I've grabbed a bunch of video frames from the best video I've seen of the GG initial roll left. These have been processed by increasing brightness and contrast, and cropping to show only the aircraft. They seem to show the port elevator deflected up. The elevators are not visible at the initiation of the problem but soon become visible as the g turns the aircraft. Compare the thickness of the stabilizer-elevator combination above and below the fuselage to see this.
The only way that there would be significant elevator split is if the elevator torque tube system failed.
The third and fourth pictures appear to be showing the port elevator deflected upward. The second picture is not clear enough, and the attitude in the first picture puts the elevator too close to the wing to be sure of anything. Subsequent pictures appear to show the elevators back in balance.
One would assume that if the torque tube simply failed at either the port or starboard side, that the port elevator would be deflected downward by a force twice what was needed from the tab, and the starboard elevator would move upward due to the loss of down torque from the other side. Instead we appear to see the opposite effect, so perhaps the likely shearing of the torque tube is indeed due to a failure of the tab system.
The Ghost should have had an elevator bobweight in the system to increase stick force per g. If so, the mass of the elevator actuation system would act to dampen torsional shock loads being sent from the port elevator, so the torque on the inboard side of the port elevator could be higher than the torque on the inboard side of the starboard elevator.
The pictures of the trim tab separating indicate that the entire tab was not lost, but that it instead detached at the inboard end and broke off at the center support. The separation at the inboard tab attachment may be due to flutter from some other initial failure (such as the linkage), or it may be the actual initiating event.