The design of the human eye was necessary to meet the competing evolutionary demands for high visual acuity and a large field of view. There is simply not enough neural real estate available in the brain to support a visual system that has high resolution over the required field of view. Even if we left no room in the cortex for any other senses (not to mention housekeeping functions like breathing or keeping the heart beating), the human cortex could not support the optimal size/resolution sensor. Some animals stay within the design limits by restricting their field of view (e.g., a hawk); others give up high resolution in favor of a larger field of view (e.g., a rabbit). Rather than picking one or the other solution, humans evolved the anisotropic retina with very high spatial resolution in the center of the visual field (the fovea), surrounded by a much lower resolution region (the peripheral retina). In the human retina, the high-resolution fovea encompasses less than 0.1% of the visual field visible at any instant, and the effective resolution falls by an order of magnitude within a few degrees from the fovea. This variable-resolution retina reduces bandwidth sufficiently, but is not an acceptable solution alone. Unless the point of interest at any moment happened to fall in the exact center of the visual field, the stimulus would be relegated to the low-resolution periphery. The 'foveal compromise' was made feasible by the evolution of a complementary mechanism to move the eyes. In order to ensure useful vision, the eyes must be moved rapidly about the scene.

The first job of an eye movement system is to move the eye quickly from the current point of gaze to a new location. Vision is blurred during an eye movement, so the length of time that the eye is moving must be minimized. In order to minimize the time during which no clear image is captured on the fovea, eye movements that move the fovea from one object/point to another are very rapid. These saccadic eye movements are among the fastest movements the body can make; the eyes can rotate at over 500 deg/sec, and subjects make well over one hundred thousand of these saccades daily. These rapid eye movements are accomplished by a set of six muscles attached to the outside of each eye. They are arranged in three pairs of agonist-antagonist pairs; one pair rotates the eye horizontally (left - right), the second rotates the eye vertically (up - down), the third allows 'cyclotorsion,' or rotation about the line of sight.

The second class of eye movements maintains clear vision by stabilizing the retinal image. This stabilization assures that the image of an object or region in the center of the field-of-view is kept over the fovea. Sophisticated mechanisms exist to accomplish this goal in the face of eye, head, body, and object motion. These eye movements are often grouped into four categories:

The vestibular-ocular reflex (VOR) rotates the eyes to compensate for head rotation and translation. Rotational and linear acceleration are detected by the semicircular canals and otolith organs in the inner ear. The resultant signals are used to command compensating eye movements.

Optokinesis stabilizes the retinal image caused by large-field motion. Retinal slip induced by field motion is used to initiate eye movements at the appropriate rate to cancel out image motion.

Smooth-pursuit eye movements are similar to optokinesis, but allow arbitrarily sized targets to be stabilized instead of large-field motion. A moving target is required for smooth eye movements; the eyes cannot move smoothly across a stationary object.

Vergence eye movements counter-rotate the eyes to maintain the images of an object at a given depth to be maintained at corresponding locations on the two retinae.