For most of human history, when someone made a promise, you knew who had made it. They were standing there. Their body was the evidence. If they broke the promise, you knew where to find them, who their family was, what their reputation could absorb. Promise-keeping wasn't a virtue floating free of infrastructure; it was a behavior shaped by the hard fact that the person who made the promise was findable.
This sounds like a small observation, and for most of history it was so obvious it didn't need saying. It became worth saying when bodies stopped being how people met.
I came to this question through several doors at once. One was the long literature on coordination problems — Axelrod, Ostrom, the serious thinkers who took up the question of why cooperation succeeds where it succeeds and fails where it fails. Another was the Israel/Palestine discourse, which I won't pretend I came to neutrally. What I will say is that watching the conversation collapse, in the most personal places, into something that looked like communication but no longer functioned like it, made the structural question impossible to look away from. Whatever else was happening, the practice we called discourse had stopped doing what discourse is supposed to do. People were saying things at each other that no one was being held to. Words were moving without commitments attaching to them.
That experience is what made me want to know what discourse actually requires.
The philosophical answer is older than I expected. Wittgenstein, in the notebooks he wrote at the end of his life, observed that some propositions function differently from ordinary beliefs. When you say I have hands or the world existed before I was born or my name is mine, you aren't reporting the conclusion of an inference. You're not weighing evidence and coming to a verdict. These propositions sit underneath the practice of weighing and verdicting; they're what makes that practice possible. He called them hinges. A door turns because the hinges stay fixed. If you tried to test the hinges by opening the door, you'd have nothing to turn against.
Asking how a hinge is justified is a category mistake. Hinges aren't justified within the system; they're what the justification game turns on.
Robert Brandom, working in a different tradition, made a related observation about discourse. Discursive practice, he argued, is a game of giving and asking for reasons. People assert things; other people are entitled to demand reasons; the asserter is bound by what they said in ways that have consequences. This is what makes assertion different from noise. But the game has a precondition the players don't usually notice: it requires participants who persist long enough to be held to their moves.
Identity, in this frame, is a hinge. Not identity in the rich phenomenological sense — though that's also a hinge — but identity in the more workmanlike sense of being someone in particular, locatable and addressable, expected to answer for what you said. Without that, accountable discourse doesn't fail. It doesn't exist.
Communities have always understood this, and have always built the infrastructure that makes it possible. Bodies and presence do part of the work. Kinship and naming do another part — every culture I know of names its children, and naming is the primal binding act, the assignment of an utterance-source that persists. Beyond that, every durable society has developed scaffolding for what I'll call witnessed commitment: oaths sworn before named witnesses, contracts sealed and registered, signatures, ordination chains, civic registries, notarial systems, the slow accumulation of techniques for binding utterance to a locatable source. Religious traditions built some of the deepest versions of this. Civic and legal traditions built others. The work was always being done, because the work has to be done for discourse to function.
Among the cultural practices that took this most seriously, the Talmudic tradition is worth pausing on. It treats disagreement as a religious activity rather than as a problem to be settled, preserves minority opinions in the canonical text, and grounds authority in continued participation in the argument rather than in any single resolution. Eilu v'eilu divrei Elokim chayim — these and those are the words of the living God — is a remarkable formal concession that contradictory positions can both be authentic. That's not a soft tolerance. It's a structural commitment to keeping the discourse open while still being binding, and it required infrastructure to sustain across two thousand years and radically different host societies. It's one of the longest-running case studies in non-territorial accountable community on earth, and I keep coming back to it as evidence that the work is doable even under difficult conditions.
Digital civic life inherits none of this scaffolding directly.
Bodies don't carry into the network. Civic registries weren't built for it. Inherited communal and religious institutions, whatever their authority elsewhere, are not the shared substrate of pluralistic digital societies. The hinges that physical and institutional worlds provided automatically, often without being noticed, don't reach into the medium where most public life now happens.
What we got instead was the corporate answer. Platforms install proprietary identity layers as a side effect of their business models. The user becomes a credential of the platform. This is identity as plumbing, and it has the predictable result: the binding serves whoever owns the pipe, not the participants in the discourse. You can be deplatformed, shadow-banned, or have your reach throttled, and the binding between you and your words can be modified, revoked, or sold without your consent. Whatever this is, it isn't the civic infrastructure that accountable discourse requires.
The result, slowly and then quickly, is what we have now. Practices that look like discourse but don't function like it. Speech without speakers in any deep sense — speakers without persistence, persistence without accountability, accountability without addressability. Noise that resembles commitment, scaled.
The project I've been working on with Wes Kussmaul and others — Osmio, chartered as a digital municipality at the ITU in Geneva two decades ago — proposes a different answer. The premise is straightforward enough to state and difficult enough to execute: identity infrastructure should be civic infrastructure. The binding of utterance to addressable agent is not a feature to be sold but a precondition for the kind of public life we say we want. It should be chartered the way municipalities are chartered, governed the way commons are governed, grounded the way public utilities are grounded — accountable to participants rather than to shareholders.
The cryptographic stack is one part of the answer. The institutional form is the other, and the harder one. Cryptography without civic chartering is just better plumbing. Civic chartering without cryptographic teeth is just declaration. Both, together, constitute the hinge that digital discourse needs and currently lacks.
I find I keep coming back to the structural observation rather than to the project, even though the project is what I'm spending my time on. The observation is durable in a way the project isn't yet. The hinges of accountable discourse used to be installed for free — by embodiment, by named community, by the slow craft of civilizations that knew their words had to bind to people. They aren't installed for free anymore, and they aren't going to install themselves. Whoever builds them, builds them. If the people who care about civic life don't, the people who care about other things will, and we'll get the discourse those other things produce.
That, at the bottom of it, is why I'm spending my time on this. The work, if there is work, is to install the hinges. The rest follows from there.