Power: Coercion, agenda-setting, and shaping what people want
An accessible guide to power: the four senses of the word, what Hobbes, Weber and Arendt meant by it, why Dahl's New Haven study and the two faces of power that answered it disagree, and how Lukes and Foucault moved the argument.
In 1957 the political scientist Robert Dahl asked his readers to picture a street corner. A man stands there and tells himself that every driver on the street will keep to the right. The drivers carry on as before. Dahl then puts a policeman on the same corner, waving traffic right or left, and the cars do as the officer signals. 1
A has power over B to the extent that he can get B to do something that B would not otherwise do. 1
Five years later two political scientists at Bryn Mawr replied that this picture, and the research method built on it, could see only one kind of power. 2 Dahl's test required an observed decision, and much of power consists in keeping decisions from being taken at all. That exchange began an argument about the word that has still not closed.
Four things the word is doing
Power is the ability to produce the outcomes you want, but political theory divides that ability into four separate claims, and most disputes about power are really disputes about which claim is on the table. 3
| Sense | What it names | Where it comes from |
|---|---|---|
| Power as capacity — "power to" | The ability to act, alone or with others | Thomas Hobbes's Leviathan; Arendt on acting in concert 45 |
| Power as domination — "power over" | Getting others to do what they would not otherwise do | Weber's probability of imposing one's will; Dahl's definition 13 |
| Power as agenda control | Keeping some questions from being decided at all | Bachrach and Baratz on confining decisions to safe issues; Lukes on shaping the wants themselves 26 |
| Power as a productive system | Relations that shape what people want and who they can become | Foucault: power that disciplines bodies, manages populations and works through expert knowledge 7 |
Two fault lines run through that table. The first separates power-to from power-over: whether the word names a capacity that can be shared and built, or an imposition in which one party's gain is another's loss. The second runs inside power-over and asks how many faces it has. The first face counts the decisions that were taken; the second adds the issues kept off the agenda; the third adds the wants themselves.
Where the word came from
Hobbes defined power as capacity. "The POWER of a Man, (to take it Universally,) is his present means, to obtain some future apparent Good," he wrote, adding that power accumulates "like to Fame, increasing as it proceeds". 4 His greatest human power is the collective one: "that which is compounded of the Powers of most men, united by consent". 4
Max Weber put the imposition back in. Power, for Weber, is "the probability that one actor within a social relationship will be in a position to carry out his own will despite resistance". 3 Where an imposition is accepted as legitimate it becomes domination, and the state's own instrument is force: "a state is a human community that (successfully) claims the monopoly of the legitimate use of physical force within a given territory". 8
Hannah Arendt rejected the whole frame. Power, she wrote, "corresponds to the human ability not just to act but to act in concert", and "is never the property of an individual". 5 She drew the line sharply: "Power and violence are opposites; where the one rules absolutely, the other is absent." 5 Jürgen Habermas objected that this definition screens strategic power out of the analysis altogether. 3
The policeman, the mayor, and the agenda
Dahl took his definition into the field. In Who Governs? (1961) he examined New Haven, Connecticut, decision by decision, and concluded that no single elite prevailed consistently; different groups won on different issues. 9 Sociologically trained researchers kept finding the opposite, and the split followed method as much as it followed the cities studied. 2 Floyd Hunter's study of Atlanta in 1953 had asked well-placed informants to name the powerful, and produced a business-dominated "power structure". 10
Bachrach and Baratz attacked the pluralist method at its foundation. Power, they argued, works in two ways, and only one of them is a decision:
... power may be, and often is, exercised by confining the scope of decision-making to relatively "safe" issues. ... power is also exercised when A devotes his energies to creating or reinforcing social and political values and institutional practices that limit the scope of the political process to public consideration of only those issues which are comparatively innocuous to A. 2
Their name for this was the "mobilization of bias", and their charge against Dahl was blunt: a researcher's "approach to and assumptions about power predetermine their findings and conclusions". 2 A study that counts only the decisions actually taken will find power dispersed, because the exercises of power it cannot count leave no decision behind.
The third face
Steven Lukes argued that the one-, two- and three-dimensional views are "alternative interpretations and applications of one and the same underlying concept of power, according to which A exercises power over B when A affects B in a manner contrary to B's interests". 6 The first two faces left one thing out, he said: the wants themselves can be shaped.
Is it not the supreme and most insidious exercise of power to prevent people, to whatever degree, from having grievances by shaping their perceptions, cognitions and preferences in such a way that they accept their role in the existing order of things... 6
Lukes quoted Dahl's own New Haven text against it, noting that Dahl had written of leaders that they "do not merely respond to the preferences of constituents; leaders also shape preferences". 6 The third face is also the hardest to test, because it requires an observer to say what B's interests really were, and that judgement is evaluative. Lukes accepted the burden: identifying those interests, he wrote, "rests on empirically supportable and refutable hypotheses". 6 He then turned the concept on itself, warning that "how we think about power may serve to reproduce and reinforce power structures and relations, or alternatively it may challenge and subvert them". 3
Power without a centre
Michel Foucault changed the question instead of extending it. Power, on his account, exists in its exercise, and it is dispersed through practices, institutions and knowledge: "Power is everywhere; not because it embraces everything, but because it comes from everywhere." 11
His model case is the panopticon, Jeremy Bentham's prison in which one unseen watchtower is enough. The inmate, Foucault wrote, "assumes responsibility for the constraints of power ... he becomes the principle of his own subjection". 12 The same logic runs through schools, hospitals and factories, producing what he called docile bodies. 7 This consent-based power is what Antonio Gramsci had called cultural hegemony: a ruling group holds hegemony when its outlook becomes the accepted one, securing "the consent of subordinate groups to the existing social order" and no longer needing to rely only on coercion. 13 Foucault's later axis was biopower, a power that "endeavours to administer, optimize, and multiply" life rather than to take it. 7 Foucault's account still leaves power in someone's hands: he held that "certain persons exercise power over others". 3
The disagreement, on one page
| Position | Its claim | The question it answers |
|---|---|---|
| Max Weber (1919–1922) | Power is the chance of imposing one's will even against resistance; accepted imposition is domination, and the state's instrument is force. 38 | How power differs from authority, and what makes an order hold. |
| Robert Dahl (1957, 1961) | Power is visible where A gets B to do what B would not otherwise do; New Haven showed no single elite doing so consistently. 19 | Who wins the decisions that get taken. |
| Bachrach and Baratz (1962) | Power also works by confining decisions to safe issues; counting open conflict alone misses it. 2 | Which questions never reach a decision. |
| Steven Lukes (1974) | Power can shape the grievances themselves; all three views share one concept, power over another against that other's interests. 6 | Why no conflict is felt at all. |
| Hannah Arendt (1970) | Power is acting in concert, not imposing will; violence appears where power is failing. 5 | What makes collective action possible. |
| Michel Foucault (1975–1976) | Power is dispersed and productive, exercised through practices and knowledge rather than held by a sovereign. 712 | How power works when nobody is giving orders. |
The six rows answer different questions, which is why more than one of them can be defensible at once. The rows are genuinely in dialogue: Bachrach and Baratz name Dahl and quote his page numbers, and Lukes builds on both before turning Dahl's own sentences against him. 26 The sharpest disagreement is between the last two rows: Arendt's power-to and Foucault's dispersed power both break with Weber's definition, and Arendt and Foucault then disagree with each other about whether power needs a centre at all.
Power to, and power with
Arendt's capacity reading has a second life in feminist theory, where the standard equation of power with domination has been treated as a political choice rather than a neutral description. 3 Mary Parker Follett distinguished power-over from power-with, a collective capacity produced by reciprocity within a group. 3 Nancy Hartsock called power understood "as energy and competence rather than dominance" the feminist theory of power, and traced its ancestry to Arendt; Amy Allen's proposal is to reverse the priority and treat power-over as one species of the more basic capacity to act. 3
The practical payoff appears in development research. Serene Khader's account of "inappropriately adaptive preferences" gives the third face a measurable form: preferences formed under conditions that are hostile to flourishing, and selectively harmful to the people who hold them. 3
What the fight is now
The pluralist-elitist dispute has been rerun at national scale. Martin Gilens and Benjamin Page analysed 1,779 policy issues in the United States between 1981 and 2002 and found that economic elites and organised business groups had substantial independent influence on government policy, while the preferences of average citizens showed a "near-zero", statistically non-significant impact once those other actors were controlled for. 14 Their method inherits Dahl's limitation, since it can only measure issues that reached a decision; read against Bachrach and Baratz, the finding is a floor rather than a ceiling. 2
Wendy Li's study of regulatory capture carries the third face into the agencies themselves. Capture, on her account, can operate by shaping what regulators understand the public interest to be, long before any rule is drafted. 15
The third face now also has an industrial-scale instrument. The large platforms rank what users see by predicted engagement rather than by subscription alone, and they decide what is recommended; TikTok's emphasis on exploration is documented as a mechanism that can lead users to acquire new interests. 16 Agenda and preference are being set continuously, without a decision that anyone can be shown making.
A reading frame
Four questions do most of the work on the next power claim that comes your way.
- Which face is the claim using? Ask whether it points to a decision someone won, an issue that never reached a vote, or a want that was shaped. 6
- What would evidence look like? Each face sets its own test, and the third is both the most powerful and the hardest to evidence, so a claim resting on it should be held to a higher standard than a claim about an observed decision. 3
- Is this power-to or power-over? Domination and capacity point to different remedies: removing a veto and building a capacity are separate kinds of work. 3
- What could count against it? A claim should be able to name the evidence that would refute it. Lukes's own safeguard was that the interests at stake must be identifiable by the person affected, "independently of A's power". 6
Dahl's policeman can be photographed. The argument has lasted because the other faces cannot be, and because each one was introduced by someone who had noticed a real exercise of power going unrecorded.
References
- 1Robert A. Dahl, "The Concept of Power" (1957)
fbaum.unc.edu
- 2
- 3
- 4Thomas Hobbes, *Leviathan* (1651)
gutenberg.org
- 5Hannah Arendt, *On Violence* (1970)
blackwellpublishing.com
- 6Steven Lukes, *Power: A Radical View* (1974)
voidnetwork.gr
- 7Stanford Encyclopedia of Philosophy, "Michel Foucault" (2026)
plato.stanford.edu
- 8Max Weber, "Politics as a Vocation" (1919)
balliol.ox.ac.uk
- 9Wikipedia, "Who Governs?"
en.wikipedia.org
- 10Wikipedia, "Elite theory"
en.wikipedia.org
- 11Wikipedia, "The History of Sexuality"
en.wikipedia.org
- 12
- 13Wikipedia, "Cultural hegemony"
en.wikipedia.org
- 14
- 15Wendy Y. Li, "Regulatory capture's third face of power" (2023)
academic.oup.com
- 16
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.