Research
The Fragmentation Was Always the Workaround. δS_closed[Φ, Θ] = 0 is a hypothesis about coupled behaviour, not a certificate of this binary.
If a coupling is doing real work, removing it should change observable behaviour. The Experiments body is the instrument. The Claim audit is the lab notebook.
Open questions we will not close in prose: Does the distributed system exhibit measurable behaviour consistent with a coupled closed-system variational structure? How much of Ω must be remembered for another body to identify the same organism? What evidence must exist for a mobile verifier to accept Ωt+1?
ZK machinery in the source repository binds residual, threshold, and block hash. It does not yet prove the canonical state transition. Inference attestation is an Ed25519 receipt, not a zkML proof. Those sentences are copied from LIMITATIONS.md on purpose. The carried-chain statement is a byte record. It is not a proof, and adoption does not require one.