The trouble with an alarm clock is that what seems sensible when you set it seems absurd when it goes off.
—REX STOUT, THE RODEO MURDER
Burrhus Frederic Skinner heard the news about Pearl Harbor while listening to a symphony on the radio. Skinner was then a psychologist at the University of Minnesota, not yet famous, and his great subject was “operant conditioning,” or the way behavior is modified by its consequences. Skinner’s role in the war began in 1940, even before the United States entered the conflict, when he was traveling by train and noticed, through the window, a flock of birds “lifting and wheeling in formation as they flew alongside.” Inspired, he got himself some pigeons and set about training them to peck in response to certain visual cues. He knew from experience that pigeons were good subjects for conditioning research; the birds are easy to handle, have good eyesight, and can see colors. Skinner’s idea now was to stick one of them into the nose cone of a bomb and have the bird guide it to a target with deadly accuracy.
Training avian suicide bombers sounds as far-fetched today as it did in 1941, when government officials at first rejected the idea. But Skinner was certain he was onto something, and he wasn’t the only one bent on using animals to deliver ordnance. Over the years people have experimented with bats, dogs, dolphins, and other creatures for this purpose, to say nothing of the microbes that can be a kind of weaponry in biological warfare. Remember, too, that bombing accuracy in those days was poor and missile-guidance systems primitive. Skinner’s arrangement seemed crude—the pigeon was basically thrust into a sock and controlled the projectile by pecking at a steering mechanism attached around its neck—but it was probably better than most of the alternatives, and for a while the U. S. Navy funded his work.
This work, at least by Skinner’s account, went exceedingly well. He managed to inculcate an astonishing degree of single-mindedness into his target-finding birds, whose conditioning was so thorough that they were impervious to distraction by velocity, gunshots, flashes simulating antiaircraft fire, heavy vibrations, or massive g-forces. Atmospheric pressure equivalent to that at 10,000 feet—enough, at one point, to blow out a window and knock the psychologist’s glasses off—made no difference. After a while Skinner had the birds pecking with the force of a miniature air hammer. They were rewarded for these patriotic feats with hemp seed, the psychologist having read that the birds were “almost fearless” after eating it. (Like so many warriors across the ages, Skinner’s pigeons were not just conscripts but had the benefit of intoxicants to undermine their inhibitions in the face of violence.)
Skinner’s air force reached something of an apotheosis in the heroic conduct of a single bird, whose job was to demonstrate what the pigeons could do. The performance was to occur in wartime Washington, at a meeting of skeptical officials from the Office of Scientific Research and Development, the government agency overseeing R&D for the war effort. Skinner set things up so that the pigeon would do its job in a darkened box, like all the others, pecking at a target projected onto a small screen in front of it; officials could take turns looking through a narrow tube to see what was going on. But that was considered too time-consuming, and so Skinner was forced to open up the box, flooding it with target-obscuring light and forcing the bird to perform surrounded by the looming faces of the federal scientific apparatus. Yet the bird’s performance was flawless. Years later, in a famous paper recalling the episode, Skinner described the upshot: “I will not say that the meeting was marked by unrestrained merriment, for the merriment was restrained. But it was there, and it was obvious that our case was lost.”
Pigeons at Harvard
Like so much about the life and work of B. F. Skinner, his exploits training avian bombardiers make for a great story. But they’re more than just a source of historical diversion. In his war work with pigeons, we get an early glimpse of Skinner’s worldview as a psychologist and philosopher, a worldview that would eventually make him a household name—and an important figure in the science of self-control.
Skinner was a behaviorist, a direct intellectual descendant of John B. Watson, the father of this particular flavor of psychology, which turned its back on the swamp of people’s emotions, their inner lives, and the issues they were supposed to have with their mothers or their penises. Behaviorism instead put forth the straightforward idea that organisms respond to reinforcement, which conditions behavior. A behaviorist is agnostic as to what’s going on in your mind, or for that matter whether you have a mind. What matters is what you do, and to a radical behaviorist like Skinner, what you do is a result of the outward consequences of your actions. Reality is a feedback system, and each of us is conditioned by a lifetime of reinforcements. If you broke your toe on the sofa last time you walked around barefoot in the dark, next time you’ll turn on a light or put on your slippers.
Behaviorists like Skinner had something in common with economists, who also emphasized the role of incentives in people’s actions. But the economists for the most part insisted that, no matter what people claimed they’d like or not like to do, their actions reveal their true preferences. Skinner, on the other hand, had strong preferences about preferences and must have believed that other people did, too, which is not the kind of humane notion generally ascribed to him in the popular mind. His wartime pigeon experiments, with their essentially mechanical view of organisms, are more the kind of thing that made him such a notorious figure over the years. Skinner taught pigeons to play Ping-Pong and created a scary-looking box for holding (and conditioning) babies.
The implications seemed pretty creepy. People don’t like to think they’re nothing more than flesh-and-blood robots, which many concluded that Skinner considered them to be. Worse, Skinner’s chillingly grandiose aspirations went far beyond the laboratory. He summed them up in a 1967 interview with Psychology Today: “What I really expect to be known for is the application of all this to education, psychotherapy, economics, government, religion and its use in designing a world that will make us into the kind of people we would like to be.”
But what Skinner was advocating, on a vast scale, was really just the kind of environmental control that John Bargh urged in the wake of his priming experiments. If our behavior is conditioned, after all, why not set things up so that we’re conditioned to be who and what we want to be? Since Skinner saw the frequency of behavior largely as the result of its consequences, it made sense that he should see the problem of self-control in terms of establishing consequences that would produce desired behaviors. Conditioning yourself, then, wasn’t any different from conditioning a pigeon or even another human. A person controls himself, Skinner wrote, “precisely as he would control the behavior of anyone else—through the manipulation of variables of which behavior is a function.”
Unlike Freud, who dismissed the notion of raising his own children according to his ideas, Skinner avidly practiced what he preached during a lifetime of systematic self-management—and in later years could point to a long life of staggering productivity. “I have studied when I did not feel like studying, taught when I did not want to teach . . . I have met my deadlines for papers and reports . . . In short, I arrange an environment in which what would otherwise be hard work is actually effortless.”
He even laid out a taxonomy of self-management techniques in his book Science and Human Behavior, including physical restraint (biting your tongue), changing the stimulus, aversive stimulation (perhaps a photo of yourself at your fattest on the fridge), arranging rewards and punishments, and simply finding something else to do in lieu of some unwanted behavior. The list may be useful for self-management, but it can also help us place Skinner in a long and very American tradition of self-improvement, one that goes back to Ben Franklin, at least. Like Franklin, Skinner was an inventor and pragmatically oriented problem solver, and both men employed formal self-management tactics.
Franklin, whose fame as an exemplar of self-control was sealed by the enduring success of his autobiography, reports that as a young man he “conceived the bold and arduous project of arriving at moral perfection,” and as an aid in this venture developed a kind of moral spreadsheet, writing the days of the week across the top and listing thirteen virtues along the side. This enabled him to plot his failings by date and category in a grid that would make his performance easy to assess. Temperance is first on the list. Moderation (“avoid extremes”) is number nine. Like Bridget Jones, whose fictional diary records not just her romantic misadventures but her encounters with the bathroom scale, Franklin instinctively understood the usefulness of self-monitoring and record keeping in any self-improvement scheme.
In 1948, two big things happened to B. F. Skinner. His novel Walden Two was published, giving the world, in fictional form, a more digestible version of his views on human nature than had been available before. And he returned to Harvard, the institution with which he would forever be identified, and from which he had earlier taken his PhD. There, at what had been the institutional home of the great William James, he launched the famous Pigeon Lab in the vast basement of Memorial Hall, where professors and graduate students labored beneath a network of overhead pipes to push back the frontiers of experimental psychology.
The Pigeon Lab, which produced a generation of leading experimental psychologists in the second half of the twentieth century, was really more like a beehive than an aviary, humming with activity. People were always coming and going, the mechanical recording devices were always clacking away, and grad students regularly and noisily debated their work. During the Vietnam years in particular, politics were debated as well. The place bubbled with enthusiasm; every experiment seemed to break new ground. Skinner was a lifelong tinkerer, and to work in this environment it helped to be mechanically inclined. These were the days before the personal computer revolution; experiments were controlled mostly by telephone circuit equipment, and the tall, vertical racks of these relay devices replicated the environment of some other, traditional experimental subjects—the kind with whiskers—by forming a large maze for the lab’s human denizens.
By 1967, when a lean young man named George Ainslie appeared at the lab, it had migrated over to Harvard’s new William James Hall, where it occupied the seventh floor and, no longer subterranean, enjoyed a lovely view. It was Ainslie’s good fortune that by now the brilliant but aloof Skinner had mostly lost interest, and so the place was being run by one of his former students, also brilliant, named Richard Herrnstein.
For much of his career, Herrnstein was almost as controversial as Skinner. In 1971—the same year that Skinner published Beyond Freedom and Dignity—Herrnstein became a lightning rod because of an article in the Atlantic Monthly in which he emphasized the heritability of intelligence. “Greater wealth, health, freedom, fairness, and educational opportunity,” he predicted with dispiriting accuracy, “are not going to give us the egalitarian society of our philosophical heritage. It will instead give us a society sharply graduated, with ever greater innate separation between the top and the bottom.”
The article generated intense hostility, and some critics branded him a racist, a charge he would battle for years. Hounded by protesters and death threats, Herrnstein later wrote a book defending his argument (I.Q. in the Meritocracy) and then ignited further controversy by expanding on his ideas about intelligence and heritability (with coauthor Charles Murray) in The Bell Curve, the much-debated book for which he is perhaps most widely known. Herrnstein died just as it was about to appear.
Among some of his detractors, Herrnstein became known as Pigeon Man, which as an insult probably said more about its source than its target, for Herrnstein’s work with pigeons was important in understanding how creatures make choices. It built on Skinner’s work, yet departed from it. And it is crucial for understanding the problem of self-control.
Ainslie wasn’t a psychologist and by any reasonable standard should have had no spare time to spend in the Pigeon Lab, since he was a full-time student at the Harvard Medical School and married as well. But as an undergraduate at Yale, Ainslie had taken a psychology course in which the professor said that a rat in a maze would run faster to reach a bigger piece of food than a smaller one. In fact, there was a direct relationship (within reason) between how fast the rat would run and how ample the reward. What the professor was describing was something called the “matching law,” which had been discovered by none other than Richard Herrnstein.
Basically this law says that animals (including us) acting freely will spend their time doing things in direct proportion to the relative value they get from each. Herrnstein derived this law working with those Harvard pigeons; he discovered that, when offered two different objects to peck, the birds pecked them in proportion to the rate of reward delivered by each. Varying the quantity of reward—a little grain—had the same effect. And most interesting of all, preference was proportional to the inverse of delay. In other words, the later the reward, the less the birds liked it. And the less they would peck.
What the matching law did was to express mathematically the experimental evidence that organisms discount future rewards. Even pigeons, apparently, will weigh larger, later rewards against smaller, sooner ones and, if the larger reward is not sufficiently larger, pick the smaller, sooner one—as if they have some instinct that reminds them that a bird in hand is worth two in the bush. And since self-control is about deciding between now and later (perhaps even infinitely later, a period otherwise known as “never”), the matching law provided something like a formula for self-control.
The problem was that there already was such a formula—and the two conflicted. In the Pigeon Lab, George Ainslie was about to prove which one was right.
Discounting the Future
This problem of deciding between now and later is known to economists as intertemporal choice, and it has bedeviled thoughtful people for an awfully long time—at least since 1834, when the Scottish-born physician and economist John Rae published The Sociological Theory of Capital. In it, he explained that different people have different levels of desire for accumulation, desires that depend in significant part on their self-control—“the extent of the intellectual powers, and the consequent prevalence of habits of reflection, and prudence.” He recognized, in other words, that self-control and intelligence go together, as psychologists would later prove. He also saw that the desire to accumulate, which involves deferring gratification, was opposed by the desire to consume now, an impulse that (as subsequent research would affirm) gets all fired up by “the actual presence of the immediate object of desire in the mind.”
John Stuart Mill, who was influenced by Rae, made similar observations some years later, in 1871: “Men often, from infirmity of character, make their election for the nearer good, though they know it to be the less valuable; and this no less when the choice is between two bodily pleasures than when it is between bodily and mental. They pursue sensual indulgences to the injury of health, though perfectly aware that health is the greater good.”
Some saw the faultiness of our trade-offs between now and tomorrow as a failure of imagination. The Austrian economist Eugen von Böhm-Bawerk speculated that “we possess inadequate power to imagine and to abstract, or that we are not willing to put forth the necessary effort, but in any event we limn a more or less incomplete picture of our future wants and especially of the remotely distant ones.” Arthur Pigou saw the problem as a kind of myopia, just as Socrates had done in the Pro-tagoras. “Our telescopic faculty is defective,” wrote Pigou, “and we, therefore, see future pleasures, as it were, on a diminished scale.”
Economics over the years has grown relentlessly more quantitative, and in 1937, economists finally came up with a way to reduce the future to a single equation, when Paul Samuelson developed the “discounted utility” model. Samuelson, you’ll recall, was the wunderkind who embedded “revealed preference” in modern economics. He later became famous to a generation of undergraduates for his best-selling introductory econ text, and, in fact, he as much as anyone propelled his profession down the mathematical road. “In Samuelson’s simplified model,” we are told in one history of the subject, “all the psychological concerns discussed over the previous century were compressed into a single parameter, the discount rate.”
The formula looks complex, but basically it postulates that Homo economicus rationally discounts the future at some consistent rate. What does it mean to discount the future? Imagine that you’ve won the lottery and you’re offered a choice of $150,000 a year for twelve years, or a single lump sum up front. The question is, what is the single sum you would accept to forgo the payments spread over a dozen years? Assume there’s no risk the state would fail to pay either way. So think for a minute: How much would you take in lieu of the payment flow?
Would $1 million be enough? How about $750,000? The lower the number you name, the more impatient you are; name a really low number and you might go down in history alongside Esau, who in a moment of hunger traded his birthright for a meal. Let’s say you offer to take $1 million. Should the state accept? To get the answer, lottery officials will calculate the interest rate at which the payment flows would be worth the same as the lump sum; financial calculators on the Internet make solving these “time value of money” problems easy. In this case the interest rate exceeds 10 percent—a very good investment for the state, but a bad deal for you, since you can’t safely invest $1 million to reliably return 10 percent annually.
This interest rate is also your discount rate, and it is a measure of how much the future matters to you now. The higher it is, the less you’re willing to defer gratification. Life is complicated, of course, and it’s hard to compute a clear discount rate for many of the now-or-later choices that we make. Nonetheless, under certain circumstances, it’s a way to quantify a messy thing like self-control.
These aren’t esoteric calculations; hardheaded businesspeople do them every day all over the world, for the most part without knowing that they owe the basic principle to a medieval mathematician named Leonardo of Pisa—a man mostly remembered as Fibonacci. Fibonacci wrote his famous Liber Abaci (literally, book of calculation) in 1202, and it earned him lasting renown for what would come to be known as Fibonacci’s numbers: 0, 1, 1, 2, 3, 5, 8, and so forth, where each number (after the first two) is the sum of the two preceding. In the twentieth century, the Fibonacci numbers were the subject of great interest, and the sequence has been found all over the natural world.
The significance of this series, alas, has helped obscure the fact that, for the most part, Liber Abaci is devoted to the mathematics of commerce—including how to calculate the value today of sums that are scheduled to arrive over time. When we think about the future and grope toward an estimate of the value of, say, a secure retirement or good health in our old age, we’re mentally performing something like a present-value calculation, and in doing so our choice of discount rate is crucial.
It turns out that some people labor under consistently higher discount rates than others—meaning they habitually place less value on the future. The poor seem to fall into this category, along with the young, the male, and those who are prone to substance abuse. Older people, females, and those with higher intelligence seem to have a lower discount rate and, by placing a greater value on the future, tend to be more willing to defer gratification. Studies of addicts have found that those hooked on cocaine, heroin, alcohol, gambling, or cigarettes all had a higher discount rate with respect to money (that is, they devalued delayed money more) than did people who weren’t addicted. Japanese students were found to discount future sums less than their American or Chinese counterparts. And in general small future sums are discounted more deeply than larger sums.
Animals seem to discount future rewards as well. Feeding studies of pigeons, rats, and various primates rarely elicit patience lasting more than a matter of seconds, although some chimpanzees and bonobos have been able to wait up to two minutes. Yet in their natural environment, some animals show quite modest discount rates. Squirrels stashing away nuts for the winter are deferring consumption more effectively than some humans are capable of doing, yet their behavior appears to be more instinctual than consciously planned. And of course many animals appear to “invest” in the future, as when beavers build dams or chimpanzees pour time and energy into cultivating other chimps in an effort to climb the local hierarchy of dominance.
Before we give our own species too much credit for self-control, we ought to consider just how much humans tend to discount the future. Researchers have studied this question and the scientific answer is, a helluva lot. How do they study such a thing? Basically there are two ways. One is simply to ask people how much money they would demand today to forgo, say, $100 in a year. Researchers can then calculate the discount rate implied by the sums and the delay. Another method is to seize on some real-world examples of human behavior from which a discount rate can be derived. One study, for instance, looked at the decisions of U.S. military personnel confronted with a cutback in force. The government offered more than 65,000 individuals who were about to be let go a choice between a onetime lump sum and a series of annual payments. The terms depended on rank, but one typical example was $22,283 up front or $3,714 annually for 18 years. These payments were a direct obligation of Uncle Sam, so there was no risk of default—yet the present value of the annual payments compared to the lump sum in this example implied a remarkable discount rate of 18.9 percent. That was fairly typical. So the annual payments were a vastly better deal—so good that even Warren Buffett probably couldn’t invest the lump sum any more effectively. What’s more, the lump sums would be taxed more harshly. The government went out of its way to explain the choices with pamphlets, counseling, and media. Yet more than half of officers and more than 90 percent of enlisted personnel chose the lump sum. Although this episode reflects badly on people’s ability to defer gratification, there is a silver lining: the poor choices by the outbound military personnel saved American taxpayers $1.7 billion, which is how much extra it would have cost if all had chosen the annual payments.
Studies of people’s discount rates over the years have produced a range of results, but they all tend to be high. As George Ainslie reports in Picoeconomics, his nonpareil account of self-control and the problem of the future, studies of consumer behavior (say, the choice between an air conditioner with a higher price and lower operating costs and a cheaper model that gobbles more power) find implicit discount rates running into the hundreds. Employees surveyed on how much they would accept to forgo a later bonus racked up discount rates from 36 to 122 percent. Even finance students posted median discount rates as high as 60 percent, and at an English university, grad students and staff who were asked how long they would wait for £10 instead of immediately getting £5 gave answers that reflected discount rates of 5,000 percent.
Curves
Ainslie was fascinated by the idea that behavior could be expressed as a mathematical function—a nice, smooth curve that plots our preferences over time. The curve, according to Samuelson’s theory, was smooth because we’re supposed to discount the future consistently. But Ainslie was also fascinated by the ways in which each of us seems to be in conflict with ourselves. We might have a long-term goal—for instance, saving for retirement or maintaining domestic peace—that is in conflict with our short-term desire for the pleasures that spending or infidelity might bring.
People confront this kind of problem all the time. I could work on this paragraph, contributing in some infinitesimal way to my long-term wealth and happiness, or I could shop for music on the Internet, sharply increasing my short-term happiness. I have two desires. Which should I honor? Although he was young, Ainslie understood the nature of this dilemma. So for his senior thesis at Yale, he built an unusual maze—one that confronted a rat with two paths to separate food-rewards of three pellets each. Both routes were identical except that one offered a tempting shortcut leading to only a single pellet of food. The question was, would the rat realize, after running the course repeatedly, that it should take the path that didn’t offer the shortcut?
Ainslie got conflicting data from the experiment and graduated before he could press it further. But it was still on his mind when he got to Harvard, and so the medical school’s resident behaviorist, Peter Dews, sent him to see Herrnstein. It was there that he learned to translate his rat problem into a pigeon problem.
That was in 1967, the same year Herrnstein and Shin-Ho Chung reported what was perhaps their most interesting finding to date, which was that pigeons weren’t only sensitive to quantity and frequency of reward—they were also sensitive to its timing. They would peck a lot more—disproportionately more—for an immediate grain reward than for a delayed one. When Ainslie read about this, he realized something: If these birds value short-term rewards disproportionately over long-term rewards, then those nice, smooth curves depicting their preferences must be wrong. If you plot preferences against time, and preferences rise sharply as time-to-reward gets shorter, than the curve looks different. It takes on the shape of a hyperbola, meaning it is concave, rising sharply as time grows short.
What hyperbolic discounting meant was that the pigeons weren’t acting sensibly or consistently with respect to time. And as Ainslie suspected, what was true for the pigeons turns out to be true for the rest of us as well.
To explain hyperbolic discounting, the psychologist Stuart Vyse likes to offer his students two envelopes, one containing $10 and the other $12. Naturally they pick the one with $12. Then he offers $10 now and $12 a week later, at which point most students still pick the larger amount. But when he offers the larger amount two or even three weeks later, things change, and most students say they want the $10 now. That’s not the end of the story; once Vyse has established the point at which students switch to the smaller, sooner reward, he freezes this gap and moves both rewards out into the more distant future. For instance: Would the students prefer $10 in twenty-eight weeks or $12 in thirty weeks? At this point, or with the offers somewhere out there in some far-off tomorrow, student preferences switch back again, and most go for the larger, later sum. It’s a classic example of time inconsistency; rationally, we should choose the larger, later reward in every instance, because the extra wait for it is unchanging.
Why do we experience this crucial preference reversal? Why is two weeks of delay unacceptable in the near term but not further off in time? The answer is the human tendency to place excessive value on rewards that are nearer in time. David Hume recognized this problem more than 250 years ago: “Tho’ we may be fully convinc’d, that the latter object excels the former, we are not able to regulate our actions by this judgment; but yield to the sollicitations of our passions, which always plead in favor of whatever is near and contiguous.”
We don’t just value near more than far; we value it disproportionately more. The simple fact is that people are made to prefer immediate rewards inordinately. Perhaps instinctively they know the truth of Keynes’s famous dictum: “In the long run, we are all dead.” In the evolutionary environment, life was uncertain, and without reliable mechanisms for securing or refrigerating precious meat, for example, it may not have selected very effectively for the exceedingly future-minded. The Bible gives creationists some parallel evidence for this aspect of human nature. Adam and Eve weren’t thinking very far ahead when they ate of the forbidden fruit, after all, just as their descendant Esau was blind to long-term consequences of the deal he was making.
As it turns out, the notion that the curve of our preferences might not be consistent over time was postulated in 1956 by the economist Robert H. Strotz in a witty but daunting paper breezily entitled “Myopia and Inconsistency in Dynamic Utility Maximization.” Strotz also noticed that, when temptation was far off, people with foresight would employ what he creatively dubbed “precommitment” techniques—just like Odysseus when approaching the Sirens—to limit their own choices against the day when temptation grew near. With respect to his own self-control, by the way, Strotz was a sophisticate; this brilliant economist, who would go on to become president of Northwestern University, had his academic employer pay him in twelve monthly installments because he didn’t trust himself to save enough of his school-year salary to get through the summer.
Ainslie knew nothing of Strotz, and he notes that the economist never mentioned hyperbolic discounting in his paper. What Ainslie wanted to do was prove in a lab what the matching law implied—that effort is related to reward. His idea was that, if creatures did have a powerful change of preference when confronted by an immediate (albeit inferior) reward—a change in favor of the thing right in front of them now, as implied by hyperbolic discounting—then they could be expected to learn how to prevent themselves from making this choice if the means were available. As Ainslie explained, “this prediction would not depend on the existence of ‘higher intelligence,’ but only on whether the effect of the larger reinforcement was great enough to cause learning of the pre-commitment device at the time the device was available.”
The idea, in other words, was to teach an animal to act like Odysseus.
To accomplish this, Ainslie worked with white male Carneaux pigeons. (Carneaux were the only breed used at the Pigeon Lab because of their genetic uniformity. Males were used to avoid the estrus cycle.) Kept at just 80 percent of their free-feeding weight, these birds were hungry—and were therefore paying close attention whenever there was the prospect of some grain. For Ainslie’s experiments, the birds were put, one at a time, into a soundproof, lightproof box—a Skinner box—that measured about a foot in all dimensions. In this box, the pigeon encountered one or more translucent disks on the wall; these objects are known as keys and can be lighted or dark. The box was connected to the usual Pigeon Lab apparatus for keeping track of what was going on, but it also had a peephole on top so researchers could peer inside.
Managing pigeon experiments by hand would be tedious, so each box was wired to what amounted to an early computer system, except there were no microchips involved. Inside the box, the keys (lit by 7-watt Christmas bulbs) went on and off at intervals in a sequence programmed onto a loop of movie film, which functioned much the way punch cards did in the computers of the day. Holes in the film controlled the opening and closing of electrical circuits, and the pigeons’ reactions were recorded by a printer not unlike those which produced cash register receipts.
Ainslie would put a pigeon into the box, and periodically the key on the wall would light up red for 2.5 seconds; if the pigeon pecked it during this time, the food hopper would open up for 1.5 seconds. But if the pigeon didn’t peck the key while it was red, the food hopper would open up for 4 seconds. The pigeons thus had the chance to gain a lot by exercising a little self-restraint—yet they all pecked the key as soon as it turned red. They just couldn’t stop themselves.
So Ainslie tried a different approach. This time, about 11 seconds before the key turned red, it was lit green. Pecking it while it was green prevented the key from turning red and gave the birds 4 seconds’ worth of food—eventually, for the interval was the same as if they had merely refrained from pecking red. If the birds didn’t peck green but later pecked red, the reward would be just 1.5 seconds of food.
It was a great experiment. Ainslie was giving the birds a chance to emulate wily Odysseus by binding themselves fast against the temptation to come. The green key was a classic precommitment device—the most powerful device in the self-management tool kit. Astonishingly, two pigeons proved themselves capable of precommitment: they learned to peck the green key about 90 percent of the time it was offered.
Sound like a fluke? Ainslie ran a control experiment in which the key would turn red whether or not the birds pecked it while it was green. Sure enough, the Odyssean pigeons stopped pecking the now-useless green key—but persisted in pecking red. There was no doubt about it; the pigeons who had pecked green were making effective use of pre-commitment, having discovered that they could profitably constrain their own impatience. Howard Rachlin (another leading self-control theorist) and Leonard Green, in a somewhat different experiment, published in 1972, got even more pigeons to precommit, and two years later Ainslie replicated and expanded his own work with some more pigeons.
Getting pigeons to make like Odysseus is pretty cool, but the real significance of Ainslie’s work at Harvard was in showing that organisms—whether pigeons or humans—are driven biologically to place inordinate value on immediate rewards. While others had noticed the human tendency to favor today over tomorrow where pleasure is concerned (Saint Augustine: “Give me chastity and continence, but not yet”), it was Ainslie who plotted the shape of our preferences. His experiments—which led to a lifetime of thinking and writing about hyperbolic discounting—demonstrated that conflicting desires are baked into the evolutionary cake—and that when it comes to self-control, humans and pigeons are birds of a feather.