Mark, it sounds like this is part preference, part subjective tone reproduction. You might find this interesting. The table, from Jack Holm, shows Zone I densities for a variety of viewing conditions. The "Reflection prints in a dim room" value seems counter to my understanding, but these values might be based on viewing a single tone and not in relation to the full range of tones (gray scale).
Zone I - viewing conditions.jpg
Instead of trying to control the highlights with the shoulder of the film curve, what about print flashing? I flash whenever I'm printing a high key image. Flashing will compress the highlights as well as control them. Sounds exactly like the results you want.