A.19.CPM:6 - Bias-Annotation — informative

CPM is a comparison kernel; it does not remove bias by itself, but it prevents the most common bias‑amplifying failure modes (hidden thresholds, hidden tie‑breakers, unknown coercion).

Typical bias risks and mitigations:

• Comparator choice encodes value judgments. Weights, priority orders, thresholds, and “tie‑break” conventions can encode organizational bias. CPM forces these to live in explicit, edition‑pinned ComparatorSpec records or policy records rather than in invisible code or informal reasoning.
• Missing evidence is rarely random. If evidence is systematically missing for certain contexts or groups, naive “unknown → worse” is a bias amplifier. CPM’s tri‑state guard avoids coercion; but teams must still define policy‑bound failure behavior and be explicit when abstention is acceptable.
• Cross-scheme comparisons can embed structural unfairness. A comparison that relies on a semantic relation between two exact F.17 SchemeSenseCell values cites the obtaining F.9 Bridge and its separate bounded-use claim; together they expose the tested correspondence or difference and tolerated loss. A plane-only crossing cites the applicable ReferencePlane relation and policy. Neither branch replaces comparison scope, predicate, comparator, or time.
• Overconfidence via scalarization. Collapsing partial orders into scalars often overstates certainty and hides tradeoffs. CPM makes set‑valued outcomes first‑class, so the human or managerial decision can remain honest about tradeoffs.