The Projection-Limited Cognition Testbed (PLCT) is a structured, phase-gated research program investigating representation-theoretic limits on measurement and dynamic stability in self-referential systems. The program executed six experimental phases (Phases 1, 2, 2.x, 3, 4, 5, 6) and two amendment tracks (A-P6-1, A-P6-2) under a strict contract-first methodology, fixed-seed determinism (20260425), and no post-hoc tuning discipline. Phase 6 established that dynamic self-referential measurement stability is achievable for specific configurations (2 of 6 Φ × CPM pairs). Amendment track A-P6-1 introduced a tetrahedral state and fractional-memory operator, demonstrating that stability does not automatically lift under geometric reformulation. Amendment track A-P6-2 subsequently certified that 4 of 6 configurations satisfy point-wise admissibility under explicit closure, non-degeneracy, and boundary-stress conditions. The cross-phase synthesis establishes that:stability requires both admissibility and temporal alignment under iteration. Admissibility is a necessary but not sufficient condition for dynamic stability. This publication is a CJCI-based technical report re-published on Zenodo for archival, citation, and reproducibility purposes. All findings are configuration-bound and explicitly constrained. Non-claims:– No claim of consciousness is made.– No general solution to calibration is provided.– No generalization beyond tested configurations is asserted. This work provides a measurement and evaluation framework, not a deployable system or architecture.
Ivan Silva (2026) studied this question.