Jean Claude is correct, of course. The centrifugal effect acts in the plane of the blades' rotation, even if that plane is not square to the rotor spindle. Centrifugal effect can't "lift" the low blade. It simply pulls spanwise on the blade.
The phenomenon we (informally) call "flapping" is retreating-blade stall in different circumstances than those experienced in helicopters. From the retreating blade's viewpoint, a stall is a stall, whether it occurs at high or low aircraft airspeed. The angle of attack of the blade exceeds its stalling AOA, and its lift decreases abruptly.
Our semi-rigid, teetering rotor system is marvelously simple, but it has its limits. By design, the teeter hinge always tries to increase the angle of attack of the blade having the lower airspeed, in order to equalize lift between the blades.
The retreating blade has a lower airspeed than the advancing blade. Increasing AOA only increases lift up to the stalling AOA, but the teeter hinge does not know this. It continues adding AOA to the retreating blade even past that blade's stalling AOA -- so the retreating blade stalls. Once this happens, the teeter hinge has failed in its mission; instead of equalizing lift between the two blades, it has created a large INequality of lift. The un-stalled blade rises wildly -- or sails, if you like. The stalled blade, being attached rigidly to the rising/sailing blade, drops hard and can strike parts of the gyro.