B Hypotheses: Always Under Construction

In earlier chapters, we talked about factors that can diminish the influence of the hypothesis—misinterpretation, overt opposition, and neglect, to name a few—but our unacknowledged tendency to generate hypotheses constantly contributes to the problem.

You might have asked another question: Why did you feel surprise? It is ob­vious why you were happy, but your surprise is different. You had expected something that didn't happen, and your expectation is essentially an implicit hy­pothesis that explained your previous academic performance: namely, that you were a somewhat above average student. This predicted an unremarkable exam grade, and the exam results had just falsified the prediction. Evidently, your hy­pothesis was wrong, and, in fact, the bolder hypothesis that you are a superior student may already be taking shape in your mind (although you sense that it will be wise to test the original one more rigorously before writing home about the new one).

Hypotheses like this one are trivial and unconsciousness. We generate and test them all the time and hardly realize that we're doing it until a reaction such as surprise suddenly brings them to light. In general, we don't merely register what's going on around us: we constantly try to understand it by making conjectures and comparing their predictions with our experiences. We'll need to take this persistent drive into account if we want to understand the influence of the hy­pothesis in scientific thinking. No doubt some readers will balk at the thought that an automatic, sense-making mental process should be dignified by calling it “hypothesis generation,” and it's true that these are low-level, proto-scientific constructs.

The philosopher Frances Bacon, though most famous for championing the causes of inductive reasoning and scientific experimentation, knew that our thoughts are not always trustworthy. He identified four classes of “idols of the mind”² that, he said, cause us to commit mental mistakes, and two of them are directly relevant here: the Idols of the Tribe encourage us to believe that our perceptions represent the world truly, whereas, in reality, the mind “is like a false mirror” that “distorts and discolors the nature of things by mingling its own nature with [them].” The Idols of the Tribe are rooted in our human na­ture; nowadays we would say that they are genetically programmed and so are universal human traits. In contrast, the Idols of the Cave are those idiosyn­cratic misperceptions that we acquire individually; they result from the unique collections of preconceptions and blind spots that we pick up through our so­cial and familial interactions, our education, etc. The Idols of the Mind and Idols of the Cave cause us to misinterpret the world unless we know about them and guard against their influences. Bacon wanted to teach his readers to avoid them by adopting the Scientific Method.

In 1738, David Hume³ also called attention to psychological factors in phi­losophy, but he was more specific than Bacon. Hume, you’ll recall, stripped away the aura of logical invincibility surrounding inductive reasoning, arguing that we believe in the validity of inductive reasoning for psychological reasons. We infer a cause-effect relationship between events from the indi­rect evidence of contiguity, priority (temporal order), and constant conjunc­tion.

While Hume’s analysis of the mind is valid as far as it goes, it doesn’t go far into the active mental processes of innate hypothesis generation. The Nobel Prize­winning neuroscientist Eric Kandel⁴ credits the nineteenth-century physicist and physiologist Hermann Helmholtz and the Gestalt psychology movement of the twentieth century with the insight that we are constantly making and testing hypotheses about the world. Helmholtz realized that our sensory systems are so crude that, if we didn't actively refine the raw information that we get from the environment, we wouldn't be able to function. “In fact, if the brain relied solely on the information it received from the eyes, vision would be impossible,” says Kandel, noting that Helmholtz “concluded that perception must also be based on a process of guessing and hypothesis testing” We rely on inspired guesswork to navigate our environment, though sometimes our guesses are inaccurate. The Gestalt psychologists added that we do not sense single stimuli in isolation, but as parts of integrated wholes; we do not hear a succession of musical notes, we hear a melody. We do not merely react to stimulation, we interpret and try to understand it and we do this automatically and unconsciously. Our perceptions therefore depend on extensive lower level “bottom-up” and higher level “top­down” types of processing.

11. B.1 The Brain Is an Organ for Making the World Intelligible

For many cognitive psychologists, prediction is “an overarching principle of brain function... in the service of promoting adaptive interactions with one's environment”⁵ Much of the baseline electrical activity of the brain, observed with electroencephalographic (EEG) or functional magnetic resonance imaging (fMRI) techniques when we are physically at rest, the brain's so-called default mode, is dedicated to this activity.⁶

Probably the most dramatic illustration of the urge to interpret comes from the classical “split brain” studies⁷ on patients who have had major brain sur­gery to treat their uncontrollable epileptic seizures. Seizures are storms of elec­trical activity that travel in a bundle of nerve cell fibers (the corpus callosum) between the two sides (hemispheres) of the brain. Surgically cutting the bundle keeps seizures from spreading and reduces their severity. The operation, how­ever, essentially leaves the patient with two independent mini-brains,⁸ and each one processes both sensory and motor (muscle movements) information sepa­rately from the other. For instance, each eye gets light from both right and left sides of the world (we won't worry about the binocular overlap for the moment) and sends the left-side information to the right side of the brain, and vice versa. Normally, information from both hemispheres funnels into a single interpreter⁹ region that puts it all together to create a seamless visual experience. Difficulties arose because the interpreter is housed in only one hemisphere, so cutting the bundle of fibers meant that split- brain patients couldn't integrate all the visual in­formation coming into both eyes; their hemispheres saw slightly different images of the world. Similarly, one hemisphere could control only one arm, one leg, etc., and not help coordinate the actions of both sides as it normally did. By studying how these patients with their two mini-brains coped with certain experimental challenges, the experimenters learned how the intact brain ordinarily processed integrated information.

To study the extent of the deficits that the patients suffered, the experimenters devised clever optical devices that let them independently control the visual in­formation that got to each hemisphere. Since one hand was controlled by one hemisphere, the experimenters could find out what visual information each hemisphere received by asking the patient to point with that hand to a picture of the object that was transited to the same hemisphere.

What happened when both hemispheres saw the same picture if the patient's hands were not in sight? Even when her hands were hidden beneath the top of the table at which she was sitting, they both pointed to the correct picture, as ex­pected. What happened if the hemispheres saw different images at the same time? This is where things became really interesting. If the patient's hands were beneath the table top, then she correctly pointed with each hand to the appropriate pic­ture but, as she couldn't see her hands, she was unaware (i.e., her interpreter was unaware) that they were pointing at different pictures. Now, what happened if both eyes saw different images and the patient could see both hands? Again, each hand pointed to the appropriate picture (which was different for each hand). And, because both hands were visible, the patient became aware that they were pointing to different pictures. This was a major conundrum because the patient's interpreter had no direct visual information from the other hemisphere about what image it saw. Remember, the interpreter is located in one hemisphere and is only getting direct visual input from that hemisphere. Therefore the interpreter had no idea why her hands were pointing at different pictures. What would the patient (specifically, the patient's interpreter function) say if she were asked to ex­plain why her hands were behaving differently?

An interpreter's job is to interpret, as best it can, the information it gets. In the textbook example, one hemisphere was shown a snow shovel and the other one a chicken's claw.

While the split- brain experiments provide dramatic illustrations of our urge to make sense of our environment, there are many less dramatic ones.¹⁰ Questions remain, though: Do we ordinarily generate hypotheses consciously or uncon­sciously, and does it matter?

11. B.2 Thinking About Science Consciously and Unconsciously

We don't understand our unconscious minds, but then we don't understand our conscious ones either. While scientists are beginning to make progress in under­standing the objective properties of conscious experiences, the eeriest aspect of consciousness, self-conscious sentience, is so difficult that it is designated “the Hard Problem,”¹¹ and put aside to await future developments. Just because we can't explain consciousness doesn't mean that we can't use what we do know about it, however.

11. B.2.a Much of Consciousness May Be Unconscious

Much of our consciousness reflects information processing that takes place be­fore we know it. Consciousness is an effect, not a cause.¹² This is not a novel idea; the philosopher Thomas Huxley said exactly that in 1874. He described humans as “conscious automata,”¹³ and explained—essentially— that “we do not run from the bear because we are afraid of the bear; we are afraid of the bear because we are running from it [emphasis added].” Christof Koch reports that the virtually im­perceptible awareness of a rattlesnake near your ankles is exceptionally effective in triggering evasive maneuvers before you know what's there.¹⁴ It makes evolu­tionary sense to respond to danger first and to think about it consciously second.

Examples of reflexive responses to danger do not prove that all consciousness is a secondary cognitive effect, of course. The nervous system j uggles a lot of tasks and, while (somehow) generating consciousness is its most spectacular one, con­sciousness may be overrated as a causal agent. Innumerable examples, from the fact that we fall for unnoticed advertising gimmick, to our notorious difficulties in making rational judgments, as we'll see shortly, confirm that our unconscious is often in charge. We don't always know what makes us tick.

The official scientific demotion of consciousness from its position as CEO of brain function began more than 30 years ago with experiments done by Benjamin Libet and colleagues¹⁵ who were trying to study consciousness by measuring brain activity associated with it. The reasoning behind their basic experiment was straightforward: the hypothesis that consciousness controls your mind and body predicts that, when you decide to act, the neural activity (“brainwaves”) measured by an EEG that signals your conscious intent to act must precede the neural commands to the muscles that carry out the action. In effect, if conscious­ness is the boss, we should be able to overhear the boss giving the orders.

To test this prediction, Libet and colleagues asked their subjects to move a finger when they felt like it and to make a mental note of the exact instant that they decided to move by watching a clock-like device nearby. From past work, they knew that a fraction of a second before the subject begins to move her finger, a blip in the EEG will indicate the brain activity being sent to move the finger. The experimenters recorded the finger motion and the subjects' EEG activity. After each trial they could compare the EEG activity with the time the movement decision was made. If her conscious mind triggered the decision to move, the brain activity associated with her decision (“I'm going to move my finger... now!”) must come before the finger movement blip. That's the prediction.

Amazingly, Libet's group found that was not what happened. Subjects became consciously aware of their decision to move only after the blip signaling finger movement. The brain activity causing the finger to move had already been under way for about half a second by the time they “decided” to move it! A recent fMRI study arrived at the same conclusions as did Libet's.¹⁶ Evidently the conscious command did not start things off. Our sense of conscious control may be an illu­sion, or simply a parallel sign that, in fact, you're starting to move your finger; in any case, it appears that the CEO is not in charge.

At first, all of this appears to be intensely counterintuitive, but is it? You say that an idea “occurred to me” or that one “popped into my head.” Really? Where did it pop in from? Your ideas must happen in your brain yet remain unknow­able until they “become” conscious. Why one idea becomes conscious and an­other doesn't is a mystery, but conscious and unconscious ideas themselves may not differ qualitatively¹⁷; for instance, a gatekeeper downstream of their origin might determine which is which. While the details of Libet's experiments have been minutely scrutinized and criticized, the experiments did bring the ques­tion of conscious influence to the attention of many scientists who had previ­ously remained aloof when consciousness was seen as a subject for philosophers only. Nowadays, “most neuroscientists... believe that conscious experiences are consequences of brain activity, rather than causes,”¹⁸ and this conclusion is rele­vant to the discussion of the hypothesis.

As we've seen (Chapters 2 and 10) some critics of the scientific hypothesis consider it a serious objection that we can't specify how we arrive at a hypo­thesis or that we don't know what sorts of mental process are responsible for it. The reality is that we're able to manage quite well even if we don't know much about where our ideas come from. I'm not promoting any particular model of consciousness, only suggesting that, since we know so little about either conscious or unconscious thought processes, we shouldn't let preconceptions about one or the other get in the way when trying to analyze our own scientific thinking.

11. B.2.b Unconscious Ideas as a Source of Hypotheses

Would the cognitive source of hypotheses matter even if we knew what it was? The history of science has famous examples of hypotheses that were formed during dreams, the archetypally unconscious state. In the early 1860s, the chemist August Kekule was trying to work out the atomic structure of benzene. At one moment, while “dozing,”¹⁹ vaguely picturing rows of atoms dancing be­fore his eyes, he visualized a snake biting its own tail and realized that the atoms in benzene could be arranged in the shape of a ring; he “spent the rest of the night working out the consequences of this hypothesis” In other words, once he saw it as a possible analogy for benzene's chemical structure, the ring became a hypothesis with specific predictions about benzene's other properties. Kekule's subsequent experimental tests were consistent with the ring model, which was accepted as the best one for many years until it was replaced by Linus Pauling's theory in the 1920s.

In the early twentieth century, pharmacologist Otto Loewi hypothesized that a chemical released by the nerves controlled the heart, but he was stumped about how to test it until a vision of a critical experiment came to him in a dream one night when he was sound asleep.²⁰ He awoke enough to scribble a note about the experiment to no avail: he couldn't read his scrawled handwriting in the morning. Fortunately, the same dream came to him the next night, and he im­mediately got up and went to the laboratory, where he took the fluid solution bathing a beating frog heart and applied it to a second, quiescent heart. When the quiescent heart began to beat, as his dream had predicted, the chemical hy­pothesis of synaptic communication had passed its first critical test and Loewi was on the road to the Nobel Prize in Physiology or Medicine, which he received in 1936.

The moral is that Kekule's and Loewi's hypotheses were products of their unconscious minds and that their origins didn't matter; what the scientists did with their hypotheses did matter, and what they did was to follow the Scientific Method in testing them. You may not find these anecdotes striking enough (though they are) to be convincing evidence that the mind is generating hypothesis-generating activity during ordinary awake behavior. Let's look at some less dramatic evidence.

11. B.2.c Complex Automatic Thinking: Counterfactual Thinking and Memory

The concept “if” is complex²¹; it invites us to consider possibilities that may not have happened. Evaluating “if” statements is so crucial that we can do it auto­matically. You know that “If I put my hand on a hot stove top, I'll get burned,” is true, although you've probably never tested it—deliberately, at least. What does information processing like this mean for understanding hypothesis-based scientific reasoning? It appears that remembering the past is closely related to imagining the future.

The science of memory has been evolving; we know that our personal mem­ories are not recalled intact in a preexisting state. Instead, they are actively reconstructed. We assemble them as needed from bits of stored information within “a common neural system [that] supports our recollection of times past, imagination, and our attempts to predict the future.”²² The memory system is not like a warehouse filled with data; it is more like a factory for producing certain kinds of thoughts. And it is capable of carrying out the information processing functions that are collectively known as counterfactual thinking.²³

Counterfactual thinking is what you do when you wonder “what if” you hadn't started seeing that person in high school that your parents disapproved of, which led to your hanging out with a new crowd, going to that college, etc. Counterfactual thinking allows you to reconfigure previous experience, to break it down, rearrange, and reorder its parts to create real or potential alternative outcomes. Similarly, it allows you to consider possible future courses of action, such as what would happen if you quit your present job and tried to make a living writing poetry.

Counterfactual thinking is not restricted to personal, episodic forms of memory and thinking, but is also pressed into service whenever you “simulate internal models of events and then compare these internally generated models with external reality”²⁴ According to psychologist Ruth Byrne, the tasks of coun- terfactual thinking range from idle musing about possible consequences “if only... ” to supporting “logical, mathematical, and scientific reason, and they underpin complex deductions”²⁵ Counterfactual thinking is what you do when you hypothesize explanations for phenomena.

We know that the system responsible for counterfactual thinking is hard­wired into our brains, and we know something about its architecture. It is partly housed in our temporal lobes, the lower sides of our brains. Patients who are se­verely amnestic because of damage to their temporal lobes lose their ability to recall past events. Perhaps surprisingly, they also suffer deficiencies in their abil­ities to make future plans.²⁶

Nerve cells in the prefrontal cortex collaborate with those in the temporal lobes to conduct counterfactual thinking, and damage to the prefrontal cortex also causes distinctive cognitive handicaps. Characteristically, prefrontal lobe patients don't fully consider the consequences of their actions and have trouble learning from their mistakes or making plans, which suggests that the pre­frontal lobe normally helps carry out these functions. And, indeed, when healthy subjects relive thoughts of their past actions (e.g., what might have been if they hadn't taken gambles that they took and lost), fMRI studies show that their pre­frontal cortex is very active.

Evidently, our ability to create and play with mental images of alternate fu­tures was so important to our primate ancestors that we evolved an extensive, built-in brain network to do the job. This network also makes our lives easier by automatically processing past information and comparing it with our current experience. When our circumstances are safe, secure, and predictable as we ex­pect them to be, all is well and we don't waste a lot of energy taking note of them. Unfortunately, this system, while it simplifies routine tasks, can also make objec­tive scientific thinking more difficult.

11. B.2.d Automatic Hypotheses and Predictions

You get an inkling about your ongoing hypothesis-generating operations when you experience surprises, such as getting unexpectedly good exam results. Surprise indicates that there is a mismatch between what your ongoing uncon­scious thinking processes predicted and what your conscious mind experienced. Mismatches between unconscious and conscious thought also characterize the phenomenon of illusions, both sensory and cognitive, which also suggests that we might learn something about cognitive illusions from studying the more straightforward sensory ones. (Some theoreticians deny the existence of cogni­tive illusions, but before we can appreciate their arguments we need to under­stand what it is they're denying.) Let's start with the familiar sensory ones.

11. B.2.e Sensory Illusions

When the information that you perceive about the world conflicts with informa­tion that you get from other sources, say, measuring instruments, you say you've experienced an illusion, although there is nothing spooky or mysterious about it. The experience of an illusion provides clues about how the nervous system au­tomatically processes information, and it's worth looking into sensory illusions because there is an analogy between how our sensory systems—vision, hearing, touch, etc.—work and how our cognitive systems work.

We've all seen the classic Muller-Lyer illusion²⁷ which shows two medium­length, parallel lines tipped with shorter lines oriented either inward, arrowhead­fashion, or splayed outward. Although the parallel lines are exactly the same length, our unshakable impression when seeing them is that one is longer than the other. The Muller-Lyer illusion is visual, but illusions are a feature of all sensory modalities. An especially convincing nonvisual one is the thermal illusion you get after you've immersed one hand in very cool water and the other in very warm water for about 30 seconds and then put both hands at once into a bucket of luke­warm water. Knowing that the final water temperature must be the same for both hands does not help you to feel it that way; it feels quite different to each hand.

Note—and this is extremely important for the later discussion of cognitive illusions that we'll encounter in Chapter 12—that your experience of this illusion does not mean that there is anything wrong with your sensory perception. It does not signal an error in the system or a failure in your instinctual perceptions of touch and temperature. The illusion simply means that there is a discrepancy be­tween what your nervous system is telling you and what an external standard, in this case, measuring instruments, tell you about some aspect of the world. The fact that you experience the water temperature differently with each hand is a sign that your sensory systems are fine; they are working just as they were designed to work.

The thermal illusion is caused by sensory adaptation; the two hands “got used to” different initial temperatures and so, by comparison, the final water bath seems warmer to the hand adapted to cool water and cooler to the hand adapted to warm water. Sensory adaptation is one way that our nervous systems deals with the flood of sensory information pouring in on us every instant. Adaptation filters out information that seems harmless and inessential, and, again, the fact that you experience a sensory illusion means that your system is working the way it should, not that it is malfunctioning. This conclusion seems innocuous in the sensory realm but is far more controversial in the cognitive realm.

11.