HN39, the energy tradeoff (~60%) takes us back to instrument scan.
How would PF NOT know he's climbing? How would PNF not see the climb and not comment. (CVR data/info would be handy here ... ) Are there two sets of eyes locked on airspeed indications with other parameters dropping one by one from scan? What CVR recordings match up to the fwd stick/reduction in climb rate? There is much missing here in terms of task management leading up to what appears to have become a task overload.
Here is something I know happened in the 80's. Helicopter crew in Persian Gulf, night, a few hundred feet over water, both pilots concentrate on the tactical display because something of interest had arisen. (Note, the PF and PNF roles had been lost. Two sets of eyes were on a display that is not the flight instruments, zero sets of eyes were on flying instruments).
For one reason or another, alt hold was either not on, or deselected. The helicopter descended toward the water. The RAWS tone (radar altitude warning) went off, and PF added power. Helicopter skipped off the water, kept flying, and was returned to its ship. That event was a few fractions of a second away from aircraft and crew in the water. The incident became boilerplate CRM and task management fodder for some years after that.
Which tasks and which stimuli attracted the most attention of the AF 447 crew? This is a significant unknown.