On the Compatibility of Quantum Matter and Classical Gravity

The formulation of a consistent scheme for treating semiclassical systems is considered, giving particular emphasis to the case of gravity. A critical review of the semiclassical problem is first given, stressing the conceptual aspects of the topic; the theory usually adopted is found to be unsatisfactory, and the need to have an unambiguous interpretation of quantum mechanics before trying to give an alternative description is realized. Our way to arrive at such an interpretation is rather unusual, but it presents the advantage of being almost compelling in the choice to make: First, we reformulate the Schrodinger equation as a set of hydrodynamical equations involving quantities which formally play the role of mass density, velocity and pressure for a fluid; the problem of interpreting the wave function is thus reduced to that of interpreting these quantities. We then show how they can be derived from the Wigner distribution function exactly as in the usual formalism of kinetic theory, and discuss how this fact provides strong support in favour of the statistical interpretation of quantum theory, according to which the state vector describes only ensembles, and not individual systems. We also consider some implications of these results on the possible existence of a more fundamental theory underlying quantum mechanics. It is shown that reconsidering the semiclassical problem in the light of the statistical interpretation, one is led to distinguish between a strongly and a weakly semiclassical regime, which are essentially characterized by the size of the statistical dispersion induced in the observables of the classical subsystem by the coupling to the quantum one. It turns out that in the weakly semiclassical regime, in which this dispersion is not negligible, the concept of coupling equations cannot be successfully applied, and one has rather to define a probability distribution even for the values of the classical observables; an hypothesis which allows to specify such a distribution is enunciated. Several examples of the application of these general principles are considered, and it is shown how the treatment of semiclassical relativistic fields requires a much more sophisticated treatment of the quantum source. This is provided reformulating quantum theory in terms of a quasiprobability functional P[1] in the space of the histories of the system. It is shown how such a functional allows to reconstruct the usual phase space distributions when integrated over suitable sets of paths, in a way which clarify the relations between operator ordering, path integration and phase space treatment of quantum theory. The relativistic extension of p[f] is also constructed, and an explicitly covariant version of relativistic quantum theory is discussed in some details. It is shown how the latter allows, formally, to consider superpositions of different mass eigenstates, although such superpositions are not directly observable. Finally, the application of these new techniques to the treatment of semiclassical electromagnetism and gravity, as well as of a scalar field, are considered. It is shown how the usual semiclassical field equations, suitably reinterpreted in terms of averages of the field, are recovered either in linear cases or in the strongly semiclassical regime, but that they do not hold in general. Finally, some possible extensions and implications of the formalism are discussed.