On raising our sights

There are some very strong traditions in contemporary research on spatial vision that have arisen, in large measure, through advances in technology. This is especially true of research on the neurophysiology of vision, e.g. in recording from single cells, and in developments in computing that have promoted the study of machine vision. No one would dispute that these are major advances in the field, but one may ask whether they have unduly constrained the sorts of question that we ask about the visual processes. It is argued that they have, and that it is timely to stand back to see what other problems and questions have suffered neglect. Two items seem to be quite ripe for reconsideration: the question of how to understand the enormous adaptability of visual systems to sensory rearrangement (by means of distorting lenses, for example); and the very broad matter of keeping our perceptual processing in close and coherent touch both with stimulation, and with the properties of the physical/biological environment--what Gibson called the problem of veridical perception. It is argued that the two items are closely allied, and can be approached from a single theoretical stance. The one I favour is outlined, and I claim that it is one that has little to do, in the first instance, with either neurophysiology or computation. In that sense tackling the two questions involves raising our sights above the current traditions of vision research, but that of course does not in any way mean attempting to belittle the immense advances that those traditions have brought about.(ABSTRACT TRUNCATED AT 250 WORDS)