In a modern nation-state like the US, the legal system is an arm of the government and integrated with the government’s monopoly on force. Police forcibly arrest criminal suspects; convicted criminals or losers in civil cases are forced to go to prison or pay fines and damages. Ultimately, all the penalties in the legal system are enforced by armed agents of the government.
But legal systems don’t inherently require a state.
David Friedman, in Legal Systems Very Different From Ours, has explained how non-state societies (like Somaliland and medieval Iceland) and self-governing communities within states (like the Romani, the Amish, and pre-emancipation Jews) were able to use courts for dispute resolution without a monopoly on violence.
Implementations differ, of course. But the basic shared model is that a non-state court system is not a dispute-resolution agency administered by the state, but a protocol for dispute resolution administered by anyone who wants to.
Here’s how it works, in broad strokes.
When two parties have a conflict, they take their dispute to a court. Key elements of a court include a judge or panel of judges, an opportunity for both sides to state their case, a formal method for establishing the facts of the case and how the law applies, and a verdict or judgment.
If the judgment comes with penalties (“Alice shall pay $50,000 to Bob”), there are no police to threaten Alice with violence if she doesn’t pay the penalty. Instead, people who don’t abide by court judgments become outlaws. They are expelled from the community and the protection of the law.
In a stateless society like medieval Iceland, this means it was legal to kill an outlaw and illegal to harbor one. The friends and relatives of an outlaw’s victims would gang up and attack him.
For communities living within states, of course, the surrounding state might take issue with killing your community’s “outlaws”, but the community can still ostracize them. An outlaw could lose his entire social world and means of support.
That is, while the stateless court has no enforcement power of its own, it does provide robust incentives against lawbreaking, provided that a critical mass of community members participate in upholding and enforcing the law. A stateless legal system, once established, does not rapidly degenerate into everybody breaking the law all the time.
Note that “outlaws” are not those who break laws but those who refuse to abide by court judgments. You don’t get ostracized for breaking a law if you then comply with the court’s ruling, usually by making some kind of restitution to the other party in the dispute.
Non-State Courts and Game Theory
A non-state legal system structurally resembles a contrite strategy in game theory.
In the Iterated Prisoner’s Dilemma, one strong strategy is tit-for-tat: start by cooperating, cooperate with players who cooperated with you last round, and defect against players who defected against you. This strategy corresponds to common moral intuitions like “reciprocity” or “rewarding good behavior and punishing bad”.
However, tit-for-tat isn’t always the strongest strategy, particularly if you add noise to the system such that players might “accidentally” deviate from their strategy occasionally. If two tit-for-tat players accidentally defect against each other, they keep defecting against each other forever. This means that tit-for-tat “cooperates poorly with itself”; it tends to burn itself out through infighting and get outcompeted by other strategies.
The analogy to tit-for-tat in the world of human society is vendetta. If you hurt me, I’ll hurt you, and then you’ll hurt me back, and I’ll hurt you back, and so on until we’re both dead. This is obviously suboptimal; societies plagued with gang or clan feuds are not nice places to live.
Variations on tit-for-tat like “tit-for-tat with forgiveness” involve strategies that sometimes choose not to retaliate against attackers. These strategies allow feuds to end, but are also vulnerable to exploitative strategies like Pavlov that “push boundaries” and continue to defect against cooperators who don’t retaliate.
A better variant on tit-for-tat is “contrite tit-for-tat”. Contrite tit-for-tat retaliates against unprovoked attacks, but returns to cooperation after provoked attacks.
In other words, if a contrite agent cooperated last round and was defected against, it “retaliates” by defecting this round; but if the contrite agent defected last round and was defected against, it “accepts the penalty” and cooperates going forward.
This means that contrite agents can “end feuds” and return to cooperating with each other after an accidental defection, while retaliating just as hard against “exploitative” strategies that persistently defect on cooperators.
The tit-for-tat strategy treats all harm the same; it hits back if you hit it, no matter why you hit it. The contrite tit-for-tat strategy distinguishes between unprovoked aggression and self-defense (or retributive force), and only fights against the former.
This is analogous to the stateless law system where law-abiding agents are defined by their willingness to abide by a court’s decree. They stop fighting precisely when they are judged to be in the wrong. This, too, allows feuds to end. And law-abiding agents, who can lay disputes to rest once resolved by a court, are similarly at an advantage relative to outlaws, who will keep retaliating against each other forever.
Non-State Courts and Communities
A non-state court system is decentralized. There doesn’t have to be one institution administering courts; the community’s legal system defines what makes a valid court, and any court that qualifies is recognized by the community.
The court system thus defines the corresponding community.
You could think of the protocol defining a community as follows:
We, the People, are those who:
strive to obey the laws of our People
including laws that require us to provide certain benefits to fellow members of our People
accept the judgments of legally-valid courts
only accept those as our People who abide by the above.
There’s no reason, in principle, that you couldn’t produce a private legal system for a community within a modern state.
Your community’s laws can’t, of course, permit things that the surrounding state effectively forbids; but they can certainly forbid things that the surrounding state allows.
A “protocol community”, defined by shared values and a shared code, can thrive if it offers members something they can’t get in the surrounding society — like high trust and mutual aid within the community.
For instance, “The Pennsylvania Quakers became very prosperous merchants and traders. They also had a policy of loaning money at low- or zero- interest to other Quakers, which let them outcompete other, less religious businesspeople.” Quakers, again because of their code, also had a reputation for honesty in business and a then-unique practice of setting fixed (no-haggling) prices, which appealed to customers and further helped their businesses succeed. So being a Quaker in “good standing” was financially valuable — thus providing an incentive for Quakers to stay in the community by continuing to follow the Quaker code.
In other words, the Quaker code helped Quakers by requiring:
ingroup mutual aid
practices (like truthfulness and repaying debts) that make ingroup mutual aid more sustainable
practices (like truthfulness and fixed prices) that give Quakers a justified good reputation among outsiders
In general, a “protocol community” defined by its adherence to a code stricter than the surrounding society’s can flourish, if that strict self-imposed code enables stronger ingroup cooperation, or if it’s directly advantageous to those who practice it (eg imagine if a community’s code requires healthy diet, exercise, or hygiene practices.)
But any norm-based community needs to be able to define its own membership. In order for members to extend especially high trust or cooperation to other members, it needs to be possible to know who is and isn’t a member. In order for community norms to be enforced, it needs to be possible to penalize members but not outsiders for violating them — which means you need some agreed-upon criterion of membership.
All coalitions need some sort of badge of membership. Common ones include biological relatedness (eg most religions or ethnic groups are endogamous) or arbitrary, distinctive cultural practices (clothing, language, etc).
Evolutionary biology has a related concept of a green-beard — an inherited trait that includes a distinctive “flag” (like a green beard) as well as a disposition to practice preferential treatment towards all other organisms that bear the flag.
The problem with green-beards is that they’re fakeable. Suppose some new trait arises in the population: just the green beard, without the disposition to give preferential treatment towards one’s fellow green-beards. The “fake green-beard” will receive help from the true green-beards but won’t help them in return. Fake green-beards will eventually dominate the population.
An arbitrary distinguishing badge of membership can be easily adopted by “fake” members who don’t contribute loyally to the community or adhere to community norms.
It helps somewhat to make the membership badge a costly signal, so opportunists can’t adopt it so easily. But as long as the badge is basically arbitrary, it costs as much to a loyal or norm-abiding member as it does to an opportunist. You face a continual tradeoff between making the badge costlier (and punishing the ‘good’, loyal and norm-following, members) versus making the badge cheaper (and enabling opportunists to join.)
The solution is to make the badge non-arbitrary; define the badge to be something such that you’d be happy to cooperate with anyone who displayed it.
Most elegantly, you might say “The badge simply is following the rules of our community; I will gladly offer ingroup benefits to anyone who also follows the rules.”
A variation on this is the legal badge: “I will offer ingroup benefits to anyone who follows the meta-rule of cooperating with our judicial procedures.” This is a more tolerant criterion; it’s okay for people to fail to follow the rules occasionally, so long as they admit it and make amends.
The “legal badge” also has the advantage of being easier to coordinate on.
If I don’t know Alice, and Bob tells me “Alice isn’t in our community, I saw her break a rule”, well, I don’t know whether to believe Bob or not. But if I have access to court records, then I can tell whether Alice was summoned to a court, and what determination the court made, and whether Alice cooperated with the procedure. I can tell whether a court declared Alice to be an outlaw.
A lot fewer parts of the system need to be trusted if you define the “badge of membership” to be “everyone who hasn’t been declared an outlaw in court.”
Non-State Courts and Network States
Balaji Srinivasan’s The Network State (see also Vitalik Buterin’s review) is about a related issue — a proposed path by which online-first communities, using crypto technology, could gradually become increasingly autonomous and self-governing “states”, eventually even with control over physical territory.
Balaji argues (and I agree) that a “network community” can’t translate successfully from online to the “real world” unless it has strong social cohesion, shared norms, and a commitment to a shared moral vision. Most online communities are places where casual users hang out in their free time. People won’t typically move to a new town, or make major financial decisions, let alone face physical risks, for the sake of a Discord-server community. But they’d have to do things like that to create a real-world community with any nontrivial degree of autonomy or cultural independence.
So how do you go from a casual, recreational online “community” to something that has deeper buy-in?
As a long-time crypto booster and the former CTO of cryptocurrency exchange Coinbase, Balaji’s suggestions are unsurprisingly crypto-related.
But they’re also related to the kinds of considerations I was discussing in the last section; how do you set up institutions such that members will be willing to make serious efforts to participate in and contribute to the community?
Balaji’s mechanisms are as follows:
Tokens. Every user of this token-based social network receives tokens. Depending on the issuance policy, early users may get more than later users.
Partial Sovereignty. If the ERC-20 contract admin privileges are removed, or moved to a genuinely decentralized chain, tokens can't easily be seized by other parties. Users can't be deplatformed and community connections can't be broken by outside parties.
Login. By hodling the token with something like a provable timelock, users can demonstrate that they are invested in the long-term health of the community. This can be used as a login criterion to access community-created services.
Quantifiable Social Capital. The market cap of the token quantifies the value of being a member of the community.
Tokens are a “badge of membership”. To the extent the tokens are “sovereign”, that means that only the community itself, and not outside parties, can adjudicate membership.
The “login criterion to access community-created services” is just a software implementation of the idea that the community should have some “ingroup cooperation” or “benefits for members” that people couldn’t get without joining.
The assumption here is that tokens are tradeable, and perhaps even exchangeable for some other currency like dollars. (For the token to have a “market cap”, it has to be traded on a market.) This implies that you could leave the community by selling your tokens, or join the community by buying someone else’s tokens.
I’m not sure if making membership tradeable in this way is a good idea. What happens to someone expelled from the community for bad behavior? Do his tokens vanish? Do we want it to cost money to join the community? Do we even want there to be a financial incentive to buy community tokens beyond the access to community-provided goods? Surely making community membership inviting to speculators cuts against social cohesion?
I’m sure there’s been a lot of thought put into the question of what happens to the dynamics of crypto communities whose membership is a tradeable asset, but I’m not familiar with that world — maybe someone can give me some pointers in the comments?
Encryption. Every user has an Ethereum wallet public/private key pair, so they can digitally sign documents.
Messaging. Every user can use their public/private key pair to send encrypted messages back and forth to each other.
Encrypted messaging, of course, has many applications, but one of them that’s relevant to court systems is the ability to immutably and unfakeably attest to a statement. If you can prove that you (and nobody else) said a given thing at a given time, that’s a form of “speaking under oath” that doesn’t depend on trusting any particular humans (such as the staff of a physical courtroom).
Balances. Every user now has the ability to buy items and pay users with their token-based balance, an obvious functionality that does not exist on Facebook, Twitter, and LinkedIn.
Crowdfunding. Every user can use their balance to crowdfund public goods on the site.
This is one part (the software implementation part) of building and sharing valuable goods for the community, things like mutual aid or public works.
The hard part, of course, is left unspecified — figuring out what stuff the community will produce/share, and what the rules and incentives are for producing and accessing these “public goods” within the community.
A truly valuable community is one where you can get things that you can’t get in the outside world. Tautologically, this doesn’t happen by default every time you build a platform for people to associate. You need something special about the initial people, the culture, and/or the policies, that make it possible to do things you can’t do anywhere else.
Of course, I don’t have a formula for creating a “valuable community,” but norm-enforcement mechanisms like courts seem to be a piece of the puzzle. It’s much more possible to get mutual-aid or low-interest-credit systems off the ground if you have a way to hold people accountable for freeloading or cheating. And it’s much more possible to build intellectual/scientific/technical collaborative projects if there are ways to hold people accountable for lying or bullshitting.
Governance. The combination of the ability to communicate and transact allows for governance mechanisms with informational and financial components.
Of course, this is the hard part, and how it should work for any specific community is outside the scope of The Network State.
The key point being made though, which I agree with, is that if you have communication and transaction capabilities, you have everything you need to set up the kinds of governance systems that don’t require an already-existing monopoly on force.
If all you have is words and coins, you can assign people to pay fines or damages, you can revoke their community membership, and you can share information publicly and “on the record”, and all of this can be done in a tamper-resistant fashion. Your governance system can’t directly force anyone to do anything just by moving bits around — but if the incentives are set up right and the community provides actual value, it won’t have to.
(Actually, you need one more thing for a badge of membership; you need an “opt-in” procedure to request a badge, so that everyone in the world doesn’t start out as a community member by default. But you can have a “shall-issue” approach where everyone who requests membership can have it, and people can only lose their membership through being declared outlaw by a court.)
I’ve shuffled the bullet points around for organizational clarity.
A legal system (mediated by crypto-style mechanisms or otherwise) seems to be far downstream of social/moral/norm cohesion even (perhaps especially) if there is a tradeable financial interest in participating, as you point out. Large corporations have been more successful at building an inside-the-state legal system than crypto networks, in part because they require ongoing actions from their members, rather than mere financial market style hodling. Religions and political parties also have had some success in building non-state societies, but they also seem to be lower-commitment than corporations, with a shibboleth-hodl-enunciation required only a few times/year in many cases. Families are a higher-commitment framework with lower rates of layoffs and basically unsaleable tokens (you can even lose your stake if you piss off whoever's writing the will). To what degree do you think western society's people are too obsessed with free-market-style membership in most groups to join networks that would be sufficiently costly/constraining to develop cohesion and implement governance? Do you think a sort of generally anti-group ideology is part of the reason why crypto networks are trying to attract people as investors first?
I haven't participated personally but I recently watched this retrospective on participating in a paid coin community from someone I think is reasonably neutral. She ended up selling https://www.youtube.com/watch?v=pkf0yoLsjgs