WHAT IS A PAINTING

There is a strange painting, covering the vault of the church of Saint Ignazio in Rome. It is the work of the Jesuit Andrea Pozzo, done about the turn of the seventeenth century. The painting shows, among a number of figures, a set of columns that appear to continue the pilasters supporting the vault. But these subjects of the painting can be seen in their normal shape only if the viewer stands in the center of the aisle. If he moves away from that point even by a few yards, the columns appear to be curved and lying down at an angle to the structure of the church. If you walk around the center of the aisle, the painted columns keep moving around, always lying down away from your position. In a paper published in Journal de Psychologie normale et pathologique in 1963, Mr. M. H. Pirenne offered an interesting explanation of these facts, and in a book entitled Optics, Painting and Photography, recently published by the Cambridge University Press, he has extended this argument further. I think his ideas have important consequences. At first sight the Pozzo phenomenon may seem to present no problem to speak of. We know that a perspectivai painting represents its subject from one central position; hence, when viewed at an angle to this direction, the painting must appear distorted. Pozzo himself gave this as the reason that his painting is distorted when seen at an angle to its perspectivai axis. But this explanation settles Pozzo's case at the cost of raising a much wider question. For it follows from it that all perspectivai paintings must be distorted to a similar degree when viewed at an angle to their perspectivai axis. And this does not happen. One can walk past a painting, for example, in a picture gallery, with-