In The Utopia of Rules, Graeber repeatedly argues that bureaucracy is not neutral coordination but a form of structural violence that depends on legibility, rules, and abstraction. Bureaucracy operates by reducing people to categories, allowing power to appear as neutral procedure rather than coercion.

In Debt, he makes a related argument: turning social obligation into numerical debt enables violence by creating equivalence where none morally exists. Measurement, this way, is not innocent. It allows force to masquerade as fairness. Formal abstraction, for Graeber, risks inviting power.

Graeber saw measurement as the tool of the creditor. I’m interested in whether it can also function as armor for the debtor, because without some way to make the system’s load legible, claims that a debt is unpayable are too easily dismissed as moral or subjective.

Graeber’s resistance to formalization was, at heart, a resistance to the administrative gaze. Where I diverge is not on the danger, but on the consequences of refusal. In my opinion, the absence of formal language does not prevent domination; it merely ensures that domination is analyzed only in moral or narrative terms.

Because contemporary power operates through models, risk frameworks, and quantified incentives, refusing formalism cedes explanatory authority to those systems by default. To me, math becomes not a tool of governance, but a means of making governance accountable.

We can also build in safeties.

♥︎ In state governance, the model is “truth.” If the model says you are a “risk,” you are treated as a risk. The Agency-Expansion Loop is treated as a hypothesis. If a community uses the framework and it doesn’t result in people feeling safer, the model is considered broken, not the people. The power stays with the “felt sense” of the participants.

♥︎ The state’s models (like credit scores or risk assessments) are “black boxes.” You don’t know how they work, so you can’t argue with them. By contrast, this framework is intended to remain transparent, revisable, and subordinate to human judgment. Its role is not to replace judgment, but to support it — to visualize load, identify failure modes, and justify human-centered resources in environments that otherwise recognize only quantifiable outcomes.

♥︎ In bureaucratic systems, models replace judgment. In the Agency-Expansion Loop, models serve judgment. Formalization here is not an attempt to manage people, but to demonstrate when systems themselves are becoming unsafe: when hysteresis is occurring, when threat is compounding, and when continued intervention would cease to be voluntary. The work of repair remains human, relational, and empathetic. The math will exist to protect that work, not to supersede it.

And finally, formalization allows for de-escalation without the need for shaming.

This is still unfinished, and I expect it to evolve with learning, use, and critique. I’m sharing these as my intentions at this stage.