The Three Faces of Power: How You're Controlled Without Knowing It

Between 1996 and 2017, American doctors wrote roughly 245 million prescriptions for opioid painkillers in a peak year. That is close to one bottle for every adult in the country. The pills were prescribed for back pain, dental surgery, sports injuries, the ordinary aches of ordinary lives, and for a long stretch almost nobody in a position of authority treated this as a national emergency. Doctors believed they were treating pain responsibly. Patients believed they were taking medicine their physician had judged safe. Regulators saw nothing requiring a dramatic intervention. By the time the crisis was named, hundreds of thousands of Americans had died.

What makes that story so unsettling is not that a few bad actors broke the rules. It is that a vast outcome unfolded with the apparent consent of nearly everyone involved, and for years it produced almost no political fight at all. If you want to understand how that is possible, the sharpest tool available comes from a slim book the British political theorist Steven Lukes published in 1974, Power: A Radical View. Lukes argued that power is not one thing but three, and that the kind we are best at seeing is the kind that matters least.

The Quiet Argument That Reframed a Whole Debate

To grasp what Lukes was doing, you have to know what he was arguing against. In the middle of the twentieth century, the most influential American account of power came from the political scientist Robert Dahl, whose 1961 study Who Governs? examined who actually ran the city of New Haven, Connecticut. Dahl was suspicious of grand claims that a hidden elite secretly controlled everything, because such claims were hard to test. So he proposed a rigorous, observable definition. Power, on this view, is the ability to win specific, contested decisions. If person A can get person B to do something B would not otherwise have done, A has power over B, and you can demonstrate it by watching who prevails when interests openly clash.

This was an attractive standard precisely because it was measurable. You could identify a decision, identify who wanted what, watch the contest, and record the winner. Lukes did not say this was wrong. He said it was incomplete. It captured, in his phrase, only one face of power, and a debate that stopped there would miss the more important ways that power shapes our lives. His book worked by absorbing the strongest version of his opponents' view and then showing that two further faces lay beyond its reach.

The First Face: Winning the Open Contest

The first face of power is the one Dahl described, and it is genuinely real. This is decision-making power, the observable capacity to prevail in a specific dispute where the sides are known and the conflict is out in the open. The method for studying it is straightforward in principle. You pick a concrete decision, such as whether a city builds a highway through a neighborhood or whether a legislature passes a bill. You determine what each party preferred. Then you look at whose preference won.

Most of what we recognize as politics lives at this level. A lobbying campaign that defeats a regulation, a vote that goes one way rather than another, a negotiation in which one side extracts more than the other, all of these are first-face exercises. In the opioid story, the first face is visible in the steady stream of victories that drug manufacturers and their allies won in the open arena. They funded research, cultivated relationships with regulators and medical bodies, and shaped the rules that governed how their products could be marketed and prescribed. When a restriction was proposed and beaten back, or a favorable guideline adopted, that was power in its plainest, most countable form. The first face explains how the producers won the fights that actually took place. What it cannot explain is why so few fights took place at all.

The Second Face: Controlling Which Fights Even Happen

In 1962, the political scientists Peter Bachrach and Morton Baratz published an essay called "Two Faces of Power" that put a crack in Dahl's framework. They pointed out that the most effective use of power often leaves no contest to observe, because the powerful arrange things so that threatening issues never reach the table. This is agenda-setting power, the capacity to keep certain questions off the political agenda entirely, so that they are never debated, never voted on, and never resolved against you.

The crucial insight is that an absence can be an exercise of power. If a question that obviously deserves attention is reliably ignored, that silence is not neutral. It is a result, and somebody benefits from it. Bachrach and Baratz called the quiet management of the agenda the second face precisely because it operates by what does not happen rather than what does. Detecting it is harder than studying open contests, because you cannot simply watch a decision and record the winner. Instead you have to compare what was actually debated against what could and arguably should have been debated, and then ask why the gap exists.

The opioid case is almost a textbook illustration. Through much of the period from 1996 to roughly 2010, the question of a prescription-opioid epidemic was systematically absent from serious federal political discussion. The data on rising overdose deaths existed and accumulated, yet the matter rarely surfaced as a genuine policy question demanding a decision. That absence of decision points was not an accident of inattention. It was itself the operation of power, sustained by the framing of opioids as the responsible treatment of undertreated pain, by the institutional credibility lent to that framing, and by the absence of any organized force pressing the issue onto the agenda. Nothing had to be voted down because nothing reached a vote.

The Third Face: Shaping What People Want in the First Place

Lukes's own distinctive contribution was a third face, and it is the one that gives his book its bite. Beyond winning open contests, and beyond controlling the agenda, power operates most fundamentally by shaping what people come to want. If you can influence someone's very desires, beliefs, and sense of what is normal, then you never have to defeat them in a contest, because they never form the preference that a contest would require. This is preference-shaping power, the capacity to secure compliance by molding perceptions and wants so that conflict does not arise in the first place.

The first two faces both assume a conflict of interest that is at least latent. Someone wants something, someone else blocks it, and the question is who prevails or whether the blocked party even gets to raise the issue. The third face goes deeper. It asks how an arrangement comes to seem so natural, so obviously correct, that the people disadvantaged by it actively endorse it. In the opioid story, this is the cultivation, over years, of a genuine and widespread belief, among physicians and patients alike, that these drugs were a safe and appropriate answer to everyday pain and that worrying about addiction was itself a kind of cruelty to suffering people. When patients asked for the pills and doctors felt they were practicing good medicine by prescribing them, no one needed to be coerced. The relevant preferences had already been shaped. That is the third face at work, and it is why the crisis could grow so large with so little resistance.

Why the Deepest Face Is the Hardest to Prove

There is an honest difficulty buried in the third face, and Lukes did not hide it. To claim that someone's preferences have been shaped against their own interests, you have to be able to say what their real interests are, on grounds independent of what they currently say they want. Otherwise the argument collapses into a tautology, in which any preference you disapprove of gets relabeled as manipulation. Detecting third-face power therefore requires comparing a person's observed preferences against an account of their real interests that the analyst can defend on independent grounds, and that requirement is both methodologically demanding and politically contested.

The objection is serious. Who is the analyst to say that people do not really want what they say they want? The worry is that the third face hands intellectuals a license to override ordinary people's stated wishes in the name of interests those people supposedly have but cannot see. Lukes accepted that this made the third face the most difficult and the most arguable part of his framework. He did not think the difficulty was a reason to abandon it, because some of the most consequential workings of power genuinely operate at this level, but he was clear that claims about it carry a heavier burden of proof than claims about the first two faces. This is not a flaw to be hidden. It is a real intellectual cost that any careful use of the idea has to pay.

Matching the Face to the Question

A key point that is easy to miss is that the three faces are not competitors, with one correct answer to be chosen. They are different tools suited to different questions, and each requires its own kind of evidence. If you are studying many ordinary legislative battles, the first face and its method of tracing contested decisions will usually serve you well. If the puzzle is why some obvious issue is never addressed, then the absence of a decision is your clue, and the second face directs you to study the agenda and the forces that police it. If the puzzle is why an arrangement that harms people is nonetheless accepted by those it harms, then the third face and its hard comparison of preferences against interests is the relevant lens.

This is also why the framework is integrative rather than a replacement. Lukes was not arguing that the first face is an illusion to be discarded. The first face is real and important, and most of everyday politics is conducted there. His point was that the second and third faces add to it, capturing dimensions of power that a decision-focused method cannot see, rather than overturning the value of studying open contests. Applied together to the opioid case, the three faces produce an analysis no single face could match. The first captures the lobbying and regulatory victories, the second captures the long suppression of the issue from the national agenda, and the third captures the cultivation of the beliefs that made the whole structure feel like good medicine. Other thinkers have pushed third-face analysis further still, with Pierre Bourdieu writing on symbolic power and Michel Foucault on the productive power of discourse, but those extensions belong to more advanced study and need not be settled here.

Key Takeaways

Steven Lukes argued in 1974 that power has three faces rather than one. The first face, drawn from Robert Dahl's study of decision-making, is the observable capacity to win specific open contests, and it is measured by identifying decisions, preferences, and winners. The second face, named by Bachrach and Baratz in 1962, is agenda-setting, the power to keep threatening issues off the table so that no contest ever occurs, detected by comparing what was debated against what could have been. The third face, Lukes's own contribution, is preference-shaping, the deepest and most contested form, in which power molds what people want so that they never form an opposing preference, and which can only be claimed by defending an independent account of people's real interests. Each face demands its own evidence and answers a different question, and the framework is integrative rather than a replacement of the first by the others. The American prescription-opioid crisis illustrates all three at once: open lobbying and regulatory wins, the years-long suppression of the epidemic from the federal agenda, and the cultivated belief that the pills were simply responsible care, which together let an enormous harm unfold with the apparent consent of nearly everyone it touched.