12 million years ago, Antarctica wasn’t located at the South Pole?
Oddly enough, there was no south pole back then. ;’)
There are two things going on; one, there’s the continental drift idea, which has been semantically dressed up under the name plate tectonics. It’s based on evidence (magnetic reversals found as “fossil” magnetism in lava flows. The lava flows are dated various ways, and the average period of time between reversals is something like 600K years I think. This is used (via the syllogism) to show that some of these lava flows froze with their trace magnetic fields pointed in the then-prevalent direction, but have since moved this or that way.
It’s assumed that the shift of the magnetic pole reflects something; some believe it follows the axis of rotation, but since the magnetic pole shifts continuously (it’s shifted quite a bit since I learned how to use a compass in Boy Scouts) and the axis of rotation manages only a couple of slight wobbles, most don’t buy into the idea that the axis of rotation shifts, even though there’s evidence that it has. For that matter, there’s evidence that the magnetic field of the Earth not only shifts, but also is much stronger from time to time. That’s been interpreted different ways, more later (after I sleep).
Hapgood originated (source?) the crustal displacement model, which allowed the surface to shift but avoided the problem of redirecting Earth’s momentum during a full change to the axis of rotation. The friction problems appear to be unsuperable.
Anyway, short story long, the magnetic pole of the Earth has been all over (it was once in the middle of Arabia, for example) while any shifts to the axis of rotation would be expected to leave behind evidence only of having happened, rather than amount and direction.