Bear, you lost me. The "locked" descriptive seems not to mesh with the design feature that leaves open control of the THS, to move up or down, via the trim wheels regardless of elevator position or AoA. (If that isn't the case, it means my understanding is deficient).
I believe that Lyman was trying to say somewhat less than concisely, was that with the aircraft established in the stall and dropping at over 100 kts, that the g could have been less than .75g leading to the THS software logic locking in position.
or to put it another way
If the software driving the THS uses g sensimg
and moves the THS with the intent of maintaining 1 g then
0.9 g would be enough for the software to continue commanding full NU.