According to O'Regan and Noë (2001), all awareness is based on mastering sensorimotor associations. They suggest that exercising these associations explains the perception of constancy and the difference between auditory and visual experience. As an example for the first, they argue that a line cannot be recognized solely by its sensory representation. The neural pattern in the visual cortex does not resemble a line anymore because on the retina, the photo receptors are distributed inhomogeneously, and furthermore, the mapping onto the visual cortex is non-linear. The line can be recognized, however, by exploiting its invariance; namely, a line shifted along its extension does not change its appearance. Thus, we perceive a line by knowing how the sensory input changes when we move along the line.
Visual and auditory sensory input vary in different and specific ways when the body moves. Therefore, knowing how the image of an object changes when we rotate it before our eyes might explain visual experience (O'Regan and Noë, 2001); ``... the visual quality of shape is precisely the set of all potential distortions that the shape undergoes when it is moved relative to us, or when we move relative to it'' (p. 942). Support for this claim comes from visuotactile devices made for the blind to `see' (Bach-Y-Rita, 1972). Here, an image from a camera mounted near the eyes of a blind is transformed into a tactile pattern on an electrical stimulus matrix, which is mounted on the back or abdomen. After some training, the blind can recognize objects or step back from obstacles approaching the eyes (even if the stimulus occurs actually on his back). O'Regan and Noë (2001) therefore argue against special neurons or neural properties that cause visual awareness.
Overall, it is an appealing concept. So far, however, it does not provide a formalism explaining how the process works. Recently, Philipona et al. (2003,2004) provided a formalism, but it only addresses the perception of the dimensionality of space, and qualities like object recognition are not addressed. Chapter 7 will come back to the idea about perceiving constancy.