NoD and Roseland both make persuasive arguments. I do think this is an issue that wasn't given sufficient weight in the AF447 report.
Granted, the crew in that incident had
up six ways from Sunday before the incident got that far; I grant you that without reservation.
I don't know which of you is right. I do know I'm still boggled at the existence of a system that, when things have gone to excrement in every conceivable way, gives you a stall warning when you push forward, and stops the stall warning when you pull back.
You can engineer as many different control laws as you like; Murphy's will trump them all eventually. I do believe the clever chaps and chapesses at AB never envisaged this one, and who can truly blame them?