Abstract

The general objective of the dissertation is to examine critically the abstract description of measurements in quantum mechanics, looking in particular at the "measurement problem" and at the proofs that this problem cannot be solved using the unitary evolution law of quantum mechanics--the so-called "insolubility proofs". The critical analysis is performed from two standpoints: experimental and formal. ;The experimental approach considers the way in which concrete measurements are performed in the laboratory. For measurements on quantum systems, this means examining the procedures employed for obtaining eigenvalues and for determining quantum states. We argue that all measurements involve both a determination of position and of either number, intensity, or energy. For measurements on an ensemble of object systems, we develop a classification according to the different ways in which elementary "filter measurements" can be combined. We also give an experimental argument for the so-called "ignorance interpretation" of mixtures. ;The formal part of the dissertation develops a definition of a "measurement operator", which represents the measured observable solely in terms of parameters characterizing the measuring apparatus and the interaction between apparatus and object. Using this formalism, a new insolubility proof is offered which presents two novel features. First, it applies to a more general class of interactions than the "spectral measurements" to which previous proofs applied. Second, one is able to determine precisely which of the postulates that define measurements, unitarity, and solubility, must be taken as premisses of the proof. It turns out that only two of the four principles defining unitarity are needed. This result allows us to study the solubility of "non-unitary" measurements, and to speculate about simple deterministic mechanisms of solubility in open systems