by Morton Hunt
—Some experiments have explored how we recognize a shape we are searching for, particularly when it is lost in a jumble of other shapes. One important process is “feature detection”—conscious searching for known and recognizable elements of a particular figure so as to distinguish it from similar shapes. In each of the following columns there is a single X. If you time your own search for it with a sweep-second hand, you will find that you locate it far faster in the first list than in the second:25
ORDQCG WEFIMZ
CRUDOQ EVLMZW
QUORDC VIMWZE
CUORCD ZIVFEW
DROCUD VIZELM
DOCURD MFWIVZ
DRGCOD ZVXIEW
ORCDUQ WVLZIE
ODQRUC EWMZFI
DRXOQU MEZFIV
DUGQOR IWEMVZ
RGODUC WEZMFV
GCUDOC EFLMIV
DGOCDR WZIEFV
The task of matching the pattern of the X retrieved from memory to what we are looking at is far easier and quicker when the hidden X is among rounded letters than among letters made up of straight lines and angles, like the X itself, in which case we must pay close attention to minor features. Or, as another explanation has it, we often identify visual images by “preattentive” processes—automatic ones concerned with overall image—but when that does not suffice, we shift to “focused attention” and consciously search for minor distinguishing features of the sought object.26
—In 1954 Fred Attneave, of the University of Oregon, asked subjects to represent certain figures by a series of ten dots; they tended to place the dots on those points where the direction of the outline changed most sharply. Attneave’s conclusion was that one way we recognize patterns is by means of analysis of its “points of change.”27 He also created some figures, greatly simplified from reality, by drawing straight lines from one point of change to another. Although this reduces curves to straight lines, the figures are still immediately recognizable, as in this example:
FIGURE 29
No curve exists, yet one sees a curved object.
—Skilled readers see words as wholes, without identifying them letter by letter, as beginning readers do. But even in rapid reading, a great deal of high-speed feature detection goes on, as shown by experiments conducted by Eleanor J. Gibson (the wife of James Gibson, mentioned above) and colleagues at Cornell in the 1960s. They made up a batch of nonsense monosyllables, some of which obey English rules of spelling and therefore are pronounceable (“glurck,” “clerft”) and then switched the consonant groups around to make others, with the same letters, that violate the rules and are not pronounceable (“rckugl,” “ftercl”). When skilled readers saw the words in tachistoscopic flashes, they could read the legal combinations far more easily than the illegal ones, even though none of the letter groups was a known word. One possible explanation was that they pronounced the words to themselves and were better able to hold pronounceable ones in short-term memory than unpronounceable ones. But Gibson repeated the experiment at Gallaudet College with deaf students who had never heard words pronounced, and she got the same results. This could only mean that in perceiving each pseudo-word, readers distinguished all the letters and instantly recognized which groups of them obeyed the rules of legitimate patterns of spelling in English and which did not.28
—Researchers working with visual illusions found that if subjects were instructed to look long at an illusion, and in some cases to let their eyes wander back and forth over it, the force of the illusion would wane. Even though the cues in the illusion mislead the mind, attentive looking enables the mind to extract much of the reality from the cues.29
—In the late 1950s and early 1960s, Irvin Rock, a young psychologist who would become a leading figure in perception research, showed subjects a square tilted at 45 degrees and asked them what it looked like; they said a diamond. He then tilted them by 45 degrees, causing the figure to be projected as a square on their retinas. But they saw it in a room with respect to which it was tilted and could feel themselves tilted with respect to that room; these two sources of information, processed by the mind, caused them still to see the square as a diamond. This simple experiment profoundly influenced Rock’s thinking about perception and led him to conclude that until perceptual phenomena have been analyzed from a psychological viewpoint, it is premature to do so on a neurophysiological level.30
These findings, and many more from studies made in the succeeding decades, made it clear that form is the most important cue for object recognition. Early in life toddlers learn to identify objects by their shape; they quickly gain the ability to distinguish between a dog and a cat, and having learned what an apple is, they recognize green ones, yellow ones and red ones as apples. Not long ago the psychologist Barbara Landau showed three-year-olds a meaningless shape and told them it was a “dax”; then she showed them other objects with the same shape but made of different materials, sizes, and colors, but the children identified each of them as a “dax.”31
And yet to this day, say Michael Gazzaniga and Todd Heatherton, “how we are able to extract an object’s form from the image on our retina is still somewhat mysterious.”32 They cite such commonplace mysteries as our ability to recognize objects from different perspectives and in unusual orientations, and to tell where one object ends and another begins, as in the case of a horse and rider. Hypotheses about how we do it are plentiful; proven theories are nonexistent.
From the 1940s on, neurophysiologists were making discoveries about visual perception that were as significant as those of the cognitivists. As early as the 1930s, they were able to record the electrical activity of small groups of nerve cells, and by the 1940s laboratory researchers had perfected glass probes containing electrodes so fine—the hairlike tip might be a thousandth of a centimer in diameter—that they could be inserted into a single cell of the retina, geniculate body, or visual cortex of a cat or a monkey that had been given local anesthetic. With this kind of apparatus, researchers could observe the individual cell’s electrical discharges when the animal was shown a light or some other display.33
This technique produced a historic discovery about form perception. In the late 1950s, David Hubel and Torsten Wiesel, two brilliant neurophysiologists at Harvard Medical School, were testing the responses of visual cortical cells in cats. They would implant a microelectrode in a cell of a cat’s visual cortex; although they could not pick a particular cell, by inserting the probe at about the right spot and right angle they knew what area they were reaching. Wiesel once likened the process to spearing cherries from a bowl with a toothpick; you may not be able to see which cherry you’re spearing, but you’re sure to hit something. The cat would be restrained in a harness while the researchers shone spots or bars of light and other figures on a screen. By securely fixing the position of the cat’s head, the researchers could know which part of the retina the image fell on and link this with the cortical area being probed. Through an amplifier and loudspeaker, they could hear the cell fire; at rest, it might produce a few “pops” per second, but chatter away at fifty or a hundred pops per second when stimulated.34
Since both the retina and the cortex are complicated structures, it took great patience to discover which cells, at what location and in which layer of the cortex, respond to messages from different areas of the retina.35 One day in 1958, this excruciatingly fine-detailed work yielded an astonishing and half-accidental finding. Hubel and Wiesel had positioned an electrode tip in a cell but for hours couldn’t induce rapid firing. As Hubel recalled, a few years ago:
We tried everything short of standing on our heads to get it to fire. (It did fire spontaneously from time to time, as most cortical cells do, but we had a hard time convincing ourselves that our stimuli had caused any of that activity.) To stimulate, we were using mostly white circular spots and black spots. After about five hours of struggle, we suddenly had the impression that the glass with the [black] dot was occasionally producing a response, but the response seemed to have little to
do with the dot. Eventually we caught on: it was the sharp but faint shadow cast by the edge of the glass [slide] as we slid it into the slot that was doing the trick. We soon convinced ourselves that the edge worked only when its shadow was swept across one small part of the retina and that the sweeping had to be done with the edge in one particular orientation.36
In short, the cell responded strongly to a horizontal line or edge but only weakly or not at all to a dot, a tilted line, or a vertical line.
Hubel and Wiesel (and other researchers) went on to show that some other cells are specifically responsive to lines at an angle or to vertical lines or to right angles or to distinct edges (where there is a contrast between an object and what surrounds it). It became clear that the cells of the visual cortex are so specialized that they respond only to particular details of images on the retina. Hubel and Wiesel won a Nobel Prize in 1981 for this and related brain research.
One bizarre offshoot of the Hubel and Wiesel work was the notion of the “grandmother cell”—a parody, by J. Y. Lettvin, at the time of Hubel and Wiesel’s work, of what he considered the simplistic notion that single neurons in the brain might detect and represent every object, including one’s grandmother. The parody had enough appeal to be seriously considered by some perception specialists but actually became shorthand for all the overwhelming practical arguments against a one-to-one object coding scheme.37
In any case, Hubel and Wiesel’s line-detector cells are a proven reality. Interestingly, this response is partly acquired, even though it is neurological. In a 1970 experiment kittens were reared in a vertical cylinder lined with vertical stripes and never saw horizontal lines. When they were tested for vision, at five months, they were blind to horizontal lines or objects. The neural explanation is that the cortical cells that respond to horizontal lines had failed to develop during the early stages of the kittens’ lives.38 Similarly, people reared in cities have more exposure to vertical and horizontal lines during early childhood than to lines oriented otherwise, and develop a greater sensitivity to the former. A research team tested a group of city-reared college students and a group of Cree Indians who grew up in traditional tents and lodges with few verticals and horizontals. The city-reared students exhibited the oblique effect; the Crees did not.39
You can also experience the specificity of the vertical, horizontal, and oblique detector cells of your retina by staring fixedly at the center of this pattern:
FIGURE 30
A pattern that confuses the retina’s line-detector cells
The whirling and vibrating you see are probably due to the fact that when you look at the center, where rays of varied angles are close together, the eye’s continual movements cause the image on the retina to shift from one kind of angled line to another, sending a jumble of signals that confuses the cortical receptors of specialized directional sensitivity.
The line detection ability of specific neurons is also exemplified by the following two displays, in each of which one object “pops out” because its lines have a unique stimulus property for those neurons:
FIGURE 31
Line Detection: The one inconsistent figure in each set “pops out.”
The microelectrode technique enabled neurophysiologists to decipher the architecture of the visual cortex—the neurons are arranged vertically, about a hundred in a column, and in layers that run through the columns—and to measure the responses of neurons in every part of the visual cortex to a broad variety of stimuli. The result was a detailed picture of how different cells in different parts of the visual cortex distinguish among all sorts of shapes, contrasts in brightness, colors, movements, and depth cues. A neuron-to-neuron and column-to-column synaptic hook-up of immense complexity links the responses of all these cells, presenting the brain with a composite message of the coded information of what had been a retinal image.40
Where and how that assembled message is “seen” by the mind was not apparent, although it was clear, from much of the cognitive perception research, that the specialized responses of the visual cortical cells are not the final product, at least not in human beings. In simple animals the neuronal responses may be enough to produce appropriate action (either flight or attack). In human beings, the neural messages are often meaningless until they are interpreted by cognitive processes. In the case of the illusory triangle, the viewer’s mind, not cortical cells, supplies the missing parts of the figure. The same is true of many other incomplete or degraded images, where the viewer, consciously invoking higher mental processes, fills in the missing parts and sees what is not there. A case in point:
FIGURE 32
A degraded image. What is it?
At first, most people see this figure (by Irvin Rock) as a meaningless array of dark fragments. How the reversal to the white regions and to perception of the hidden word comes about is not known, but once it has been seen, the mind is almost unable to see the figure again as meaningless fragments.
Seeing Movement
The metaphor of the eye as camera implies that we see the world in snapshots, but our visual experience is one of unbroken movement. Indeed, the perception of our movement through the environment and the movement of things in the environment is one of the most important aspects of seeing. Vision without perception of movement would be almost valueless, perhaps even worse than no vision, to judge from a rare case reported in the journal Brain in 1983.
The patient was a woman who was admitted to a hospital after experiencing severe headaches, vertigo, nausea, and, worst of all, a disabling loss of the perception of movement. A brain scan and other tests showed that she had suffered damage to a part of the cerebral cortex outside the primary visual receiving area that is known to be crucial to movement perception.41 From the report:
[She had] a loss of movement vision in all three dimensions. She had difficulty, for example, in pouring tea or coffee into a cup because the fluid appeared to be frozen, like a glacier. In addition, she could not stop pouring at the right time since she was unable to perceive the movement in the cup (or a pot) when the fluid rose…In a room where more than two other people were walking she felt very insecure and unwell, and usually left the room immediately, because “people were suddenly here or there but I have not seen them moving”…She could not cross the street because of her inability to judge the speed of a car, but she could identify the car itself without difficulty. “When I’m looking at the car first, it seems far away. But then, when I want to cross the road, suddenly the car is very near.”42
Even without such evidence, we can tell that movement perception is of paramount importance. Perception of our own movement guides us in making our way through our environment; perception of objects coming toward us enables us to escape harm; perception of the movement of our hands provides data vital to control when we are reaching for an object or doing fine manual work; perception of our minute bodily movements when standing keeps us from weaving or losing balance. (If you stand with your feet close together and shut your eyes, you will find it difficult to remain perfectly steady.)
Much research on movement perception for the past half-century has dealt with external variables: how the size, speed, location, and other characteristics of moving objects affect the way they appear to us. Such research is akin to psychophysics: it gathers objective data but says nothing of the internal processes responsible for the experiences. Still, it has provided important clues to those processes, both of the innate neural and the acquired cognitive kinds.
A typical finding about an innate low-level process: Researchers projected a shadow or boxlike figure on a screen in front of infants, then made the shadow or figure rapidly expand. When it did, the infants reared back as if to avoid being hit. The reaction is not a result of experience; a newborn who has never been hit by an approaching object will react in this fashion, as will many young and inexperienced animals. The avoidance response to a “looming” figure is evidently a protective reflex built into us by evolution; the visual impre
ssion of an object coming at us triggers escape behavior without involving higher mental processes.43
A typical finding about an acquired high-level process: In 1974 the psychologists David Lee and Eric Aronson built a floorless little room that could be slid one way or another across an unmoving floor. When they placed in it a toddler of anywhere from thirteen to sixteen months and slid the room in the direction he or she was facing—that is, away from the child’s face—the child would lean forward or fall; if they slid it in the other direction, the child would lean backward or fall. The explanation is that when the walls moved away, the child felt as if he or she were falling backward and automatically tried to compensate by leaning forward, and vice versa. This seems to be acquired behavior. The child learns to use “optic flow” information when beginning to walk. (Optic flow is the movement of everything within our visual field when we move. As we walk toward some point, for instance, everything around it expands outward toward the limits of our vision.)44
These and other fruitful studies of movement perception revealed additional defects of the long-held notion that the eye is a kind of camera. One such defect is that although the eye has no shutter, moving objects do not cause a blur, nor do we see a blur when we move our eyes as we do on a photograph if the camera is moved during exposure. Accordingly, much research on motion perception has sought to discover why there is no blurring. One hypothesis that gained favor was based on the finding by Ulric Neisser and various others that when we view an image flashed on a screen by a tachistoscope for even a tiny fraction of a second, we can briefly see it afterward in the mind’s eye. In 1967 Neisser used the term “icon” for this very brief visual memory, measured its duration as about half a second (later research reported as little as a quarter of a second) to two seconds, and found that it is erased if a new pattern is presented before it has faded.45 Other vision researchers then suggested that since the eye sweeps across the field of vision or follows moving objects in a series of jumps known as “saccades,” it sees nothing while moving but at every momentary stop sends an iconic snapshot to the brain. The snapshots are assembled there into a perception of motion, somewhat as if one were watching a movie.46
