HN39 (#859), I had considered that the first stall warning could have been an artifact of the ADC changeover, but ‘real’ as you describe does not alter the thrust of my view (#816).
The important inference was that (at least) the PF decided that this alert represented a real stall and commenced the erroneous action, even though we know (in hindsight) that a warning would not be given in normal law. I surmise that the PF did not consider this, nor re-evaluate the situation even when the PNF called the Alt Law condition. The PF may never have appreciated the change in control law as his attention was fixed on the erroneous ‘stall recovery’ pull up manoeuvre.
Do you mean they did not associate the UAS procedure with the UAS situation?
Yes, this supposition is supported by the crew reports from previous incidents. The loss of an airspeed display in itself is not ‘unreliable’ – it’s not there, reliability as such cannot be judged (unreliable might mean it’s not there very often).
This may be playing with semantics, but it is an exploration of how the PF might have interpreted the situation, or how this procedure might have been taught which then influenced the assessment.
IIRC the distant history of the UAS drill probably came from accidents involving flight without airspeed, was this taught. Or perhaps the UAS training was done against the backdrop of ice crystal icing, a completely different context. Therefore there may not have been an immediate association between loss of an airspeed display and a drill relating to ‘ice crystals’ and aircraft handling.