Vision first evolved as a system for the distal control of movement. The later emergence of mechanisms for the perceptual representation of the world has enabled animals to be more flexible in their behaviour – with respect to both goal selection and action planning. But even in species with well-developed visual perception, the final programming of the movements still depends on separate action pathways. As it turns out, this distinction between vision-for-action and vision-for-perception can be seen in the organization of the visual pathways of the primate brain. Beyond primary visual cortex in the primate cerebral cortex, visual information is conveyed to a bewildering number of extrastriate areas. Despite the complexity of the interconnections between these different areas, two broad "streams" of projections from area V1 have been identified: a ventral stream projecting eventually to the inferotemporal cortex and a dorsal stream projecting to the posterior parietal cortex.

Evidence from neuropsychological studies in humans suggests that the dorsal stream is a visual ‘action’ pathway and that it contributes little to the contents of our perception of the world. It has been known for a long time, for example, that patients with damage to the dorsal stream, particularly the superior parts of the posterior parietal cortex, have difficulty reaching in the correct direction to objects (even though they have no difficulty reaching out and grasping different parts of their own body indicated by the experimenter). Some patients with damage to this region show an inability to rotate their hand or open their fingers properly to grasp an object placed in front of them, even when it is always placed in the same location. Nevertheless, despite their problems in visuomotor control, patients with damage to the dorsal stream are usually able to describe the orientation, size, shape, and even the relative spatial location of the very objects they are unable to reach towards, grasp correctly, or foveate.

The ventral stream projecting to the temporal lobe has been more closely linked with visual perception than with the visual control of action. Patients with brain damage in this region show the complementary pattern of deficits and spared visual abilities from patients with damage in the posterior parietal cortex. Thus, patients with ventral-stream damage have great difficulty recognizing common objects on the basis of their visual appearance, but have no problem grasping objects placed in front of them or moving through the world without bumping into things. In short, they can act on objects in the world despite the fact they cannot perceive the very object features controlling their movements.

The idea of two visual systems in a single brain might seem initially counter-intuitive. Our visual experience of the world is so compelling that it is hard to believe that some other quite independent visual signal – one that we are unaware of – is guiding our movements. After all, it seems obvious that it is the same subjective image that allows us both to recognize the coffee cup on our desk and to pick it up. But this belief is an illusion. As work with neurological patients shows us (and this has been confirmed in recent brain imaging studies), the visual signals that give us our experience of the cup are not the same ones that guide our hand as we pick up it up!