The Metaverse Doesn’t Need a Blockchain. It Needs a Constitution.

Every few years the word metaverse comes back, usually with tokens, NFTs and roadmaps attached. The implicit claim is that a metaverse is a technology: something that will arrive once enough of the world is on a blockchain and assets can move freely between platforms.

I think this gets it backwards. A metaverse is defined not by the technology that powers it, but by the world it makes possible: a persistent space with shared rules and possibilities for action, inhabited by a community that gives those rules meaning. By that definition, metaverses already exist, and some have existed for more than twenty years. One could argue about the level of immersiveness, but that is ultimately an issue of peripherals and network infrastructure coverage.

So the interesting question is not how to put them on-chain. It is whether the societies living inside them could one day recognize, negotiate and cooperate with one another across worlds. The usual pitch for interoperability is about assets: a sword that travels from one game to another. I want to argue for a different layer. Interoperable swords are easy to imagine and hard to care about. Interoperable societies are the real frontier.

The essay has three parts. The first starts from worlds that already work and the organizations inside them, and asks why those organizations would want to meet at all. The second proposes how they could: a federation and the minimal constitution it would need, what operators get from it, and the smallest protocol that could make it work. The third takes on the problems that could break it: who counts as a member, what happens when members aren’t human, how value crosses between worlds, and how to keep the biggest groups from dominating everyone else.

Part I — The Metaverse Already Exists

1. Worlds That Already Work

Consider three very different worlds.

World of Warcraft has geography, history, hierarchies, professions and economies. Its guilds are organized communities with rules, ranks and internal politics, and many outlive the content they were formed to play. Players stay for the people, not the raids.

EVE Online is a political and economic universe. Alliances wage wars that last months, run supply chains across star systems, sign treaties and break them. Very little of this is scripted by the developers. It emerges from thousands of players acting under shared constraints: scarcity, distance, risk and the permanence of loss.

Sea of Thieves is smaller in scale but no less social. Its world is built from encounters between strangers: an alliance at sea, a betrayal at the last moment, a reputation that precedes a crew into the next server.

None of these worlds needs a blockchain. A world can run on a ledger and remain socially empty, a marketplace with no society around it. Another can be fully centralized and host an extraordinarily rich political life. Blockchain is neither necessary nor sufficient for a metaverse.

What these worlds have in common is that they are layered. Part of their rules lives in code: what you can build, what is scarce, what happens when you die. But much of what makes them alive lives elsewhere: in conventions about fair play, in reputations that travel by word of mouth, in institutions players invent on top of the code. The code defines what is possible. The community decides what is done, what is acceptable and what is remembered. Treating metaverses only as infrastructure, as a question of servers and asset formats, misses most of what matters.

2. Autonomous Organizations Within Worlds

In crypto, a DAO is usually an organization that coordinates resources through smart contracts, governance tokens and on-chain voting. Its rules live in code no one can unilaterally change.

Now look at an EVE alliance. It has collective goals and distributed roles: fleet commanders, logistics officers, diplomats, recruiters. It has decision-making rules, shared treasuries and procedures for joining, leaving and being expelled. It manages cooperation and conflict, internally and with rivals.

It is not a DAO in the technical sense, but it does much of what we expect an autonomous organization to do. The difference is that its coordination infrastructure is not a set of contracts. It is the world itself. Territory, scarcity, travel time and the economics of production shape what the alliance can decide and enforce.

This suggests a different design space. Instead of building organizations on an abstract ledger and then looking for something for them to govern, we can start from organizations whose infrastructure is a world, with its own economy, geography and consequences. Their power is real within that world, and legible to everyone who lives there. The question is not which label applies. It is which kinds of collective agency a given world makes possible.

3. Power Without Sovereignty

There is a real limit to all of this. If a single developer controls the world, no organization inside it is sovereign over its infrastructure. Blizzard controls World of Warcraft, CCP controls EVE, Rare controls Sea of Thieves. However powerful an alliance becomes, the operator can change the rules, ban accounts or shut down the servers. The alliance governs in the world, never over it.

This tension between social autonomy and infrastructural dependence is the central fact any serious proposal must start from. But lacking sovereignty is not the same as lacking power, and a current example shows the difference.

At the time of writing, Olympus is the guild network built around the streamer Asmongold in the beta of World of Warcraft: Forever. Players have formed more than forty guilds named Olympus, Olympus II and onward, most with hundreds of members or already at capacity. The game caps guild size, so the community did what political communities have always done when they outgrow their institutions: it federated. Olympus is not one organization. It is a network of organizations sharing a name, a leader and a purpose.

Olympus owns no server, yet it coordinates thousands of people, sets expectations, distributes responsibilities and mobilizes collective action. Its decisions stick partly through the game’s mechanics, partly through social incentives, and partly through the practical consequences of being inside or outside it.

This is where a distinction the debate often blurs becomes essential. Governance determines who makes decisions. Enforcement determines whether those decisions have consequences. Smart contracts can enforce some rules mechanically, but they cannot replace social coordination or guarantee that an organization stays legitimate.

Enforcement also runs the other way. As thousands of members followed the streamer around, non-members complained about crowded cities. Some built their own response: a client-side mod that hides chat from anyone in an Olympus guild and auto-declines their invites. Because the game groups players by ruleset rather than by realm, moving server wasn’t an option, so they built a partial exit enforced on their own screen.

The point is not to judge Olympus. It is that in any world, members, rivals and bystanders all have enforcement tools. The power of a large organization is bounded less by formal rules than by what others can do in response.

The mod works because Olympus labels itself: it filters by declared membership. A network that didn’t want to be filtered would only need different names. Recognizing an organization is one thing; establishing who controls it is another.

4. Why Worlds Would Meet

Before proposing any structure, there is a question most interoperability visions skip: why would a guild in one world want anything from an alliance in another? If the answer is “because it would be cool”, everything that follows is a solution looking for a problem. I think there are real answers, and some of them are already being lived informally.

Communities outlive their worlds. Games end, decline or change beyond recognition. Many long-running gaming communities already span several titles, moving from game to game while keeping their leadership, culture and history. Today they carry that continuity in Discord servers, spreadsheets and memory. When they arrive in a new world, nothing there knows who they are. A portable organizational identity, with a verifiable record of what the group has done, is the first thing a federation could offer, and the one with the clearest demand.

The Syndicate is the best example. Founded in February 1996, before Ultima Online had even launched, it has played Ultima Online, EverQuest, Shadowbane and World of Warcraft under a single charter that forbids stealing, cheating and non-consensual player-killing. Guinness World Records lists it as the oldest continuously operating online gaming community. It has protected its identity across worlds the only way available: by trademarking its name and working with game companies to rename guilds that used it. That is a portable organizational identity, enforced through intellectual property law because nothing else would enforce it. Goonswarm shows the same pattern from another angle: Goons from the Something Awful forums were already in EVE by 2005 and formed the Goonswarm alliance in 2006, a community that existed before the game and was, from the start, more about itself than about EVE.

Trust is expensive to rebuild. Recruiting, trading and alliance-making all rest on reputation, and reputation is local. A guild that has vetted its members for years, honored dozens of agreements and run a fair loot system starts from zero every time it crosses a border. A shared, verifiable record of past commitments lets trust travel without forcing anyone to accept it blindly.

When that trust breaks, there is rarely anywhere to take it. In 2009 the CEO of EBank, EVE’s largest player-run bank, embezzled around 200 billion ISK of depositors’ funds and sold it for real money. CCP banned him for the sale, not the theft; inside the game’s rules, betraying thousands of depositors was legitimate play. The depositors had no court, and no record that would follow him into the next world.

Organizations already act across worlds. Charity events, tournaments, joint streams and multi-game leagues are coordinated across titles all the time. These arrangements involve real commitments: prize pools, schedules, rules, who gets credit. They are held together by goodwill and private chat logs.

Smaller worlds need allies. A small or new world has no way to attract established communities except marketing. A federation where organizations can arrive with their identity intact, and where the world decides on what terms to recognize them, gives small worlds a way to compete for communities rather than just players.

None of this requires moving a single character or item. In a sense the informal federation already exists: it is Discord, plus trust, plus memory. What it lacks is any way to make recognition, commitments and judgments verifiable once they cross a border.

It should also be said plainly that demand today is modest, and concentrated in multi-game communities. That is not a weakness of the idea. It is the size of the problem worth solving first. A federation has to start where that demand is, not where the vision is largest.

Part II — A Federation of Worlds

If some communities have a concrete reason to meet across worlds, what would a structure for that meeting look like? This part proposes a federation and the minimal constitution it would need, explains what operators would get from it, describes the smallest protocol that could make it work, looks at where blockchain fits, and asks how it could start.

1. A Constitution in Three Layers

Olympus federated because one world’s rules couldn’t contain it. If federation can emerge spontaneously inside a single world, why not between worlds?

There is no reason to imagine metaverses as isolated forever, or as destined to merge into a single universe. I’d propose a third option: a federation of autonomous worlds that agree on minimal protocols for cooperation while keeping their differences intact. Azeroth does not need to become compatible with New Eden. Their difference is the point.

Such a federation would need a constitution, and its first job would be to separate three kinds of authority:

• World governance decides the rules of each environment: its physics, economy and permissions.
• Organizational governance decides how communities inside a world behave, who leads them and how they use their resources.
• Federal governance decides only what cooperation requires: how organizations are recognized, how agreements are made, how disputes are resolved and how the shared rules can change.

These layers do not have to belong to the same actors. A federation could include centralized commercial worlds, open-source worlds and fully decentralized ones, as long as it is clear where each participant’s authority begins and ends.

Taken together, these layers are a constitution in the minimal sense. Not a founding myth or a code of law, but an answer to two questions: who decides what, and which guarantees hold for everyone, wherever they play. The three layers answer the first question. The guarantees are harder. A first draft of them might read: the right to leave, the right not to see, recognition granted rather than assumed, and a place to be heard when someone breaks their word. The rest of this essay tests whether they can hold.

Each layer also raises the question Olympus made concrete: who can enforce the rules, and what keeps that power from becoming domination? A federation’s rules matter only if its members have credible ways to enforce them, through technical mechanisms, economic incentives, reputation or the ultimate right to exclude. And every enforcement power needs its own limits, including the right of those excluded, or who want no part of it, to leave.

The federation also has to be honest about what it promises. A federation of communities can run on top of centralized infrastructure, much as international associations operate across countries they don’t control. A federation that promises technical sovereignty to its members has to confront the infrastructure problem head-on. Both are legitimate projects. They are not the same project, and confusing them is how most metaverse visions end up overpromising.

2. What Operators Get

Any proposal that touches existing worlds has to answer a blunt question: why would Blizzard or CCP cooperate? Most commercial games prohibit trading in-game assets for outside value, and for good reasons. Real-money trading brings gold farming, account theft, fraud and regulatory exposure, and it erodes the designer’s control over scarcity. An operator asked to plug its economy into a cross-world currency will, rightly, say no.

So the federation has to be built in two tiers, with very different demands on operators.

The organizational tier needs no operator at all. Identity, reputation records, agreements and dispute resolution can run alongside a world without touching its code or its economy. Where an operator is willing, it can attest that an organization exists in its world. Where it isn’t, other recognized organizations can act as witnesses. Nothing here breaks a terms-of-service agreement, because nothing in-game is being sold or moved. This tier is where the demand described in Part I lives, and it can be built first.

The economic tier is opt-in. A neutral currency, and the controls needed to keep it from draining worlds, would apply only to worlds that choose them: worlds built with an open economy from the start, community-run worlds, or operators who decide the trade-off is worth it. A federation that required economic convertibility to join would exclude almost every world worth federating. Part III looks at this tier in detail.

What do operators gain from even the first tier? Retention, for one. A community that can carry its identity and history out of a world is, paradoxically, more willing to invest in it, because the investment is no longer a hostage. Acquisition, too: communities leaving a declining game need somewhere to go, and a world that recognizes their record on arrival is more attractive than one where they start as strangers. And some disputes between organizations, which today land on overloaded support teams or nowhere at all, would have a venue.

The honest caveat is that the largest incumbents have the least to gain. Lowering the cost of leaving helps challengers more than leaders. That suggests where adoption would start: small and new worlds, indie and community-run servers, and multi-game communities themselves. Many open standards have spread this way, adopted first by those with something to win from interoperability and only later by those who could no longer afford to stay out.

3. A Protocol for Meeting

All of this stays abstract until we ask how two organizations from different worlds, with different operators and identity systems, would actually recognize each other, make a commitment and settle a dispute. A minimal protocol might have five parts.

Organizational identity. An organization’s federal identity is controlled by the organization itself, through keys held by its officers under its own internal rules, not by the operator of its world. In-world existence can be backed by an operator’s attestation or, where operators won’t cooperate, by recognized organizations acting as witnesses. If a world shuts down, the organization’s identity, agreements and reputation survive. This is the piece that serves communities moving between worlds. The hardest case is not an operator or a rival, but an internal split. When officers disagree, the protocol must define in advance how keys are rotated and who inherits the record, including whether the guild forks or the federation arbitrates the schism.

Membership without exposure. Members prove they belong to an organization, and what role they hold, without revealing which character they play or linking identities across worlds. Zero-knowledge proofs make this possible; Part III returns to what they can and cannot prove.

Reputation as a record, not a score. A single portable reputation number would be easy to game and would erase context. Instead, reputation is a history of verifiable attestations: this organization fulfilled this agreement with that one. Each world and organization decides how much weight to give foreign records, translating reputation rather than transferring it.

Commitments as structured agreements. An agreement states the parties, the obligations, how fulfillment will be verified, any collateral posted, and when it expires. Verification is the hard part. Some in-world events can be attested by operators, some by witnesses, and some not at all. Those should be marked honestly as resting on trust. In practice, every attestation should carry its provenance and its strength: who observed the event, how, and what that observation can and cannot establish. A panel weighing an operator’s log against a rival’s testimony is doing what any investigator does with mixed evidence, and it can do it well only if the quality of each piece was recorded when the agreement was made, not reconstructed after it broke.

Disputes resolved by peers. When an agreement breaks down, a panel drawn from uninvolved organizations hears the case, much as the Hanseatic Diet settled disputes between member cities. Sanctions escalate: a mark on the record, loss of collateral, suspension, and finally exclusion. There is a right of appeal and, as always, a right to leave.

The Hanseatic comparison carries a warning as well as a model. The League’s ultimate sanction, expelling a city from its trading network, worked because no city could afford to lose access to that network. A federation of worlds has the same dependency. Its sanctions will bite only as far as membership is worth more than leaving, and at the start it will be worth very little.

None of this requires every world to adopt the same technology, or any operator to change its game. It requires agreeing on what a recognition, a commitment and a judgment look like when they cross a border.

4. Where Blockchain Fits

Blockchains offer real properties: shared records, verifiable execution of certain rules, and coordination between parties who don’t want to trust a single intermediary. For a federation of worlds run by different companies and communities, those properties matter. In the language of this essay, they are enforcement tools.

The protocol above is the most natural place for a ledger: a shared record of identities, agreements, collateral and reputation attestations that no operator can alter. If the economic tier comes to exist, a neutral currency would be the other: a unit no single world controls, with conversion rules enforced in code. Dispute panels and recognition decisions, on the other hand, are human institutions. A chain can record their outcomes, but it cannot replace them.

Blockchains also bring costs: fees, scalability limits, poor ways of identifying people, and dependence on off-chain systems for most interactions with the real world. Above all, a decentralized ledger is not a decentralized world. A game can issue tokens while giving players no meaningful power, and every large guild proves that sophisticated institutions can exist without touching a chain. A ledger records that value moved between two keys; it says nothing about who holds them. Anyone who has traced funds across a public chain knows the record is where an investigation starts, not where it ends.

If cryptography ends up essential to this project, the ledger may not be the most important part. Zero-knowledge proofs might matter more. They already let members prove they belong without exposing who they are, and they could one day verify that a member is a unique human and that an agent answers to someone. As Part III argues, that is what decides how far a federation can govern itself.

So the right question is not “how do we bring games on-chain?” It is “which tools let societies in different worlds meet, while respecting what makes each world distinct?” That is a problem of institutional architecture before it is a technological one.

5. How It Starts

A protocol like this has a chicken-and-egg problem. A reputation record is valuable only if others read it. Witnesses are credible only if they are already recognized. A dispute panel needs uninvolved members to sit on it. On day one there is no one to recognize anyone.

The way out is to begin with what already exists. A handful of multi-game communities that already know and trust each other, and already run tournaments, charity events or joint leagues together, can form a founding cohort. They vouch for each other, so the first recognitions are simply existing social trust made explicit. Their next shared event becomes the first structured agreement, with stated obligations and a named panel in case it goes wrong.

The first version can be small: a registry of organizational identities controlled by officer keys, signed agreements between them, and attestations when those agreements are fulfilled. At this stage its value is not portability across worlds. It is giving commitments these communities already make a place to be written down and a place to be judged. Portability becomes valuable later, once the record is long enough to be worth carrying.

The test is blunt. If the founding cohort doesn’t use it for something it already does, nothing else will make it useful.

Even this small version raises the questions that could break a larger one. Who counts as a member, when one person can run a dozen characters, or a hundred AI agents? How should a federation decide anything, if it can’t tell people apart? And how do you keep the largest communities from carrying their dominance from one world to the next?

Part III — The Hard Problems

This part is about what could break the federation. It starts with the most basic question any shared institution faces, and ends with the problems I can’t yet solve.

1. Who Counts as One?

The moment a federation makes shared decisions, it faces a basic question: who counts as one?

Inside a single world, the operator acts as a de facto identity provider, and even then sybils are everywhere. In EVE, one person may run a dozen characters. Inside a world this is tolerable, even part of the game. At the federal level it is fatal: if influence is distributed per participant, whoever controls the most accounts, or the most fake organizations, controls the federation.

The principled answer is a zero-knowledge proof of personhood. With scoped nullifiers, a person could prove “I am a unique human and haven’t voted on this proposal yet” without revealing who they are or linking their characters across worlds. That allows two kinds of identity to coexist: within worlds, plural, pseudonymous and playful; in the federation, unique, proven and unlinkable to any in-world persona. Players keep their masks. Uniqueness applies only where it matters. The cost is real: the federation cannot follow a thief like EBank’s CEO from world to world as a player. It can follow the organization he ran, and, as an officer, him. Ordinary members keep theirs; officers, who hold other people’s trust and treasuries, answer under a persistent key, bound to their proof of personhood, so that a new world does not mean a clean record.

Two caveats remain. Zero-knowledge proofs don’t stop someone from selling or renting their identity, so vote-buying stays open. And existing personhood systems have real costs in privacy, exclusion and, for biometric approaches, trust.

Then there is a harder version of the problem, arriving fast. AI agents will soon play, negotiate and form relationships in real time well enough to pass as human. This breaks more than voting. A natural sybil defense is reputation earned through play, since every fake identity should cost real time. But an agent plays around the clock at almost no cost, so one person could raise a hundred credible characters. An organization the size of Olympus could be mostly agents controlled by one actor, and no one would notice. It also threatens what makes these worlds meaningful: trust, loyalty and betrayal all assume someone is on the other side. A betrayal by a script is not a story. It is a bug.

Even zero-knowledge personhood doesn’t fully solve this. It proves a unique human holds a key, not that the human is the one acting. The question shifts from “are you human?” to “which human answers for you?”

Banning AI characters is neither enforceable nor desirable; agents could populate worlds and keep them alive at 4 a.m. What a federation can do is require disclosure and accountability: every agent declared, every agent traceable to a unique human, and a cap on how many agents a person can field in federal affairs. Writing these rules is easy. Verifying them is not yet possible. Until it is, a federation cannot be governed person by person.

2. A League Before a Federation

If a federation can’t count people, it has to count something else. Today, two kinds of governance hold up against sybil attacks, human or artificial, and neither is “one person, one vote”.

The first is admission by vouching. The federation’s members are organizations, not individuals, and an organization is admitted only when existing members vouch for it and share part of the consequences if it turns out to be fake or hidden. Creating a thousand accounts is cheap. Persuading established organizations to put their own standing behind a thousand fake guilds is not. This is how the organizational tier can govern itself from the start, with no token at all.

This defense is only as strong as the federation’s sanctions. If losing standing costs little, vouching for a fake guild costs little too, which is why the enforcement problem at the end of this essay is not separate from this one.

The second is economic weight. Splitting your stake across ten accounts, or a hundred agents, doesn’t increase your influence. Stake-weighted voting resists sybils precisely because it gives up on counting people. Even quadratic voting collapses without unique identity, because it rewards splitting funds across many wallets. This only becomes relevant where there is something to stake, which means in the economic tier.

So I want to be honest. Today, a federation of metaverses can realistically be a league, not a democracy.

The Hanseatic League, already the model for dispute resolution in Part II, is the precedent here too: autonomous cities coordinated by shared interest rather than democratic principle, with no emperor and no parliament of citizens, that still acted together for centuries.

A league doesn’t have to mean pure plutocracy. It can be balanced by design:

• Exit rights. Members who disagree can leave with their share, where there is one, and always with their record. In Hirschman’s terms, where voice is limited, exit must be guaranteed. The mod built to mute Olympus showed what happens when exit is missing: people improvise one.
• Non-transferable reputation. Standing earned through verifiable contribution can’t be bought, though its value as a sybil defense will erode as agents improve.
• Conviction voting. Influence grows with the duration of commitment, so stake bought at the last minute counts for little.
• Two chambers. One represents organizations admitted by vouching. The other, weighted by stake, exists only once the economic tier does, and the first holds a veto over it. Capital alone never decides, and at the start it doesn’t decide at all.

The path runs in stages: a league of organizations first, honest about its limits; an economic chamber as worlds opt into the economic tier; a political federation once personhood and agent accountability can be verified. The league is not the destination. It is the version we can build without lying about what our tools can guarantee.

3. The Economic Membrane

This section applies only to the opt-in economic tier. But for worlds that join it, it is where the league’s main counterweight lives or dies.

Exit rights are only as strong as what you can take with you. A member who leaves with in-world gold, ships or land may be leaving with assets worthless outside that world. When the cost of leaving is losing everything, exit stops being a right and becomes a threat nobody can afford to carry out.

The obvious fix is a neutral currency: a unit of value shared across the federation, issued by no single world, into which in-world assets can be converted. If members can sell their share for something that keeps its value elsewhere, leaving becomes credible, and large organizations have to keep earning their members’ loyalty.

But convertibility has a price. Once in-world assets have a stable external price, every world’s economy is connected to every other, and to the economy outside the game. Players optimize for extraction rather than play. Farming becomes a job, one AI agents can do around the clock. Capital can flee a world in an afternoon. The play-to-earn experiments of recent years showed how quickly a world turns into a mine, and how quickly it collapses when the price falls.

So the boundary between each world’s economy and the neutral currency should be neither a wall nor an open door but a membrane, letting value cross at controlled rates: conversion caps so that no world suffers a bank run, exit delays that make panic and speculation costly, exit fees that compensate those who stay, and some goods deliberately left non-convertible, so that parts of each economy exist only for play.

There is a tension here that should be named rather than hidden. Each of these tools limits exit, and the whole reason for the neutral currency was to make exit cheaper. The membrane does not resolve that trade-off; it chooses a point on it. Set too loose, worlds get drained. Set too tight, the right to leave is back to being a threat no one can carry out, and the league loses its main protection against capture. There is no correct setting, only a setting that is visible, bounded and contestable.

Nations have managed this problem for a century through convertibility rules and capital controls. The postwar Bretton Woods system tied currencies together while letting each country shield its domestic economy from sudden flows. It is also a warning: it broke down in the early 1970s, when the pressures on it outgrew the fixed settings it was built around. A league of worlds needs something similar, scaled down and made programmable, and its settings must be adjustable before they break.

Who sets the membrane? Each world sets its own, because how open its economy is to the outside is part of its identity. Each organization decides how its treasury and its members’ shares are redeemed, since that is where exit is actually exercised. The federation only bounds the trade-off: a minimum permeability, so no world can turn its members into hostages, and a maximum total cost of leaving, so fees and delays can’t quietly add up to the same thing.

4. What Exit Can’t Fix

The membrane and the right to leave protect members against economic capture. They are much weaker against something subtler: social dominance.

An organization doesn’t need to control a world’s economy to define what living in it feels like. Olympus shapes the experience of non-members through sheer density, the volume of its chat and the pressure of its recruitment. None of this breaks a rule, and the people most affected aren’t members, so they have nothing to exit from.

A federation could make this worse. If an organization can carry its recognition and alliances across worlds, a dominant group in one world arrives in the next one already powerful. Interoperability amplifies whoever is already biggest.

A few principles can limit this without punishing popularity itself:

• Networks count as one. Forty guilds under one name and one leader are one organization for federal purposes. Otherwise splitting becomes the organizational version of a sybil attack.
• Caps on federal weight. No single network holds more than a fixed share of either chamber, however large its membership or stake. However, this creates an incentive for Sybil attacks, with dominant networks potentially splitting into multiple identities to circumvent the caps.
• Recognition is granted, not automatic. Each world decides whether, and on what terms, to recognize an organization arriving from elsewhere. A world can remain a place where a powerful outside group starts from zero.
• The right not to see. The mute mod built against Olympus points to a real need. A federation could guarantee interoperable block and filter lists, so opting out of an organization’s presence doesn’t depend on a volunteer writing an addon.

The first principle is harder than it looks. Olympus is easy to recognize as one network because it wants to be recognized: one name, one leader, one brand. A network that wanted to evade a cap would do the opposite, with different names, front leaders and no visible link between its parts. Detecting that is the sybil problem again, one level up, and it has no clean solution. The realistic tools are indirect: patterns of coordinated voting, shared treasuries, overlapping officers, and the shared liability of vouching, so that organizations which vouched for a hidden network pay part of the price when it is exposed. These raise the cost of disguise. They don’t eliminate it.

This is a familiar difficulty in adversarial investigations: the evidence an observer relies on is part of the environment an actor can learn to manipulate. Coordination can be concealed, legitimate activity can look like collusion, and independence can be manufactured. The hard part is not detecting patterns but deciding when converging signals are strong enough to support attribution. A system that treats every correlation as proof will produce false accusations; one that demands certainty before acting will be easy to evade. A panel should therefore name the threshold before it looks: which independent signals, from which sources, must converge before a set of guilds is treated as one network, and what an accused network can show to rebut it.

The test of legitimacy is not size. It is whether people outside an organization can still play the game they came for.

5. Open Questions

I want to end the argument where it is weakest. Three problems remain genuinely open.

Personhood and agents. A proof can show that a unique person holds a key. It cannot show who, or what, is acting with it. Until that can be verified, rules about agent accountability are only good intentions.

Enforcement between worlds. Inside a world, enforcement leans on the game’s mechanics and the daily presence of other players. Across worlds, most of that disappears. What remains is reputation and collateral, both weak tools. An organization that doesn’t need its partners can shrug off a bad record, and collateral only covers what can be priced. As the Hanseatic comparison already warned, exclusion only bites when membership is worth more than leaving, and a young federation’s is not. Nothing guarantees it ever will be. Without stronger enforcement, a federation risks becoming a forum rather than an institution.

Cultural capture. Counting networks as one, granting recognition instead of assuming it, and guaranteeing a right not to see are reasonable defenses. But they probably underestimate network effects. Large communities attract members because they are large, and no cap on federal weight stops one culture from becoming the default everywhere players gather.

These three problems are connected: each is a version of not knowing who is really acting, or what they will do once they no longer need you. I think they will have to be answered together rather than one at a time, and that is the subject of a future piece. For now, I would rather leave them open than force an answer that makes the essay look more complete and the idea less honest.

Conclusion

Metaverses are already here. They live in guild halls, alliance channels and pirate crews, not in token whitepapers. Their organizations already exercise power in worlds they don’t own, and some, like Olympus, have already federated to grow beyond a single structure. Communities already move between worlds, carrying their history in chat logs and memory because nothing else will carry it for them.

What doesn’t exist yet is a way for these societies to recognize, negotiate and cooperate across worlds without dissolving into a single universe. The place to start is modest: an organizational layer that needs no operator’s permission, serving the multi-game communities that already want it. An economic layer can follow, opt-in, for worlds willing to accept the trade-offs of a membrane. And the whole thing can be honest about being a league held together by mutual vouching, economic interest and the right to leave, until proofs of personhood and accountable agents make something closer to a true federation possible.

This is where the constitution sketched in Part II was heading. Its three layers say who decides what. Its guarantees, first drafted there, are what the hard problems leave standing: not a founding document for a single universe, but a short list that holds across worlds that share nothing else. The right to leave, the right not to see, recognition granted rather than assumed, and a place to say you broke your word and be heard.

A federation of metaverses would not need to merge these worlds. It would need to make their societies capable of meeting each other.

Next: Episode #16: Four Firebreaks Between a Mixer and an Exchange