the lamppost, gas pipe or whatever will have to provide a return path to the substation of a similar impedance to that of the PEN conductor in order for the lights to stay on as you suggest. The loop impedance would have to be a couple of ohms at most for the loads in the installation to function as normal, after that the reduced voltage across the normal loads as a result of the resistances in series would become a significant factor.
As a rough calculation If the normal loads in the installation at the time of the break in the PEN occurring total 20A at 240V then their combined resistance would be 12ohms, if the loop impedance of the return path via the lamppost or gaspipe etc was also 12ohms then the potential Dropped across the installation will be 120V at which point a lot of loads will stop working.
Whichever way you look at it I don't see how the very simple science can support your ideas of melting cables and fires starting.