Spiral dive ...
Spiral dives generally end in a broken airplane due to excessive G loads during recovery (HS or VS failure in flight) unless you get to FL 0 first that is. The graveyard spiral seems ruled out in this case by the available evidence. Don't we all practice recovery from unusual attitudes on instruments? With loss of one or more primary instruments?
The BEA report indicates the pilots never lost inertial attitude reference information. Power and attitude is how you fly the plane and I see no indication the pilots ever lost power or attitude information on their flight displays even though they did lose airspeed and vertical speed info. I strongly suspect that the pilot and flight computer were working at odds with each other (as someone suggested here due to change in stick feel in direct law) or else the pilots allowed the A/P to fly the plane with two incorrect but agreeing airspeed inputs (note the new AD) trusting the A/P in IMC at night with the computer putting the aircraft outside the flight envelope due to invalid airspeed and altitude data.
Being a software engineer, I am curious about the approach Airbus used for boundary testing of the flight control computer software. It is always the combination of inputs that nobody thought was possible that gets you in the end. The X31 article that was posted here is highly relevant in that regard. I believe that a software design error will be one of multiple contributing factors to this accident. The new AD that was issued suggests that Airbus and the FAA already believe this to be true. You can design better air data sensors but never one that is 100% reliable in all possible situations. Therefore, you better design software that can handle the loss of all air data input without loss of control of the aircraft. And computer systems are not 100% reliable either. You see the basis for my distrust.