Active Alignment Dominance in Three-Fibre Coherence Recovery

Within a three-fibre toy architecture modelling the leakage-alignment structure of the Q5 transport framework, active re-alignment after phase disturbance produces substantially higher coherence recovery than passive evolution under the unmodified generator. The toy-model program is introduced as a structural stress test of the leakage-alignment architecture, not as a claim of physical realization. The three-fibre space \( span{u_2, u_3, u_4} \) implements the Q5 leakage sector structure (T124, T127): outer fibres u_2 and u_4 carry coherent transport and are coupled only through the middle fibre u_3, which acts as a mandatory balancing channel. Phase disturbances are applied to the first outer fibre via \( ψ_φ = (1/√3)(e^{iφ}, 1, 1) \) , with recovery quality measured by fidelity \( F(t) = |⟨ψ_align, ψ(t)⟩|² \). Key numerical results (tested sweep): Moderate disturbance (φ = π/2): passive recovery F_max ≈ 0.623; active re-alignment F_max ≈ 0.916. Middle fibre strongly recruited. Hard inversion (φ = π): passive recovery F_max ≈ 0.180; active re-alignment F_max ≈ 0.731. Passive reconstruction strongly suppressed. Broad state family (60 asymmetric initial states): re-alignment outperformed passive recovery on maximum fidelity in 98.33% of cases at φ = π/2 and 96.67% at φ = π; on mean fidelity in 100% of cases in both regimes. Three corollaries are established: effective irreversibility under restricted return (recovery requires information, not time); information- demanding reversibility (recovery quality scales with protocol information content); and a balancing-to-rupture regime split with crossover near φ_c ~ 0.7π–0.8π, not yet proven universal. The theorem proves that within the tested architecture, coherence behaves as an actively maintained alignment condition rather than a passively self-restoring property. No claim is made about physical laboratory decoherence, quantum error correction, or gravitational stabilization.