MB8 — Extrapolated Volition (gravestone)
Retired matrix bridge. CEV factorizes into ConstitutionalRule → AlignmentTarget; construction is separate from certification. Not a live backup route to correction.
What decision changes?
CEV is an AlignmentTarget special case, not a second live route to correction integrity. Do not treat an appeal to process preservation as an independent certificate.
Gravestone — MB8 retired from the field matrix (Krym Phase 5).
Opening the CEV black box showed that volition, referent, extrapolation, coherence, correction, and preservation largely map onto existing TSA bridges (MB2–MB4, MB3, grounding, successors). What remains of CEV is a constitutional rule parameterizing an AlignmentTarget, not an independent CEV → Alignment bridge.
In Lean (AlignmentConstruction.lean):
ConstitutionalRule+cevConstitution→constitutionalTargetRealizes/TargetRealizable— open construction crux (not axiomatized)CertifiedAsRealizing— separate certification claim
The legacy axiom MB8_cev_process_convergence is kept only for the Chokepoint special-case instance (why a packaged backup arrow is not independent under a shared instrument). It is not in BridgeAssumptions or the live certification derive path.
Where agendas agreed: MIRI CEV process/legitimacy language. Where this project diverges: live path is MB4 + MB4a; legitimacy theater diagnostics tag capture resistance there, not a separate matrix column.
For the old summary of what MB8 used to claim, see git history of this card before Phase 5.
What would count as evidence?
Evidence would include an independently verified account of what the value-update process converges to, and confirmation that preserving the process does not itself get gamed (legitimacy theater).