Scale-dependent diagonal nonnegative extensions

The approach to Friedrichs is slow when $\nu$ is small: relative to $r^{-2}$ or $r^{-4}$, the correction is of order $r^{-2\nu }$. This is the power-law analogue of the logarithmic tail in (eq:logPV). At fixed regular points the limits $\lambda \downarrow 0$ and $\lambda \to \infty$ give ${\mathsf{E}}_0$ and Friedrichs, respectively, by dominated convergence.