Often cited dictums in Quantum Mechanics include “observation disturbance causes loss of interference” and “ignorance is interference”. In this paper we propose and describe a series of experiments with modified Mach-Zehnder interferometers showing that one has to be careful when applying such dictums. We are able to show that without interacting in any way with the light quantum (or quanta) expected to behave “wave-like”, interference fringes can be lost by simply gaining (or having the potential to gain) the which-path knowledge. Erasing this information may revive the interference fringes. Delayed choice can be added, arriving to an experiment in line with Wheeler’s original proposal. We also show that ignorance is not always synonym with having the interference fringes. The often-invoked “collapse of the wavefunction” is found to be a non-necessary ingredient to describe our experiments.