backtracking-retracts-least-entrenched

IN premise

`add_nogood` resolves contradictions via dependency-directed backtracking: `find_culprits` traces to premises, scores by `_entrenchment`, and retracts the least-entrenched premise to minimize disruption.

Summary

When the system detects a contradiction between beliefs, it doesn't just pick an arbitrary one to throw out. Instead, it traces back through the chain of reasoning to find which foundational assumptions led to the conflict, then removes the one considered least important (lowest entrenchment score), preserving as much of the existing knowledge as possible.

Dependents

These beliefs depend on this one:

Details

Sourceentries/2026/04/23/reasons_lib-network.md