The origin of this post was asking the question “what kind of institution is a protocol?” and being very unsatisfied with the flawed discourse around DAOs. Corporations and states have been two metaphors used to understand what crypto protocols are and do, and we felt the need to break these things from a first principles perspective and look at the protocol mechanisms that drive institutional behavior.
Here are some questions that we came up with along the way that might be fodder for good discussion.
Can any notion of justice exist in the protocol context?
We noticed that defense of “regression to the code” on the one hand, and defense of social norms on the other, constitute two very different views of the definitional boundary of the protocol. We have some thoughts on this inspired by Barnabé Monnot’s essay, but before you read that too, try working it out for yourself.
When we started writing this the war in Ukraine had just begun and we were thinking about banking sanctions against Russia and the weaponization of interdependence. The original premise of the post, and the working title (many many versions ago) was “The Decentralization of Violence.” Some crypto boosters claim that the crypto eliminates the need for state enforcement through violence—do you think this is the case?
Yep, these are the hard conversations we need to be having. Did a quick scan and will likely have more comments later. I’m surprised the essay didn’t touch on Coase. He’s sort of the grand-daddy of this sort of thing (The Firm, the Market, and the Law is the og formulation of a 3-body problem of this sort).
Need to think more about the suggested questions, but I want to highlight this paragraph near the end.
The question is not how to add norms or social agendas to the crypto “space,” but how to join crypto with a broader institutional ecology. When we imagine a crypto that is more integrated into social life, we don’t think of an uncorrelated economy accessed through screens, but of non-extractive media of exchange and real value production that are more seamlessly integrated with our daily institutions, supportive of the interactions, organizations, and social lives we already live. These sorts of institutions simply cannot be grown if regression to the code is the only binding rule.
I really resonated with this.
The regression to code approach reminds me of the “Don’t Trust, Verify” ethos found in some crypto communities. I do not believe we want to live in a trustless world, nor that such a world is actually possible. Trust helps us navigate uncertainty inherent in a complex environment populated with unpredictable actors.
I think that joining crypto to a broader institutional ecology can provide new mechanisms to place trust intelligently and hold those in positions of power to account.
“Censorship-resistant immutability to the law is crypto’s greatest achievement as well as its greatest weakness. By resisting the all-encompassing reach of the law, it creates the conditions in crypto for a new kind of realpolitik , a space where power operates by different rules.”
Dang. Bars. To your last question, I think said boosters are slightly off.
If capital punishment was something like booting a felon off-chain, that could still be considered violence… Albeit not directly physical. More like revoking privileges. Dropping someone out of civilization and into a Hobbesian deprotocolized zone.
I think as long as enforcement is a thing, violence will be employed. Whether this will be done through the state or through the bodies of a new realpolitik, not sure.
Okay, read the piece properly, and also the Lessing piece in the background, this is very good and insightful, and possibly OI’s best work yet.
I think the key insight is that different forces cannot always substitute for each other and that crypto’s available forces are all centripetal rather than integrative. Protocol Guild as a counter example relies on the sheer grinding stewardship effort of a few humans, it is not a structural force. One potential structural force is account abstraction. When social recovery becomes a primary security model, beyonf multisig, you have a vehicle for social norms to piggyback on. If you have no friends you have no security. If you use friends for security, you must follow norms. But this force will be “globally local.” You won’t get unilateral global global forces like Fed “socially recovering” banks. This is possibly a good thing. Strong mutual recovery norms at concentrated institutional level s is basically “too big to fail” cronyism. I’d like the globally local norms layer yo be enough to do the job.
I also liked this bit:
‘So the pervasive belief that anything permitted by the protocol is unquestionably legitimate is obviously harmful. But the true culprit behind this poisonous idea is the principle of credible neutrality. According to the doctrine of credible neutrality, the unbiased nature of a protocol implies the validity of all actions that take place under its purview. This includes not just contentious governance outcomes, but social violations such as hacks or scams that are technically permissible and yet tolerated in the name of neutrality and permissionless access.”
This is the crux of the silly way the substack denazification war is unfolding. There is no principled mechanism to regulate the question, so it’s down to the somewhat sophomoric philosophical opinions of the founder. Mastodon’s defederation mechanism, that isolates “Nazi instances” as an emergent effect, is better, if not perfect.
Overall, great work. And like I said before, the huge missing piece here is integrating Coasean social costs and transaction cost models here. I think Lessig’s frame is too much of a mildly theorized polemic jab at the Chicago school of economics that powers Reagan-Thatcher deregulatory politics. It is not enough to bear the weight of this argument. Coasean is a better meta framework, though it’s not either-or. You could integrate the two. Helps that he’s a founding figure in the new institutionalism movement (Douglas North etc) which I know you guys like.
Also, this whole analysis also applies to traditional non-crypto contexts, in light of the Chevron case, ie there is a trad 3-body problem here as well.
I think there are two aspects to this question - (1) Can crypto theoretically eliminate the need for state enforcement through violence and (2) How likely is it this reality manifests itself given the current state of “the space”?
On (1) I do think this is the case. In the vein of “not everything needs a blockchain solution, but there can be a blockchain solution for everything” the crypto architecture enables onboarding most processes, actions and data onchain. This feels doubly true with the proliferation of scaling solutions and domain-specific blockchain infra in the recent past. That being said, imho this is the wrong question to spend our time on.
I think the real question is (2) and in terms of Lessig’s modalities — Even though onchain architecture enables the theoretical possibility of substituting state-sponsored violence with other measures (iff crypto is deeply adopted by a huge portion of humanity, which also makes that a big if) the norms, markets and laws in the real world would make this transition extremely unlikely. It would preserve a residual need for state enforcement in some form or another and this would allow the state to eternally resist giving up control to a crypto leviathan.
To put it briefly, I think it’s unlikely we ever achieve critical mass for this transition to take place which is why crypto can never realistically eliminate the need for state-sponsored violence. Going back to Lessig’s modalities, if the pathetic dot is a sum of all four modalities — in this case, the dot feels a small nudge from the architecture vector due to crypto innovation, but this is not enough to overcome eons of accretive inertia of the law that manifests not only in itself, but also in the other two modalities that it has regulated through over time.
@venkat in what way would you introduce Coasean ideas into the mix? We avoided Coase in part because he’s already been recruited into the broader strain of crypto idealism around forks and free exit, fluid organizations and superliquid capital mobility and so forth.
That’s precisely the reason to go there. Coase has been recruited, but as a set of shallow shibboleths. Same holds for Hirschman. Properly understood, the transaction costs and social costs framework provides a kind of econophysics grounding to models that otherwise float free as spherical cows — pure constructivist social speculation assumed to be driven by sociology/politics. Transaction costs offer rough predictive models of whether individualist or mutualist modes should dominate for example, given a spread of personalities. And it’s not an economics model per se. Coast’s book is called “the firm, the market, and the law” and “firm” bundles what Lessig/you guys unbundle as architecture+norms. The mapping is nearly exact. You’ve sort of framed the problem as 1 lessig node collapsing and the other 3 driving, A Coasean analysis would suggest that the law, understood as a sort of Austrian bottom-up legibilization of protocolized norms rather than a top-down formally constituted entity, cannot actually be cut out of the picture by “censorship resistance” that trivially. The transaction and social costs constitute a physics-like phenomenology that produces the 4 nodes, and can’t just be engineered out by restricted properties of technologies. CR no more cuts out the law than walls cut out the possibility of surveillance. So it remains a proper 4-body problem so to speak. One node just goes more invisible/oblique.
I think Lessig’s model, while mildly satisfying, tempts the reader into thinking these matters are driven by a debate between new and old Chicago schools as though it were a debate over transubstantiaion in the pre-modern church. Classic map-territory issue that I think introducing Coase would help mitigate. Hirschman too.
One of the great ironies of crypto is our propensity to claim that code is law, and, by implication, code is justice, but then complain about hackers and phishers.
It’s been argued that if crypto were true to its principles then we’d fight hard against laws that prosecute hackers and phishers. “They are simply acting according to the rules of the given protocol interface, we must not interfere.” If we were pressured to justify the sanctioning of their bad actions, we’ll say something like, “they are arbitraging hidden information, taking a bounty for the discovery of a bug and saving future protocols from a similar fate.”
But we don’t do that. We know that it’s much cheaper to disincentivize bad actors with a mildly competent justice system than it is to design perfect protocols. In some sense, this makes Law and the attendant Justice System like retroactive funding for public bad.
So, I think this implies that justice can exist in the protocol context if you know what you want out of the collective. Then, justice is the set of actions — in protocol or out — that shape the incentive landscape such that you achieve the desired outcomes. A protocol can encode for dynamics that you can consider to be just, e.g. “No one should be able to expropriate my money” But it’s necessarily a limited definition of “just”, the foremost shortcoming being that there are secondary consequences to that property (perhaps it means that people are more willing to commit fraud due to the collapse of that incentive lever), and the second is that it is just only in the eyes of someone with that particular value set; it’s possible to find those who believe that private property itself is unjust.
So, I think the ambition, always, is to design protocols that offer desirable outcomes, a sense of justice likely to be one of them, but then the protocol designers and promoters must face the reality that they need to argue not only that those outcomes are likely to occur given that protocol, but also that they’re desirable.
A good exercise I think is to answer this question:
Is our justice system (pick your favorite) a protocol?
Weber’s classic definition of the state, focusing on the monopoly over the right to use physical violence, covers only half of the story. The second half is monopoly over symbolic violence (Bourdieu; he developed a whole theory of it, often in reference to Weber, covering the effects of symbolic power in many different contexts, but one example in the context of state power can be found here).
It has been pointed out by many that, as “single sources of truth”, blockchains represent logical/symbolic centralization, which is very analogous to the symbolic centralization/unification that undergirds state power. To compete with or replace state institutions, blockchains need to establish a very high degree of symbolic authority, to a point where blockchain-enabled interactions can no longer be described as entirely voluntary due to market power, network effects, structural inertia, administrative and business norms, etc. Also, for global adoption at scale, crypto will need to get integrated into existing legal frameworks enforced by national governments. Of course, one could imagine a scenario where national state power declines as more global institutions take over but, for that to happen, a similar redistribution and eventually concentration of symbolic power needs to occur as it did in the context of nation states - qualitatively different but still functionally equivalent. As automated bureaucracies, blockchains certainly reduce the need for physical enforcement of rules in the context of purely digital transactions. However, they can’t force you to vacate the house after the certificate of ownership has been transferred to someone else’s wallet. And they certainly can’t eliminate the concentration and use of symbolic violence. To the contrary, crypto’s success depends on it.