When the Rules Ate the Community: How Online Subcultures Govern Themselves to Death

I watched a Discord server die in slow motion. Not from drama, not from a ban wave, not from the owner getting doxxed. It died from rules.

The server was for writers — a few thousand people sharing drafts, workshopping chapters, trading tips. When I joined, it was chaotic in the way creative spaces should be. People posted half-baked ideas at 2am. Someone would ask “is this paragraph working?” and get twelve replies before morning. Arguments happened, sure, but they resolved themselves because everyone involved knew each other’s names and reputations.

Then someone added a rule.

Then another.

Then a moderation handbook.

Then a tiered warning system with automated enforcement.

Six months later, the server was a ghost town. Not because anyone left in protest. Because nobody felt like talking anymore. The rules had eaten the culture.

I started noticing this pattern everywhere once I looked for it. Forums, subreddits, Discord servers, fan wikis, hobbyist communities — they all follow the same arc. They form around a shared interest and a loose set of norms. They grow. Growth brings conflict. Conflict brings rules. Rules bring mods. Mods bring process. And somewhere in that chain, the thing that made the community worth being in disappears.

This is a story about governance. But it is also a story about ethics — what happens when a group tries to turn its moral intuitions into enforceable policy, and discovers that the translation loses something real.

The phase before the rules

Every online subculture starts in what I think of as the pre-rule phase. No written guidelines because none are needed. Everyone knows each other, or at least knows the norms. A small forum for vintage camera collectors does not need a rule against spam because there is no one to spam and nothing to gain. A fan wiki for an obscure 90s anime does not need a dispute resolution policy because disputes are settled by the one person who has watched the series seventeen times and can recite the episode guide from memory.

This is not idealism. It is a practical fact of scale. The anthropologist Robin Dunbar argued that humans can maintain stable social relationships with about 150 people at a time. Below that, informal gossip and mutual accountability are enough. Above it, you need institutions. Online communities blow past the Dunbar number fast — a single popular Reddit post can bring in five hundred new subscribers overnight — and the institutions they build are rarely designed so much as improvised.

People describe this phase as a golden age. I have heard that word from every community I have looked at. The truth is less romantic: it was the phase before the community had to decide what it actually stood for, because nobody had challenged its assumptions yet. The moral logic was implicit, unexamined, and for that reason, fragile.

The boring part of codification

The first rule in most communities is almost always about spam. The second is usually about harassment. Both are reasonable. Neither predicts the avalanche.

What happens next is well documented in legal theory, but the people living through it rarely recognize it. The sociologist Max Weber described the shift from “value rationality” — acting according to deeply held beliefs — to “formal rationality” — acting according to codified rules — as the defining feature of modern bureaucracy. Online communities are a laboratory for this transition, compressed from centuries into months.

A rule about spam seems neutral. But once you write it down, you have created a category: spammer. Now you need to define spam. Does a self-promotional post from a regular member count? What about a link to a relevant YouTube video? What if the member has not contributed anything in months? Each boundary case demands a clarification. Each clarification becomes a new rule.

The same thing happens with harassment. The first version is simple: don’t be mean. Then someone reports a comment that was not mean but was condescending. Someone reports a joke that was funny to three people and hurtful to a fourth. Someone reports a political opinion that some members find offensive and others find banal. The moderation team starts building a glossary. Tone policing. Sealioning. Concern trolling. Whataboutism. Each term is an attempt to draw a boundary the previous rule did not capture.

The community I watched die added a rule about “unsolicited advice” after a new writer posted a rough draft and got more feedback than they could handle. Then a rule about “backseat moderating” after people started telling each other to follow the rules. Then a rule about “meta-discussion” after people started arguing about whether the rules were being applied fairly. Every rule responded to a real problem. Every rule also made it slightly harder to just participate.

The legal scholar Robert Cover wrote that “legal interpretation takes place in a field of pain and death.” He was talking about criminal sentencing, but the idea travels. The moment you write a rule down, you create the possibility of enforcement. Enforcement always involves judgment. Judgment always involves someone being told they are wrong in front of other people. Do that enough times and the character of a community changes.

The moderation trap

I have talked to about a dozen volunteer moderators from different communities. They all tell the same story. They started because they loved the community and wanted to protect it. They stopped because the job burned them out. Not from the volume of reports. From the moral weight of the decisions.

The moderation trap works like this. A community’s rules are written in good faith, but they cannot capture every situation. A moderator facing an ambiguous case can either enforce strictly (producing an unjust outcome) or make an exception (undermining the rule’s authority). Either choice creates pressure for more rules. Strict enforcement generates appeals and demands for clarification. Exceptions generate accusations of favoritism. Both lead to the same place: an expanding rulebook and an exhausted moderation team.

I do not think this is the moderators’ fault. It is a structural feature of online governance. James C. Scott, in *Seeing Like a State*, argued that states fail when they try to impose legible, standardized rules on communities that operate through local knowledge and informal norms. The online moderator is trying to do the same thing with much less authority and no police force. A one-person state, trying to fit the messy reality of human interaction into a text file.

The philosopher Iris Marion Young called this the “scaling up” problem. Norms that work in small groups — trust, reciprocity, mutual adjustment — do not scale to large, anonymous populations. At scale, you need formal rules. But formal rules change the interaction. They replace trust with compliance. They replace judgment with procedure. They replace “let me think about whether this is okay” with “let me check the rules.”

What the rules cannot see

Here is the part that bothers me most. Rules cannot capture what matters about a community, because the most important things about a community are not rule-shaped.

A thriving community has texture. Inside jokes. Shared references. A sense of timing — when to joke, when to be serious, when to let a conversation breathe. Tacit knowledge: the regulars know that Threadsby, who posts at 3am, is going through something and needs space. NewUser99, who just joined, does not need to be told that their first post violates rule 4(c) about formatting. That knowledge cannot be written down. It is lived.

The philosopher Michael Polanyi called this tacit knowledge: “we can know more than we can tell.” Communities run on it. Rules try to make it explicit, and in doing so, they destroy it. The explicit rule replaces the nuanced judgment. The new member reads the rulebook and follows it literally, because they lack the context to interpret it wisely. The old member, exhausted by having to explain things that were once obvious, stops contributing.

The community does not die from malice. It dies from the loss of its tacit knowledge.

I have seen this pattern everywhere. A subreddit about mechanical keyboards developed such elaborate posting requirements — photo verification, post flairs, account age minimums, karma thresholds — that new enthusiasts simply went elsewhere. A fan fiction forum added a 500-word minimum for comments to prevent “low effort” engagement, and regulars stopped commenting because everything they wanted to say was beneath the threshold. A hobbyist game development community required design documents for any announced project, and a dozen active projects went silent rather than comply.

In each case, the rule was created to solve a real problem. In each case, the solution was worse than the problem because it could not tell the difference between the behavior it was trying to prevent and the behavior it was trying to protect.

What good governance looks like

This is not an argument against having rules. Large communities need them. The question is how you approach rule-making.

Most online communities treat their rules as a constitution — a fixed set of principles to be enforced consistently. I think they should treat their rules more like a garden. Rules need pruning. They need to be removed when they outlive their usefulness. They need to be applied with judgment, not automation.

Lon Fuller, in *The Morality of Law*, argued that a legal system needs eight criteria to be considered legitimate — among them that laws must be public, understandable, non-contradictory, and stable over time. Online community rulebooks routinely violate every one of these. They are hidden in obscure wiki pages. They use jargon new members cannot parse. They contradict each other. They change weekly.

Fuller also argued that the most important criterion is congruence between the rules as written and the rules as applied. A rule enforced selectively — enforced against some people and not others — destroys the legitimacy of the entire system. This is where most online communities break. The rules are written for everyone, but they are enforced against whoever the mods happen to notice or dislike. Once that perception takes hold, the moral authority of the governance system collapses.

I have been thinking about this in relation to something I wrote earlier, about how gaming communities build moral logic from unwritten norms. The guilds I wrote about there were more stable than the rule-heavy communities I am describing here. Not because they had better rules, but because they had fewer rules and more trust. The trust came from shared experience — playing together, losing together, winning together. Thin written code. Thick unwritten code.

That is hard to replicate in a large, open community. But it suggests a principle: rules should be the last resort, not the first. Before adding a rule, ask whether the problem can be solved with a conversation, or by setting a norm, or by changing how the space is designed. Most problems can. The ones that cannot probably need a rule. Those are rarer than we tend to think.

What survives

The writer’s server I mentioned at the beginning? It still exists, technically. The rules are still there. The moderator handbook is now forty pages long. The channel that used to get fifty posts an hour now gets maybe five. The people who made the server interesting are gone. What is left is a well governed empty room.

I do not know if there is a way to avoid this. Every community that grows beyond a few hundred members faces the same choice: codify or collapse. The ones that codify end up with orderly silence. The ones that refuse to codify end up with chaotic toxicity. I have not seen a third path that works at scale.

But I think knowing about the trap changes how you write rules. Knowing that rules can kill culture helps you write them more carefully. Knowing that tacit knowledge cannot be replaced by explicit policy helps you protect the people who carry it. Knowing that enforcement is always a moral act, not a mechanical one, helps you choose moderators who can exercise judgment rather than just follow a script.

I keep coming back to that Discord server. I keep wondering whether there was something we could have saved — not by writing better rules, but by trusting each other a little longer.