional philosophers may be disappointed. He remains too much captive to the scruples of his guild. He levels his fire against the formal arguments used by realism's deprecators and deniers, but he does not touch the anxieties of which those arguments are no more than the confused expression. The arguments may crumble; the anxieties remain. Why, after all, did so many artists in the 20th century withdraw from the task of depicting reality? Was it that photo graphy and cinema could do the job better? Did the scientific disenchantment of nature compel them to retreat into their own imaginations? Or was it the horror of the modern world, as Kandinsky once remarked, that forced art to become abstract?

Hyman has no answer to these questions; indeed, he does not even ask them. He can show that realism is a meaningful endeavour, and that it is not the same as illusionism. But he cannot address the more fundamental question of its value, for that would take him outside his logical and epistemological brief, into the treacherous territory of ethics, theology and cultural history. For all his subtlety, he cannot tell us why art should depict the world at all.