Cover Story

Resisting Temptation

Every year the holidays put us grownups through the same wringer. Sometime late in December, we remember that the “holiday spirit” is really not about eager anticipation and indulgence; it is about controlling ourselves, resisting temptations, mastering our urges. All that food, occasions to drink a little or a lot too much, the pressure cooker of families and relatives … so many opportunities to say, do, or consume things we might regret.

And now comes January, that cold month when we all lay in our warm beds and reconsider our resolutions. Get up early and go to the gym before work? I must have been drunk when I resolved to do that.

The problem of how we master our impulses and follow through on our goals is not new, of course — it has engaged philosophers for as long as there have been people in togas. Controlling our passions and cravings was emphasized by the ancient Stoics and Buddhists, for example, and virtue derived from self-restraint is a cornerstone of the Judeo-Christian tradition. In one way or another, all cultures have regarded the ability to discipline ourselves as central to what defines us as human. Animals appear to obey their appetites in the moment; people — at least, adult people — can say “no” to what may be immediately tempting, for the sake of the greater material rewards or moral virtue that come from reining ourselves in.

Yet most of us fail to rein ourselves in as much as we would like, at least in some part of our lives. In one way or another, failures of willpower are at the root of countless problems in our society — obesity, addictions, violence, relationship problems, consumer debt, to name just a few. For this reason, the science of willpower — as a subset of the larger domain of self-regulation (see Baumeister & Vohs, 2004) — is emerging as a major priority in the psychological sciences. Where does willpower come from? Why do some people have more of it than others? How can it be strengthened?

Hot Marshmallows

The litmus test of how much willpower you have is whether you can resist a marshmallow. At least, this is the paradigm in the classic deferral-of-gratification studies conducted at Stanford in the early 1970s by current APS President Walter Mischel (see Mischel, 1996). In one version of his famous experiment, four-year-olds were left in a room with a bell, with which they could summon an experimenter who, they were promised, would give them a single marshmallow. But they were also told that if they could hold out and wait for the experimenter to return on his own, they would receive two marshmallows. In other variants, the children had the first marshmallow in front of them from the start and had to resist this immediate temptation if they were to reap the larger reward. In such studies, some kids managed to wait up to 20 minutes for the grand prize of two marshmallows, and some caved early for the lesser reward.

The principle underlying the challenge faced by those four-year-olds is sometimes called temporal discounting: To a small child, one marshmallow right away may seem more valuable than two marshmallows in some indefinite future. Transposed to adult behavior, it is easy to see how temporal discounting leads otherwise rational people to compromise their long-term health and happiness for short-term gains. That big-screen TV on sale in Best Buy might seem more tangibly rewarding as you stand gazing at it than some vague future free of credit card payments. Or tonight’s dessert special — chocolate mousse cake — may seem more important (as you observe another lucky diner savor it) than a slimmer waistline down the road.

A single marshmallow, like a single trip to the gym or a single impulse purchase, may seem trivial, but over the course of life, our successes and failures in the area of self-mastery add up to predict a lot about our success and failure in many areas. People high in self-control are healthier, have better relationships, and are more successful in school and work than those low in the trait. The children in Mischel’s marshmallow study, for example, were tracked through adolescence, and the researchers found that the four-year-olds who successfully delayed gratification to reap a larger reward were better adjusted later and scored higher on their SATs, than did those who hadn’t been able to hold out (Mischel, 1996).

Sigmund Freud thought that young children developed the capacity for deferring gratification by forming mental images of desired stimuli (the mother’s breast was Freud’s exemplar, but you could think in terms of marshmallows too); they endured frustration and laid the foundations for later self-mastery by enjoying their mental pictures in place of the real objects they represented. Mischel used his sweet-treat paradigm to test Freud’s theory and found that the Viennese psychoanalyst was only partly right. Keeping a reward in mind, and thereby maintaining an expectation of it, did seem to be important. But some mental marshmallows worked considerably better than others.

Specifically, encouraging kids to create vivid, highly arousing mental images of the sweet, gooey treat led them quickly unto temptation, whereas encouraging them to form abstract, non-arousing (or “cold”) cognitive representations (thinking of the marshmallows as “puffy clouds,” for instance) facilitated waiting. Children who were able to wait for two marshmallows did so by distracting themselves from the aversive wait by thinking about something else instead — or, in some versions of the experiment, by forming arousing, or “hot,” mental images of alternative, unavailable rewards, such as a warm fresh pretzel.

Mischel’s discovery leads to real-life strategies for self-control: Viewing temptations abstractly — “cooling” immediate stimuli — helps redress the here-and-now bias produced by temporal discounting, enabling us to take a longer-term perspective. By the same token, making long-term priorities hot adds weight to those goals, helping them defeat our short-term impulses.

Many psychologists view self-control in terms of the interplay of two distinct self-regulation systems, one that is planful and takes more mental effort, and one that is quick and impulsive and responsive to stimuli in the moment. In the model proposed by Mischel and APS Fellow Janet Metcalfe (Columbia), the more emotional “hot” system dukes it out with the reasoned and rational “cool” system (Metcalfe & Mischel, 1999). These systems are part of a larger cognitive-affective processing system — a constellation of goal representations, expectations about our self-efficacy and competence (i.e., it matters for our willpower if we believe we have it), and situational appraisals that interact and moderate each other (Mischel & Ayduk, 2004). Research in neuroscience is lending support to such dual-process theories by showing the brain systems involved in both impulsivity and self-control.

Willpower in the Brain

Most Psychology 101 students know the story of Phineas Gage, the Vermont railroad worker who received the first known prefrontal lobotomy when an explosively propelled iron rod passed clear through the front of his head, destroying parts of his frontal lobe. That Gage survived this freak accident at all is miraculous, particularly considering it happened in 1848. But it was the dramatic change in his behavior after his accident that earned Gage his permanent place in psychology textbooks. Once an “efficient and capable” foreman, his accident rendered him, according to his physician, irreverent, profane, impatient, obstinate, vacillating, and unable to follow through with his plans. He became an impulsive and ineffectual shadow of his former self – in short, “no longer Gage.”

Gage’s accident was hardly a controlled experiment, and some of the anecdotal evidence in the case is now viewed with a critical eye by scientists. But the take-home point has held for over a century and a half:  Be careful when tamping dynamite with an iron rod. The other take-home point has been pretty much just as durable, at least in psychology: Willpower and other executive faculties like decision making have a lot to do with the prefrontal cortex, or PFC.

Neuroimaging and other tools are rapidly adding to our understanding of the various executive processes that unfold in the human forebrain. The front-most or anterior portion of the PFC, which is implicated in working memory capacity, is linked to the kinds of deferral of gratification tasks studied by Mischel. This may be because overriding temporal discounting requires keeping long-term goals in mind (see Shamosh et al., 2008). Other major structures within the PFC include the dorsolateral PFC, associated with forethought and inhibition of impulses, and the ventromedial PFC, involved in regulating emotions and sensitivity to punishment and reward. The latter area is what was probably damaged in Gage’s case (see Wagar & Thagard, 2004).

Information about punishment and reward is vital to making sound decisions. Patients with damage to the ventromedial PFC have trouble taking the long view and considering the future consequences of immediate payoffs — what University of Iowa neurologist Antoine Bechara and his colleagues have called “myopia for the future” (Bechara, Tranel, & Damasio, 2000). This is shown in a risk-decision paradigm known as the Iowa Gambling Task, in which participants overturn cards from an array of decks that vary in the size of the monetary rewards offered. Some of the decks appear highly rewarding at first but also contain catastrophic losses that ultimately bankrupt players who preferentially choose them. People with ventromedial PFC damage tend to persist in picking from these high-payout but ultimately treacherous bad decks. Bechara (2005) suggests that failures of willpower such as drug addiction can be understood as a failure of the prefrontal, long-term reward/punishment system to suppress an impulsive, short-term reward/punishment system involving the brain’s limbic system, particularly the amygdala.

Whether the cool-headed prefrontal system is able to override our emotional, amygdala-based impulses depends on various factors. Age is one of them: The ventromedial PFC matures later than other brain systems (it is still developing in early adulthood); as a result, children and adolescents also perform poorly on the Iowa Gambling Task (Hooper, Luciana, Conklin, & Yarger, 2004), and generally make poorer decisions than adults do (see Reyna & Farley, 2006). Stress is another factor: Exposure to stressors such as loud noises, crowds, bureaucratic frustrations, or being discriminated against has been found to reduce people’s performance in subsequent self-control tasks (see Muraven & Baumeister, 2000).

Flex Your Muscles

One of the main limits on willpower, though, turns out to be … willpower. Exerting self-control in one domain makes it harder to exert self-control in another, at least right away. Over the past decade and a half, a large number of studies have shown that executive processes by which we control our impulses behave a lot like muscles that get tired through use.

In a study led by APS Fellow Roy Baumeister (Florida State University), a group of hungry participants was forbidden from eating freshly baked cookies sitting on a plate in front of them and made to eat radishes instead. These participants gave up faster on a subsequent frustrating task than did a control group who had been freely allowed to indulge their sweet tooth (Baumeister, Bratslavsky, Muraven, & Tice, 1998). And in a study led by Mark Muraven (SUNY-Albany), participants made to suppress all thoughts of a white bear for five minutes consumed more beer afterwards in a “taste test” than did those in a control group, even though they knew they would subsequently be taking a driving test (Muraven, Collins, & Nienhaus, 2002).

Numerous variants of this paradigm — making a group of participants exercise restraint in one situation and then comparing their performance with that of a control group in a subsequent self-control task — have shown the same pattern: Self-control is a limited resource that can be drained through exertion (Baumeister, Vohs, & Tice, 2007). Baumeister and colleagues call this fatigued state “ego-depletion,” and it is a significant discovery because it explains why many of our specific willpower failures occur when our strength has been taxed by other self-control demands. Refraining from blowing up at one’s boss during the day may make it hard to resist a big meal of comfort food that evening. The constant effort of sticking to a diet may cause us to make more impulsive purchases at the mall.

The muscle that controls willpower does much more than just keep our impulses in check. It is part of a larger set of executive functions involved in self-monitoring, coping with stressors, weighing alternatives, and making decisions, all of which draw on the same limited energy source. In a recent series of studies led by Kathleen Vohs (University of Minnesota), participants who were asked to choose among various consumer products subsequently showed diminished pain tolerance and consumed less of a nasty-tasting vinegar drink (even though they were paid a nickel per ounce consumed); college students asked to choose among various college courses subsequently studied less for a math test, opting to play video games or read magazines instead (Vohs et al., 2008). Non-ego-depleted persons are also better at logical reasoning and intelligent thought (see Baumeister, Vohs, & Tice, 2007) and at dealing with setbacks that are unexpected (see Baumeister, 2008).

It is even possible to become ego-depleted by watching other people exert willpower. In a new study by psychologists at Yale and UCLA, participants were asked to put themselves in the shoes of a fictional hungry waiter or waitress in a gourmet restaurant who was forbidden from eating on the job. They then viewed pictures of various products like watches, cars, and appliances and rated how much they would be willing to spend for them. Exercising vicarious self-control led people to be willing to spend more on the consumer goods, as compared with a control group (Ackerman, Goldstein, Shapiro, & Bargh, in press).

Impulse purchases and eating binges may let us know our willpower is tired, yet until recently psychologists have not known how to assess self-control effort independently of such behavioral indicators. But based on the fact that self-regulation overlaps with brain systems governing the autonomic, “fight-or-flight” stress response, University of Kentucky psychologist Suzanne Segerstrom hypothesized that physiological responses shared with stress might also work as a measure of effort at self-regulation. In 2007, she and her collaborator Lise Solberg Nes published the results of a study showing that participants’ heart-rate variability (speeding and slowing of heartbeats), a common stress response, also increased when participants resisted cookies in favor of carrots and when they worked at solving tough anagrams (Segerstrom & Solberg Nes, 2007). Although people often know it when they are stressed, they generally aren’t directly aware of exerting self-regulatory exertion, especially for such trivial stakes as a cookie. The new finding points to one way future researchers may directly measure flexing and fatigue of the willpower muscle.

But the $64,000 question is this: If willpower acts like a muscle, can we strengthen it through exercise? Evidence so far suggests the answer is yes: Using self-control in specific areas such as spending or exercise can gradually increase one’s resistance to ego-depletion, even in unrelated self-control tasks (Baumeister, Gailliot, DeWall, & Oaten, 2006). For example, in one study, participants suppressed all thoughts of a white bear during an initial task subsequently gave up sooner on a strenuous handgrip task. They all then returned to the lab two weeks later to undergo the same two ordeals; however, in the intervening time, one group engaged in some form of self-regulatory exercise — tracking their food consumption, controlling their mood, or working on their posture. On their second lab visit, those who had exercised in the interim were not as worn out by not thinking of a white bear. Other studies have shown the same benefit of doing daily exercises like using one’s nondominant hand for routine tasks or working to improve one’s language (such as trying not to curse).

Free Will Hunting

Psychological scientists have often shied away from using terms like “will” or “willpower,” preferring the less philosophically or morally loaded “self-control” or the broader “self-regulation.” Since the Enlightenment, science has generally accepted that our bodies, and our brains, are machines, subject to the laws of mechanistic causality. If mental processes unfold like clockwork, where could will — that is, free personal choice — enter in?

Attempts to reconcile free will with mechanistic determinism tend to result in the homuncular fallacy: Medieval physicians blamed certain mental and physical ailments on homunculi, miniature people working mischief inside our heads and bodies; such a “mini-me” has served psychologists and logicians as an apt metaphor for the hall of mirrors that often results when attempting to explain unconstrained choice in a deterministic universe: If I’m not the one consciously deciding my destiny, something inside of me is. But if I’m just the vehicle for an inner decider, then how does the decider decide? And how is that decided? Simply, how can you account for free will without endlessly deferring the question to a homunculus inside a homunculus inside a homunculus, on to infinity? Just thinking about it is enough to make your homunculus spin.

Twentieth-century psychology tended to deal with the homunculus problem by severely restricting the scope of free will. Freud, for example, argued that much of what we think of as willed behavior is actually governed by instincts and drives that we are unaware of. But to many of his critics, an unconscious self (especially one that expresses not-ready-for-prime-time desires through dreams, symptoms, and slips of the tongue) sounds like just another homunculus, not a real solution to the problem of personal agency. The more empirically grounded Behaviorists of the 1950s, such as B.F. Skinner, advocated that we stoically banish the notion of freedom altogether: Human behavior was purely mechanistic, as obedient to discernible causal laws as any other physical process, and we would create a better world if we gave up our childish belief in free will.

Findings in the neurosciences and cognitive science have lent considerable tacit support to this position. The neurobiologist Benjamin Libet found that motor actions subjects experienced as being consciously willed were actually initiated in the brain as much as a half second prior to subjects’ awareness of their own intention (Libet, 1985). APS Fellow and Charter Member John Bargh (Yale) has conducted numerous studies showing the degree to which our actions are governed by automatic, nonconscious processes (see Bargh & Chartrand, 1999; Bargh & Williams, 2006). In light of these kinds of findings, some researchers, like APS Fellow Daniel M. Wegner (Harvard), argue that the feeling of conscious will is an illusion that arises when our automatically guided actions happen to coincide with our internal monitoring of them (Wegner, 2002).

One problem with eliminating free will from the picture is that with it goes the concept of personal responsibility, the philosophical underpinning of modern societies and the root of all ethics. For medieval theologians, human freedom was what makes humans capable of salvation — we can choose rightly and wrongly, and thereby help determine our fate not only here on earth but also in the hereafter. Jean-Paul Sartre went so far as to maintain that every second of our lives we are “radically free” to choose our destiny, fully responsible for every single thing we do, say, or think. (The French philosopher would have little patience for my excuses for not going to the gym.) In 2006, an anonymous editorial-writer in The Economist worried what would happen if neuroscience, by exposing the mechanisms of decision making, caused the ideas of free will and responsibility to disappear (Free to Choose?, 2006).

Bearing out such fears, recent research suggests that a world that disbelieved in free will would be a worse place, not better. In a study by Vohs and APS Fellow Jonathan W. Schooler (University of California, Santa Barbara) participants who read a passage about free will’s nonexistence by the biologist Francis Crick (the discoverer of DNA) were more likely to cheat on a subsequent arithmetic task than were controls who read a neutral essay (Vohs & Schooler, 2008). In another study, participants read a series of passages that either affirmed or denied the existence of free will and then answered a set of GRE questions; they checked their own answers and rewarded themselves monetarily for their number correct. Again, there was a significant effect: Belief in determinism promoted cheating.

Are the notions of free will — and by extension willpower — simply convenient illusions that somehow keep us from misbehaving?

Sweet Freedom

Ironically, it may be the empirical mindset of psychology that rescues free will from oblivion. Complete determinism is just as unproven and scientifically unprovable as complete freedom is, Baumeister (2008) points out. Moreover, the fact that belief in and desire for freedom are enduring and universal features of human societies makes little sense if there is not some way in which free will is a relevant and useful construct. Baumeister argues that it is most productive to think of it as a continuum: Some people have it more than others, sometimes we are able to hold out for two marshmallows and sometimes we aren’t, and it’s the difference that is revealing and informative in a science of mind.

The ability to exercise conscious will over our impulses, Baumeister argues, was a late-evolving system that conferred enormous advantages on our ancestors by enabling them to live in complex cultural groups. The executive gray matter occupying the proud position over our eyes expanded rapidly after our hominid ancestors rose up on two feet and began making tools, communicating, and cooperating for the common good. Other social animals exert self-control too, but submitting that control to conscious, reasoned guidance — what most people mean by free will — seems to be a hallmark of humans, whose intricate, rule-bound societies far surpass in complexity those of other species. Yet free action was, and remains, limited by its high energetic costs, as revealed by the ego-depletion research already mentioned.

Recent research by one of Baumeister’s students, Matt Gailliot (University of Amsterdam), has revealed just how apt the term willpower — once regarded by psychologists as merely a metaphor or folk concept — really is. The brain is powered by glucose, the same thing that powers your muscles. And though the brain makes up only 2 percent of the body’s mass, it consumes an amazing 75 percent of the glucose in the blood; executive functions like reasoning, decision making, and impulse control may be especially demanding (Gailliot, 2008). Increased blood glucose has been linked to improved executive processing, working memory, and reaction time; low blood glucose or problems utilizing it have been linked to such self-control problems as aggression and criminality, impulsivity, poor attentional and emotional control, trouble coping with stress, and difficulty quitting smoking.

While working with Baumeister at Florida State University, Gailliot conducted a series of studies showing just what a glucose guzzler willpower can be (Gailliot et al., 2007). Participants in ego-depletion experiments were found to have significantly lower blood sugar following willpower-demanding tasks. And those who replenished their blood sugar with a sweetened drink after exercising self-control in an initial task were better able to master temptation in a subsequent task than a control group of subjects who instead quenched their thirst with an artificially sweetened beverage. Gailliot has also gone on to examine the role of glycogen, the chemical form in which glucose is stored in the body and brain for future use (Gailliot, 2008). The brain’s glycogen stores become recharged during sleep and gradually become depleted over the course of the day. It is likely that this pattern partly accounts for the diminished willpower (and other aspects of executive functioning) commonly experienced in the evening. (Other factors that can drain glycogen stores are stress, alcohol, and possibly even high summer temperatures.)

The seeming irony for dieters in the willpower–blood glucose link is not lost on Gailliot and his colleagues: Future research will need to examine if there is a Catch-22 effect whereby limiting sugar intake through dieting could actually impede our ability to resist, say, sugary food (there’s something else to make your homunculus spin). In any case, the authors emphasize that the lesson in their research is not to start sucking down soda or candy bars in an effort to boost willpower: Such foods give a quick, temporary boost to blood sugar, which is why the researchers used sweet drinks for their laboratory studies, but the effect is short lived. Over time, high sugar consumption can lead to insulin resistance or even diabetes, conditions in which the ability to metabolize glucose is seriously impaired (Taubes, 2007), with all the willpower problems that may entail. Fortunately, the glucose fuel needed for the body and brain are metabolized from many other, better food sources; protein and complex carbohydrates (e.g., vegetables and fruit) maintain a steady supply of glucose throughout the day and are surely a better bet for optimal brainpower and willpower (Gailliot et al, 2007).

Saving Gas

Willpower is a highly energy-demanding process in an already energy-hungry organ, so it makes sense that we only use it when we have to (Baumeister, 2008). Most of the time, for many of our daily-life activities, relatively automatic processes and reflexive responses to situational cues are good enough to guide us.

Given willpower’s scarcity, an effective strategy for following through on your New Year’s resolutions may be to circumvent it altogether. APS Fellow and Charter Member Peter M. Gollwitzer (NYU) has studied the difference between mere goals (such as “I want to lose weight”) and implementation intentions, which actually specify a plan of action (“When the waiter comes, I’m going to order a salad”; Gollwitzer, Fujita, & Oettingen, 2004). The latter prove to be much more effective in helping us keep our promises to ourselves. Implementation intentions help us notice opportunities for working toward our objectives (e.g., the waiter taking my order) and even anticipate likely obstacles, prescribing ways of dealing with them. In a number of experiments, Gollwitzer has found that creating mental links between situational cues and predecided courses of action — “if–then plans” — helps enlist the force of habit to alleviate willpower’s brain burden, allowing a predetermined behavior to unfold without too much effortful deliberation.

Whether we view willpower in terms of its muscle-like qualities or in terms of competing (e.g., hot vs. cold) circuits in a complex cognitive control system, a positive message in the psychological science of willpower is that self-mastery, however limited, is also responsive to intervention and improvement. By adjusting how we think of the things that tempt us in the short term and the goals we are striving for in the long, we can alter the balance of these influences. We can also build self-control the same way we build our biceps, not only improving its strength but also intelligently allocating it to high-priority challenges rather than wasting it on things that are less important. Future research may show that we can even optimize our willpower through diet. And knowing willpower’s limitations helps point us in the direction of strategies that can help our goals run off as planned even if our inner decider is fatigued.

This leaves us with no more excuses. I’m off to the gym.


APS regularly opens certain online articles for discussion on our website. Effective February 2021, you must be a logged-in APS member to post comments. By posting a comment, you agree to our Community Guidelines and the display of your profile information, including your name and affiliation. Any opinions, findings, conclusions, or recommendations present in article comments are those of the writers and do not necessarily reflect the views of APS or the article’s author. For more information, please see our Community Guidelines.

Please login with your APS account to comment.