A computer-implemented semantic progression layer maintains per-user progression state over a base graph of nodes and typed edges including prerequisite edges. For a user, signals including familiarity and optionally mastery, recency, uncertainty, and goal alignment are computed and superposed onto the base graph to form a per-user overlay. From the overlay, the system derives progression-related constructs including predicate-defined region memberships and/or compact representations usable by downstream components. Region memberships may include a forward region and a backward-looking/skipped region and optionally a lateral/adjacent region, with operation in single-region, two-region, or three-or-more-region modes. Membership is determined by predicate tests under traversal constraints rather than by visualization geometry. In some implementations, constructs are stored as per-user bitsets keyed by node identifiers and are incrementally maintained via localized delta updates in response to user events without full-graph recomputation. Outputs include recommendations, structured guidance, navigation cues, visualizations, assessments, and scheduling.