Elites, Pluralism and Who Actually Governs

The last lesson (11.1.2) left us with four theories of whose interests the state serves — pluralist, elite, Marxist, autonomous — and the promise that people once tried to settle the argument by close observation. This lesson is the story of that attempt: a decades-long fight among researchers about who really runs things, fought out over a few ordinary American cities, that ended up teaching the discipline something far larger than the answer to its own question.

It is one of the most useful stories in all of sociology, because it is not really about cities. It is about how the method you choose to study power quietly determines the answer you will find — and how a research programme can be rigorous, honest, and completely blind at the same time. If you learn to see the trap they fell into, you will never again take a study of "who has power" at face value, and you will catch the same trap in a hundred other places.

Two researchers, two cities, two opposite answers.

In the 1950s a sociologist named Floyd Hunter went to a large Southern city — he called it "Regional City" — to find out who really governed it. His method was simple and clever. He asked knowledgeable people — civic leaders, journalists, insiders — a question: who are the most influential people in this town? Then he took the names that came up most, went to those people, and asked them the same question. The lists converged. A small circle of names kept appearing.

What Hunter found was an elite. A few dozen people — mostly business leaders, bankers, big employers — sat at the top. They knew each other, met informally, and the important decisions of the city flowed, one way or another, through them. Elected officials carried out policy; they did not originate it. Power in Regional City was concentrated, private, and stable. This became the textbook picture for the elite theory.

A few years later a political scientist named Robert Dahl went to New Haven — the small Connecticut city that was home to Yale — to find out the same thing. But he rejected Hunter's method entirely. Asking people who has power, Dahl objected, only tells you who has a reputation for power — and reputation is not the same as power. To find who actually governs, he said, you must look at actual decisions: take the important issues the city faced, find out who proposed what, who opposed whom, and who prevailed. Follow the conflicts and see who wins.

What Dahl found was the opposite of Hunter. In New Haven, power was dispersed. Different people prevailed on different issues — one set of players dominated schools, another urban redevelopment, another party nominations — and the sets barely overlapped. No single elite ran everything. The mayor was a broker, assembling different coalitions for different fights. Power was plural, fragmented, and issue-specific. This became the textbook picture for the pluralist theory, and Dahl's book — Who Governs? — was a landmark.

Same question. Two able researchers. Opposite answers. And here is the thing that should nag at you: it was not mainly that the cities differed. It was that the methods differed — and each method, it turned out, could only find the kind of power it was built to look for.

The puzzle: how can both be right and both be wrong?

Line the two studies up and the disagreement looks like a simple empirical dispute — maybe Regional City really was run by an elite and New Haven really was plural, and both researchers were correctly describing different places. That is possible, and partly true.

But something deeper was going on, and a pair of critics — Peter Bachrach and Morton Baratz — put their finger on it in an argument so important that we already met its conclusion in 11.1.1 (it is the origin of the "second face of power"). Their point was this:

Each method could only detect the face of power it was designed to detect — and each was blind to the other.

Hunter's reputational method (ask who has power) picks up the standing of the elite — but it cannot distinguish real power from mere reputation, and it cannot show power working on any particular issue. It risks seeing an elite that is not actually deciding much.
Dahl's decisional method (study who wins the fights) picks up power in visible action — but it can only study issues that became issues, decisions that got made, conflicts that actually happened. It is structurally blind to the power that keeps things from ever becoming issues at all.

And that second blindness is the fatal one. Recall the town by the poisoned river (11.1.1). Dahl's method, applied there, would examine the decisions the town council actually took — and find no conflict over pollution, because none reached the agenda. It would conclude, correctly by its own lights, that no one was exercising power over the pollution question. But the absence of conflict was itself the achievement of power — the second face, keeping the issue off the table. The method that studies decisions cannot see the power that operates through non-decisions.

So the puzzle resolves like this: Dahl was right about what he could see and wrong to think it was everything. Where visible conflict occurred, New Haven really was plural — different winners on different issues. But by studying only the decisions that got made, he mistook the absence of visible conflict for the absence of power — and so missed whatever was being kept off New Haven's agenda entirely. And Hunter, though his method was cruder, was groping toward a real thing his critics' method could not reach: standing power that need not fight because it rarely has to.

Going deeper: the three faces, mapped onto three methods

This is the payoff, and it is worth seeing cleanly, because it ties the whole debate to the framework of 11.1.1. Each face of power calls for a different method, and each method sees only its own face.

First face — power in visible decisions. Method: study who wins the open conflicts (Dahl's decisional approach). Detects: A prevailing over B when they clash on an issue that reached the agenda. Blind to: everything that never became a visible clash.

Second face — power over the agenda. Method: study non-decisions — what gets kept off the table, which grievances never reach a vote (Bachrach and Baratz). Detects: the "mobilisation of bias," the way some issues are organised into politics and others organised out. Blind to: power that has shaped what people want in the first place. Hard to study, because you are looking for the dog that did not bark.

Third face — power over preferences. Method: hardest of all — you must establish what people's interests would have been absent the shaping, and show they were led to want otherwise (Lukes). Detects: the deepest and most effective power — quiescence produced not by suppression but by consent that was itself manufactured. Blind spot of both earlier methods, and the hardest to prove, because it requires a claim about a counterfactual — what the powerless would have wanted in some freer condition — that can slide into telling people you know their interests better than they do (the false-consciousness trap, 11.1.1).

Now the community-power debate falls into place. It was not a fight about New Haven. It was a fight about which faces of power a study is even capable of seeing — and the field advanced not because someone finally counted correctly, but because the critics made visible a whole region of power that the leading method was constitutionally unable to detect.

The trap generalises far beyond cities — watch for it everywhere.

The lesson here is a portable tool, and it catches errors all over the place. Any time someone claims to have measured power, influence, bias, or fairness by looking at outcomes, ask: which face is this method built to see, and what is it therefore blind to?

A study finds that a hiring process is fair because, among the people who applied, selection was unbiased. But who was discouraged from applying at all? (The applicants are the visible decisions; the non-applicants are the non-decision.)
A platform says its users are free to see what they like because it gives them what they click. But the menu of what there was to click was set by someone. (First face: your choice among options. Second face: who set the options — 11.1.1's agenda.)
A market is called efficient because it clears at a price both sides accept. But were the preferences the buyers brought shaped by the seller's decades of advertising? (Third face.)

In each case the visible-conflict method returns a clean, honest, incomplete answer — and the incompleteness runs all in one direction, hiding the power that operates before the visible contest begins. Learn to ask "what would this method miss?" and you have learned one of the most useful habits in all of social analysis.

An honest accounting: what pluralism got right

It would be a mistake — the opposite mistake — to conclude that the pluralists were simply wrong and the elite theorists simply right. That is not where the evidence lands, and the symmetry we owe every debate (the standard from Part 7) requires saying so plainly.

On the first face, the pluralists were substantially correct, and this matters. In open societies, power over visible issues really is more dispersed than the elite theorists claimed. Different groups really do prevail on different questions; powerful interests really do lose, sometimes badly; there is no single committee quietly deciding everything. The elite theory, in its strong form, overstated the unity and coordination of those at the top — mistaking a shared social position for a shared plan (the class-in-itself / class-for-itself distinction from 8.1.3 again). Hunter's method, by asking about reputation, could not tell a genuinely ruling elite from a merely prominent one.

So the honest verdict is not "elite theory wins." It is: pluralism correctly described the first face and wrongly assumed it was the only face; elite theory correctly sensed a deeper, standing power but described it too crudely and too conspiratorially; and the second-face critics were right that the biggest thing was what none of the visible studies could see. The truth is not a midpoint between the camps. It is a layered picture in which each camp mapped a different layer — and the field's real progress was learning that the layers exist.

Who governs your street?

Bring it down to something you can check. Ask "who has power over what gets built, allowed, or banned on your street?" and watch the three methods give three answers.

Decisional (first face): look at the actual planning meetings. Who proposed the new development, who objected, who won the vote? You will find real contests with real winners — and often the residents win some and lose some. Pluralism, confirmed at this layer.
Non-decision (second face): now ask what never came up. Was there an option — a kind of housing, a use of the land — that was never on any agenda because the rules, the procedures, or "what everyone knows is realistic" ruled it out before anyone proposed it? That is where the deeper power sits, and no meeting minutes record it.
Preference-shaping (third face): harder still — do the residents even want what would most serve them, or have decades of assumption about "the sort of place this is" shaped their wants to fit the arrangement? Here you must tread carefully (the false-consciousness trap), but you must at least ask.

Notice that you cannot answer the second and third questions by attending more meetings. More of the first-face method will never reach the other faces. You have to change what you look for — and that change of gaze is the whole lesson.

Why this matters

Because "who really governs?" is a question you will meet constantly — about your workplace, your town, your country, the platforms you use — and you now have something better than a side to take. You have a diagnostic.

Most public argument about power is stuck on the first face: who won the vote, who got the job, who prevailed in the visible fight. That argument is not wrong, but it is shallow, and both sides of it usually share the same blindness — they are fighting over the decisions that got made and ignoring the ones that were prevented, the options that were never offered, the wants that were shaped upstream. When you can ask "what was kept off the agenda?" and "who set the menu of choices?" and "were these preferences formed freely?", you are no longer a spectator to the visible contest. You are studying the ring itself, and who built it.

And there is a humbling methodological lesson folded inside the political one, and it is the deepest thing here. Two honest, careful researchers reached opposite conclusions because their methods could only find the power they were built to find. This is not a failure of rigour; both were rigorous. It is a permanent feature of studying anything: your instrument shapes your finding. The reputational method finds reputations; the decisional method finds decisions; neither, alone, finds the whole. The advance came not from a better single method but from someone asking what the leading method could not see — which is, in the end, the reflexive discipline this whole course keeps returning to (Part 7): applied to the study of power itself, the first question is always "what is my way of looking making invisible?"

The next lesson closes this topic's loop by turning to what happens when the powerless stop accepting the arrangement — when the agenda that was kept closed is forced open. That is the story of social movements (topic 11.2), and it begins exactly where this lesson ends: with people making an issue of something that power had worked to keep off the table.

What to carry out of this lesson:

The community-power debate: Hunter's reputational method (ask who's influential) found a concentrated elite in "Regional City"; Dahl's decisional method (study who wins actual conflicts) found dispersed pluralism in New Haven. Same question, opposite answers.
The methods, not just the cities, produced the difference. Each method can detect only the face of power it was built for — and each is blind to the others.
Bachrach and Baratz's decisive critique: Dahl's decisional method is structurally blind to non-decisions — the power that keeps issues off the agenda entirely, leaving no conflict to study. This is the origin of the "second face of power" (11.1.1).
Three faces, three methods: study decisions for the first face, non-decisions for the second, preference-formation for the third. Each later face is harder to study and more effective in operation; the third risks the false-consciousness trap.
The honest verdict is layered, not a winner. Pluralism correctly described the first face and wrongly assumed it was all of power; elite theory sensed a deeper standing power but drew it too crudely. The real advance was realising the layers exist — and that your instrument shapes your finding.

Reputational method — studying power by asking knowledgeable people who is influential, then following the names that recur (Hunter). Detects standing and reputation; cannot separate real power from mere prominence, or show power acting on a specific issue.

Decisional method — studying power by examining actual decisions: who proposed what, who opposed, who prevailed (Dahl). Detects visible, first-face power; blind to non-decisions.

Non-decision — the prevention of an issue from ever reaching the point of decision, by keeping it off the agenda or making it unraisable. Leaves no visible conflict, and so is invisible to the decisional method.

Mobilisation of bias — Bachrach and Baratz's term (from Schattschneider) for the way a system's rules and procedures organise some issues into politics and others out, systematically and not by accident.

Community power debate — the mid-twentieth-century argument among researchers over whether local power is concentrated (elitist) or dispersed (pluralist), which ended by revealing that method determines finding.

Your instrument shapes your finding — the general methodological lesson: a research method detects only what it is built to detect, so the first reflexive question about any study of power is "what is this way of looking making invisible?"

Try it yourself:

One — Catch the missing method. Find a claim in the news that someone "has" or "lacks" power, or that a process is fair or unfair. Identify which face the claim is about (usually the first — who won). Then ask the two questions the claim's method cannot answer: what was kept off the agenda? and who set the options in the first place? Write down what you would need to look at — not more of the same evidence, but a different kind.

Two — Run Hunter and Dahl on one place you know. Pick an organisation you belong to (a workplace, a club, a family even). First do Hunter: ask yourself and others who's really influential here? and see whose names recur. Then do Dahl: recall the last few real decisions and ask who actually prevailed? Do the two lists match? Where they differ, you have found the gap between reputation and visible power — and possibly a clue to the second face.

Three — Find a non-decision. Think of something that is simply never discussed in a group you are part of — an option no one raises, a question that feels out of bounds, a "that's just not how we do things." That silence may be a non-decision: power operating by keeping something off the agenda. Ask, carefully, who benefits from its never coming up? — and notice that you could attend every meeting for a year and never see it.