Syd, thank you for your answers but surely there has to be an axis of rotation if the body rotates.
Consider just the translating tendency of the TR thrust before the disc is tilted to compensate - that lateral force must act either through or at some distance from the vertical C of G. If it acts through it the body will simply move laterally but if it acts at some distance (ie TR vertically separated from the vertical C of G, it must also cause a rotation.
If my sentence is correct then applying lateral cyclic gives a force acting at the rotor head which, again, if not aligned with the vertical C of G, will produce a rolling moment as well as a lateral force to oppose TR drift.
Thoughts?