Hypothesis
It is a curious adverse affect that the constant/continuous use of the modern flight deck seems to have on pilot behavior and awareness of the airplane's state. For the most part, the increasing application of automation has yielded the desired and intended improvement of nominal performance and operational efficiency. I can't prove that with data but I'm sure there are marketing types around with plenty of evidence for that.
But the automation brings attention to itself, its set up, its management and its own unique indications - of how the automation is doing, not how or what the airplane is doing. The pilot has gradually become a systems manager in spite of the title "pilot" and by habitual practice fostered by the modern flight deck become less attentive to and aware of the airplane state, leading to a lower state of readiness to recognize, interpret and react to hazardous non-normal conditions.
The automation performs so well in normal conditions and is so reliable that "normal" conditions are predominant. The airplane's state, while not deliberately neglected, is normally so benign that perceived need to monitor it closely is diminished. Other more apparent needs for the pilot's attention, whatever they might be, can seem to be safely attended to without any obvious consequence - normally. Like a constant dripping of water, the modern flight deck is, in effect, retraining the pilot, changing the pilot's routine behavior. The water is slowly but continuously getting warmer without notice, until it comes to a boil and things quickly become messy.
The non-normal condition occurs, the pilot keeps expecting normal system behavior, but it's not normal and the pilot is confused, not quite believing the indications. Not used to connecting the dots because the modern systems have been doing that for the pilot, the ability and likelihood that the dots can be reconnected in the midst of missing and disparate indications is unsurprisingly low.
The unintended consequences of flight deck modernization due to a failure to consider the effects on the human component of the system, that component which all the safety analyses assume will be there to put humpty-dumpty back together again, but just isn't up to the task - too little experience dealing with THIS situation.
Consider the Schipol 737 accident, the Buffalo Dash-8 accident, AF447, even Cali 20 years ago, and so many others. I do not presume that there are any silver bullet solutions to the many factors that feed this unfortunate state of affairs. I really don't believe that the typical response - to add another alert to compensate for inattention and lack of awareness - really addresses the problem. Why? Because it is not sufficient to presumably bring the pilot's awareness to its necessary level only when the situation has become hazardous and when an immediate, correct intervention is necessary. There are too many examples of disbelief, confusion, and delay for such unanticipated conditions.
The solution(s) must address the real problem - lack of routine, continuous, and deliberate awareness of the state of the airplane and associated systems - EVEN WHEN things are normal and the automation is actually handling the airplane. We need a "post-Modern" flight deck - designed to draw the pilot's attention regularly and continuously to where it belongs and where our system safety analyses have been assuming it was. The pilot should have timely awareness of even small unintended deviations from the normal desired conditions, that in their own right are not hazardous, but if left to grow unchecked could become hazardous. Such "minor" deviations should be corrected sooner, before they become major. A pilot (or flight crew) that has successfully maintained continuous awareness of the airplane is best prepared to do this.
What can system design do to help achieve this? Like I said, I have no silver bullets. I do think that more could be done to more positively distinguish normal, within desired tolerances, conditions from those where minor but certainly unintended deviations occur. Distinguished well enough that a quick glance will at least draw a double take and lead to heightened vigilance. The need for correction is not urgent, and means of correction need not be heroic. The vigilant pilot (flight crew) can assess the situation and take timely action, without the airplane ever reaching a condition in which safety was jeopardized. Whatever the cause of the deviations, including unexpected automation behavior, the pilot (flight crew) is aware and has been intervening (or at least thinking about it) before they ever reach an unsafe level.
I suspect, but have no hard data, that such "near normal" deviations occur more frequently than we think. That the pilots becoming aware of these will naturally become more vigilant - because their expectations will have changed. The result? The "post-Modern" flight deck will train pilots in the opposite direction that the modern flight deck has - toward more vigilance and awareness, not less. And THAT is addressing the problem. Building hours on the modern flight deck is not giving them the experience and wisdom we want from airline pilots. Building hours on the post-modern flight deck will have the opposite and desired effect.