chrispenycate
Sage
Because they are not actually orbiting the planet, right.If the moon was cycling through the L2 Lagrangian point it could be in shadow for weeks at a time. Although objects at the Lagrangian points are typically not thought of as moons.
But, apart from 'Trojan asteroids', how would you specify such bodies?
Unlikely, maybe. But I can quite easily specify a stable orbit for a 'satellite of a satellite' (stable for a few million years, at least). I wouldn't like to do it for the Jupiter system, littered with moons, but on a simple 'one star, one planet, one moon, one sub moon' system I could specify orbits and make a reasonable explanation of how it condensed like that out of the original protoplanetary discsomeone there who had great knowledge of astronomy who stated that it was impossible for a satellite of a planet to have a satellite of its own.