Governance · Principle 5

Coordination Without a Center

A thousand parts, each following a simple local rule and responding to what's around them, can produce coherent, coordinated behavior that no single one of them planned or could have planned alone.

Try this

Before assuming a group needs someone 'in charge' to coordinate a task, try giving everyone the same simple, visible information and see if coordination emerges without you directing it.
In any system you help design, a team process, a community group, a project workflow, identify what local signal people are actually responding to, and ask honestly whether that signal is trustworthy or could mislead people who are each behaving reasonably.
Build in at least one deliberate pause point for any fast-moving, decentralized process you're responsible for, not to control it, but so a cascade has somewhere to stop and be reassessed before it runs away.

Explain it to a child

Think about how a big crowd of people naturally forms a path around a puddle on the sidewalk. Nobody stands there directing traffic, each person just looks at the people right in front of them and steps around, and a smooth path appears all on its own. That's how a flock of birds flies together too: no bird is the boss, each one just watches the birds closest to it. But sometimes this can go wrong, if one person suddenly starts running because they think something's chasing them, other people might start running too, just from watching each other, even if nothing is actually there. The same thing that makes a crowd move smoothly can also make it panic together by mistake. That's why it's healthy for a group to have at least one moment to stop and check what's really happening, instead of just following whoever moved first.

What it means

Many people assume coordination requires a coordinator, someone or something at the center collecting information and issuing instructions. But some of the most coherent, adaptive behavior we can observe, in nature and in human systems, emerges without any central planner at all, from many independent parts responding to simple local signals. This principle exists to test whether a system actually needs top-down direction to stay coordinated, or whether it's been assumed to need one out of habit.

Where this comes from

Open what you want. Nothing below is needed to use the principle; it is here because a claim without its working is an assertion.

In nature

Two clean, well-documented biological examples illustrate this from different angles.

Starling murmurations, the huge, swirling flocks that move as one shape-shifting cloud at dusk, have no leader at all. Each bird simply tracks and mirrors the movement of about seven of its nearest neighbors, and this purely local rule, repeated across thousands of birds simultaneously, produces the whole flock's complex, coordinated movement, including its ability to respond to a predator's attack in a fraction of a second. Researchers modeling this since the 1980s found that a flock following three simple local rules, keep some distance from nearby birds, match their direction and speed, and move toward more distant ones, reproduces the same lifelike collective motion, with no bird ever receiving or issuing an instruction to the group as a whole.

Ant colonies achieve something similar through what biologists call stigmergy, coordination through the shared environment rather than through direct communication between individuals. An ant doesn't tell other ants where the food is; it deposits a pheromone trail as it returns, and other ants independently respond to that trail, reinforcing the most successful routes while weaker trails simply evaporate and disappear. No ant has a map of the whole colony's foraging effort, and no ant is in charge of directing it, the coordination doesn't live in any individual, it lives in the shared trace they leave behind for each other. The same principle now names a formal concept, stigmergy, with applications well beyond ants, including how Wikipedia's decentralized editing process converges over time without any single editor directing the whole encyclopedia.

What the science says

The clearest human-scale articulation of this principle in economics is F.A. Hayek's 1945 essay "The Use of Knowledge in Society." Hayek argued that the knowledge needed to coordinate a complex economy is never available to any single person or planning body, it exists only as fragments, dispersed across millions of individuals who each know their own immediate circumstances. His central claim was that market prices function like the starlings' seven neighbors or the ants' pheromone trail: a simple, local signal that lets each participant coordinate their own actions with everyone else's, without anyone needing the full picture.

Honesty flag, kept in the record

Hayek's argument is highly influential economics, but it's also a foundational text of a specific and contested political-economic tradition (classical liberalism, Austrian economics, anti-central-planning), used historically to argue against forms of collective and state coordination. This document presents the coordination mechanism itself, price as a distributed information signal, as a genuine, well-documented insight, while treating "therefore markets should replace planning generally" as the contested political conclusion some draw from it, not a settled fact. Real markets also coordinate imperfectly around things prices don't capture well, pollution, care work, long-term risks, which is worth holding alongside the insight itself.

History

Wikipedia offers a genuinely instructive case, because its own history contains both the success and the limits of this principle within one project. It began in 1999 as Nupedia, a centrally managed, expert-reviewed encyclopedia, and produced only 15 articles in its first year, strangled by its own formal review hierarchy. In 2001, its founders deliberately abandoned centralized editorial control and let anyone edit directly, with no formal review chain. The result was explosive: by late 2005 the English-language edition alone held roughly 750,000 articles, grown almost entirely from decentralized, uncoordinated volunteer effort with no central author ever directing which article should be written next.

But the honest part of this history matters just as much: Wikipedia's "leaderless" coordination didn't stay leaderless. Academic research studying its editing patterns found that decentralized activity naturally gave rise to emergent, informal hierarchies among editors, and a separate 2014 study concluded that Wikipedia's contributor community had self-organized into oligarchic structures over time, though later research in 2021 found real counter-tendencies pushing back against that concentration of editorial power. Participation itself peaked and then declined after 2006, partly because the same decentralized openness that fueled its early growth made it harder to manage conflict and toxicity among editors at scale. Coordination without a center produced Wikipedia's remarkable early growth, but it didn't produce permanent, self-sustaining equality of influence; power and structure re-emerged inside the decentralized system on their own, without anyone designing it in.

In practice

Technological/organizational scale

Open-source software development runs on the same logic as ant stigmergy in a human context: contributors don't receive central assignments, they respond to what's visibly needed, an open bug report, a stalled feature branch, a pull request awaiting review, and the shared, visible state of the project (its code repository, its issue tracker) functions the way a pheromone trail does, coordinating thousands of independent contributors who never need a project manager assigning their next task.

Everyday human scale

Traffic at a four-way stop, or pedestrians navigating a crowded sidewalk, coordinate largely without any central controller: each person follows simple local rules (yield to whoever arrived first, walk around whoever's in front of you) and coherent, mostly collision-free group movement emerges from thousands of these small local decisions, the same way starlings track only their seven nearest neighbors rather than the whole flock.

Economic scale

A local farmers' market or a neighborhood barter economy coordinates supply and demand the way Hayek describes prices doing at larger scale, no one person decides who should grow what or sell to whom; each participant responds to what they see and what price signals tell them, and the overall pattern of who's selling what and to whom emerges from many small, local, independent decisions rather than a central plan.

What argues against it

Coordination without a center isn't automatically good coordination, it can converge on failure just as readily as it converges on success, because the mechanism itself is neutral about which direction it amplifies.

The clearest modern example is the 2010 "Flash Crash." At 2:32pm on May 6, a single large automated sell order, executing without regard to price or time, entered the market. High-frequency trading algorithms, themselves independent, decentralized actors with no central coordinator, each simply responding to local price signals exactly the way this principle describes, passed the same contracts back and forth in what regulators called a "hot potato" effect, generating enormous trading volume without providing any real liquidity. As prices fell, more algorithms hit their own local risk limits and withdrew or began selling too, each acting on its own local information exactly as intended. In under five minutes, the market lost roughly a trillion dollars in value, before a temporary trading pause allowed the cascade to break and prices largely recovered. Every individual actor in that chain was doing exactly what decentralized coordination is supposed to do, responding rationally to local signals, and the aggregate result was a near-catastrophic failure, not an emergent success.

This sharpens the principle rather than undermining it: decentralized coordination reliably amplifies whatever local signal it's built around, for better or worse. Starlings' local rule (mirror your seven neighbors) produces predator evasion because the signal being amplified is genuinely useful information. The Flash Crash's local rule (respond instantly to price and inventory signals) produced a cascade because, under those specific conditions, the signal itself became misleading faster than any individual actor could recognize it. A decentralized system is only as trustworthy as the local signal every part of it is responding to, which means part of designing a decentralized system well is making sure the local signal itself stays honest, has some dampening built in, and doesn't get amplified into a runaway cascade with nothing to interrupt it. This is precisely why exchanges added circuit breakers after 2010, not to recentralize the market, but to add one deliberate point where the decentralized cascade could be paused long enough for the system to reassess.

Where that leaves us

Across starling flocks, ant colonies, Wikipedia, markets, and algorithmic trading, the same underlying mechanism produces radically different outcomes depending on what it's built around: many independent parts, each responding only to a simple local signal, can produce coherent large-scale behavior with no central planner ever directing the whole.

This works remarkably well when the local signal genuinely carries useful, honest information, a starling's neighbor's movement, an ant's pheromone trail, a price reflecting real scarcity. It fails, sometimes catastrophically, when the local signal becomes misleading faster than any individual part can recognize it, a cascading sell order that looks like real demand until it's too late, or a Wikipedia article's hierarchy-free coordination growing beautifully while the same absence of center makes it harder to protect from being quietly captured by whoever invests the most sustained effort.

The deeper lesson connects directly to Principle 3 (Networks Outlast Hierarchies): decentralization isn't automatically self-correcting just because it lacks a center. Wikipedia's leaderless system still grew informal hierarchies over time; a "decentralized" market still needed a designed circuit breaker to interrupt a cascade nothing inside the decentralized system itself could stop. Coordination without a center is a powerful mechanism, not a guarantee of good outcomes, its value depends entirely on whether the local signal being amplified is trustworthy, and on whether something is built in to interrupt the cascade when, inevitably, it sometimes isn't.

Open questions

How can a system be designed to keep the benefits of decentralized coordination (speed, resilience, no single point of failure) while still having a way to interrupt a bad cascade before it runs to completion, without that interruption mechanism becoming a new, permanent center of control?
Wikipedia's informal hierarchies emerged even though the system was explicitly designed to prevent them. Is genuinely permanent, non-hierarchical coordination possible at any real scale, or does every sufficiently large decentralized system eventually grow some form of informal center regardless of its original design?
The Flash Crash shows that individually rational local behavior can produce collectively disastrous outcomes. Is there a way to test, in advance, whether a given local signal is the kind that stays honest under stress, versus the kind that can flip into a runaway cascade, or can this only be discovered after a failure has already happened once?
This principle and Principle 15 (self-correction) both depend on a trustworthy signal. Is "coordination without a center" really a distinct principle, or is it best understood as a special case of self-correction applied across many actors instead of one system?

Evidence and references

Draft. Several citations here are secondhand and are being checked against the primary sources. Where the underlying science is contested, the dispute is described rather than settled.

Nature

Starling murmuration self-organization, National Geographic; The Conversation, "Why do flocks of birds swoop and swirl together in the sky?"; Young, G. et al., PLOS Computational Biology / "scale-free correlation" study via All About Birds; Reynolds, C., "boids" simulation, 1987
Ant colony stigmergy, Gordon, D.M. et al., "The Regulation of Ant Colony Foraging Activity without Spatial Information," PLOS Computational Biology, 2012; Grassé, P-P., stigmergy concept, 1959, as summarized in multiple secondary academic sources

Modern Science

Hayek, F.A., "The Use of Knowledge in Society," American Economic Review, Vol. 35, No. 4, 1945