The Puzzling Puzzles ofHarry Harlow and Edward Deci
In the middle of the last century, two young scientists conductedexperiments that should have changed the world— but did not.Harry F. Harlow was a professor of psychology at the Universityof Wisconsin who, in the 1940s, established one of the world’s firstlaboratories for studying primate behavior. One day in 1949, Harlowand two colleagues gathered eight rhesus monkeys for a two- weekexperiment on learning. The researchers devised a simple mechanicalpuzzle like the one pictured on the next page. Solving it requiredthree steps: pull out the vertical pin, undo the hook, and lift thehinged cover. Pretty easy for you and me, far more challenging for athirteen- pound lab monkey.
Harlow’s puzzle in the starting (left) and solved (right) positions.
The experimenters placed the puzzles in the monkeys’ cages toobserve how they reacted— and to prepare them for tests of theirproblem- solving prowess at the end of the two weeks. But almostimmediately, something strange happened. Unbidden by any outsideurging and unprompted by the experimenters, the monkeys beganplaying with the puzzles with focus, determination, and what lookedlike enjoyment. And in short order, they began figuring out how thecontraptions worked. By the time Harlow tested the monkeys ondays 13 and 14 of the experiment, the primates had become quiteadept. They solved the puzzles frequently and quickly; two- thirds ofthe time they cracked the code in less than sixty seconds.
Now, this was a bit odd. Nobody had taught the monkeys howto remove the pin, slide the hook, and open the cover. Nobody hadrewarded them with food, affection, or even quiet applause whenthey succeeded. And that ran counter to the accepted notions of howprimates— including the bigger- brained, less hairy primates knownas human beings— behaved.
Scientists then knew that two main drives powered behavior. Thefirst was the biological drive. Humans and other animals ate to satetheir hunger, drank to quench their thirst, and copulated to satisfytheir carnal urges. But that wasn’t happening here. “Solution did notlead to food, water, or sex gratification,” Harlow reported.1But the only other known drive also failed to explain the monkeys’peculiar behavior. If biological motivations came from within,this second drive came from without— the rewards and punishmentsthe environment delivered for behaving in certain ways. This wascertainly true for humans, who responded exquisitely to such externalforces. If you promised to raise our pay, we’d work harder. If youheld out the prospect of getting an A on the test, we’d study longer.If you threatened to dock us for showing up late or for incorrectlycompleting a form, we’d arrive on time and tick every box. But thatdidn’t account for the monkeys’ actions either. As Harlow wrote, andyou can almost hear him scratching his head, “The behavior obtainedin this investigation poses some interesting questions for motivationtheory, since significant learning was attained and efficient performancemaintained without resort to special or extrinsic incentives.”What else could it be?
To answer the question, Harlow offered a novel theory— whatamounted to a third drive: “The performance of the task,” he said,“provided intrinsic reward.” The monkeys solved the puzzles simplybecause they found it gratifying to solve puzzles. They enjoyed it.The joy of the task was its own reward.
If this notion was radical, what happened next only deepened theconfusion and controversy. Perhaps this newly discovered drive—Harlow eventually called it “intrinsic motivation”— was real. Butsurely it was subordinate to the other two drives. If the monkeyswere rewarded— with raisins!— for solving the puzzles, they’d nodoubt perform even better. Yet when Harlow tested that approach,the monkeys actually made more errors and solved the puzzles lessfrequently. “Introduction of food in the present experiment,” Harlowwrote, “served to disrupt performance, a phenomenon not reportedin the literature.”
Now, this was really odd. In scientific terms, it was akin to rollinga steel ball down an inclined plane to measure its velocity—only to watch the ball fl oat into the air instead. It suggested thatour understanding of the gravitational pulls on our behavior wasinadequate— that what we thought were fixed laws had plenty ofloopholes. Harlow emphasized the “strength and persistence” of themonkeys’ drive to complete the puzzles. Then he noted:It would appear that this drive . . . may be as basic and strongas the [other] drives. Furthermore, there is some reason tobelieve that [it] can be as efficient in facilitating learning.2
At the time, however, the prevailing two drives held a tight grip onscientific thinking. So Harlow sounded the alarm. He urged scientiststo “close down large sections of our theoretical junkyard” andoffer fresher, more accurate accounts of human behavior.3 He warnedthat our explanation of why we did what we did was incomplete. Hesaid that to truly understand the human condition, we had to takeaccount of this third drive.
Then he pretty much dropped the whole idea.Rather than battle the establishment and begin offering a morecomplete view of motivation, Harlow abandoned this contentiousline of research and later became famous for studies on the scienceof affection.4 His notion of this third drive bounced around the psychologicalliterature, but it remained on the periphery— of behavioralscience and of our understanding of ourselves. It would be twodecades before another scientist picked up the thread that Harlowhad so provocatively left on that Wisconsin laboratory table.In the summer of 1969, Edward Deci was a Carnegie Mellon Universitypsychology graduate student in search of a dissertation topic.Deci, who had already earned an MBA from Wharton, was intriguedby motivation but suspected that scholars and businesspeople hadmisunderstood it. So, tearing a page from the Harlow playbook, heset out to study the topic with the help of a puzzle.
Deci chose the Soma puzzle cube, a then popular Parker Brothersoffering that, thanks to YouTube, retains something of a cultfollowing today. The puzzle, shown below, consists of seven plasticpieces— six comprising four one- inch cubes, one comprising threeone- inch cubes. Players can assemble the seven pieces into a few millionpossible combinations— from abstract shapes to recognizableobjects.
The seven pieces of the Soma puzzle unassembled (left) and then fashioned into one ofseveral million possible configurations
For the study, Deci divided participants, male and female universitystudents, into an experimental group (what I’ll call GroupA) and a control group (what I’ll call Group B). Each participated inthree one- hour sessions held on consecutive days.
Here’s how the sessions worked: Each participant entered a roomand sat at a table on top of which were the seven Soma puzzle pieces,drawings of three puzzle configurations, and copies of Time, The NewYorker, and Playboy. (Hey, it was 1969.) Deci sat on the opposite endof the table to explain the instructions and to time performance witha stopwatch.
In the first session, members of both groups had to assemble theSoma pieces to replicate the configurations before them. In the secondsession, they did the same thing with different drawings— onlythis time Deci told Group A that they’d be paid $1 (the equivalentof nearly $6 today) for every configuration they successfully reproduced.Group B, meanwhile, got new drawings but no pay. Finally,in the third session, both groups received new drawings and had toreproduce them for no compensation, just as in session one. (See thetable below.)
The twist came midway through each session. After a participanthad assembled the Soma puzzle pieces to match two of the threedrawings, Deci halted the proceedings. He said that he was going togive them a fourth drawing—but to choose the right one, he neededto feed their completion times into a computer. And— this being thelate 1960s, when room- straddling mainframes were the norm anddesktop PCs were still a decade away— that meant he had to leavefor a little while.
On the way out, he said, “I shall be gone only a few minutes, youmay do whatever you like while I’m gone.” But Deci wasn’t reallyplugging numbers into an ancient teletype. Instead, he walked toan adjoining room connected to the experiment room by a one- waywindow. Then, for exactly eight minutes, he watched what peopledid when left alone. Did they continue fiddling with the puzzle,perhaps attempting to reproduce the third drawing? Or did they dosomething else— page through the magazines, check out the centerfold,stare into space, catch a quick nap?
In the first session, not surprisingly, there wasn’t much differencebetween what the Group A and Group B participants did duringthat secretly watched eight- minute free- choice period. Both continuedplaying with the puzzle, on average, for between three and ahalf and four minutes, suggesting they found it at least somewhatinteresting.
On the second day, during which Group A participants were paidfor each successful configuration and Group B participants were not, theunpaid group behaved mostly as they had during the first free- choiceperiod. But the paid group suddenly got really interested in Soma puzzles.On average, the people in Group A spent more than five minutesmessing with the puzzle, perhaps getting a head start on that thirdchallenge or gearing up for the chance to earn some beer money whenDeci returned. This makes intuitive sense, right? It’s consistent withwhat we believe about motivation: Reward me and I’ll work harder.Yet what happened on the third day confirmed Deci’s own suspicionsabout the peculiar workings of motivation— and gently calledinto question a guiding premise of modern life. This time, Deci toldthe participants in Group A that there was only enough money topay them for one day and that this third session would therefore beunpaid. Then things unfolded just as before— two puzzles, followedby Deci’s interruption.
During the ensuing eight- minute free- choice period, the subjectsin the never- been- paid Group B actually played with the puzzlefor a little longer than they had in previous sessions. Maybe theywere becoming ever more engaged; maybe it was just a statisticalquirk. But the subjects in Group A, who previously had been paid,responded differently. They now spent significantly less time playingwith the puzzle— not only about two minutes less than duringtheir paid session, but about a full minute less than in the firstsession when they initially encountered, and obviously enjoyed, thepuzzles.
In an echo of what Harlow discovered two decades earlier, Decirevealed that human motivation seemed to operate by laws that rancounter to what most scientists and citizens believed. From the officeto the playing field, we knew what got people going. Rewards—especially cold, hard cash— intensified interest and enhanced performance.What Deci found, and then confirmed in two additionalstudies he conducted shortly thereafter, was almost the opposite.“When money is used as an external reward for some activity, thesubjects lose intrinsic interest for the activity,” he wrote.5 Rewardscan deliver a short- term boost— just as a jolt of caffeine can keepyou cranking for a few more hours. But the effect wears off— and,worse, can reduce a person’s longer- term motivation to continue theproject.
Human beings, Deci said, have an “inherent tendency to seekout novelty and challenges, to extend and exercise their capacities,to explore, and to learn.” But this third drive was more fragile thanthe other two; it needed the right environment to survive. “Onewho is interested in developing and enhancing intrinsic motivationin children, employees, students, etc., should not concentrate onexternal- control systems such as monetary rewards,” he wrote in afollow- up paper.6 Thus began what for Deci became a lifelong questto rethink why we do what we do— a pursuit that sometimes puthim at odds with fellow psychologists, got him fired from a businessschool, and challenged the operating assumptions of organizationseverywhere.
“It was controversial,” Deci told me one spring morning fortyyears after the Soma experiments. “Nobody was expecting rewardswould have a negative effect.”
This is a book about motivation. I will show that much of whatwe believe about the subject just isn’t so— and that the insights thatHarlow and Deci began uncovering a few decades ago come muchcloser to the truth. The problem is that most businesses haven’tcaught up to this new understanding of what motivates us. Too manyorganizations— not just companies, but governments and nonprofitsas well— still operate from assumptions about human potential andindividual performance that are outdated, unexamined, and rootedmore in folklore than in science. They continue to pursue practicessuch as short- term incentive plans and pay- for- performance schemesin the face of mounting evidence that such measures usually don’twork and often do harm. Worse, these practices have infiltrated ourschools, where we ply our future workforce with iPods, cash, andpizza coupons to “incentivize” them to learn. Something has gonewrong.
The good news is that the solution stands before us— in thework of a band of behavioral scientists who have carried on the pioneeringefforts of Harlow and Deci and whose quiet work over thelast half- century offers us a more dynamic view of human motivation.For too long, there’s been a mismatch between what scienceknows and what business does. The goal of this book is to repair thatbreach.
Drive has three parts. Part One will look at the fl aws in ourreward- and- punishment system and propose a new way to thinkabout motivation. Chapter 1 will examine how the prevailing viewof motivation is becoming incompatible with many aspects of contemporarybusiness and life. Chapter 2 will reveal the seven reasonswhy carrot- and- stick extrinsic motivators often produce theopposite of what they set out to achieve. (Following that is a shortaddendum, Chapter 2a, that shows the special circumstances whencarrots and sticks actually can be effective.) Chapter 3 will introducewhat I call “Type I” behavior, a way of thinking and an approach tobusiness grounded in the real science of human motivation and poweredby our third drive— our innate need to direct our own lives, tolearn and create new things, and to do better by ourselves and ourworld.
Part Two will examine the three elements of Type I behavior andshow how individuals and organizations are using them to improveperformance and deepen satisfaction. Chapter 4 will explore autonomy,our desire to be self- directed. Chapter 5 will look at mastery,our urge to get better and better at what we do. Chapter 6 willexplore purpose, our yearning to be part of something larger thanourselves.
Part Three, the Type I Toolkit, is a comprehensive set of resourcesto help you create settings in which Type I behavior can fl ourish.Here you’ll find everything from dozens of exercises to awakenmotivation in yourself and others, to discussion questions for yourbook club, to a supershort summary of Drive that will help youfake your way through a cocktail party. And while this book ismostly about business, in this section I’ll offer some thoughts abouthow to apply these concepts to education and to our lives outside ofwork.
But before we get down to all that, let’s begin with a thoughtexperiment, one that requires going back in time— to the days whenJohn Major was Britain’s prime minister, Barack Obama was a skinnyyoung law professor, Internet connections were dial- up, and a blackberrywas still just a fruit.