To speak of the environment being rotated relative to the observer in these cases (instead of the body being rotated relative to the environment) would be simply nonsense. The environment, in the sense of the persisting environment, is that with reference to which objects move, animals move, and surfaces deform. There has to be an underlying nonchange if change is to be specified. The principle of the relativity of motion cannot be applied to rotation of the body.
Visual Kinesthesis of the Limbs and HandsChapter 7, on the optical information for perceiving one's body and its movements, contained a section on the limbs and hands. Certain shapes protrude into the field of view, or else the field sweeps down to reveal them. If they squirm restlessly and are five-pronged, they specify hands. Every manipulation is specified by a corresponding change in the five-pronged silhouette. Reaching, grasping, letting go, plucking, and twisting are controlled by the ongoing optical motions that specify them, as I shall emphasize in Chapter 13.
There are no experiments, however, on this kind of visual kinesthesis. Only so- called eye-hand coordination has been recognized, as if sensations from the eye and the hand had to be associated and that were the end of it.
Evidence for the direct perception of changing layout in the environment and evidence for the direct perception of the movement of the self relative to the environment have been summarized. The awareness of the world and the awareness of the self in the world seemed to be concurrent. Both event motion in the world and locomotion of the self can be given by vision, the former by a local change in the perspective structure and the latter by a global change of the perspective structure of the ambient optic array.
The visual perception of motion in general has been taken to depend on a set of discrete motions of stimuli over the retina. If this is so, an explanation is required of how they are made to cohere in the process of perception. Experiments on the "grouping" of spot motions are inspired by this requirement, as are theories of so-called kinetic depth. But if a change in the optic array is already coherent, its elements do not have to be made coherent.
Experiments with progressive magnification and experiments with progressive transformation suggested that a coherent change in the optic array could be picked up by the visual system. The first kind yielded a direct perception of an approaching object and the second kind that of a turning surface. The perception of these two events was