In 1927 Einstein 1 presented a precursor of the much better known argument of Einstein-Podolsky-Rosen ͑EPR͒ 2 that the quantum mechanical description of physical reality is either incomplete or violates locality by tacitly assuming superluminal causality in the measurement process. He proposed a thought experiment in which the support of the wave function of a ball is contained within two boxes, B 1 and B 2 , that are carried arbitrarily far from each other. When a measurement is performed to detect whether the ball is in B 1 , not only is a positive or negative answer found regarding that box, but simultaneously a negative or positive answer is obtained regarding box B 2 . He concluded that either the wave function is an incomplete description of the ball, requiring supplementation by additional information concerning its location, or else an experimental intervention in B 1 immediately causes a determination of its presence or absence in B 2 , which would constitute non-local causation.Norsen 3 has deplored the neglect of Einstein's ''box'' argument, claiming not only that it is valid, but is superior to EPR because of its simplicity: it involves only a single system, considers only the position of that system rather than a pair of non-commuting properties, and above all avoids the need to invoke counterfactual reasoning. ͑EPR not only infers the position of particle I from the position of particle II, or vice versa if position is measured on one of the particles, but they infer the momentum of one particle from the momentum of the other if momentum is measured on either.͒ However, there is a serious flaw in Norsen's defense of the box argument in my opinion. If Q 1 is the projection operator representing the physical proposition that the ball is in B 1 and Q 2 is the projection operator corresponding to the ball's location in B 2 , then the logical structure of the lattice of projections in the quantum mechanics of localizable systems ensures that Q 1 and Q 2 are orthogonal, that is,independently of the quantum state of the ball. Hence any state that is an eigenstate of Q 1 with eigenvalue 1 or 0 will also be an eigenstate of Q 2 with respective eigenvalue 0 or 1. ͑The relation of Q 1 and Q 2 to B 1 and B 2 , which is accepted intuitively by physicists, is generalized and presented with mathematical rigor in Mackey's discussion 4 of projectionvalued measures on Borel spaces.͒ Because Eq. ͑1͒ is independent of the quantum state of the ball, the inference from a measurement yielding the location or non-location of the ball in B 1 to the respective non-location or location in B 2 does not rely on any contingencies, and therefore it is fair to say that it is a matter of logic rather than of causality. Hence if one rejects Einstein's contention that quantum mechanics gives an incomplete description of the ball, in the sense that its location is definitely in B 1 or B 2 even when the wave function does not vanish identically in either box, one is not forced to accept non-local causality as the only alternative. The ...