It has been suggested that fracture-induced MHz–kHz electromagnetic emissions (EME), which emerge from a few days up to a few hours before the main seismic shock occurrence permit a real-time monitoring of the damage process during the last stages of earthquake preparation, as it happens at the laboratory scale. Despite fairly abundant evidence, electromagnetic (EM) precursors have not been adequately accepted as credible physical phenomena. These negative views are enhanced by the fact that certain "puzzling features" are repetitively observed in candidate fracture-induced pre-seismic EME. More precisely, EM silence in all frequency bands appears before the main seismic shock occurrence, as well as during the aftershock period. Actually, the view that "acceptance of "precursive" EM signals without convincing co-seismic signals should not be expected" seems to be reasonable. In this work we focus on this point. We examine whether the aforementioned features of EM silence are really puzzling ones or, instead, reflect well-documented characteristic features of the fracture process, in terms of universal structural patterns of the fracture process, recent laboratory experiments, numerical and theoretical studies of fracture dynamics, critical phenomena, percolation theory, and micromechanics of granular materials. Our analysis shows that these features should not be considered puzzling.