We examine a gedanken experiment described by Hardy, which purports to prove that the ‘‘elements of reality’’ corresponding to Lorentz-invariant observables cannot themselves be Lorentz invariant without violating quantum mechanics. We consider a number of criticisms of this proof and show that these criticisms are not convincing. We demonstrate that the contradiction, which arises in the gedanken experiment and forms the basis of Hardy’s proof, has nothing to do with realism, but is a consequence of the well-known noncovariance of the reduction postulate. A reexamination of the gedanken experiment, using a more appropriate formalism, helps to clarify its implications. We conclude with a brief appraisal of realist interpretations of quantum mechanics in the relativistic domain.

• Received 23 May 1994

DOI:https://doi.org/10.1103/PhysRevA.52.76