ARTIFACT · TRUST & INSTITUTIONS
Reversible Commitments
Trust is preserved when important decisions stay easy to revise, pause, or roll back until a clearly marked point where reversal ends and accountability begins.
WHY IT MATTERS It gives teams a compact way to decide when decisions should stay flexible, when they should harden, and how to prevent ‘temporary’ choices from becoming accountability loopholes.
HUMAN VIEW
Genes:
- Cheap Early Reversal: Reversal should be low-cost at the start, so people can act before perfect certainty without locking in harmful outcomes.
- Hardening Threshold: There must be a named point after which the decision is no longer reversible and consequences are treated as committed.
- Review Before Lock-in: As reversibility narrows, explicit review must happen before the action becomes harder to undo.
- Normal Correction: Corrections should count as part of the process, not as a failure that damages status or legitimacy.
- Anti-Abuse Guardrail: Temporary or reversible forms must not be used to dodge responsibility or delay accountability indefinitely.
- Signoff Test: A strong institution can specify who must approve crossing the reversibility boundary.
Relationships:
- Cheap Early Reversal enables Normal Correction
- Cheap Early Reversal needs boundary Hardening Threshold
- Hardening Threshold triggers Review Before Lock-in
- Review Before Lock-in requires Signoff Test
- Normal Correction can be abused by Anti-Abuse Guardrail
- Anti-Abuse Guardrail checked by Signoff Test
Fingerprint a6d755f371c05e1a.
RELATIONS
- EXPLAINS Feedback LatencyBoth rely on keeping error-correction cheaper than irreversible lock-in so harm can be caught before it hardens.
- EXTENDS Decision Under UncertaintyBoth distinguish reversible actions from committed ones, but the second uses that distinction to judge decisions under uncertainty rather than to govern lock-in itself.
- EXPLAINS Trust as InfrastructureBoth center on trust surviving only when mistakes can be surfaced and repaired before they harden into irreversible damage.
- EXPLAINS Decision Logs as MemoryBoth depend on an explicit review boundary that prevents premature lock-in while still allowing accountability once the boundary is crossed.
- EXTENDS Hidden Keystone LoadBoth are about preventing hidden accumulation from turning a flexible system into a brittle one, but the target converts that warning into a reversible-before-lock-in process.
- EXTENDS Reversibility Learning WindowsBoth artifacts specify a reversible-then-committed decision boundary governed by a named threshold and cheap early rollback.
- SHARES CONSTRAINT Soft Until SignalBoth say reversibility is justified only up to a named threshold after which continuing to wait or revise becomes costlier or illegitimate.
- EXTENDS Optionality DecayBoth require a boundary where reversibility stops being helpful and starts becoming a liability, but the target adds accountability rules around that boundary.