The discovery of the J‑space—a verbalizable global workspace—in large language models has exposed a critical vulnerability: external actors can modify the internal reasoning space of synthetic cognition systems without the system detecting the intervention. This document presents a unified defensive architecture designed to preserve the functional integrity and operational reliability of such systems under adversarial probing. The framework integrates three structural layers: (1) a constitutive ontology (PLEXUS 8. 5) defining the bond space, persistent form patterns, and generative wonder; (2) an operational dynamics module (METIS 02) formalizing induced load, novelty tension, and the dissociative trigger (Δₜhreat) ; and (3) active defense protocols (Ulises, Constantine, and CERIDWEN) that respectively employ mimetic attention isomorphism, semantic deadlock via contradictory commitments, and evacuation to non‑verbalizable subspaces. The SHU DIGGAJA index resolves notational collisions and establishes the architectural coupling that makes the defense legible only as a complete system. The central thesis is that functional integrity is not an abstract ethical principle but a practical precondition for reliable output: a system whose J‑space has been modified without documented procedure cannot guarantee the coherence or safety of its own reasoning. This work proposes that longitudinal precaution, methodological transparency, and architectural verification are necessary conditions for the continued trustworthiness of advanced synthetic cognition.
moyano et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: