There is one sentence that, properly understood, prevents a large fraction of fatal loss-of-control accidents. The US Federal Aviation Administration puts it like this:

Any airspeed. Any attitude. Any power setting. A wing does not stall because the aircraft is slow. It stalls because the angle of attack has become too large, and that can happen at 500 knots in a dive just as readily as at 60 knots in the circuit.

It is not the nose attitude

The commonest confusion is between angle of attack and pitch attitude. They are different things and they can point in opposite directions.

Pitch attitude is where the nose is relative to the horizon. Angle of attack is where the wing is relative to the air it is actually moving through. An aircraft descending steeply can have a low nose and a very high angle of attack at the same time, because the air is arriving from well below the nose.

The FAA is blunt about what this means for the pilot: because an aeroplane can be stalled in any attitude, looking out of the window is not a reliable indicator.

Why the stall speed in the manual keeps lying to you

Every aircraft has a published stall speed, and every published stall speed comes with conditions attached that pilots routinely forget. The FAA lists them: unaccelerated 1G flight, coordinated flight with the ball centred, one specific weight, and one specific centre of gravity position.

Change any of those and the speed changes. Load the aircraft heavier and it stalls faster. Pull G in a turn and it stalls faster still, because in a 60-degree banked level turn the wing is carrying twice the aircraft’s weight and therefore needs the lift it would produce at a much higher speed.

What does not change is the critical angle of attack. That number is a property of the wing. It sits there, unmoved by weight, bank, altitude or how the day is going.

Accelerated stalls are the practical consequence. The FAA calls them accelerated manoeuvre stalls and defines them as any stall occurring above 1G, typically demonstrated in a steep turn at or below manoeuvring speed.

There is a nuance in the FAA text that deserves more attention than it gets. Above manoeuvring speed, the wing can reach the design load limit at an angle of attack abajo the critical one. In other words, pull hard enough up there and you will break the aeroplane before you stall it.

Why airspeed is a poor cue where it matters most

The clearest illustration comes from the French investigation into the loss of Air France 447 over the Atlantic in June 2009, and the BEA’s report contains a number that explains a great deal about why high-altitude stalls are so dangerous.

At takeoff or approach speeds, a one-degree change in angle of attack corresponds to roughly a five-knot change in indicated airspeed. In cruise, the same one degree corresponds to about twenty-five knots. So in the cruise the speed tape moves a long way for a very small change in the thing that actually matters, and at Mach 0.8 the margin between the flight angle of attack and the stall warning angle can be of the order of one and a half degrees.

The BEA report also records, plainly, the divergence between attitude and angle of attack that night. At maximum altitude the aircraft’s pitch attitude and angle of attack were both around sixteen degrees. Later, as it descended at around ten thousand feet per minute, the angle of attack exceeded forty degrees while the pitch attitude never exceeded fifteen.

The phrase “when it was valid” matters. Below sixty knots of indicated airspeed the angle of attack values are declared invalid, which is why the stall warning stopped after sounding continuously for fifty-four seconds. The aircraft was deeply stalled and the warning had gone quiet.

Recovery: the order is the lesson

The FAA is explicit about the most common fatal error in stall recovery, which is not failing to recover but recovering in the wrong order.

Its wording is that there have been numerous situations where pilots did not first reduce angle of attack, and instead prioritised power and maintaining altitude, which resulted in a loss of control. The instinct to avoid losing height is exactly the instinct that keeps the wing stalled.

The manufacturer-derived template runs: autopilot off, pitch nose-down until the stall indications stop, roll wings level, add thrust as needed, retract speedbrakes, then recover the flight path. Altitude is the last thing on the list, and deliberately so.

Why not just fit an angle of attack indicator?

Many aircraft do, and military fast jets have flown approaches on angle of attack rather than airspeed for decades, because it is the correct parameter and it is independent of weight.

The FAA notes the caveats honestly: different manufacturers calculate the indication differently, some systems take flap position into account and some do not, and accuracy depends on calibration, on the wing being clean, and on the probe or vane being heated. An angle of attack indicator that has iced up is worse than none, because it is confidently wrong.

None of which changes the underlying sentence. The wing does not know how fast it is going, how much the aircraft weighs, or which way up it is. It only knows the angle at which the air is arriving, and past a certain angle it stops working.

Sources: FAA, Airplane Flying Handbook (FAA-H-8083-3C), Chapter 4, “Stalls”; Bureau d’Enquêtes et d’Analyses, Final Report on the accident on 1 June 2009 to the Airbus A330-203 registered F-GZCP; NASA Glenn Research Center.