Inequalities for complementarity in observed statistics
We provide an analysis of complementarity via a suitably designed classicalmodel that leads to a set of inequalities that can be tested by means ofunsharp measurements. We show that, if the measured statistics does notfulfill the inequalities it is equivalent to the lack of a joint distribution for theincompatible observables. This is illustrated by path-interference duality ina Young interferometer.