It's a Jungle in There: How Competition and Cooperation in the Brain Shape the Mind (18 page)

What accounts for this
waterfall illusion
, as it’s called? The basis for the effect is similar to the one that produces color negative effects. Neural elements specially tuned to an aspect of visual stimulation—downward motion rather than specific hues—get excited, but because they’re excited for a long time, their responsiveness declines. The continual signaling of downward motion ceases to be informative, so the outputs of the relevant cells diminish or are discounted by other neural subsystems. The result is that neural pools sensitive to motion in other directions contribute more to the global visual experience than they otherwise would. Neural pools for rightward and leftward motion continue to fire, but they oppose each another and their outputs cancel, but neural pools for upward motion get to have a much stronger voice than they normally would. As a result, there is an illusion of upward motion, a result that would be expected only if the contribution of the up-motion population carried more weight than the contribution of the down-motion population.
⁹

Motion aftereffects are not limited to linear motion. They also arise after prolonged observation of rotary motion. Suppose you watch a swirling spiral that seems to move away from you. Later, when you look at a stationary object, it will seem to approach you. The opposite outcome occurs if you watch a swirling spiral that seems to move toward you and then you look at a stationary object. That stationary object will seem to recede.
¹⁰

The neural dynamics of such effects are similar to the opponent processes for up and down motion. Cells for approach “bounce back” when cells for recession have fired for a long time and then stop being supported by external inputs. Similarly, cells for recession recover when cells for approach have been activated for a long time and then get no support from the external world.
¹¹

In the waterfall illusion, you have the sense of upward motion, but you can tell the hill you’re looking at isn’t really rising. You have no basis for believing the hill’s on the rise, and you have plenty of basis for believing it’s staying put. Your gaze angle doesn’t increase as you look at the cliff beyond the stream. Your head doesn’t tilt up more between successive glances at the slope. You don’t
feel the ground rumbling before your feet or see rocks and pebbles sliding down the hill as they would if the earth were heaving.

Inference lets you know that motion hasn’t occurred, but in other situations, inference lets you know that motion
has
occurred. A familiar example is noticing that the minute hand on an analog clock is not where it was before. You may not see the minute hand move, but you can infer that it must have because it’s not where it was before.

Why am I speaking of inference? What does inference have to do with perception? Should inference be taken seriously as a factor in perceptual experience? How does perceptual inference, if there is such a thing, relate to the larger claim that it’s a jungle in there?

Drawing inferences about visually sampled events entails drawing conclusions based on more information than first meets the eye. If you infer that a man is unmarried because he’s a Roman Catholic priest, you’ve reached a point you weren’t at initially. You got there without being told the priest is a bachelor.

The ability to draw inferences is what the Greek philosopher Aristotle pointed to when he drew attention to logic as a way of learning. Via logic, Aristotle argued, you can know more than you would from data alone. If Aristotle had commented on mechanical clock-watching, he would have said that you can infer from a clock that time has passed, because the clock’s minute hand occupies a different position that it did a while back.

Inferring that a clock’s hand has moved is one way of perceiving motion, but it’s not the only way. The other is perceiving motion directly. If an object passes from one place to another before your eyes at a sufficiently high rate, its passage through space and time are immediately apparent.

Returning now to the example of the mysteriously rising hill, you can decide that the hill doesn’t rise after watching a stream cascade downward because you don’t get other information to support the hill-rising inference. By contrast, you
can
decide that the water is really descending because you can see bits of debris being washed down the waterfall, you can hear the water falling, you can stick your hand into the stream and feel the water running through your fingers, or you can ford the stream at the base of the waterfall and feel the water race around your toes.

The idea that perception relies on inference was advanced by an intellectual giant of nineteenth-century science, Herman von Helmholtz (1821–1894). Helmholtz invented the ophthalmoscope, the device used by doctors to view the inside of the eye. As mentioned in the last chapter, he also measured the speed of nerve conduction.
¹²

Helmholtz argued that inference helps perception. He also suggested that perceptual inference is largely unconscious. In saying this, he pointed to cognitive processes beneath conscious awareness, much as Freud did, but Helmholtz’s suggestion is less well known, perhaps because the subject of seeing is less sexy than the subject of sex. Helmholtz suggested that perception entails bringing together, without awareness, cues that enable us to make the best guesses we can about what’s in the external environment.

To help Helmholtz think through the nature of such perceptual detective work, he contemplated illusions such as the ones discussed here. Consider another illusion associated with a special room built years after Helmholtz’s passing. The room was built by an American ophthalmologist named Adelbert Ames Jr.
¹³
The design of the room was meant to trick you. It looked, for all intents and purposes, like a perfectly ordinary room, but it was far from ordinary. Whereas a normal room is rectangular (or square), this room was trapezoidal. Owing to its design, if you peered into the room and saw two people of objectively equal size standing in it (one in the front and one in the back), one person would appear gigantic compared to the other. If the two individuals traded places, the one who looked large before would now look tiny, and vice versa.

The basis for the illusion is that the optical cues from the room indicate that the person who is farther away is closer than he or she really is. To make sense of the apparently greater proximity of that individual, the brain “says” in effect, “If the individual is only so high when he or she is so close, then he or she must be extremely short.”

Cues like those used for the Ames room are exploited by artists to convey the impression that some depicted objects are farther away than others. This is a neat trick considering that the objects shown in drawings and paintings lie on flat planes. That the trick is of use to you in your daily life is shown by the fact that, at least to a first approximation, the retinas of your eyes are flat planes too.

The way you see depth is by relying, in part, on the visual cues just mentioned, which are monocular (one-eye) cues to depth. There are also
binocular
(two-eye) depth cues. The main binocular cue to depth is the disparity between the retinal projections of objects on your two retinas. The fact that you have two eyes not only provides you with cues to depth based on retinal disparity. Your two-eyedness also provides you with a whole raft of fascinating
phenomena subsumed under the rubric of
binocular rivalry
. Rivalry is, of course, another word for competition. Inputs from your two eyes converge to let you see. Along the way, however, there is fierce antagonism.
¹⁴
Seeing as you do is due to the pushes and pulls of neural gnomes vying for a piece of the visual action.

Scientists who are attracted to Helmholtz’s perspective have sought to further his inferential approach to perception.
¹⁵
It is worth listing a few examples of perceptual phenomena that convey the value of this approach because they illustrate how putative agents in the brain can exert considerable influence on perceptual experience.

Focus first on seeing more than is there. One example is related to the
blind spot
. It turns out that one part of your retina has no photoreceptors. This is the part of your retina where your optic nerve leaves your eye and courses inward toward your brain. Oddly, your optic nerve—the nerve that carries signals from your photoreceptors to your brain—doesn’t begin
behind
your retina and travel back to your brain. Instead, it begins
in front of
your retina and then makes a U-turn, passing back through the retina toward structures within your brain. Because there are no photoreceptors where the optic nerve passes, there is literally no way to see whatever light might be projected there. Nevertheless, you have no awareness of such invisibility.

You can demonstrate the blind spot for yourself by stretching your arm out before you and holding your thumb up while staring at it with one eye open. As you slowly vary the angle of your outstretched arm, you’ll notice that your thumb vanishes at some critical angle. When this visual amputation occurs, the image of your thumb is cast on the part of the retina where your optic nerve passes through your retina. The fact that you normally don’t experience this blind spot indicates that your brain “fills in” the hole. Neural elements receiving inputs in the areas surrounding the blind spot extend their influence to the area where no visual input is received. The no-input area remains vulnerable to whatever its visually powerful neighbors signal. The result is seeing more than is there, as would be expected from massive cooperation and competition among neural elements involved in visual processing.
¹⁶

Now consider nother exmple of seeing more thn’s there. This is illustrted in the wy this prgrph is written. You probbly hve little trouble reding this. Your brin fills in the missing informtion though the first letter of the lphbet hs been removed.

This next paragraph—the one you’re now reading—returns to proper spelling, and illustrates yet another example filling in. If you’re focusing on the meaning this paragraph, you probably don’t have much trouble reading it. You may not even notice that anything’s wrong with it, but something is.

What was wrong with the last paragraph? You may have to go back to check. The answer is that I removed the word “of” twice. Quite possibly, you noticed only one the missing ofs. To keep up the mischief, I deleted yet another “of” just now, one sentence ago. Did you notice? If you didn’t and if you had no trouble understanding the text, that’s because your brain filled in the missing information. In terms of the metaphor running through this book—the metaphor of little demons seeking evidence for their favorite input—it wasn’t critical that the letters “o” and “f” appeared. The contexts supporting “of” allowed their neural representatives to get so excited that they yelled loudly enough for you to be fooled into thinking the “ofs” were there.

This phenomenon is so robust that writers must be on guard for it. Otherwise, they fall prey to the
proofreader’s error
, the tendency to miss textual mistakes because of expectations (conscious or unconscious) about what should appear. The proofreader’s error can be maddening. You may be as vigilant as possible when you check for typos, but they escape your notice, as if there’s a conspiracy in your head not to notice those bloopers. There
is
such a conspiracy! Often, it’s only when you come back to text you’ve worked on for a long time and read it “with fresh eyes” that you notice mistakes you missed before. The reason why there’s a higher chance of spotting errors later is that the dynamics of the inner ruckus have settled down.
¹⁷

Filling in linguistic information isn’t limited to reading. It also occurs in listening. You rarely hear speech in perfectly pristine acoustic conditions, with each syllable being pronounced with Shakespearean clarity. Imagine a short-order cook yelling to his helper, “Gimme a nummer 8!” The helper gets the message in the midst of clattering dishes, clanging spatulas, sizzling bacon, and whistling kettles. If some part of the message is masked by extraneous noise, it doesn’t matter. The helper fills in the missing information.

The ability to fill in missing speech sounds is called the
phoneme restoration
effect. It’s the auditory analogue of the proofreader’s error. Like the proofreader’s error, the phonemic restoration effect is clearest when the material that is removed is highly predictable.
¹⁸

There are many other examples of perceiving more than is there. Imagine having a conversation with someone who happily chews bubblegum while speaking to you. At some point, your chewy chum blows a big bubble. The
bubble conceals much of her face. Do you think the person’s mouth and nose have vanished? Of course not. Your brain fills in the missing parts.
¹⁹

The same thing happens if you see someone with his hand in his pocket. You don’t see an arm ending in a stump. If the person pulled his arm from his pocket and no hand emerged, you’d be astonished. Only if you had reason to expect a lost hand would you be unsurprised by its absence.

A final example is less visually arresting. (Not all visual phenomena need to dazzle, by the way; part of doing science is learning to recognize that small, humble things are significant.) Here’s the example. Are you ready? A child sees her dad pass behind a tree. That’s the end of the demonstration.

The point of this humble example is that though the image of the dad is momentarily interrupted, parts of him being hidden by the trunk, the child doesn’t see parts of her dad fade away and then come together again. What she sees is a coherent person. This is so unremarkable that it takes a moment to appreciate the sophistication permitting it.

Phenomena like these reflect
top-down
processing. The term refers to high-level interpretations biasing perception, so perception is not just dictated by immediate sensory data—
bottom-up
processing—but is also shaped by expectations.
²⁰
A child expects to see her dad reappear after he, or part of him, momentarily vanishes behind a tree.