A book by William H. Calvin UNIVERSITY OF WASHINGTON SEATTLE, WASHINGTON 98195-1800 USA |
HOW BRAINS THINK
A Science Masters book, to be available in 12 translations (BasicBooks in the US) copyright ©1996 by William H. Calvin |
Evolving a Good Guess
The short answer is that the single number seems to tell us something additional while hazardous when overgeneralized, its an interesting bit of information. Heres why: Doing well on one kind of intelligence subtest never predicts that youll do poorly on another; one ability never seems to be at the expense of another. On the other hand, an individual who does well on one subtest will often perform better than average on the other subtests. Its as if there were some common factor at work, such as (dare I mention the word?) test-taking ability. The so-called general factor g expresses this interesting correlation between subtests. The psychologist Arthur Jensen likes to point out that the two strongest influences on g are performance speed (such as how many questions you can answer in a fixed amount of time) and the number of items you can mentally juggle at the same time. Those analogy questions (A is to B as C is to [D, E, F]) typically require at least six concepts to be kept in mind simultaneously and compared. Together, they make high IQ sound like a job description for a high-volume short-order cook, juggling the preparation of six different meals at the same time, hour after hour. Thus high IQ might be without significance for the kind of lives that most people lead, or only important on those occasions demanding a quick versatility. A high IQ is usually necessary to perform well in highly complex or fluid jobs (for example, being a doctor), its an advantage in moderately complex ones (secretarial or police work), but it provides little advantage in work that requires only routine, unhurried decision making or simple problem solving (for example, for clerks and cashiers, whose reliability and social skills are likely to be far more important than intelligence). IQ is certainly one fascinating aspect of intelligence, but it doesnt subsume the others; we shouldnt make the mistake of trying to reduce the subject of intelligence to a simple number on a rating scale. That would be like characterizing a football game in terms of one statistic, say the percent of passes completed. Yes, over the football league as a whole, winning does significantly correlate with that statistic, but theres a lot more to football than just percent-passes-completed; some teams win without completing a single pass, by emphasizing other strengths. IQ does correlate with winning in many environments, but its not what the intelligence game is all about, any more than successful passing is what football is all about. I think of intelligence as the high-end scenery of neurophysiology as the outcome of many aspects of an individuals brain organization which bear on doing something one has never done before. We may not be able to explain intelligence in all its glory, but we now know some of the elements of an explanation. Some are behavioral, some are neurophysiological, and some are evolutionlike processes that operate in mere seconds. We even know something about the self-organizational principles that lead to emergent stuff those levels-in-the-making, as when (to anticipate a later chapter) categories and metaphors compete for cerebral territory. The big issue for understanding intelligence isnt who has more but what it is, when its needed, and how it operates. Some of what intelligence encompasses are cleverness, foresight, speed, creativity, and how many things you can juggle at once. More later. |
Did our intelligence arise from having more of what other animals have? Just looking at the
brain and judging it by its size, as if it were a cantaloupe, is apt to be misleading. Only the outer
shell, the cerebral cortex, is markedly involved in making novel associations. Most of the brains
bulk comes from the insulation around the wires that connect one part of the brain to another;
the more insulation, the faster the messages flow. As animals become larger and distances
greater, more of the bulky insulation is needed to speed up transmission and keep the reaction
times short; this insulation increases the bulk of the white matter even if the number of cortical
neurons were to stay the same. An orange peel is only a small part of an orange, and our cerebral cortex is even thinner: about 2 mm, the thickness of two dimes. Our cortex is extensively wrinkled; were it to be peeled off and flattened out, it would cover four sheets of typing paper. A chimpanzees cortex would fit on one sheet, a monkeys on a postcard, a rats on a stamp. Were we to mark off a fine grid on the flattened surface, wed find about the same number of neurons in each little grid square in all cortical regions (except primary visual cortex which, in all animals, has got lots of additional small neurons). So if you need more neurons for a particular function, you need more cortical surface area. We tend to talk of demanding visual tasks for food-finding as enlarging monkey visual cortex in later generations but not its auditory cortex, that evolution tends to first produce a bulge here and then, when some other selection pressure comes into play, a bump there. But there is now a strong suspicion that any nonolfactory natural selection for more brain space, say vision, results in more brain space for all the other functions as well, that it is often developmentally difficult to make regional enlargements of the brain. So enlarge one, enlarge them all may be the general rule, rather than an exception. And if one evolutionary route to a free lunch isnt enough, heres another: new functions often first appear by making spare-time use of some pre-existing part of the brain. Brain regions are, to some extent, multifunctional, resisting our attempts to label them. So what pre-existing functions might be most relevant to the quantum leap in cleverness and foresight during hominid evolution from the apes? Most would say language. I will argue that a core facility common to language and to planning hand movements (and used in our spare time for music and dance) has even greater explanatory power than a special facility for only language functions.
|
Intelligence is sometimes described as a patchwork of know-how and know-what areas in
the brain, all those perceptual mechanisms so sensitive to expectations. That is surely true, but if
your definition of intelligence is so broad as to include most things that the brain does, such a
formulation doesnt advance your understanding any more than extending consciousness to cover
plant life does. Catalogs are not explanations, however interesting the list or however much the
topics may need inclusion in an introductory course. Its not my purpose to eliminate perceptual
mechanisms from intelligence but to illuminate the underpinnings of guessing well and those
levels of self-organization that produce stratified stability. The Spanish physician Juan Huarte defined intelligence in 1575 as the ability to learn, exercise judgment, and be imaginative. In the modern literature, intelligence often connotes the capacity for thinking abstractly, for reasoning, and for organizing large quantities of information into meaningful systems. Not only does this sound like academics trying to define themselves, but it aims too high to be a definition that is readily extended to other animals. A better place to start for the what aspects is the animal behavior literature, where good operational definitions of intelligence center on versatility in problem solving. Bertrand Russell once wryly noted, Animals studied by Americans rush about frantically, with an incredible display of hustle and pep, and at last achieve the desired result by chance. Animals observed by Germans sit still and think, and at last evolve the solution out of their inner consciousness Besides being a British commentary on the scientific fashions of 1927, Russells quip about problem-solving cleverness illustrates the usual false dichotomy between insight and random trial and error. Insight is, beyond argument, intelligent behavior. Mere randomness is not, in the usual scheme of things; but we are thereby misled of which more later. I like Jean Piagets emphasis that intelligence is what you use when you dont know what to do. This captures the element of novelty, the coping and groping ability needed when there is no right answer, when business as usual isnt likely to suffice. Intelligent improvising. Think of jazz improvisations rather than a highly polished finished product, such as a Mozart or Bach concerto. Intelligence is about the process of improvising and polishing on the timescale of thought and action. The neurobiologist Horace Barlow frames the issue a little more tightly, and points us toward experimentally testable aspects, by saying that intelligence is all about making a guess not any old guess, of course, but one that discovers some new underlying order. Guessing well neatly covers a lot of ground: finding the solution to a problem or the logic in an argument, happening upon an appropriate analogy, creating a pleasing harmony or a witty reply, correctly predicting whats likely to happen next. Indeed, you routinely guess what comes next, even subconsciously say, in listening to a narrative or a melody. Getting a crying child to fill in the last word of each song line is an amazingly effective distraction, seen in many cultures. Subconscious prediction is often why a jokes punch line or a P.D.Q. Bach musical parody brings you up short you are surprised by the mismatch. Being a little wrong can be amusing, but substantial environmental incoherence is unpleasant, as when a day filled with job insecurity, noise, erratic drivers, and too many strangers leaves you frustrated, because of the frequent mismatch between what you expected and what actually happened.
[Calvins Cure for such Environmental Incoherence? Scale back the predictive challenges to a more comfortable level, not all the way into the boredom of sure-fire predictability but to where youll be right half the time. That way, you reassure yourself that you are still competent at predicting. Perhaps thats why, after a hard day awash in unpredictability, you tend to seek relief in ritual, music, or sitcoms anything where you can again take pleasure in frequently guessing what comes next!] |
One of the beginners errors is to equate intelligence with purpose and complexity.
Elaborate, complex behaviors initially seem like a reasonable place to look for signs of
intelligence. After all, our language and foresight behaviors are surely aspects of intelligent
behavior and theyre quite complex. But many complex behaviors in animals are innate: no learning is needed as theyre wired in from birth). Such behaviors tend to be inflexible and often difficult to perform at will, such as sneezing and blushing. These stereotyped movement patterns exhibit no more insight or understanding of purpose than does a computer program. Theyre a set piece. Both innate and learned behaviors can be long and complex. Consider, for example, the performance of an idiot savant, a person with enormous detailed recall but poor ability to make good use of the recollected information in a new context, by breaking the pattern into meaningful parts and recombining them. Whale song and insect nest building may be equally unintelligent. That whales and birds chain song sequences together is also not evidence of versatility. The most mindless of behaviors are often linked, the completion of one calling forth the next. Courtship behavior may be followed by intricate nest building, then segue into egg laying, then incubation, then the various stereotyped parental behaviors. Indeed, the more complex and purposeful the behavior is, the further it may be from intelligent behavior, simply because natural selection has evolved a sure-fire way of accomplishing it, with little left to chance. Learning, after all, is usually focused on far simpler things than the complex chains of all-important behaviors. The animal might understand its own behavior no better than we understand our yawn, or our tendencies to hug and kiss (clearly seen in bonobos and chimpanzees). Most animals in most contexts dont appear to have much need for understanding in our sense of appreciating the underpinnings and they dont attempt innovations except by modest variations and a slow learning process. Its as if thinking were a little-used backup, too slow and error-prone to be depended on in the normal course of things. The best indicators of intelligence may be found in the simpler but less predictable problems that confront animals those rare or novel situations for which evolution has not provided a standard response, so that the animal has to improvise, using its intellectual wherewithal. While we often take intelligence to mean both a broad range of abilities and the efficiency with which theyre done, it also implies flexibility and creativity in the words of the ethologists James and Carol Gould, an ability to slip the bonds of instinct and generate novel solutions to problems. That narrows the what field quite a lot.
In tests of convergent thinking there is almost always one conclusion or answer that is regarded as unique, and thinking is to be channeled or controlled in the direction of that answer.... In divergent thinking, on the other hand, there is much searching about or going off in various directions. This is most obviously seen when there are no unique conclusions. Divergent thinking... is characterized... as being less goal-bound. There is freedom to go off in different directions.... Rejecting the old solution and striking out in some direction is necessary, and the resourceful organism will more probably succeed. |
Arent-they-clever stories are what many people recall when the topic of conversation turns
to intelligence. Surely a dog qualifies as intelligent, they will insist. Most such stories turn out
to hinge on how well a dog understands English or reads his owners mind. Ethologists and animal psychologists will patiently reply that dogs are very social animals, expert in reading body language. A dog is always looking up to his owner, in the same way that a wild dog looks to the pack leader, asking Whats next, boss? or emotionally seeking reassurance in a juvenile manner, hoping to elicit benevolence. Talking to domestic dogs plays into these innate tendencies, though your words per se may not carry the message. People just dont realize how much information is conveyed by the tone of voice and body language of the substitute leader (thats you). If you read todays newspaper headline to your dog in the same tone of voice, and with the same glances and postures, as you use to ask him to fetch your slippers, it might work just as well in evoking the desired behavior. In many cases, there isnt much to confuse the dog. The setting itself (people, places, situations, objects present) provides most of the information the dog needs to respond appropriately to a command. Most dogs have limited repertoires, and its therefore easy for them to guess correctly. Training a dog to fetch a dozen different items on command is a more difficult proposition, simply because it becomes harder for the dog to guess your intentions. If you are confident that your dog understands words per se, you might try getting someone else to speak the words from another room over an intercom; this will eliminate most of the situational cues. Many smart animals cannot pass this severe a test of understanding spoken words, not even some extensively tutored chimpanzees who readily respond to graphical symbols. But dogs do pass the lesser test of performing the desired action most of the time, when the situation is familiar and the choices are obvious from the context. The size of the response repertoire is one important factor in intelligence. Dogs have many instinctive behaviors, such as herding and alarm barks; they can learn many more. Even their communicative repertoire can reach impressive numbers with extensive training, as the psychologist Stanley Coren observes.
[My pet] dogs have a receptive language of about sixty-five words or phrases and about twenty-five signals or gestures for a total receptive vocabulary of about ninety items. They have a productive language of about twenty-five vocalizations and about thirty-five bodily gestures for a total productive vocabulary of about sixty items. They show no evidence of syntax or grammar. If they were human children, they would be demonstrating the level of language customary at around eighteen to twenty-two months of age. [Bonobos] that have learned [a sign or other symbolic] language can obtain [comprehension] scores equivalent to a child of around thirty months of age. [WHC amendments]Speed of learning is also related to intelligence; one reason that dogs and dolphins achieve a wider repertoire of behavior with training is that they learn faster than cats usually do. So intelligence is quite a composite of other things and many mental abilities are relevant. Perhaps it is making effective combinations of them that better constitutes intelligent behavior.
|
An animals selection of appropriate behavior may be the key to sorting out claims of
animal intelligence. In many of the arent-they-clever animal stories, the animal isnt thinking for
itself but merely responding to a command. Piagets element of creativity, in the face of an
ambiguous task, is usually missing except during their playful antics. The scientific literature on nonhuman intelligence tries to cope with innovation, but since most putatively intelligent animal actions are not repeated actions, its hard to avoid a series of anecdotes (indeed, theres a wonderful book of them about apes, Machiavellian Intelligence). The usual scientific hazards of anecdotal evidence can be somewhat reduced by emphasizing comparisons between species. For example, most dogs cant untangle their leash from around a tree, but a chimpanzee seems to have what it takes. A leash-style snap fastener will suffice to keep most small monkeys inside their cage, even if they can reach the fastener to fiddle with it. But the great apes can figure the fastener out, so you must use padlocks and not leave the key lying around! Chimpanzees can practice deception: a chimp can guess what another animal is likely to be thinking, and can exploit this knowledge. But most monkeys dont seem to have the mental machinery to deceive one another. To many people, the essence of intelligence is such creative cleverness. When an animal is especially versatile at solving problems or inventing new moves, we consider that behavior to be particularly intelligent. But human intelligence is judged by additional standards.
|
When I tried out this creative cleverness definition of intelligence on one of my colleagues,
he was dubious and started citing examples of the terminally clever. You know, someone asks you how intelligent a certain person is, and you say, Well, hes certainly clever. By this, you mean that he talks a good line hes versatile at improvising tactics in the short run but doesnt follow through on his projects and lacks longer-term virtues, such as strategy, perseverance, and good judgment. Okay, I agreed, it also takes foresight to be truly intelligent. And chimps dont think much about tomorrow, as far as anyone can tell from their behaviors, even if they occasionally do some planning on the half-hour timescale. So maybe the flexible future is a human addition to ape intelligence. Intelligence also involves some imagination, I continued, remembering a high IQ group for whom I once gave an after-dinner speech. I had been surprised in view of the fact that everyone in the audience had scored high on intelligence tests at how unimaginative one of them was, and then I abruptly realized that I had always thought that IQ and imagination went hand in hand. But imagination contributes to intelligence only when shaped up into something of quality. Patients with hallucinations are pretty imaginative, too, but it doesnt necessarily make them highly intelligent. It just goes to show that IQ measures only some aspects of what we more commonly understand as intelligent behavior. The very nature of IQ exams tends to preclude tests of creativity or the ability to make plans.
|
Innovative behaviors are usually not new units; instead, they are composed of a novel
combination of old elements: a different stimulus evokes a standard behavior, or some new
combination of movements is used in response. How is sensory/movement innovation related to
intelligence?
Many animals have sensory templates, which they try out for size (and shape) on what they see, rather like a child trying out a number of cookie cutters on the baked assortment of Christmas cookies, to see which (if any) fits a particular cookie. Baby birds, for example, crouch when a hawk flies overhead, a behavior suggesting that they were born with the image of a hawk wired into their bird brains. The reality is quite different: initially, they crouch when any sort of bird flies above them. They then come to recognize the sorts of birds they see every day; as a shape becomes familiar, they cease the response to it. Because of such habituation, they eventually crouch down only to infrequently seen shapes, such as exotic birds that are just passing through and to predators, such as hawks, which are infrequent because there arent very many of any species at the top of the food chain. So the crouch is a response to novelty, not to a prewired alarm search image. Its as if the child found a misshapen cookie that none of the cookie cutters fitted, and was thereby distressed. Composers note that while pure overtones (as from the flute) are relatively soothing, random overtones (as in heavy metal or the raspy voices of some singers, such as Mick Jagger) seem to signal threat or alarm, and Ive long thought that the disordered sensations produced by nerve injuries were often perceived as painful (rather than merely nonsensical) for the same reason. Besides sensory templates for familiar sights and sounds, animals also have familiar movement schemas, among which they pick and choose. A cormorant can decide whether to cruise around underwater in search of another meal, or to fly away to another pond, or spread its wings to dry (cormorant feathers lack the oil that ducks have), or just stand around presumably by consulting the weightiness of its wings, the fullness of its stomach, its sexual drives, and so forth. Decision making is something that all animals do; it is usually an economistlike weighing of sensations and drives, followed by a standard behavior from its repertoire, as modified by the present circumstances. Of course, we humans often do something similar in deciding on a restaurant, taking into account its menu, parking, cost, travel and waiting time, and ambiance and somehow comparing all these factors with those of other restaurants. While such weighing of choices seems especially conscious, purposeful, and intentional, choice per se does not imply an extensive mental life not of the kind we associate with creating novel additions to the list of choices for what to do next (Suppose there are any northern Vietnamese restaurants in town?).
Curious, I took a pencil from my pocket and touched a strand of the [spider] web. Immediately there was a response. The web, plucked by its menacing occupant, began to vibrate until it was a blur. Anything that had brushed claw or wing against that amazing snare would be thoroughly entrapped. As the vibrations slowed, I could see the owner fingering her guidelines for signs of struggle. A pencil point was an intrusion into this universe for which no precedent existed. Spider was circumscribed by spider ideas; its universe was spider universe. All outside was irrational, extraneous, at best raw material for spider. As I proceeded on my way along the gully, like a vast impossible shadow, I realized that in the world of spider I did not exist. |
Sometimes an animal tries out a new combination of sensory template and movement during
play, and finds a use for that combination later on. So perhaps we should add play to our list of
intelligence attributes. Many animals, however, are playful only as juveniles. Being an adult is a serious business, with all those mouths to feed, so adults dont have the time or inclination to fool around. A long juvenile period, characteristic of apes and humans, surely aids versatility because of the accumulation of useful combinations. In addition, some evolutionary trends, including domestication of animals, tend to carry over juvenile traits into adulthood so that, too, might increase versatility. You dont learn just from your own experiences. You can copy the actions of others, as Japanese monkeys were observed to copy one inventive females technique for washing the sand off food. You may avoid what seems to spook others, even if you havent been personally threatened by it, and such superstitious behavior can be passed on. The original reason for Dont step on the crack in the sidewalk may be lost, but the cultural transmission between generations continues for centuries, sufficient unto itself.
|
A wide repertoire of good moves, of course, makes foresight a lot easier. Foresight
initially seems simple, almost too simple to be a requirement for high intelligence. But thats
because we confuse foresight with species-specific seasonal behaviors. Squirrels hoarding nuts for winter seems to be the standard example of planning ahead in the animal kingdom. And we now know how such things work. The hormone melatonin, released from the pineal gland during the hours of darkness, serves to warn of the approach of winter. Longer and longer nights result in the release of increasing amounts of melatonin each week, which in turn triggers food hoarding and new fur coats. It doesnt take much of a brain, to do that kind of planning. There are, of course, other behaviors created by the brains initial wiring which serve to set things up for months ahead. Mating behaviors have the effect of producing offspring after a considerable delay. Seasonal migrations come with the innate brain wiring or are learned as a juvenile to become mindless adult rituals. Of course, such behavior isnt the result of planning at all. Seasons are eminently predictable; and over the millennia, plants and animals have been shaped by evolution to sense the signs of approaching winter by means of the innate surefire mechanisms: hoarding nuts probably feels good as the days shorten, much as does following the gradient of a sexual pheromone in the air. Planning on the timescale of a few minutes is seen in several senses but, as youll see, neither should probably be called planning. Keeping a movement plan on hold as when monkeys remember where food is hidden for the next 20 minutes until allowed out of their cage to seek it, is sometimes called planning. But is it simply the memory of an intention? Another disputed type of evidence comes from spatial maneuvering. When bees are kidnaped and carried in a windowless container to a random location several kilometers distant and then released, they quickly set off on the optimal paths to an unseen favorite food source. Is this planning, or are they just referencing memories of horizon profiles? Before setting off in the correct direction, they fly a few circles first to get oriented; they may well be scanning the horizon for clues. Perhaps we should say that planning involves something novel, closer to the way in which we procrastinate, figuring out what can safely be put off until tomorrow (or avoided altogether). Indeed, Id reserve the term for something that requires multiple stages of the move to be assembled in advance of action, rather than when the later stages are organized after getting the initial moves in motion, which goal-plus-feedback can accomplish. Alas, there is surprisingly little evidence for this kind of multistage planning in the great apes. None of the termite-fishing chimps, as Jacob Bronowski once pointed out, spends the evening going round and tearing off a nice tidy supply of a dozen probes for tomorrow. Although wild chimps often seem to arrive at a distant fruit tree just as the fruit is ripening, how much of that is migration ritual and how much is all-in-advance planning of a unique route? For most of your movements, such as raising a coffee cup to your lips, there is time for improvisation en route. If the cup is lighter than you remembered, you can correct its trajectory before it hits your nose. Thus, a complete advance plan really isnt needed; a goal and periodic piecewise elaboration will suffice. You get started in the general direction and then correct your path, just as a moon rocket does. Most planning stories about animals fit into that mold. Multistage planning is perhaps best seen in an advanced type of social intelligence: making a mental model of someone elses mental model, then exploiting it. Imagine a chimp that cries food in a place where there is no food, and then quietly circles back through the dense forest to where it actually saw the food earlier. While the other chimps beat the bushes at the site of the food cry, the chimp that uttered it gets to eat all the food rather than having to share it. Whats really difficult is to make a detailed advance plan in response to a unique situation like those leftovers in the refrigerator and what might go with them. It requires imagining multiple scenarios, whether you are a hunter plotting various approaches to a deer or a futurist spinning three scenarios bracketing what an industry might look like in another decade. Compared to apes, we do a lot of that: we are even capable of occasionally heeding the eighteenth-century admonition of Edmund Burke, The public interest requires doing today those things that men of intelligence and goodwill would wish, five or ten years hence, had been done. So multistage planning for novel situations is surely an aspect of intelligence indeed, one that appears greatly augmented in the transition from the ape brain to the human brain. But knowledge is, I think, a commonplace.
|
A base of existing knowledge is, of course, required for versatility, foresight, and creativity.
You cant be a poet or scientist without a good vocabulary, but definitions of intelligence that
stress knowledge or memorys synaptic mechanisms really do miss the mark; theyre mistaken
reductionism the practice of reducing something to its fundamental constituents, which for
present purposes is carried a few steps too far. This is the mistake, as I explain in the next
chapter, that the consciousness physicists often make. For example, Shakespeare didnt invent the vocabulary he used. He invented combinations of those words, most notably the metaphors that allow relationships to be imported from one level of discourse to another. In a similar manner, much intelligent behavior consists of new combinations of old things. Deductive logic is another what aspect of intelligence, at least, of the human variety. Philosophers and physicists have, I suspect, been unduly impressed with the human faculty for logical reasoning. Logic might consist of guessing the underlying order of things, à la Horace Barlow, but only in situations where there really is an unambiguous underlying order to be guessed (mathematics being the prime exemplar). Piecewise approximation, as with the guessing needed for long division, could operate subconsciously so rapidly as to seem like a leap to the finished logical product. Could it be that logic is more a property of the subject matter than of the mental process that guessing is the name of the game during mental calculations as well as during creative thinking? The what list can be extended further, both for what is and what isnt. But I am going to focus hereafter on Barlows guessing-at-order aspect and more generally on Piagets improvisation problem of how to proceed when the choice isnt obvious. I realize that this excludes certain uses of the word intelligence, as when we talk of intelligent design or military intelligence, but the guessing aspect buys us such a broad range of intelligence connotations that we will do well to organize analysis around it provided we can avoid consciousness confusions and inappropriate levels of explanation.
|
You are reading HOW BRAINS THINK.
The paperback US edition |