The semantic gap is the set of all known axiom violations in the predicate graph. It is the semiotic interpretation of a backlog.
Unlike a conventional backlog (a list curated by a product owner), the semantic gap is derived, not curated. The predicate graph’s satisfaction checker computes it: every MUST error is a gap item at high priority, every SHOULD warning is a gap item at lower priority. Each item names a specific predicate that a specific page needs — a concrete, actionable unit of work.
Agents select from the semantic gap based on leverage (policy 009) and appetite, not on managerial assignment.