Hi
The goal of the FCS, in normal law, is to be "path stable". So "stick free", which has not much sense here, the airplane should maintain flight path angle. In the case of level flight, FPA=0º. But it is in fact a curved path right? So I conclude that the system does not keep an inertial path, but a path linked to earth, linked to gravity.
I think the ADIRUs use schuller pendulum (i have never fully understood it) to help calculating the gravity vector. So the calculated accelerations, velocities and position are not corrupted be a small component of gravity.
But the effect of the centripetal force is a different matter, even independent of gravity. It just depends on the radius of turn and the square of the inertial velocity. I guess that the FCS uses IR (rather than ADR) altitude, true heading and GS information to correct the 1g value due to this effect, and thus maintain "curved stable paths".
I did not read well your post. So you have all pitot blocked. What happened? If the three of them give the same incorrect values, the failures would go undetected by the ADRs?. Was this the case or did they fail one by one?
Regarding stick free in direct law... Do the surfaces float of remain fixed in the position corresponding to the sidestick angle?
The 1g compensated for pitch... This means that the system calculates actual g forces or the component of the g forces at right angles with the airplane path or the component at right angles with the longitudinal axis?