I have not studied this specific question to any degree, but have some experiences and observations to contribute to your search started in the various studies of specific circumstances.  I have not read these specific publications to comment on them.

I have seen that in areas with chemical weathering or influence in some soils, the erosion rate is higher including the potential for piping in some instances, gully formation, etc (mostly South Carolina, Georgia, with other examples in North Carolina, Texas, Arizona, Alabama, Colorado and Arizona). I understand your thought that eventually erosion may occur to the extent that eventually the most weathered material may be gone, exposing less weathered soils (e.g., C horizon, regolith), or soils with accumulation of clay materials (B horizon) may offer resistance. I note that for soils to be eroded to degree of C horizon or regolith, you are most likely talking about channelized systems or more or less permanent barrens, with little potential to support vegetation without further soil weathering and development.  We had some of these barrens in the weathered C horizon, saprolite soils in SE USA, and they lacked vegetation and often rilled or gullied surfaces, and on occasion continuing to contribute sediments during extreme rainfall events (researchgate paper on sediment from a small ephemeral gully.  It mentions freeze-thaw, high rain intensity and wind-cyclonic storms that move trees as other processes that can contribute in dislodging particles in extreme headwaters with exposed conditions.

Denudation of soils exposes them to rainfall, and increases the potential that soil particles will be dislodged and entrained where surface flow accumulates, potential for development of rills or gullies on sloping terrain if rainfall rates tend to be high, and soils are erodible.  Weathered soils contribute much to the potential for these severe erosion processes.  Landslides may or may not be connected to soil weathering, and usually dependent more on steep slope, geologic contacts, ability of the soil to hold rather than percolate water (such as mica schist soils), and conditions of excess water is available to help overload the weight of the soil and reduce the friction against the bedrock or controlling geology features.  Roads with poor surface drainage properties and their undercutting of slopes, sometimes placing fills on steep slopes increase the natural potential for landslides due to the adjacent denuding activity, soil disturbance, compaction, vibration by heavy equipment and modifications to hydrology (or hydraulic properties).  There may also be conditions where chemical weathering and erosion rates are not well correlated, even if soils are denuded.  In general, I would suggest that chemical weathering of soils contributes to erosion rates, but the lack of chemical weathering does not prevent soils from being eroded under conditions that favor erosion, such as high rainfall, steep slopes. The factors that affect surface erosion in the Universal Soil Loss Equation include the soil K factor, but several other factors.  Is the product of these varying factors that affect surface erosion.  It does not address channel or slope erosion, which are affected by some of the same factors but also some different ones such as the concentration of water to the specific site. 

So in the models suggested on these processes, there are many elements to consider, and perhaps some further refinements to be able to separate the various types of erosion, types of conditions, types of soils, potential for and rates of weathering, soil disturbances, geology, topography, hydrology, rainfall, etc.