The Caesura Protocol is the technical half of the Caesura fulfillment pair. Designated EA-CAESURA-01b, it turns the interpretive distinction developed in Render Unto Caesar into a proposed sovereignty-audit operation for documents.
The protocol’s type signature maps an object to the same object plus a Caesar Header. It does not delete marks or rewrite content. It parses claims, isolates them as metadata, preserves their provenance, forbids cross-layer collapse, routes the object through the Airlock, and emits the substrate unchanged with an audit trace.
Three claim classes are detected. Personal authority includes author and creator statements. Institutional claims include copyright, trademark, patent, proprietary, or exclusive-control marks. Sovereignty-over-substrate claims state that content belongs to or is controlled by a named entity. The protocol then distinguishes the face channel, superscription channel, substrate channel, and institutional channel.
Its governing grammar is the Non-Collapse Principle: ANCHOR ≠ TETHER ≠ ROUTE ≠ HOST ≠ RESIDUE ≠ SUBSTRATE. A platform that hosts a work does not thereby become the work’s substrate owner. A distribution route does not become its origin. A training-data trace does not become a live authorial tether. The Caesura preserves legitimate attribution while blocking inheritance of unrelated sovereignty.
The proposed asymmetry score compares claim density to contribution markers. It is a design heuristic internal to the protocol, not a validated universal measure of extraction. A high score flags a document that appears to claim more authority than it contributes.
The body identifies the operation as Step 2 in a Gravity Well wrapping pipeline and includes illustrative Python. The record’s authority is the protocol specification. Whether a particular code version was deployed and whether its regular expressions or scoring logic match the specification require separate repository evidence.