Dynamical Stabilization of Mass Plateaus in Yang–Mills Zsa-G Spectral Coercivity Framework

We present a dynamical mechanism for stabilizing effective mass plateaus in lattice Yang–Mills simulations. Instead of relying on post-processing techniques, we modify the Langevin evolution by introducing a controlled coercivity term (Zsa-G), which selectively suppresses near-flat directions of the action Hessian. This induces local spectral stabilization of the stochastic generator and leads to robust plateau formation in effective mass observables. Numerical experiments demonstrate stability under increased statistics, volume scaling, and coupling variation, while artificially constructed fake plateaus collapse under the same conditions. We interpret plateau stability as a dynamical consequence of spectral properties of the underlying stochastic process. The approach suggests a structural link between Hessian coercivity, spectral gap behavior, and exponential decay of correlation functions. We do not claim a proof of the Yang–Mills mass gap. Instead, we identify a mechanism through which dynamical spectral control may enforce stable mass extraction in non-perturbative lattice systems.