Regime-Dependent Stability in Trust-Mediated Governance Networks: Disturbance Primacy, Control Masking, and Governance Insensitivity

We investigate stability dynamics in a trust-mediated governance network model across single-node and multi-node configurations. The model tracks four state variables — Trust T(t), Disturbance D(t), Cognitive Distortion C(t), and Stability S(t) — evolving under discrete-time dynamics with continuous control intervention. Through systematic falsification testing, stability envelope mapping, regime analysis, and multi-node coupling experiments, we report three principal findings. First, high-control regimes constitute a non-identifiability zone for structural inference: control dominance suppresses internal variable dependencies, rendering standard stability testing uninformative about model structure. Second, disturbance is the primary binding variable under constrained control, with trust degradation as the mediating failure pathway and cognitive distortion as a secondary amplifier. Third, a governance-insensitivity phenomenon is observed across a 3x capacity range (0.30 to 0.90): stable run count remains zero and average final trust remains flat at 0.149 across all four bracket conditions, while cooperation increases by only 0.055 total. We interpret this as evidence that governance capacity is insufficient to override disturbance-driven dynamics when the trust floor is critically degraded, and identify it as a structural research frontier requiring investigation of the trust-governance interaction pathway. Multi-node extension reveals that stability propagation is conditional on recovery coupling exceeding an emergent threshold k_T*, while failure channels are structurally always active within the tested linear additive coupling class.