Tom Jonard's Perceptual Filling In PageOur senses are the fundamental mechanism by which we are connected to our world. They sit between the interior world of our minds and the exterior world through which we move. In the history of the Philosophy of the Mind much attention has been paid to this interface and rightfully so. Our experience is so intimately mediated by our senses that we easily take them for granted and seldom consider their workings. We see with the eyes without perceiving the very act of seeing. So too we hear, taste, smell and touch without perceiving any of the mechanisms of the senses. Yet how we perceive is likely to effect our relationship with the world and also with the body which is our physical presence in that world. Just ask any blind person.
Filling in is one of those perceptual phenomenon that is just out of view most of the time. Most of us are familiar with it in the form of the visual blind spot. Hopefully in a grade school science class you were told about the fact that the retina in each of your eyes has a blind spot where there are no light receptors. This occurs at the point where the optic nerve exits the eye. It is to one side of the central line of sight in each eye so that when using both eyes the entire visual field is covered with receptors in at least one eye. Yet when you close an eye the blind spot in that eye is not readily visible. You need to know where to look but you can discover it.
The odd thing about the blind spot is that no one sees it even though at 6o it is 36 times the size of the full moon! Even if you do one of the exercises to discover it you still don't actually see it. What you see is that a spot or mark that enters the blind spot disappears. As an astronomer people ask me what a black hole is like. Well the blind spot is sort of like a black hole in the visual field. You can only see it indirectly when something disappears into it. There is no corresponding black spot in your visual field. As a matter of fact your visual field appears to be continuous right across the blind spot. There is no sense of anything missing in the blind spot because there is nothing there -- no visual receptors and therefore no visual field. We don't just ignore these two holes in our vision -- they simply aren't there.
The retina of the eye is not uniformly populated with light receptors. A spot in the center of the line of sight called the fovea has the densest aggregation of light sensitive cells. Moving away from that point the number of these cells per unit area falls off. The type of cell also varies over the surface of the retina. In the fovea the color sensitive cone cells predominate. White light sensitive rod cells predominate in the periphery surrounding the fovea. All of these variations in the visual field are ignored by the visual system. We think that we see the periphery of our vision with the same acuity and color content as the central portion. This is not true. In fact much of what we "see" there is filled in from our visual memory of what we have seen or what we expect to see there based on what we can see.
Sometimes what we see in our peripheral vision is filled in after the fact. This is shown by experiments in which subjects read text from a computer screen while the movements of their eyes across the text were tracked by the same computer. When the computer detected the start of a move from one word to another it erased the second word and replaced it with another of the same length. Subjects did not notice this change. Other observers saw the text flicking as the subject read. For subjects the text appeared as stable as a printed page. For other observers it was hard to believe that this could be so. Remarkably subjects who observe the experiment first and therefore know what is going on are still unable to detect that the experiment is working.
This experiment touches on another phenomenon of vision that is totally transparent to us -- the saccade (from the French for a sudden movement or jerk). It turns out that when we think we are gazing steadily at anything in our field of view we are not. The eye exhibits continuous jerky motions even during the steadiest gaze. The reason maybe that because the area of the fovea where highest visual acuity is possible is rather small constant shifts in the position of the retina allow the visual system to scan the visual field for detail and build up a visual memory which allows us to fill in areas at which we are not directly looking.
Another reason for saccades may be that our visual system primarily detects motion. If an image is fixed on the retina for a few seconds it disappears. This was originally demonstrated by using a very small projector attached to a contact lens so that shapes it projected into the eye moved with the eye. Subjects wearing this apparatus reported that the shapes disappeared over several seconds. You can come close to the same result (as I have done) by picking a point and staring at it. Almost everything in your field of vision surrounding the point will shortly disappear into a white fog -- at least until an involuntary eye movement brings it all back.
Apparently saccades also serve to prevent the visual system from losing sight of the stationary world. In a system that developed through evolution this functionality makes sense. Vision that brings to our attention any motion in our visual field would be adaptive for sensing environmental threats. If that system had the side effect that anything that didn't move would disappear from our attention then saccades would be an adaptive modification that corrects for it. The result is a visual system that is adaptively tuned to detect movement and yet also not lose sight of the fixed background.
From our perspective the most interesting thing about saccades is that we don't notice them. If we did our vision might be something like those handheld camera shots you see in movies or on TV. But vision is not like that. The jerky motions are ignored by the visual system. In fact as eye tracking experiments like the one described above show something even more unexpected happens. During a saccade visual perception is inhibited. This phenomenon is known as saccade blindness. The subjects in the eye tracking reading experiments do not see the word change before reading it. In fact in their visual field it either does not change or some sort of fill in editing occurs so that they perceive that the word has always been the one they see at last even though other observers know this is not true.
Visual perception has been shown to be inhibited over the whole visual field during a saccade. In testing general vision instead of just reading other experiments show that any disruption in the visual field (blinking, flashing or a sceen cut in a move) can produce a similar change blindness (see also here and some examples). The size of a background building in a scene or the number of people in a crowd can change under these circumstances without the viewer noticing. Instead of being like a handheld video vision is more like a series of still photos each taken from a slightly different angle. As these still photos are presented to the mind the previous one is hidden from view. With no other evidence we believe that the last view was substantially like the current one and if we notice some detail in the current view we believe it to have been in the previous one as well as all its predecessors. Even if this is not true.
We say, "Seeing is believing", and that would appear to be true but not in the way commonly meant. For what is commonly meant by this saying is that seeing is a way of verifying the way things are. In fact experiment shows that seeing is quite literally believing -- we do believe what we see is the way things are but what we see may not actually be the way things are. We believe that what we see here and now is what we saw then and there or what we will see when and where. But it may be that what we believe we see is not really anywhere. We simply do not see the whole picture the way that we believe that we do.
Though I and the researchers in cognitive science have concentrated on vision do not suppose that these phenomenon are unique among all the senses. In hearing for instance we all perceive stereo sound as three dimensional when we listen to recorded music even though the true source is one or more pairs of speakers. Many other examples exist in this and the other senses that we won't go into here. The evidence from vision is enough to make a point.
We have avoided referring to these phenomenon as illusions because that might be misleading. The word illusion suggests that we sense something that is not there or are otherwise deceived. It does not seem right to call any normal process of perception an illusion. It might be better to say that our perceptions are constructions that allow us to make the most out of our sensory input. This input is often less than we think it is. In everyday life where clever experimentalists are not trying to trick our perceptual apparatus into revealing its secrets this may not be a problem. The construction is a normally reliable representation of the external world as long as that world continues to meet our expectations -- expectations that are the result of evolutionary adaptations.
There is not a direct connection between our perceptions and the external world. When the world ceases to conform to our evolutionarily formed expectations the perceptions constructed by our minds can be seriously in error. We use the term eyewitness to emphasize the veracity of some report about the world. But the phenomenon we have discussed show that what we see cannot be universally naively trusted. Whatever we otherwise perceive may likewise be questioned. Our sensory connection to the external world has built-in tricks, shortcuts and edits that most of the time facilitate our perception of reality. But they may also inhibit it in ways that are hidden from us because that is one of the tricks. What we witness with our eyes and other senses may in fact not be what is.
.Return to Tom Jonard's Consciousness and Mind page
Created October 4, 2002,
© 2002, Thomas A. Jonard