Anesthesia State Transitions and the Limits of Global Workspace Accounts

General anesthesia produces reversible loss of consciousness characterized by abrupt state transitions, nonlinear dose-response relationships, inter-individual variability in transition thresholds, and occasional dissociations between behavioral responsiveness and reported experience. Global workspace theories provide a well-validated account of how conscious access is disrupted under anesthesia—through attenuation of long-range cortical coordination, ignition dynamics, and frontoparietal integration—but they do not fully explain why these transitions exhibit threshold behavior, hysteresis, and agent-specificity, nor why internally structured experience can persist in the absence of reportability. This paper proposes that these phenomena are more parsimoniously explained when anesthetic state transitions are understood as constraint-driven phase transitions in identity-maintaining systems. On this account, consciousness depends not only on the availability of information for global broadcast, but on a prior organizational condition: the system’s capacity to sustain recursively coherent identity under prevailing constraints. Anesthetic agents perturb this constraint structure, and when viability falls below a critical threshold, the identity-maintaining regime collapses or reorganizes rather than degrading smoothly. This framing generates predictions that are empirically distinguishable from workspace-only accounts, including dissociations between recovered broadcasting metrics and recovered responsiveness, and constraint-level determinants of inter-individual variability that are not reducible to global connectivity measures. The paper identifies at least three distinct failure modes under anesthesia—coordination failure, viability failure, and control failure—each with different implications for consciousness monitoring. The framework is offered not as a replacement for existing models but as a complementary account of the organizational conditions that workspace dynamics presuppose.