I think the talks of hypoxia are irrelevant. The simulator re-enactments of the crash have failed to produce better flight time to crash compared to what actually happened. And simulator is running at sea level pressure.
Actually, for some reasons UA232 pilot was able to nearly suppress phugoid cycle while JA123 was unable. Most likely, the pitch momentum of the 747 with all control surfaces on zero was too large for engine throttle to keep pace with it. Smaller phugoid cycle period = unability to control.
May be a silly idea..reducing pitch torque by moving passengers more to nose first.
Next, organize most physically fit passengers in rows in aisles, and make them run on command to nose or tail if pitch control is needed. Make them change aisle if roll control is needed. Control yaw with differential throttle.
It should work..what makes control is the distance between centre of mass and centre of pressure. If centre of pressure cannot be controlled, only one option remains - controlling the centre of mass.
Example:
-group A, switch to left aisle!
-5 seats back! No, 10 seats back!
-3 seats forward, steady, steady..
Also, some trimming may be achieved by using fuel pumps. Not enough for positive control, but at least reduce the pitch torque. I do not know though if it was possible with the fuel pumps configuration of 747.
Of course, it was impossible to even think for Japanese pilots. Delegation of control to passengers..unthinkable.