On Protective Measurements in Quantum Mechanics

Recently proposed idea of ``protective'' measurement of a quantum state is critically examined, and generalized.Earlier criticisms of the idea are discussed and their relevance to the proposal assessed.Several constraints on measuring apparatus required by ``protective'' measurements are discussed, with emphasis on how they may restrict their experimental feasibility.Though ``protective'' measurements result in an unchanged system state and a shift of the pointer proportional to the expectation value of the measured observable in the system state, the actual reading of the pointer position gives rise to several subtleties.We propose several schemes for reading pointer position, both when the apparatus is treated as a classical system as well as when its quantum aspects are taken into account, that address these issues.The tiny entanglement which is always present due to deviation from extreme adiabaticity in realistic situations is argued to be the weakest aspect of the proposal.Because of this, one can never perform a protective measurement on a single quantum system with absolute certainty.This clearly precludes an ontological status for the wave function.Several other conceptual issues are also discussed.