Edit: reading some more about it, cages that are close to the radiating element may experience capacitive coupling, and this is what can cause an ungrounded cage to serve as an antenna. A larger cage, with the radiating element farther away from the cage, is less likely to experience this. In either case grounding should reduce this risk.
If a Faraday cage blocks interstellar signals only if one part of it is stuck in a ball of mud and rock... well, I have some questions.
There is the possibility of the ground being a return path to the transmitter, but if that were effective, radio infrastructure would interfere world-wide, and you could transmit through the earth's core. And even that argument would suggest that the Faraday cage should be floating, not grounded.