Abstract

If it is possible to detect the path of an atom in a double-slit atom interferometer, then interference will not be observed even if path detection causes no change in the atom's momentum. This fact was explained by Scully et al and Duerr et al by using a principle ofcomplementarity that could not be formulated mathematically, together with a rather peculiar notion of entanglement. We provide an explanation based on considerations of Galilei invariance, without recourse to any notion of complementarity or entanglement. However, we do have to assume that, under conditions of the experiment, composite systems such as atoms and molecules may be regarded as single particles. This assumption is the basis of modern atom and molecule interferometry, but it lacks justification from the first principles of quantum mechanics. In our world-view the welcher Weg problem merely highlights this fact. The empirical validity of this assumption seems to reveal an aspect of quantum mechanics which can justly be termed anti-reductionist, and we conclude that use of unnecessarily strong hypotheses to explain observed phenomena can hide important physical -- and occasionally philosophical -- problems.