Well, we talked about earplugs a few times on Finnbass, but I don't think this issue ever came up.
Is there hearing damage that can occur via another mechanism than excessive sound arriving at the cochlea, wahtever its transmission route (and not including physical damage to tympanum, malleum, stapes, or the other one)? Because if it's sound arriving at the cochlea that causes the damage, you'd hear the sound. So if something made the sound a lot quieter, it would reduce the potential hearing damage.
As an experiment, I tried talking to myself. When riding a motorcycle at 90mph on a closed private test track, failing to use earplugs results in excessive and painful noise levels, thus demonstrating that a helmet does not act as ear defenders. When wearing a pair of Howard Leight's finest Maxes, the level of noise is considerably reduced. This allows me to talk to myself at a volume lower than normal conversational volume and still hear myself without having to ask myself to repeat myself. Obviously this is possible through bone conduction. As the noise level is considerably reduced, surely this demonstrates that earplugs, even of the 50p a pair 33dB orange bell variety, are effective.
All tests conducted on a closed private test track by a rider who hears no internal voices.