Speaking of brains and our human response to various stimuli -- computer aided or not -- what would the "seat of the pants" part of the pilots' brains be experiencing?
In this type of stall, would they feel increased positive G's which would then seem to decrease (or even go negative) as they rapidly descend in the continued stall? I gather the BEA will provide a graph with G forces plotted with a full report of FDR data, but no doubt it could influence the pilot's actions or lack thereof.
If they stalled it at a load factor of less than one g, that would result in a stall at a considerably lower speed. The Airbus Safety newsletter mentioned that buffetting would occurr as the aircraft approached stall.
But in this case, if the aircraft stalled at less than a G, the speed would be lower and there may not have been any or much buffetting. If that is the case, the PF may not have felt enough of a stall or break to make him think he had actually stalled it. Stall horns go off when you approach stall, not after you stall.
What very well could have happened is that as the stall warning started and the PF applied TOGA. The aircraft didn't buffet, shake or break into a conventional stall, it stopped flying and started falling at a lower speed.
Absent the break or buffett, when the stall horn stopped the PF likely thought that he flew out of the approach to stall and was flying again. All speculation, of course, but if that is what happened it would offer an explanation of what the PF thought and would explain a lot better his actions.