The Captain made a few careless errors - he was not familiar with the qualifications of the co-pilot he left in charge, he did not properly assess the weather system they were approaching, and he did not leave clear instructions when he called the 2nd co-pilot into the coughpit. A possible explanation is that he was preoccupied with being somewhere else at the time. Then when things went pear-shaped less than 10 minutes later it took him nearly a minute to respond to the increasingly frantic calls for his return and did not act decisively when he came back into the coughpit.
Quite honestly, he was entitled to believe that AF had given him suitably qualified FOs. Presumably he had also flown the first leg of the trip with them as well. There is
always weather at the equator. ANY FO qualified to act as relief should be able to handle it, otherwise he simply should not be there. If the captain had to be on duty around any weather he'd never get a break.
In fact as he walked back in the stall warning had just gone off, they were at around 35,000 feet, the engines were at TO/GA , they were nose up (approaching 15%) and falling rapidly (10,000ft/min). There is only one scenario that fits those parameters but the captain at no time identified the stall they were in or tried to take control from the flying co-pilot. I think the press (and Airbus) will have a field day with him and conjecture that he was thinking with the little head instead of the big one, regardless of the fact that we will never know.
Well, if he entered the coughpit over a minute after it all went pear shaped (and that is quite a quick time), the aircraft cannot have still been at FL350.
Taking control means getting into the seat. This wouldn't have been a smooth event, and doing so would not have been easy. Perhaps not even possible. The problem is that there was no easily way for him to see that the joystick was being held full aft. If it had been where it should have been, the attitude, power and performance displayed would have made no sense at all...we understand it simply because we know what control inputs were being made. Remember too, that much of the displayed data was blank.
Airbus spent a lot of effort convincing people that their aircraft can't be stalled. In normal law, that's almost true (there is one way that I think it can be done, but I'll refrain from testing it). The switch to either alternate or direct, does bring up an ECAM, but that's quite likely to be hidden in an avalanche of warnings. The PFD change that goes with it is quite subtle.
I see this whole accident as another version of the Buffalo Dash 8 crash. Deskilled pilots, coupled with aircraft that change character dramatically after quite minor failures, will be an increasing part of the toll.