I think we can all agree that when we are shooting many photons one by one, through an interferometer, we can eventually land up with the interference pattern. This can be explained by saying that two photons combining in some areas to give four photons and in some places annihilating each other. This violates the conservation of energy so a more plausible reasoning would be to assume that every photon interferes with itself only to preserve local conservation of energy. However, I have a problem with this for in the interferometer set up where we are splitting one beam into two parts, we still get an interference pattern with arms of different lengths. This would mean that no single photon that gets "split" into two states of motion ( two beams travelling in different arms ), arrives at the same time, at the screen or sensor. So how is it even possible to get the interference pattern? ( which we do get experimentally ) In his book, principles of quantum mechanics by Dirac, he mentions that when the beam is split in two different components we should still regard them as being one entity with a superposition of two wavefunctions (for two beams); But doesn't the very act of splitting disturb the photon and cause it to already collapse or make it's mind to go in one of the arms? Is it just that we're unaware of the direction the photon choose and so we use superposition to address our uncertainty of the situation? In case of the double slit experiment I think a similar situation arises since the most points on the screen are at different path lengths from the slits from which wavefunction gets split. TLDR: How do we explain single photon interference in an unequal arm length interferometer? Do we simply accept that it is a fact of nature that when we do the experiment with a large no. of photons we get the familiar interference pattern because single photon interference has no meaning?