Twisting the stall inboard
Because a root-first stall is so much safer than a tip-first one, designers actively move the stall inboard, and the geometry lesson's washout is the primary tool, now with its full purpose visible:
Geometric twist (washout) — the wing is built with less incidence at the tip than at the root: whatever the aircraft's angle of attack, the tip flies a few degrees below the root, reaches its critical angle later, and keeps the ailerons alive through the root's stall.
Aerodynamic twist — the same end reached through section shape instead of angle: root sections that give up at a lower angle, and tip sections that hold on longer: a larger nose radius raising their stalling angle, extra camber raising the maximum lift they can reach before letting go.
Most real wings blend both. The purpose is one sentence: make the root stall first, whatever the planform wanted to do.
Fences, teeth, vortilons and strakes
Swept wings need more help, because their tip-stall mechanism, the outboard drift of the boundary layer, is built into the sweep itself. The fixes all attack that drift, mostly by standing something in its way or by stirring the flow it has tired:
Wing fences — chordwise walls standing on the upper surface, physically blocking the outboard drift of the boundary layer so the tip flow stays fresh to a higher angle of attack.
Saw teeth (dog-tooth leading edges) — a step in the leading edge whose discontinuity sheds a vortex back over the wing at high α; the vortex acts as an invisible fence, blocking the spanwise drift and energising the outboard boundary layer.
Vortilons — small fence-like plates under the leading edge; at high angle of attack, with the stagnation point far aft below the nose, the local flow strikes them and rolls a vortex over the upper surface: a fence that only exists when it is needed.
Vortex generators — the familiar rows of small vanes, re-energising the outboard boundary layer directly so it survives to a higher angle.
Strakes on engine nacelles — large vanes on the nacelle sides; at high α, with flaps out, the nacelle disturbs the flow onto the wing behind it, and the strake's strong vortex re-energises exactly that stretch, protecting the local stall behaviour.
Why warning is necessary
The quantity that stalls the wing, the angle of attack, is the one quantity the pilot cannot see. Attitude is visible, speed is indicated, but the angle between chord and relative airflow appears on no basic instrument, and the stall can arrive at any speed and any attitude once the load factor and the flight path conspire. A stall warning exists to convert the invisible approach of the critical angle into something unmistakable (shaking, sound, or both) early enough for an ordinary crew, not a test pilot, to react before the wing actually lets go. Certification writes the same philosophy into the speeds: CS-23 and CS-25 require take-off and landing to be flown with a margin above the stall: the scheduled speeds are fixed factors above the stalling speed (CS-25: V2 at least 1.13 times the reference stall speed, the landing reference speed at least 1.23 times it; CS-23 classically 1.3 times the landing-configuration stall speed), so that gusts, turns onto final and ordinary handling scatter never carry the wing to its critical angle in normal operations. Warning and margin are the same idea at two layers: keep the stall at a distance in normal flying, and warn unmistakably whenever it approaches.
Natural and artificial warning
The best stall warning is the aeroplane's own: the natural pre-stall buffet of the separation lesson, the separated wake drumming on the tailplane. Where a design provides it strongly and early (the root-first straight wings of light aircraft) the aerodynamic warning suffices, sometimes helped by a stall strip tuning where and when the buffet starts. But many modern designs do not co-operate: swept wings stall tip-first, far from the tailplane that would have felt the wake; T-tails (the tailplane mounted atop the fin) lift it out of the wake altogether; and some clean wings simply give too little shake, too late. There the regulations demand an artificial warning: a sensor watching the approach to the critical angle, driving a shaker or a horn. The family, sensor by sensor:
Buffet — the natural warning itself: the separated wake striking the airframe. Free, unmistakable, and only trustworthy where the stall pattern sends the wake somewhere the crew will feel it.
Stall strip — a small sharp-edged wedge on the inboard leading edge. At high angle of attack its edge trips separation early at the root, manufacturing a clear, well-placed buffet a moment before the real stall: shaping the natural warning rather than sensing anything.
Flapper switch (leading-edge stall-warning vane) — the light-aircraft classic: a small hinged vane just below the leading edge. As the stagnation point slides aft beneath the nose with rising α, the local flow direction reverses over the vane and lifts it, closing a circuit and sounding the horn: the stagnation-point migration of the separation lesson, used as a sensor.
Angle-of-attack vane — a small pivoting aerofoil on the forward fuselage that weathercocks into the local flow, measuring α directly and continuously; the standard transport sensor, feeding warning systems and displays alike.
Angle-of-attack probe — the same measurement by pressure instead of pivot: a probe with slotted openings senses the flow direction from the pressure split between them. No moving vane, same α signal.
Stick shaker — the transport-category warning output: motors with eccentric weights on the control columns, fired by the α sensors a few degrees before the stall, shaking the column hard and loudly: artificial buffet delivered straight into the pilot's hands, impossible to miss and impossible to confuse with turbulence.