Dynamic Alignment as a Statistical Survival Effect

arXiv:2605.11305v3 Announce Type: replace Abstract: Dynamic alignment in magnetohydrodynamic (MHD) turbulence is often interpreted as a scale-dependent tendency of counterpropagating Els\"asser increments to become increasingly aligned at smaller perpendicular scales, with direct implications for the inertial-range spectrum of space and astrophysical plasma turbulence. We show that this is not the correct physical interpretation of the standard amplitude-weighted measurements. Using high-resolution incompressible MHD simulations from the Johns Hopkins Turbulence Database and near-Earth in situ solar-wind measurements from the Wind spacecraft, we separate angular statistics from Els\"asser-amplitude weighting and measure the finite-time retention of amplitude--angle states. In the simulations, the unweighted folded angle remains only moderately below the random three-dimensional baseline and shows no monotonic scale-dependent decrease over the resolved inertial-range separations. The much smaller angles inferred from weighted diagnostics arise primarily from large-\(A_r=|\delta_r z^+||\delta_r z^-|\) events, producing a negative covariance between \(A_r\) and \(\sin\theta_r\) that is removed by shuffled controls. Time-resolved transition measurements show that high-amplitude large-angle states deplete faster than high-amplitude small-angle states. The measured source--depletion balance reconstructs the second-order Els\"asser amplitudes without fitting their scale dependence and gives an effective root-mean-square increment scaling close to \(\ell_\perp^{1/4}\), although the typical folded angle is nearly scale independent. Wind measurements reproduce the same amplitude--angle hierarchy and negative covariance under Taylor sampling. Conventional dynamic-alignment diagnostics therefore measure selective retention of intense Els\"asser fluctuations, not volume-filling progressive alignment of typical fluctuations.