How much can be seen in a blank, stretched canvas? Three men stand in front of one. The first describes it as 'an arena in which to act', a place where 'the painter thinks by charging a surface with paint'. The second thinks it 'already exists as a picture' because it acknowledges the norms of painting, 'flatness and the delimitation of flatness'. The third finds 'a pre-existent minimal structure', one that absorbs the viewer because it represents the idea of sleep and the infant's view of the flattened maternal breast. (CWS, III, p. 217, pp. 159-60).

Unlike Harold Rosenberg and Clement Greenberg, Stokes didn't write much about abstract expressionism. But linking Stokes's writings to theirs is a good way to learn something about the theory of art.

Rosenberg's essays were the occasion of 'fiery disputes . . . about the relations between process and product in . . . painting'. Rosenberg emphasized process in ways a Stokesian may at first find suggestive: 'the fragmentary art of . . . action painting engages itself within the fragmentary inner world of contemporary man and the fragmentary outer world . . .' But that he can imagine painters approaching canvas without prior ideas shows that Rosenberg's concept of action is undeveloped. Where Stokes's modelling/carving distinction allows elaborate description of the artist's mental processes, for Rosenberg action seems to be only a gesture. Wollheim's argument that psychology shows why artworks are valuable is pertinent here. Isolating actions from a ...