Mr Optimistic:
I am a bit puzzled about the concern about AoA indication.
While true that pitch and power would likely have worked something else -did- go wrong and the pilots ended up in a confusing situation.
At that poin a raw AoA indicator would have likely helped them to sort out
the counterintuitive return of stall warning as they were applying nose down.
In fact the stall warning in this case was a tragically misleading on/off AoA indicator that had 3 states:
Normal : silent
close to stall or stalled : STALL STALL STALL
severe stall : silent
Severe stall results in <60kt sensed speed and inhibition of warning.
This is not related to pitot ice/failure, it is a combination of true speed plus AoA effects on pitot as others have pointed out.
Had the raw data been avaiable they likely would have seen the AoA improving as the warning returned and pressed on with the correct recovery.