A high wing airplane will correct itself (lateral stability) when disturbed because its c.g. is below the c.p. (looking at the plane from its side), according to this forum post and the book below.
So my question is, how does this "pendulum effect" affect a biplane with two centres of lift?
It will affect a biplane just as much as a monoplane.
Not at all.
The pendulum effect does not exist in airplanes. It does in airships, but not in heavier than air craft.
For a proper discussion, we should first define what a pendulum is. Only then can be established if such an effect can exist in airplanes.
Let's base the definition on Wikipedia. It says that
A pendulum is a weight suspended from a pivot so that it can swing freely.
Maybe it is also worth to look closer what a pivot is: A thing on which something turns.
So the pendulum is fixed to a fulcrum which keeps it suspended and allows it to swing freely. The ideal pendulum has all its mass in its massive bob and, therefore, pivot and center of gravity are not in the same place. If the center of gravity and the pivot would fall together, a pendulum could only rotate but not swing. And that swinging motion is what the pendulum is all about.
Now for airplanes: Here we have no pivot. All rotation can only happen around the center of gravity. This is equivalent to the pendulum with no length which is no pendulum any more.
Lift is the sum of all pressure acting perpendicular to the direction of movement. Lift on a banked wing will also bank with it. The lift vector will still be in the plane of symmetry of the banked aircraft and will have no lever arm with the center of gravity, thus causing no uprighting rolling moment. The Figure 34 of your copied book page is simply wrong. The author did not know what he was talking about.
I added an answer here about the rolling maneuver of paragliders. The description should make obvious that no pendulum effect is involved. Rather, the whole is quite similar to roll control in hang gliders where the lift is shifted sideways in order to create an imbalance, but then again different in wonderful ways. The discussion below would not allow me to explain my thoughts in such detail.
Disturbed - but how? A fixed wing aeroplane in a stationary, zero sideslip turn is roll neutral. There is no tendency from gravity to upright the roll. Not for a monoplane. Not for a biplane. Not for any number of fixed wings.
The last picture in the OP with the vertical lift vector for the rolled aircraft is wrong: the lift vector deflects with the wing and is always perpendicular to it, therefore always points though the CoG. The picture only considers the stabilising moment of the vertical component $L_v$. and conveniently disregards opposite moment of the horizontal lift component $L_h$, which magically counteracts the rolling effect of the vertical component.
Disturbed in sideslip caused by $L_h$: yes, this causes an aerodynamic rolling moment, from several mechanisms.
Pendulum effect is a bit of a misnomer. What is thought of as "pendulum effect" is actually just a favourable rolling moment that can be generated during sideslip if the center of mass is a large distance from the lateral aerodynamic center, which doesn't really exist on normal aircraft.
Paragliders however, which operate in a kind of topsy turvy alternate world of control, exploit this effect to achieve lateral stability and to turn. A paraglider turns by skidding, the skid being created by increasing the lift and drag on the into-turn side when you pull the trailing edge down with brake application (you are only interested in the drag increase, not the lift increase, which is working against you - you are only interested in the lift increase when using both brakes together to slow down and flare).
Doing this (applying brake on one side to turn that way) actually creates a small aerodynamic rolling moment in the opposite direction of the turn (like trying to turn an airplane to the right by lowering the right flap and aileron only - doesn't work so well in that case), but because the center of mass is more or less down where the pilot is, and the lateral aerodynamic center is up by the wing someplace, the lateral force on the pilot to the outside of the skid massively overpowers the differential lift created by the brake application and rolls the glider to the right. And you can say that it's sort of acting like a pendulum, kind of.
You could say that paragliders exploit this effect to use adverse yaw to turn the wrong way, allowing control with inputs that are seemingly opposite to the normal world (turning right by lowering right aileron as it were).
As well, the mass way below the wing creates a strong centering effect (you're basically hanging from a parachute that is able to glide forward) and if spontaneous sideslip occurs the restorative rolling moment is immediate. It's also how paragliders are to somehow magically be able to achieve very strong yaw stability without any weathervaning element like a fin or sweep. Yaw and roll are strongly interconnected due to the mass 15 feet below.
So you could say that there is a pendulum effect but it only works for paragliders, or maybe some crazy airplane with most of its mass in a concentrated bob weight at the bottom of a long rigid pole extending below it, with most of its surface area at the top. In any normal airplane, the lateral aerodynamic center and the center of mass are too close together for this effect to overpower the other forces and are insignificant if they exist at all.
