Third day with the lift broken. Twenty-odd people in the building's group chat are complaining, and complaining in unison. Nobody has phoned the management company.
You have probably seen the variants. Everyone privately says a process is broken, and when the meeting comes round nobody raises it. Everyone knows nobody reads that weekly report, and it still goes out.
The old explanation was that they were not angry enough — let the grievance build and people will move. The main finding of this field over the past sixty years runs the other way: grievance is almost never in short supply. What is missing is three other things, and any one of them can stall the whole thing on its own.
In 1965 Mancur Olson wrote a thin book called The Logic of Collective Action. Once the lift is fixed, the person who made the call and the twenty who did not are riding the same lift. A thing you cannot stop other people from using is a public good. Since you cannot stop them, the smart move for each person is to wait for someone else to call — free riding. The awkward part is that nobody here is being selfish: each person's arithmetic is correct, and the sum of it is that nobody moves. "We share an interest" does not get you to "we will act together". There is a gate in between.
Two things follow for Olson. First, small groups act more easily than large ones: three flatmates each get a third of the benefit, and if one of them does nothing the other two know at once; in a block of twenty flats the benefit is a twentieth each and nobody knows the neighbours' names. Second, a large organization gets moving on the back of things handed out only to participants — he called these selective incentives: a union's hardship fund and legal service, a trade association's market data, a congregation's help with weddings and funerals. His sharpest line is that many members of large organizations are in it for those, and the public cause is thrown in.
Be clear about how strong the evidence is: this is a theoretical argument plus historical examples, not an experimental result. The laboratory has an analogue, the public goods game — several people each get a sum of money, and whatever they put into a common pot is multiplied and split evenly. In the first round people typically put in about half; over more rounds contributions slide steadily down. Allow people to spend their own money punishing low contributors and cooperation holds up, a result that has been replicated many times. But in 2008 Benedikt Herrmann, Christian Thöni and Simon Gächter compared sixteen comparable participant pools around the world in Science and found places where people punish the high contributors instead, strongly enough to wipe out the benefit of punishment altogether. So "let people police each other" is not a switch you can throw anywhere.
In 1977 John McCarthy and Mayer Zald published a paper, since cited to death, that turned the question over in its opening lines: grievance is the constant, movements are the variable. Every era and every group carries a stock of complaints, and explaining the rise and fall of movements with something that does not vary will not get you far. The question is not why they are angry. It is why they managed to organize this time.
The Montgomery bus boycott makes the point, because the version everyone knows removes exactly this layer: one person was tired and would not give up a seat. Aldon Morris's The Origins of the Civil Rights Movement (1984) gives the other side. Rosa Parks was arrested on 1 December 1955; that night Jo Ann Robinson wrote a leaflet calling for a boycott at Alabama State College and, with three other people, ran the mimeograph machine all night — the usual figure is thirty-five thousand copies, and some sources give more — and the next day they went out across the city. Robinson belonged to the Women's Political Council, an organization of Black professional women that had been working on treatment on the buses for years and had lists, contacts and a printing machine.
The three hundred-odd days that followed also ran on things already in place: churches were the venues, church rolls were the address book, Black taxi drivers and private cars were assembled into a lift network. Doug McAdam's 1982 book takes the scale out to decades: Black churches, Black colleges and local branches of the National Association for the Advancement of Colored People had been growing for a long time before the protest broke out, something he called indigenous organizational strength. The view has a cost: written carelessly, people turn into containers for resources. Later work brought grievance back, but in a different position — grievance shapes what a movement looks like, organization decides whether it gets off the ground. On evidence: this strand is mainly historical comparison, and its weak point is picking cases after the fact.
Charles Tilly and his assistants did a laborious thing: they coded, one by one, the "contentious gatherings" reported in British periodicals between 1758 and 1834, more than eight thousand of them. Going through the cards you can watch a gear change.
In the mid-eighteenth century the ready-made ways of expressing a grievance were these: stopping a grain wagon and selling its load at what the crowd held to be a fair price; pulling down enclosure fences; carrying someone who had offended the neighbourhood through the streets to be shamed; breaking one particular household's windows. All local, aimed at a specific person, and often needing someone of standing to pass the claim upward. By the 1830s a different set was in use: public meetings, signature petitions, marches in column, founding a dedicated association, starting a periodical — national, aimed at Parliament, speaking for itself.
Tilly called this limited set the repertoire of contention. The word repertoire is deliberate: a company only knows a few plays, and people do not protest in whatever way they please. Why only a few? Because the form has to be legible to both sides. Participants need to know where to stand, what to shout and when to disperse; the other side needs to know whether this is a bid to talk or a fight. Inventing a new form is expensive, and the expense is that nobody knows how to play along — but once a new form has been performed successfully a single time it spreads astonishingly fast. Sit-ins, occupations and human chains all went that way.
Tilly also left a checklist: a movement has to display four things outward — worthiness, unity, numbers, commitment. Why a march forms up neatly, why everyone fights over the crowd count, why people being beaten at a sit-in do not hit back, are each one of those four. On evidence: this is description resting on coded historical material, not causal identification.
The word frame is borrowed from Goffman — the man from the piece on dramaturgy — and means the account people use to see a pile of loose events as "what is going on here". The same event told as "the system filtered these people out" and told as "he was unlucky" mobilizes entirely different things. David Snow and colleagues wrote a 1986 paper on frame alignment, later tidied into three jobs that have to be done together: diagnosis (what the problem is, whose account it goes on), solution (what should be done) and the call (why you, why now). Leave one out and it jams. Diagnosis alone leaves people angry with nowhere to push; diagnosis and solution without a call takes you straight back to the gate in "'Everybody wants it' is exactly why nobody does it".
William Gamson ran a set of focus groups in 1992, letting ordinary people talk about nuclear power, affirmative action and issues of that kind. His observation: saying "the system has a problem" is not enough to mobilize, because you need an agent that can be named — an injustice frame needs a "who". But a frame has to be carried by people. Snow and colleagues studied recruitment to a Buddhist group in the United States in 1980 and found that approaching strangers in the street worked miserably; the overwhelming majority of new members came in along ties of family, friends and acquaintances that were already there.
The finest material comes from Mississippi Freedom Summer in 1964. More than a thousand students applied to do voter registration, and McAdam later found 1,068 application forms: 720 went, 239 withdrew at the last moment. Both groups had filled in the same form, and on attitudes you can barely tell them apart — the ones who withdrew did not care less. The difference is elsewhere: whether they had strong ties to someone already taking part, and whether they had prior experience inside an organization. He had also expected that people without full-time work or family commitments would be more likely to go (this is called biographical availability); it did not come out cleanly in this pool, and his own explanation is that the applicants had already been filtered once. This is a rare piece of material: the records were made in advance.