The Mechanics of Constitutional Enforcement

Governed systems frequently represent policy as runtime data evaluated by predicates: counters, flags, permission checks, validation functions, and conditional branches. This paper argues that the placement of governance semantics is itself a security and identity decision. A requirement represented as data depends on the completeness and integrity of every check that interprets it. The same requirement represented as state-transition topology constrains which executions are validly expressible. When that topology is canonically serialized and cryptographically bound into a constitution's identity, a canonical change produces a different identity except with negligible collision probability; custody verification can detect substitution against an expected identity, while a semantic court must still decide whether the change is a weakening. The argument arises from a bounded repair in which a requirement for seven ordered, unique stages was initially represented by a counter. Seven repetitions of the same stage satisfied the quantity while violating the intended order and uniqueness. The repair added no new constructs: the same guards, effects, invariants, and kernel remained. The requirement was instead encoded as a state chain. The change did not enlarge the vocabulary; it relocated the semantics from data into topology. From this observation, the paper develops three propositions. First, the Semantic Placement Principle: data-borne semantics are protected by the completeness of checks, while topology-borne semantics are protected by the closure of valid transitions. Second, the Identity-Bearing Topology Principle: when semantically relevant enforcement topology lies inside a declared canonical identity boundary, changing that topology changes the governed artifact's identity under the stated hash assumptions; a separate semantic judgment determines whether the change weakens the rule. Third, the Ratchet Governance Conjecture: a substantial class of governance requirements may be compilable into finite, monotone, role-gated, threshold-bounded transition systems. Structural semantics do not eliminate the problem of a defective or malicious executor. A nonconforming executor may abandon the automaton and impersonate its results. The paper therefore distinguishes structural adequacy, executor conformance, custody fidelity, and effect integrity. It argues that consequential effects should be mediated by independent authority brokers, external witnesses, transactional isolation, and absence-sensitive monitoring. The irreducible limit is not that malicious execution cannot be countered, but that no component can prove its own faithful behavior solely through evidence it alone controls. The result is a theory of constitutional systems in which the kernel need not contain every governance concept. Its task may instead be to provide the smallest trusted mechanics into which governance can be compiled without losing identity, auditability, bounded refusal, or externally observable custody.