Visual recognition, navigation, tracking, and imagery are posited to involve some of the same types of representations and processes. The first part of this paper develops a theory of some of the shared types of representations and processing modules. The theory is developed in light of neurophysiological and neuroanatomical data from non-human primates, computational constraints, and behavioral data from human subjects. In developing theories of some of the high-level computations that are performed by the visual system, three problems which must be solved by any visual system are considered: (1) position variability; (2) figure/ground segregation; and (3) non-rigid transformations. The second part of the paper develops a mechanism for the development of lateralization of visual function in the brain. The theory rests on a set of relatively simple, uncontroversial properties of the brain, including: (1) processing components; (2) exercise; (3) selectivity; (4) "central" bilateral control; and (5) reciprocal innervation. This theory leads to predictions about the lateralization of the putative processing modules. The third part of the paper examines critical tests of these predictions, and reviews relevant empirical findings in the literature. Eight pages of references and several figures are included. (Author/LMO)