On the initial value problem for semiclassical gravity without and with quantum state collapses

We explore how the initial value problem may be formulated for globally hyperbolic, Hadamard, solutions of the semiclassical Einstein-Klein-Gordon equations. Given a set of data on an initial 3-surface, consisting of the values on the surface of a spacetime metric and its first 3 time derivatives off the surface, we introduce a notion of 'surface Hadamard' state on the CCR algebra of the surface. We conjecture that, for a given such set of classical Cauchy data with a surface Hadamard state satisfying the semiclassical constraint equations, the initial value problem will be well posed. We present similar conjectures for a semiclassical scalars model and semiclassical electrodynamics. Moreover, partly inspired by work of Parker and Simon in 1993, we define semiclassical gravity `physical solutions' to be those that are (jointly smooth) functions of $\hbar$ and of coordinates continuous in $\hbar$ at $\hbar =0$. We conjecture that for such solutions the second and third time derivatives of the metric off the surface need not be specified, but rather will be determined by that continuity condition. Assuming the initial value conjecture for such physical solutions holds, and that a stochastic rule were available which leads to quantum state collapses occurring on (non-intersecting) random Cauchy surfaces, we discuss the well-posedness of semiclassical gravity with stochastic quantum state collapses. We also discuss two notions of approximate physical semiclassical solutions (both with and without collapses): Namely solutions to order $\hbar$ (first discussed by Parker and Simon in 1993) and solutions to order $\hbar^0$. We point out that the latter do not require higher derivative terms or Hadamard subtractions, but that nevertheless order $\hbar^0$ semiclassical gravity is a distinct theory from classical general relativity capable of incorporating quantum interference phenomena.