Offense and Enforcement
Every state keeps a sharp edge, and every constitution is finally judged by what happens at it. Axiacracy narrows that edge to a per-act license, binds it to designated harm, and puts the officer who wields it on the same ledger as everyone else.
It is two in the morning, and an officer stands at a door. In the states we know, she stands there as the bearer of a general power: a standing authority to detain, to search, to compel, granted once at her swearing-in and bounded afterward mostly by her judgment and her department's culture. In the value state she stands there with something narrower and stranger: a license computed for this act, against this person, at this hour, from three published factors, carrying the judicial authorization that issued it, and expiring the moment the act ends. Every step she takes is logged under her own name. If she uses more force than the least that would suffice, the excess is booked against her, not absorbed by her uniform. And if she had been ordered to breach the one floor no order can reach, her duty, personal and non-delegable, would have been to refuse. This chapter is about that door: who may stand at it, with what in hand, and under whose watch.
The sword, honoured and bounded
Hobbes's enforcement lemma is the oldest hard truth in political science: covenants without the sword are but words. A polity that measures harm exquisitely and then cannot act on the measurement is not a gentle state but a failed one, and the design takes the lemma with full seriousness. Under-enforcement is priced as heavily as over-reach: an officer who looks away, tips off, or under-enforces for a favored offender commits aggravated anti-value, with no shelter in good faith or a superior's order. A powerful offender must not be able to buy out a rights violation, launder guilt through a dissolving shell, run out the clock by obstruction, or push guilt onto a hired scapegoat; the enforcement architecture is written, clause by clause, to close each of those exits. The sword is real.
But it is bounded in a way Hobbes never imagined. Force here is not a standing power held by any organ; it is a per-act capability, licensed act by act. The ceiling on any coercive measure is the output of a single published function of three factors: the confidence of the evidence, the objectivity of the harm across value frames, and the severity of the floor breach at stake. The function is uniform, versioned, and computed with no per-case discretion, and until it has been ratified and published, the system fails closed: no hard measure issues at all. Nor is there a single sword-bearer. Enforcement is polycentric, plural organs under distinct mandates, mutually watching, with no apex agency, and the courts, which sit structurally outside the enforcement organs, resolve their conflicts. Hobbes himself supplied the boundary in spite of himself: his list of what is "not Punishment, but Hostility", force without authority, without antecedent condemnation, after the fact, beyond the prescribed penalty, is adopted whole and turned on the ruler. A coercive act outside its license is not strong enforcement. It is hostility, void, and chargeable.
Designated harm, not improvised morality
What the officer at the door enforces is a designation, never her own moral sense. The constitutional rule from earlier chapters, coercion only on democratically designated harm with countable effect, becomes, at the street level, a rule about what an enforcement organ may even open a case on. Every offense must be defined in advance, with enough precision that a subject could ascertain what conduct it covers; a vague offense is refused at intake, before any measure issues. An intra-frame moral objection, distaste for a person's speech, belief, or way of life, may never be relabeled as objective cross-frame harm; the burden of proving that a harm reaches across frames lies on the asserting organ, attested by parties not aligned with it. The state that annotates rather than censors in the epistemic commons keeps the same character at its sharpest edge: its officers are executors of published designations, not missionaries of order. The constable is not a moralist, because the constitution has removed morality-at-discretion from the set of things a constable can wield.
This is the deep difference between a value state and a virtue state. Both care about how life goes; only one authorizes its agents to enforce their reading of it. In Axiacracy the reading belongs to the polity, through the weighting vote and the anti-value designations, and the agent inherits a closed list. Everything off the list is, for the enforcer, simply none of the state's business, however loudly a metric or a minister might wish otherwise.
Policing a society with a floor
What does the docket look like in a polity with a dividend, a real-capability floor, and captured rent? Book I argued that a large share of what police forces actually process is engineered desperation: theft for subsistence, the survival economies of people the formal economy has discarded, the cascading offenses of debt and homelessness. Where the floor is real, that share of the docket thins, not because the state grew lenient but because the condition that produced the conduct was corrected upstream, which is the design's first instrument everywhere. The doctrine is careful not to overclaim: no floor abolishes malice, appetite, or cruelty. What remains is what desperation never explained, predation, violence, fraud, concealment, and the characteristic offenses of the powerful: corrupting the measurement, hiding the flow, buying the referee. Hayek's observation that deception is coercion's twin, the manipulation of the data on which a person counts, does heavy work here: in a state whose governance runs on sensed value, feeding the sensor false data or corrupting the ledger is not a paperwork offense but an attack on the polity's nervous system, coercible on the same footing as force. The center of gravity of enforcement shifts, in short, from the desperate to the deceitful. A police force designed for that docket looks less like an army of the streets and more like a guild of auditors with a narrow, licensed capacity for physical intervention, and the design means it to.
Investigation without fishing
The most dangerous power the polity holds is not the truncheon but the read: the ability to look into a person's affairs before any court has found anything. Here the constitutional character of the sensor governs. The sensor is aggregate and statistical, never a register of individuals' private lives; the state's lawful window onto persons-in-general is cohort-shaped, floored at sizes that identify no one. From that aggregate vantage, no investigation may be conjured against a name. A profile, a cohort statistic, an anomaly score may never direct an inquiry at a particular person; person-directed investigation requires an independent, articulable predicate, logged before the first question is asked, and even then its cumulative duration and intensity against any one subject are capped. There are no fishing expeditions, because the license to fish is structurally absent.
When a case is properly opened, every read of evidence is itself a lawful instrument: a judicial access contract, issued per subject and per scope on a finding of necessity, with each act logged, timestamped, and attributed to the organ and officer who performed it. An unlogged investigative act is void as evidence and books anti-value against the organ that performed it: the state cannot profit from its own dark reads. Intrusion is laddered from open sources through voluntary interview to compelled production and beyond, each rung requiring a predicate no weaker than the last, with prior judicial authorization from the search-analog rung upward; and from the first coercive contact, not from the charge, the subject has a funded, independent advocate present. One more wall completes the room: statistical attribution, market shares, exposure fractions, cohort excesses, may ground restitution from an organization, but may never supply the confidence basis for a hard measure against an individual human being. To restrict a person, the state must show that person's own act, individually evidenced and corroborated. The cohort mathematics that runs the rebalancer stops at the individual's door.
Proportion, force, and the officer's own account
Where force is finally used, three disciplines bind it. Proportionality: every coercive measure must be the least restrictive means that achieves its restorative or preventive object, and proportionate to the severity at stake; any excess beyond that finding is over-coercion, aggravated anti-value against organ and officer alike, even when the measure stayed inside the officer's formal authority. Exigency: prior judicial authorization is the rule, and the only exception is an enumerated imminent threat, registered by an organ independent of the one invoking it, ratified by a court within a fixed short window or lapsing into chargeable wrong; no organ certifies its own emergency, and even under a surge of genuine emergencies the requirement escalates in witnesses rather than disappearing. And the absolute limit: the dignity floor, life, bodily integrity, subsistence, contact, freedom from degrading treatment, sits outside the force calculus entirely. No confidence level, no severity, no license reaches it. An act that touches it is void, excluded, and charged.
Behind these rules stands a distinctive theory of the officer. There is no qualified immunity: the person who wields the state's force answers on the same ledger as the citizen, with only a narrow good-faith defense for genuinely reasonable factual error, and none at all for excess, conflict, benefit, or dereliction. A superior's order is no defense; each officer carries a personal, non-delegable duty to refuse an unlawful one, and obeying it is her own anti-value, not her commander's alone. And the accountability is symmetric: every coercive act and every declination to act is logged under a name. The design refuses the two ways enforcement traditionally escapes judgment, the uniform that absorbs the individual's excess, and the shrug that hides the individual's inaction. At the sharp edge, there are only named people, doing licensed acts, on the record.
Guarding the instruments, and the feared drift
All of this presumes that the instruments themselves, the registry, the ledger, the sensor, are sound, and the design does not presume it politely. The security architecture in front of them is default-deny: nothing in the polity's records is readable except through an authorized contract, every projection is shaped and minimal, and, decisively, the state logs its own reads, even inside its own namespaces. An organ that controls a register is never exempt from per-person read logging in it; the watchers' looking is itself watched, structurally rather than by policy. The keys, the audit anchors, the compromise procedures are specified with the same severity as the rights they protect, because a guarantee resting on an unattested substrate is a hope. And the whole apparatus is priced against both failure modes at once: the surveillant state that reads too much, and the stonewalling power that hides behind privacy to obstruct discovery. A security architecture that guards only one flank, the doctrine observes, is an accomplice of the other.
The enforcement failure this design fears most is not the cinematic one, the coup, the massacre, but a quieter drift: enforcement sliding from designated harm to convenient order. Every police power in history has felt the pull, from offenses against persons toward offenses against tidiness, from protecting the citizen toward managing him, until the criminal law becomes an inventory of the ruler's irritations. The counters are structural, and by now the reader can name them: the vagueness bar that voids any elastic offense of "conduct against systemic balance"; the low-legitimacy flag and higher confidence bar on any coercion with no named victim; the anti-revenue rule, under which no enforcing organ keeps or directs the proceeds of the penalties it seeks, so the machinery has no budgetary appetite for guilt; the absence of an apex organ that could be captured whole; the courts outside the enforcement chain; and the log that no enforcer can suppress, in which the drift, if it comes, is itself a countable, chargeable pattern. The sharp edge stays tolerable for the same reason it stays sharp: because it is narrow, and because everything it does is written down where the polity, and the next chapter's failure analysis, can read it.
In the doctrine
The full enforcement specification, harm by kind, the coercion threshold, and the officer's account, is developed in Rights, Crime and Justice; the aggregate-only sensor is set out in Measurement, the drift scenarios in Failure Modes, and the sword's oldest argument in Hobbes.