When the Heat Is On

Coalitions fracture under pressure. Our job is to build the infrastructure underneath them.

Organizing conversations across the pro-democracy field continually return to the same anxious question: how do we hold the coalition together when defecting starts to look like safety?

The honest answer is that we mostly will not. Not by trying harder to hold the coalition. Not with stronger statements of unity. Not with better-worded memoranda of understanding. Coalitions fracture under sustained authoritarian pressure for reasons that are structural, not moral, and the response the moment requires is not a better coalition agreement but a different kind of infrastructure - built before the crisis, underneath the coalition, so that the work survives even when the coalition does not.

This is not a counsel of despair. It is a more honest, and more strategically useful, place to start.

Why coalitions break

Traditional coalitions aggregate organizations around shared interest and shared position. Under low pressure, shared interest holds. Under high pressure - when the question is no longer whether you agree with a position statement but whether you will stand publicly beside a community being targeted - shared interest is frequently not enough.

This isn’t a failure of character. It’s a structural feature that shapes how we respond this situations of high pressure. Authoritarian movements understand it and act on it. They identify the most institutionally exposed organizations in a coalition and apply pressure directly to them, converting what looked like a unified coalition into a series of private, bilateral negotiations the coalition itself never gets to have. Each organization makes its calculation in isolation, and the coalition dissolves one quiet withdrawal at a time.

The antidote is not stronger Memoranda Of Understanding. Documentation doesn’t possess the coercive power of defunding, litigation, or violence. People of good faith defect when the cost of staying becomes higher than the cost of leaving. The strategic question is not whether they’ll defect under sufficient pressure. It’s whether the underlying infrastructure has been built deeply enough that their defection doesn’t destroy the work.

What follows is a sketch of that infrastructure - four elements, an operating logic, and a set of historical examples that suggest it is not just an idea. It’s what has actually worked when other things have failed.

Springs as relational infrastructure

The springs-to-river framework names the kind of infrastructure the field actually needs. The river is the broad, coordinated movement of mass mobilizations, noncooperation campaigns, and cross-sector coalitions at scale. The springs are the community-rooted, trust-based organizing units that feed the river.

A coalition brings organizations together. A spring brings people together. That distinction matters enormously when the heat is on.

Springs are built on authentic relationships - people who have done real work together, shared genuine risk, navigated conflict and survived it. The cost of defection from a spring isn’t organizational. It’s personal. You aren’t withdrawing from an affiliation; you’re betraying people you know, people whose children you know, people who would show up for you. That’s a different calculation.

The security a spring offers is moral and psychological, not operational. It is the context in which people can hold their clarity and their commitment when the external world is doing everything it can to make capitulation seem reasonable. This is built deliberately, over time. It’s not an attribute people either have or lack. It’s infrastructure people build.

Distributed architecture

Centralized networks fail systemically. Distributed networks fail locally.

Hub-and-spoke organization is efficient under ordinary conditions and, precisely because of that efficiency, vulnerable to targeted disruption. Decapitate the leadership, defund the central hub, generate internal conflict at the organizing center, and the whole structure can seize.

A network of relatively autonomous nodes, connected by shared frameworks and relationships rather than by hierarchical authority, doesn’t fail this way. When nodes are disrupted - when organizations lose funding, leaders are targeted, local coalitions fracture - the damage is local rather than systemic. Other nodes continue to function.

The design principle that makes distributed networks work is this: springs need enough connection to each other to reconstitute, but not so much connection that the compromise of one spreads to all. The river provides the coordination. The springs maintain their structural independence. The flow between them is not the same as the vulnerability of being merged.

This means resisting one of the most familiar temptations of crisis: the urge to centralize for efficiency. Centralization feels like decisiveness. It is also what makes the whole network susceptible to a single targeted disruption. Coordination is communication, shared frameworks, and aligned timing. Centralization is hierarchical authority and operational dependency. You want the first and not the second.

Solidarity as cost imposition

Solidarity is usually described as a sentiment. The strategic framework treats it as a structural practice that does specific work.

Solidarity’s primary function is not as sentiment but as cost imposition. When communities and organizations act publicly in defense of a targeted group - when they show up physically, refuse institutional cooperation with enforcement, make the targeting of one community visible as an attack on all - they raise the cost of authoritarian advance. They convert a bilateral assault (regime against targeted communities) into a multilateral one (regime against an expanding coalition). They create the social friction that slows the normalization of cruelty.

There is a deeper logic here. Erica Chenoweth’s civil resistance research is clear that potential defectors in the opposing coalition decide their long-term interests lie with the resistance, rather than with the regime, when they see concrete evidence that the movement is large, determined, and willing to defend its own. Solidarity is what makes a movement credible to those defectors - diverse, real, and reaching across social networks to people who are not yet sure which way to move.

This points to a critical distinction. Performative solidarity - the statement, the post, the letter of support - can serve a signaling function but cannot build the cross-community trust that makes coalitions of necessity possible when crisis comes. Relational solidarity is built through actual presence, actual risk taken together, actual care expressed when it costs something. It’s built before the crisis, so that when the crisis comes, the people who show up already know each other.

Depth and breadth, held together

The framework contains a real tension. The qualities that make springs resilient - depth of trust, strength of relationship, and shared history - can also make them closed, self-referential, and inwardly focused. A spring that is too tight to let new people in is reproducing itself rather than building a democratic movement. A network of springs that are each internally coherent but have no real relationships with each other is not a river. It’s a collection of puddles.

The movement needs both depth and breadth. Springs provide depth. The river needs breadth. What bridges them is leadership that models both - people deeply grounded in their spring who are genuinely oriented toward the broader movement. People who maintain their distinct identity and analysis while building real relationships across difference. People who understand that the spring is infrastructure for something larger, not the movement itself.

Martin Luther King Jr. was a Christian minister, deeply rooted in that spring, who built and maintained working relationships with socialists, labor organizers, and secular activists who did not share his theology. He didn’t stop being a minister. He became a movement leader. Spring depth and river orientation, held together.

The operating logic

A handful of practical things follow.

Build before the crisis. The worst time to build deep-trust, high-commitment springs is when the pressure is already on. Build them now, through actual shared work on actual problems. Don’t wait for a crisis to discover who your people are. This is not just advice that will serve you in a fight against authoritarians. It’s advice that will serve you in every aspect of your life.

Treat the spring as infrastructure, not the movement. A spring that has stopped feeding the river is a social club. The test is orientation, not tightness.

Calibrate connection carefully. Too few ties between nodes leaves springs isolated and the river unformed. Too many ties create the vulnerability of shared compromise. The goal is flow without merger.

Distinguish performative from relational solidarity. The statement is not the practice. Both have uses, but only the practice builds the infrastructure.

Name the depth–breadth tension out loud. Frameworks that pretend their tensions don’t exist tend to fail when the tensions show up under pressure. Naming the tension is the first step toward managing it.

What history says

This is not theoretical. The pattern of distributed, relationally grounded networks surviving targeted pressure shows up repeatedly when you look for it.

After the catastrophic arrests of 1943, the French Resistance reorganized into smaller, more distributed cells. The reconstituted networks had better operational security than the original and contributed decisively to liberation.

Polish Solidarity under martial law built its own newspaper, its own schools, its own self-governing community institutions. Altogether, these constituted a parallel infrastructure that could absorb repeated assaults and reconstitute in new forms when conditions allowed.

The Civil Rights Movement survived the King assassination because it wasn’t actually dependent on a single leader. Decades of network-building across Black churches, HBCUs, local civic organizations, and labor networks kept the work alive.

Serbia in 2000: Chenoweth’s research identifies the moment when police officers declined to fire on demonstrators because they recognized their cousins in the crowd as a direct consequence of the kind of mass, cross-cutting participation solidarity produces. The cousins were there because the movement had been built to put them there.

The 2017 airport mobilizations put lawyers, faith communities, civil liberties organizations, and ordinary people in the same physical space taking risk together. The speed of mobilization reflected relationships that had been built before the moment required them.

Each of these cases shows the same logic. When the pressure came, what held was not the most visible coalition. What held was the infrastructure underneath it.

The way through

The way through is not to build better coalitions. It’s to build the infrastructure underneath them that makes us less dependent on formal coalitions - and therefore, paradoxically, makes those coalitions more likely to hold when they’re needed most.

A coalition that fractures under pressure doesn’t necessarily mean the movement has failed. It may mean the coalition was bearing weight it was never built to bear, and the actual work was always going to live somewhere else. If we’ve built the somewhere else, the fracture is survivable. The work continues. The springs feed the river. The river finds its course.

The next phase of this work is not louder unity. It is quieter, deeper, more patient building - in relationships, in distributed networks, in solidarity practiced rather than performed, in leadership that holds depth and breadth at once. It’s the unglamorous work of preparing for pressures we can’t fully predict, in ways that don’t depend on predicting them.

The heat will come. The question is what we have built underneath the surface to hold when it does.