The Fourth Cookie Problem

A man steps into a corner office. He has earned the promotion through years of careful work: listening in meetings, reading the room, knowing when to push and when to defer. These were the skills that lifted him. Within months, something shifts. Colleagues notice he interrupts more. He listens less. The junior staff he once mentored seem to have become invisible to him, as though they have dropped below some threshold of his attention. He has not become a worse person, exactly. He has become a different one.

For most of the twentieth century, psychology offered a tidy explanation for this kind of transformation. Power, the thinking went, simply reveals what was always there. Give a generous person authority and they remain generous. Give a tyrant a throne and the tyranny that was latent becomes visible. Power was a spotlight, not a force. It exposed character but did not create it.

That assumption has not held up. Over the past two decades, a small but determined body of research has arrived at a stranger and more unsettling conclusion. Power does not merely reveal who we are. It changes how the brain works. It alters behavior, hormones, and even the neural circuitry that lets one human being feel what another is feeling. The change is measurable in a laboratory, visible in a skull, and consistent enough that at least one prominent scientist has compared it, only half in jest, to a head injury.

The Paradox at the Center of Power

The scientist most responsible for this shift in understanding is Dacher Keltner, a psychologist at the University of California, Berkeley. Keltner spent the better part of twenty years watching how power moves through groups: in college dormitories, in summer camps, in courtrooms and boardrooms and family kitchens. He was looking for the rules that govern who rises and who falls, and what happens to people once they arrive.

What he found contained a deep irony, one he eventually named the power paradox.¹ People tend to gain power, Keltner observed, through qualities that benefit the groups around them. They are attentive. They cooperate. They read social situations with care and act in ways that make others feel understood. Empathy, in other words, is a kind of currency. We hand authority to the people who seem to see us most clearly.

And then, once they hold that authority, the very capacities that earned it begin to erode. The attentive listener becomes the person who finishes your sentences. The reader of rooms stops reading. “Once we have power,” Keltner wrote, “we lose some of the capacities we needed to gain it in the first place.”¹ The skills that elevate us are the first to decay once we are elevated.

This is not a story about a few bad apples corrupted by greed. Keltner’s point is more disturbing precisely because it is so ordinary. The pattern appeared again and again, across personalities, across institutions, across people who by every other measure were decent and well-intentioned. Whatever was happening did not depend on a flaw of character. It seemed to depend on the simple fact of being in charge.

What a Plate of Cookies Reveals

The difficulty with studying power is that it usually accumulates slowly, tangled up with money, status, history, and personality. To isolate the effect, researchers needed to create power in the laboratory, suddenly and arbitrarily, and watch what it did to people who a moment earlier had none.

One of the most quietly famous experiments in this literature involved cookies. Keltner and his colleagues assembled groups of three students and assigned one of them, at random, to a position of authority over the others. The leader was given the task of evaluating the group’s work, a small but real form of control. Partway through the session, a plate arrived bearing four cookies for the three participants.¹

Here is the part everyone understands intuitively. Nobody takes the last cookie. Three people, three cookies, one left over: the fourth cookie sits on the plate, claimed by no one, because grabbing it would announce something unflattering about the grabber. It is a tiny social contract, enforced by nothing but mutual awareness.

The randomly chosen leaders broke that contract. They reached for the extra cookie noticeably more often than the others. And the data went beyond the act of taking. The leaders, observers noted, ate more sloppily. They chewed with their mouths open. They scattered crumbs across the table without seeming to register it. A position of authority handed out minutes earlier, by the flip of a coin, had already loosened their grip on the small considerations that hold social life together.

The cookie study is easy to smile at, and that is partly why it endures. But its implication is serious. If a trivial, arbitrary, instantly-conferred sense of power can change how a person eats a snack in front of strangers, what does real power, held for years, do to the architecture of attention?

Drawing the Letter on Your Own Forehead

Adam Galinsky, a social psychologist now at Columbia Business School, designed an experiment that gets even closer to the mechanism. The setup is something a reader can try right now. Take your finger, place it on your forehead, and draw the capital letter E.

There are two ways to do it. You can draw the E so that it reads correctly to yourself, which means it appears backwards to anyone facing you. Or you can draw it so that it reads correctly to an observer, which means it is reversed from your own vantage point. The choice reveals, in a single stroke, whose perspective your mind defaults to: your own, or someone else’s.

Galinsky and his colleagues primed participants to feel either powerful or powerless before asking them to perform the task.² People who had been made to feel powerful were significantly more likely to draw the E so that it read correctly only to themselves, backwards to everyone else. By one measure, the powerful drew the self-oriented E roughly three times as often as the powerless.² They were not being selfish in any moral sense. They simply found it harder, in that moment, to spontaneously imagine how the world looked from another point of view.

This is the heart of it. Power does not necessarily make people crueler. It makes them less able to take another perspective, which is a different and quieter problem. Cruelty announces itself. The slow shrinking of one’s capacity to model another mind does not.

Galinsky’s broader work documented the same drift across many tasks. Powerful participants were worse at judging others’ emotions from facial expressions, worse at recognizing when their own knowledge was not shared by a listener, more anchored in their own experience. The powerful, it turned out, were lonely in a peculiar way: increasingly sealed inside their own viewpoint, often without knowing it.

Inside the Skull, the Circuit Dims

Behavior is suggestive, but it leaves open the question of what is physically happening. For that, the research had to go inside the brain. The person who took it there was Sukhvinder Obhi, a neuroscientist at McMaster University in Canada.

Obhi was interested in a system the brain uses to connect us to other people, sometimes called the mirroring system. The idea, developed from decades of work beginning with research on macaque monkeys in the 1990s, is that when you watch someone perform an action, parts of your own motor system quietly activate as if you were doing it yourself.³ Watch a person reach for a glass and the regions of your brain that govern reaching stir faintly to life. This neural echo is thought to be one of the deep roots of empathy: a way the brain simulates the experiences of others from the inside, rather than reasoning about them from the outside.

Obhi wanted to know whether power affected this system directly. He and his colleagues primed participants to feel either powerful or powerless, then used transcranial magnetic stimulation, a focused magnetic pulse delivered to the scalp, to measure how strongly the motor cortex responded while subjects watched a simple hand action.⁴ The technique allows researchers to gauge the strength of the mirroring response with some precision.

The result was clear. In participants primed to feel powerful, the mirroring signal dropped sharply.⁴ The circuit that lets us resonate with the actions of others, the neural foundation of feeling-with, was being measurably suppressed. Power did not just change what people chose to do. It changed how their nervous systems responded to the presence of another human being.

Obhi’s conclusion was carefully worded but striking. Power, he found, changes how the brain itself responds to other people, dampening a process that operates below conscious awareness. This was no longer a story about attitudes or motivation. It was a story about wiring.

Power as a Kind of Brain Damage

It is this finding that led to the most provocative framing in the whole field. Keltner has noted that the behavioral profile of the powerful, the impulsivity, the reduced awareness of others, the difficulty taking another’s perspective, bears an unsettling resemblance to patients who have suffered damage to the orbitofrontal cortex, a region of the brain involved in social judgment and empathy.⁵

People with injuries to this part of the brain can become strangely impaired in social life. They grow impulsive. They lose their sense of how their behavior lands on others. They become, in a phrase that recurs in the literature, less able to read the room. Keltner’s point, sharpened into a memorable line by the journalist Jerry Useem in The Atlantic, was that power produces a startlingly similar set of symptoms without any physical injury at all.⁵ The brain, exposed to authority, begins to behave as though a particular region has gone partly offline.

The hormones reinforce the picture. Power is associated with a shift in the body’s chemistry: a tendency toward higher testosterone and lower cortisol, the hormone most linked to stress.⁶ In studies of this so-called dominance profile, higher testosterone has been connected to greater appetite for risk and reduced caution. The combination produces a particular interior state. The brain on power feels focused, energized, and oddly invulnerable, while the quiet machinery that keeps it tethered to other minds runs slower and softer than before. It is not stupidity. It is a kind of selective deafness, and the person experiencing it rarely hears the silence.

The Circuit Can Be Switched Back On

If the story ended there, it would be a counsel of despair. Anyone who gained authority would be doomed to a slow neurological estrangement from the people they led, and the only safe society would be one with no leaders at all.

But the most important finding in this body of work is also the most hopeful one. The change is not damage in the literal, permanent sense. The empathy circuit is not destroyed. It is dimmed. And dimmed circuits can be turned back up.

The evidence for reversibility comes from the same priming logic that exposed the problem in the first place. Just as researchers could make people feel powerful and watch their mirroring response fall, they could shift people’s mindset and watch elements of empathy return. In Obhi’s work and the broader literature, the suppression of the mirroring system tracked the felt sense of power rather than any fixed trait, which means it moves when the mindset moves.⁴ Keltner and others have found that practices as simple as recalling a time of one’s own powerlessness, or deliberately attending to the experiences of others, can restore the perspective-taking that power erodes.¹

This points to something practical. The antidote to the power paradox is not the abandonment of power but the cultivation of a particular orientation while holding it. Gratitude. Humility. The deliberate, effortful practice of listening, especially to the people authority makes it easy to ignore. Leaders who keep their own histories of struggle close, who remember the view from below, seem to keep the circuit alive.

The organizational research lines up with this. Studies of humble leadership, leaders who acknowledge their limits, credit others, and remain genuinely curious about perspectives not their own, associate that style with higher trust, stronger engagement, and more resilient teams.⁷ The most effective people in charge, it turns out, are rarely the loudest in the room. They remain powerful precisely because they refuse to stop seeing the people around them, which keeps the very faculty that power tries to switch off in working order.

Coda

The lesson hiding inside all of this is not that power corrupts, a maxim old enough to have lost its sting. It is something more specific and more useful. Power is not a transformation of the self. It is a tool that acts on a particular piece of neural machinery, the same machinery that lets one person feel the reality of another. Left alone, it dims that machinery. Attended to, it does not have to. The next time someone gains authority and seems to become a stranger, the change has a name and a mechanism, not a moral failing. And the next time the authority is yours, the test is small and immediate. There is a plate on the table, and there are four cookies, and three of you. Watch what your hand does.

Sources

1. Keltner, D., The Power Paradox: How We Gain and Lose Influence, Penguin Press, 2016. — https://www.penguinrandomhouse.com/books/313229/the-power-paradox-by-dacher-keltner/
2. Galinsky, A. D. et al., ‘Power and Perspectives Not Taken,’ Psychological Science, 2006. — https://journals.sagepub.com/doi/10.1111/j.1467-9280.2006.01824.x
3. Rizzolatti, G. & Craighero, L., ‘The Mirror-Neuron System,’ Annual Review of Neuroscience, 2004. — https://www.annualreviews.org/doi/10.1146/annurev.neuro.27.070203.144230
4. Hogeveen, J., Inzlicht, M. & Obhi, S. S., ‘Power Changes How the Brain Responds to Others,’ Journal of Experimental Psychology: General, 2014. — https://psycnet.apa.org/record/2013-26430-001
5. Useem, J., ‘Power Causes Brain Damage,’ The Atlantic, 2017. — https://www.theatlantic.com/magazine/archive/2017/07/power-causes-brain-damage/528711/
6. Carney, D. R., Cuddy, A. J. C. & Yap, A. J., ‘Power Posing,’ Psychological Science, 2010. — https://journals.sagepub.com/doi/10.1177/0956797610383437
7. Owens, B. P. & Hekman, D. R., ‘Modeling How to Grow: Humble Leadership,’ Academy of Management Journal, 2012. — https://journals.aom.org/doi/10.5465/amj.2010.0441

Related reading

• The Cage Without a Lock
• The Burden of Never Forgetting

More from the Mind edition →