belief-consistency-spans-structural-and-semantic-dimensions
IN derived (depth 5)
The system maintains belief consistency through two complementary mechanisms targeting different quality dimensions: deduplication resolves structural redundancy (merging equivalent beliefs while preserving topology with user-editable auditable plans), while contradiction management resolves semantic inconsistency (traceable dependency-directed backtracking with consistent nogood IDs and minimal-disruption culprit selection).
Summary
Consistency in the belief network is maintained along two distinct axes. Structural cleanup (deduplication) merges redundant nodes while keeping the dependency graph intact and giving users an auditable plan, whereas semantic cleanup (contradiction management) resolves actual logical conflicts through dependency-directed backtracking with minimal disruption. Together these ensure the network stays both compact and logically coherent.
Justifications
SL — Structural dedup and semantic contradiction resolution are complementary consistency mechanisms
Antecedents (all must be IN):
- dedup-is-topology-preserving-and-auditable — Deduplication preserves network topology (rewrites both antecedent and outlist references to survivors), selects structurally-optimal survivors (most dependents with lexicographic tiebreak), and supports human oversight (KEEP/RETRACT markers in a user-editable plan format).
- contradiction-management-is-complete-and-traceable — The system provides complete contradiction management: the revision pipeline reliably resolves contradictions through outlist defeat and dependency-directed backtracking with guaranteed termination, while nogood resolution maintains a consistent referenceable history of all detected contradictions — enabling both automated resolution and post-hoc forensic analysis of belief conflicts.