XMM-Newton RGS spectra of MCG À6-30-15 and Mrk 766 exhibit complex discrete structure, which was interpreted in a paper by Branduardi-Raymont and coworkers as evidence for the existence of relativistically broadened Ly emission from carbon, nitrogen, and oxygen, produced in the innermost regions of an accretion disk around a Kerr black hole. This suggestion was subsequently criticized in a paper by Lee and coworkers, who argued that for MCG À6-30-15, the Chandra HETG spectrum, which is partially overlapping the RGS in spectral coverage, is adequately fitted by a dusty warm absorber model, with no relativistic line emission. We present a reanalysis of the original RGS data sets in terms of the model by Lee and coworkers. Specifically, we show that (1) the explicit model given by Lee and coworkers differs markedly from the RGS data, especially at longer wavelengths, beyond the region sampled by the HETG; (2) generalizations of the Lee and coworkers model, with all parameters left free, do provide qualitatively better fits to the RGS data, but are still incompatible with the detailed spectral structure; (3) the ionized oxygen absorption-line equivalent widths are well measured with the RGS for both sources, and place very tight constraints on both the column densities and turbulent velocity widths of O vii and O viii. The derived column densities are well below those posited by Lee and coworkers and are insufficient to play any role in explaining the observed edge-like feature near 17.5 Å ; (4) the lack of a significant neutral oxygen edge near 23 Å places very strong limits on any possible contribution of absorption to the observed structure by dust embedded in a warm medium; and (5) the original relativistic line model with warm absorption proposed by BranduardiRaymont and coworkers provides a superior fit to the RGS data, both in the overall shape of the spectrum and in the discrete absorption lines. We also discuss a possible theoretical interpretation for the putative relativistic Ly line emission in terms of the photoionized surface layers of the inner regions of an accretion disk. While there are still a number of outstanding theoretical questions about the viability of such a model, it is interesting to note that simple estimates of key parameters are roughly compatible with those derived from the observed spectra.