This piece was originally written as a section of my upcoming proposal for designing communities as commons on the AT Protocol. However, I have separated it here for the sake of focus in the main body of text, and will reference this leaflet when it becomes relevant. This section is used to establish a necessary premise of my proposal: that online democracy is not the solution to implicit feudalism.

I’ve had my fair share of experiments with online democracy, from communities as large as tens of thousands to as small as a few dozen. These experiences will be what guide my commentary below.

Why do online democracies fail? I believe I have zeroed it down to a few, fundamental problems.

Democratic governance works best when:

• Issues are legible enough to vote on meaningfully

• Participants share a meaningful commitment to both the outcome and the process itself

• Conflict can be externalized to an authority who absorbs the social cost of enforcement

None of these conditions reliably hold true in democratic online communities.

The Shared Role of Governance

In the real world, governance has direct outcomes on your material reality. It determines whether your infrastructure is maintained, whether you receive healthcare insurance this year, whether the rule of law itself is respected in civil society. In any kind of society, we can assume that anyone who lives within it has both some knowledge of the social systems at play and a vested interest in the shape these systems take.

This is not the case within online communities. In fact, I think it’s fair to say that, in most online communities, people don’t want to govern. They just want to participate. Asking users to vote on moderation policy is asking them to do something which they did not sign up for. After all, they don’t live in these communities, do they? It’s merely one of many social spaces any one person inhabits. Even more than that, it’s an abstracted social space, one which feels fundamentally less “real” in a visceral sense than real-life social spaces.

This mere fact can potentially have detrimental effects on the composition of the electorate. After all, how can we reasonably expect every individual who votes to have a serious stake in what’s being voted on? How can we expect them to keep up with the intricacies of every social space they inhabit at all? People already have enough to worry about in their real lives. It’s unrealistic to ask them to keep up with the history and conflicts of every social space they inhabit on top of that.

What does this result in? Well, this means that people who do end up engaging with governance consistently—the people who shape the outcomes— are the ones who are the most unusually invested within the community. This might not necessarily be a bad thing; after all, these could simply be people who are acting as dedicated stewards of the community space. However, when a contentious issue arises, that voting population can very quickly tilt towards people who simply have an ax to grind.

Self-selection into governance is not random.

Outcomes and Process

In every social online space, there will come a time where a contentious decision must be made. These decisions are, almost definitionally, cases where the community is deeply divided. People who might otherwise have a high degree of trust may nonetheless disagree on what the best outcome is. Anyone who has managed a community space knows that this is an inevitable outcome of managing such spaces.

In these situations, consider; what exactly does voting do here? Holding a vote on a contentious issue does not resolve or even diffuse the social divisions within the community. In fact, it does quite the opposite; it crystallizes it into the fabric of the community. The outcome of the vote and the opinions of the community space are calcified into the trust dynamics of the social web.

With any contentious decision, the outcome will cause anger within the community. This is inevitable. Yet, consider what happens when a vote is introduced. Do you, dear reader, think that the losing side will simply let go of their anger, throw up their hands, and think “Well, there’s nothing we can do! We lost fair and square”.

…Of course not!

More often, users will externalize their anger to the community, a community that, they so believe, has turned against them and their values. In a worst-case scenario, the losing side relitigates, not just the outcome, but the process itself. “Was the vote really fair?” “Did the other side brigade the vote?” “Were the rules here applied consistently?”

I’ve seen communities tear themselves apart over these democratic moderation decisions. Factions form. Internal trust networks collapse. People are forced to choose a side, losing social trust either way. Communities split. The very social fabric of the community is torn apart. And, in the end, who loses? It’s not just the faction that lost. It’s everyone. Those who won just as much as those who lost just as much as those who did not vote at all.

The vote makes every conflict everyone’s conflict.

But… controversial moderation decisions happen all the time in any community. Why is it that , in other, autocratic communities, the community fabric can survive such controversial decisions?

The Scapegoat

Over my years of moderation experience, I’ve learned an uncomfortable truth: communities need someone to blame.

When difficult decisions happen—and they will happen— people will get angry. This anger builds up within the community like a pressure cooker—and it has to go somewhere. If there are moderators who made that decision, the anger will be channeled to the moderators. This is, of course, unpleasant for the moderators—but it’s survivable. It comes with the territory. Any good moderator knows this.

In this way, moderators act as a literal scapegoat: their social job is to absorb the anger, take the hit, and let the community continue functioning.

But consider: what happens in a community with democratic governance? When the community itself votes to implement a contentious policy? Well… the only ones left to blame now are each other. The anger that would have been directed at the moderators instead circulates viciously within the community.

Democratic governance does NOT reduce or eliminate the impact of conflict within communities. It simply internalizes it. And, in my experience, internalized conflict is much, much harder for a community to recover from than a conflict with leadership.

I’m not trying to make an argument for autocracy here. I’m simply sharing my findings with my own experiments with online democratic governance. And it’s clear to me that, in any community, absorbing blame is one of the core social functions of moderators. Any governance system needs to account for this important social role. Online democracy, quite simply, does not.