This tragic and regrettable event, among its many complexities, illustrates the danger of permitting uncertainty about who has control.
Once the commander had taken control of the throttles, while leaving the co-pilot to control pitch and roll attitudes, the distinction between PF and PNF became essentially meaningless. Unless this was a specific, drilled SOP, both of them (i.e. neither of them) were now in control. Any ambiguity in their perceptions or intentions was now potentially disastrous.
With control of the flight path now split between two persons, conflicting communications could have played a further part in the accident:
<<speculation alert>>
The CVR dialogue would indicate that the co-pilot (still nominally but no longer fully the PF having deferred control of the throttles - a psychologically stressful situation) was expecting to go around in the event the commander was not visual at minimums. They had referred to an element of the missed approach procedure at 2:54 from the end of the recording: ("Okay missed approach three thousand") and subsequently in the final descent the commander called out heights to minimums as one would when prepared to go around if landing criteria were not met. Two missed approaches had already been made, and alternate fields had been discussed.
At 0:13 the commander made the call "Okay minimum... continue".
My speculation is that the co-pilot, primed mentally for another missed approach, interpreted this call as meaning "go around", whereas the commander, anxious to get in, meant it to signify "continue on the glide slope". As evidence for this contention, FDR data (fig. 5 in the report) shows that from about 0:12 altitude was held level (indeed at 0:11 it shows a slight increase); pitch attitude is increased slightly; and both are sustained until beyond the power reduction at 0:09. My guess: the co-pilot thought they were doing a go-around, but he couldn't do it fully, because the other guy had the throttles. Energy bled off while he sustained the pitch. At about 0:08 power is abruptly restored (likely although not necessarily by the commander) and suddenly the unexpected differential thrust aggravates the situation. A second later the stall warning sounds, and now they are trapped: low, slow, turning, asymmetric, devoid of useful energy (and at odds with what each other is doing?) and with no time, height, or visibility to react effectively.
One of the many lessons we may perhaps take from this event is to ensure that it is always clear "who has control".
Edited for spelling.
Last edited by justanotherflyer; 3rd Feb 2014 at 14:41.