So auto-trim was still active throughout the stall and continued stick-down inputs, had they been applied, would have moved the THS down and thus aided the recovery?
@Lonewolf: my reading of this, and the BEA report, is that the AOA data is lumped together with other air data measurements and accepted or ignored together... either all or none. (thoughts... does this apply to altitude and ROC too?)
No abnormal law for the same reason as no stall warning.