ABSTRACT
Visual search is concerned with processes for finding specified objects of unknown position. At any given moment the retinal fovea deals with only a narrow cone of visual space outside the eye. Therefore, the fovea has to be guided to the position of interest. It is one of the functions of eccentric vision, outside the fovea, to provide such guidance. Eccentric vision, in particular the size of the functional visual field, and control of the line of sight by head and eye movements are therefore basic to an understanding of search.
If, arbitrarily, head movements and moving objects are excluded from consideration, eye movements are restricted to jumps (saccades) that are separated by eye pauses. Since the jumps are quick, useful vision occurs during eye pauses only. Vision is then based upon a series of quasi-stationary retinal images which are shifted a few times each second. These basic facts lead to the following rather straightforward division of research problems in visual tasks: (1) vision from a brief stationary image, (2) the control of eye saccades, and (3) the integration of successive images into one continuous visual impression. Of course, the relations between these three separate types of process have to be established as well.
The aim of the present contribution is to show that this view has led to fruitful research aimed at understanding both visual search and reading. The approach is advocated for further studies of visual activities.
Studies on vision from brief stationary images have to concentrate on eccentric rather than on foveal vision, and on presentation durations representative of eye pause durations. They should also preferably concern structured rather than homogeneous visual fields or backgrounds. The latter distinction is of importance because of the strong adverse interactions between neighboring objects, which prevent studies with homogeneous backgrounds from being representative of many conditions in daily life. These lateral interferences are specific in the sense that certain stimulus properties interact selectively among each other over relatively long retinal distances. Although the interactions may extend not only in space but in time as well, the lateral interferences probably limit the functional visual field more than the effects of backward and forward masking.
Control of eye saccades is a difficult research problem in its own right. There is a wide range of processing levels where eye guidance may originate. A preliminary division might distinguish between (a) semiautonomous guidance by motor routines; (b) sensory and perceptual guidance by visual, auditory, and tactual objects and; (c) cognitive guidance by knowledge and expectation on as many different levels of complexity as one might wish to distinguish, including individually unique factors. These levels should be seen as cooperating rather than as being mutually exclusive. Indeed, eye saccades may reflect the complexities of man himself. Now that eye monitoring equipment is more generally available, the task of research might be to peel off from reality reduced situations in which relatively simple and amenable types of control appear dominant.
On the integration of visible information from successive images, some data exist that suggest that the visual system does not need advance information on eye saccades in order to integrate perceptually the incoming information. Some preliminary evidence suggests that eccentric recognition may sometimes be slower than foveal recognition, so that the internal recognition process may not always be all that successive.
Applications of the above research questions to processes of search and of reading are demonstrated.
