by Bor, Daniel
When you move up in sophistication to the fruit fly, another commonly studied animal, you can create a surprisingly brainy biological machine with only about 200,000 neurons. As well as learning by simple association, like nematode worms, fruit flies sleep, have short- and long-term memories, and even have a primitive analogue of attention—all inside a brain the size of a poppy seed.
Learning is all very well, but what should I try to learn when surrounded by infinite potential facts? And why should I bother to learn to avoid the pie until it cools down? Why should the worm bother avoiding some type of toxic food? Why not learn instead that the wind makes that leaf there on the right bob up and down a little faster than the one on the left? The crucial answer is that animals are constrained by a value system—the representation of what’s good or bad, pleasant or painful. Animal behavior is closely governed by this system. And this mechanism, via evolution, has been honed to closely map that which is beneficial or detrimental to survival and reproduction.
This value system labels any remotely relevant stimulus according to whether it will aid or imperil the animal and how great the danger or benefit is. Simple animals will enshrine this in their movements—they will approach what is good (such as food or sex) and escape from what is bad (such as predators). In many animals, the speed and level of permanence of learning is also related to just how beneficial or dangerous the source is. For instance, a dog may almost immediately learn that when its owner starts screaming at it, a kick is sure to follow, but it might take many repeats of its owner calling out “new water” for it to realize that there is something to drink again, since water is in plentiful supply, and its thirst is rarely life-threatening.
Though a simple animal may learn what is good or bad about the environment, this doesn’t explain in what way it is good or bad. C. elegans will back up if it smells something paired with a toxic food source in the same way that it backs up if it senses a vibration—there in fact is little distinction in its brain between the two events. This is where emotions extend this value system. Emotions put meat on the bones of what is beneficial or harmful. The three main primitive emotions are fear, disgust, and anger. If we’re afraid of some smell, we sprint away—or freeze in cover—while also being far more alert to the danger, ready to notice its slightest detail, and actively prepared to escape again if necessary. If we find that a smell evokes memories of a disgusting meal, it’s just plain silly to sprint away through the forest as fast as we can. Instead, we will merely slowly back off and look for something else to eat. So basic emotions can shape our behavior in far more sophisticated, prepackaged ways, in relation to different categories of threat, than a crude value system that only ever has two labels: good or bad.
Some psychologists have suggested that we recognize our own emotions wholly by what we pick up about our body states. When we’re angry, they suggest, we’re actually just noticing that our heart rates have increased, our fists are clenching, and so on. Again, emotions are largely a signal to move, to change the environment to maximize our survival and reproduction within it.
Of course, humans, in contrast to simple animals, have a large range of different, more complex emotions, such as jealousy, schadenfreude, or a sense of injustice. The variety of our feelings has increased dramatically compared with many other mammals because our evolutionary heritage is that of a large, complex, hierarchical society, which places many more demands on managing social politics and a diverse group of friends. Some emotions may destroy our lives, such as the obsessive love we may feel for an unavailable woman, or the addictive rush of gambling. Nevertheless, all the feelings we experience have a clear evolutionary foundation. Our brains perceive features of the environment to be good or bad for us, even if the computation can go askew at times due to the complexities of our lives. And many complex, seemingly subtle emotions are combinations of simpler ones, each with a clear evolutionary purpose. Jealousy, for instance, is a deep desire for some object, such as a mate, and primitive anger at the threat to our possessing this object.
In fact, the more we study our chimpanzee cousins, the more emotions we find that we seem to share with them, and the more apparent it becomes that our own large set of sophisticated emotions have an evolutionary underpinning. The latest research additions, for instance, include findings on the existence in chimpanzees of a moral sense, and possibly even a degree of wonder. If a chimp encounters a waterfall, it will sometimes display in front of it, touch it, and stare at it for prolonged periods of time—even if no other chimps are around (thus helping to rule out the possibility of an alpha male showing off to assert his authority over his group). If there is a thunderstorm, both males and females have been known to play out a kind of dance. Some primatologists have speculated that this is because our simian cousins can occasionally show intense curiosity, even a kind of reverence, for dramatic displays of nature.
VAST INTERNAL WORLDS
I’m not, for the moment, assuming any awareness in any animal aside from ourselves. But at the same time, I firmly believe not only that our emotional repertoire closely links us with chimp minds, but also that there is a continuous thread running from a husband’s bursts of jealousy all the way down to the battles between Alice, Beth, and Claire earlier in this chapter.
Although all life is predicated upon capturing useful ideas about the world, there is a remarkably common tendency for information stored on one level to combine to create a richer concept at a higher level. In some cases, further layers can be constructed on these foundations of foundations, and so on, until an efficient yet towering edifice is created. It leads to bacteria using control genes, or finding computational switches for rudimentary learning, or combining forces so that all can optimize some food source. It leads to shoals of thousands of fish collectively twisting elegantly, easily away from a predator, even though each fish alone wouldn’t stand a chance. It leads to millions of ants together developing highly intelligent behavior, partly via simple chemical signaling. It also leads to evolution creating a value system as a shortcut for hypotheses about what helps or harms an animal, then building simple emotions on top of this, and then stacking complex emotions on top of the simple ones. And it leads to humans developing categories, plans, language, and ingenious strategies to serve, modulate, and enrich our emotions and motivations. This is partly because nature really is highly structured, interconnected, and hierarchical. Within each level, complex structure can be discerned—sometimes so highly ordered that simple mathematical equations can capture almost every detail.
Information in the universe is brimming with patterns. Scientists aim to discover those patterns with sufficient accuracy that they can make staggeringly accurate predictions (how else can a satellite sent on a 500-million-kilometer, half-year journey into space arrive at the precise orbit around Mars that NASA scientists wanted?). But it’s not just scientists who benefit. As biological machines with large, complex brains and pressures to innovate screaming at us from every angle, we all have good reason to spot useful patterns rather than just simple facts. The more accurately we represent the structures of the universe, the more control we gain over the environment and ourselves. This is true both in terms of scientists, research, and technology and from an evolutionary point of view. Both have benefited from the invention of incredibly powerful hypothesis testers. In humans, the connection between scientific research and our fundamental intellectual qualities seems so similar as to be trivial, but we came in the first place to these incredible mental faculties precisely because of evolution favoring accurate information processing, almost by definition, ever since the first proto-life creatures emerged in the oceans.
Exploiting patterns isn’t limited to animals. You can find patterns everywhere in nature. Many viruses adopt a regular circular shape, partly because the smallest, most efficient genetic instructions are required for such a regular structure. Flowers form highly symmetrical shapes with their petals or seeds—patterns that mathematicians immediately
recognize, and that are attractive to bees precisely because of the order the bees perceive in them. But non-animals are severely limited in terms of the regularities they can encode.
Many animals are constantly, almost desperately, looking for patterns. Occasionally, this search can go rather wrong. Burrhus Frederic (B. F.) Skinner, one of the pioneers, along with Ivan Pavlov, of the study of simple forms of learning, found that if he presented food at regular intervals to a pigeon, without any cues preceding the feeding, the bird would nevertheless manufacture some action to pair with the tasty stimulus. One bird carried out a bizarre dance, twisting anticlockwise a few times. Another carried out a repeated head nod. It was as if the bird believed that twisting around would generate food. Skinner pointed out that this behavior looked remarkably like human superstition, which can result in such things as rain dances, or beliefs in astrology. It does seem clear that one component of the extreme popularity of irrational beliefs such as religion or alien abductions is our unerring search for structure and meaning within the constant torrent of information to which we are exposed.
The prevalence of superstitious beliefs in the animal kingdom suggests that some underlying process closely related to it is useful for animals. Imagine that the extent to which an animal searches for useful patterns is like the volume control on a low-grade stereo. Turn it too low, and you can’t hear any detail. If the volume is halfway, you can hear quite a bit of the music, but some of the accompanying instruments are hard to pick out. Turn it to the max, and much of what you pick up is distortion; you can hardly hear the main tune. Animals, it seems, like their informational music pretty loud; they want to be able to pick out every interesting detail in the melody, as well as the bass and the accompanying instruments, and achieving this highish volume is worth living with a bit of buzz and distortion. It’s better to maximize your chances of exploiting opportunities to learn new, interesting features of the world than to miss these potential insights, even if, occasionally, you make false guesses and latch onto irrelevant behavioral habits. This is highly reminiscent of chaos-inducing strategies that occur on a genetic level, such as mutations, sex, and jumping genes, which may create many fatal novelties, but also some genuinely useful innovations.
But while the innovations in bacteria were impressive, when you have a large biological computer, such as a brain, capable both of storing stable ideas and learning new ones on the fly, the gains in information processing and physical control of the environment are exponentially improved.
This constant searching for tricks and patterns regularly yields solid dividends, to such a degree that we are largely blind to how vast our collection of structured knowledge is and how much it dictates who we are. I’m not alone in being a huge fan of tennis giant Roger Federer. I’m constantly astounded by his accuracy, his power, and the diversity of his shots, which seem so many orders of magnitude above what humans should be capable of. But I take comfort from the fact that he is merely mortal, and spent many thousands of hours obsessing over and practicing his art, starting not long after he was out of diapers. In almost any field, spending thousands of hours devoted to your beloved topic seems an important prerequisite for appearing to be a genius. But, really, we are all super-super-Federers in the way we mentally interact with the world, and the reasons are similar—practice. This time, though, the practice is largely on an evolutionary level. We’ve had not thousands of hours to hone our skills, but half a billion years.
Most animals have a combination of highly practiced models of the world—some are genetically determined; others are merely primed by genetics; and still others are entirely prescribed by learning and experience. Humans are unusual in the animal kingdom in that although we are born fully formed, with all the parts where they should be, we are helpless for many months—normally having to wait for over a year before we can even walk. Part of the reason for this is somewhat trivial: If we were born with our brains properly developed, we would have to be born with much larger heads, and our mothers would require pelvises so large they would hardly be able to walk. Another reason, though, goes to the heart of what it means to be human. In this seesaw between fast, brittle, uncomplicated instincts and slower, flexible, potentially complex plans, humans definitely weigh heavily on the side of the latter. We begin life embarrassingly ignorant and incapable, but have an enormous capacity to learn and optimize any motor skill and any representation of the world—and that’s exactly what we spend our lives doing.
We may not all have Federer’s physical poise, but we nevertheless find it trivial to walk about a busy town while chatting to a friend, fluidly moving our limbs and mouths in incredibly coordinated ways, picking up and moving multiple grouped objects simultaneously with undoubted finesse, recognizing obstacles automatically, and so on. All these are utter computational marvels. They are partly the product of hundreds of millions of years of evolution, which have provided us with an immensely powerful biological computer, honed in just the right way for us to understand the world and move within it. But the supremely sophisticated ways in which we navigate through our daily lives are also partly a result of the vast amount of learning we’ve accomplished by the time we reach adulthood.
In many surprising realms, our apparently effortless feats of thought and motion make those of the most cutting-edge robots look utterly idiotic. Despite many years of research on the part of their creators, any robot that is designed not to perform perfectly prescribed, repetitive movements, but instead to learn about perception, walking, and the meaning of objects, tends even now to fail at these tasks in a comical, catastrophic way. The robot ends up resembling some combination of a heavily alcoholic squirrel and a newborn child. This is despite the stupendously powerful computers we now have at our disposal and much genuine progress in artificial intelligence and engineering. We might not realize it, but almost everything we do so automatically and flawlessly is in fact frighteningly complex, requiring a vast infrastructure of carefully sculpted biological algorithms.
The brain as a computer is a fantastically complex statistical engine, constantly refining models of how the world is and will be. Everything feels so easy to us, but that is all an utter illusion, just as Federer’s skills seem so deceptively effortless (until you pick up a racket and try to emulate his expertise yourself). In almost every corner of our mental lives, we are applying a particularly powerful form of predictive statistics (known as Bayesian inference), which boils down to tweaking our model of current and future events based on related events from the past.
This combination of frenetic statistical computation and a constant, conscious, roaming search for patterns allows us to aggressively, minutely track thousands of flittering features of the world and discover insights about them to recreate so much of its beautiful structure inside our minds. We are consequently, in some ways, now heavily protected from the cruelties of biological evolution, since so much of the evolutionary battles between ideas now occur in the mental realm. A professional, scientific picture of the universe is one quite natural product of our massive brains. We have tamed and controlled nature to give us plentiful food and protect us from the elements, and we are increasingly able to fight the disease and decay in our bodies through medical advances.
At the same time, our vast capacity to learn—to extract regularities and meaning from that information stream—is intimately related to our awareness, to the rich and deep experiences that constantly punctuate our lives.
3
The Tip of the Iceberg
Unconscious Limits
A HOLIDAY FROM AWARENESS
For many years I’ve played soccer every Wednesday with my Cambridge research department. The game is taken extremely seriously—aside from the small exceptions that we never keep score, we swap players around continuously if sides become uneven, and we seem to view the pub trip afterward as an event more sacred than the actual match. A few years ago, halfway through a game, I had the ball in front of a particularly large, intimid
ating member of the opposition. In my head, I was preparing the execution of a complex, deft maneuver: I would elegantly dance around him, along with any other foes in my path, the ball seemingly glued to my feet. I would sprint impressively for the goal, scoring right in the corner with a blistering shot. In reality, on my first body-twist in this foolproof plan, I landed embarrassingly on my arse, unable even to claim the conciliatory prize of a foul, since no one had so much as touched me.
On my clumsy journey to the grass, with my backside leading the way, I’d felt two rather worrying clicks in the middle of my right knee. These resulted in an abrupt end to the game and a trip to the emergency room, with a possibly torn anterior cruciate ligament.
When the knee failed to recover after a few weeks, the surgeon recommended an operation so that he could have a look inside and fix any injuries. The next morning, I was lying in the anteroom next to the operating theater, about to experience general anesthesia for the first time in my life. It’s completely understandable that many people are rather anxious at such events. I, on the other hand, felt like a kid in a candy store, albeit a very nerdy, neuroscience-loving kid, with candies made exclusively of brain-altering medications. Earlier that morning, I’d already had an in-depth, chatty discussion with the anesthetist about the wonderful drugs I’d be injected with and which neurotransmitters they’d work on. Now, with the needles in place, I was tremendously excited, supremely curious to discover what effects this concoction of chemicals would have on me.
