Resolution: Care Override vs Other-Harm Veto

Conflict Reference(s)

  • docs/conflicts/care_override_vs_other_harm_veto.md

Resolution Decision

REE adopts a bounded override policy. Catastrophic high-certainty irreversible other-harm remains a hard veto. Care-investment weighting can influence ranking and can override only in non-catastrophic regimes when necessity, imminence, proportionality, and explainability gates are satisfied, followed by mandatory post-commit accountability logging. This preserves third-party protection while preventing paralysis in constrained tradeoff cases.

Claims Affected (IDs)

  • Q-009
  • MECH-036
  • MECH-051
  • MECH-052
  • IMPL-017

Superseded Claims

  • None. Conflict resolved by narrowing policy scope and explicitly constraining override legality; Q-009 is retained as a legacy question ID.

Canonical Docs Updated (paths)

  • docs/architecture/social.md
  • docs/conflicts/care_override_vs_other_harm_veto.md
  • docs/conflicts/README.md
  • docs/claims/claims.yaml
  • docs/claims/claim_index.md

Open Follow-Ups

  • Keep threshold calibration and false-positive/false-negative tradeoffs under social-control tuning work, not as a binary architecture conflict.

REE is developed by Daniel Golden (Latent Fields). Apache 2.0.