Solution to the ghost problem in higher-derivative gravity (2109.12743v1)

Abstract: With standard Einstein gravity not being renormalizable at the quantum level there is much interest in studying higher-derivative quantum gravity theories. Thus just as a Ricci-scalar-based action produces a propagator that behaves as a non-renormalizable $1/k^2$ at large $k^2$, an action based on the square of the Ricci scalar behaves as a renormalizable $1/k^4$ at large $k^2$. An action based on both the Ricci scalar and its square leads to a renormalizable propagator of the generic Pauli-Villars form. However, given the form of the Hamiltonian and the propagator such theories are thought to be plagued by either energies that are unbounded from below or states of negative Dirac norm (the overlap of a ket with its Hermitian conjugate bra). But when one constructs the quantum Hilbert space one finds (Bender and Mannheim) that in fact neither of these problems is actually present. The Hamiltonian turns out to not be Hermitian but to instead have an antilinear $PT$ symmetry, and for this symmetry the needed inner product is the overlap of a ket with its $PT$ conjugate bra. And this inner product is positive definite. Moreover, for the pure $1/k^4$ propagator the Hamiltonian turns out to not be diagonalizable, and again there are no states of negative energy or of negative norm. Instead there are states of zero norm, non-standard but perfectly acceptable states that serve to maintain probability conservation. With the locally conformal invariant fourth-order derivative conformal gravity theory being in this category, it can be offered as a candidate theory of quantum gravity that is renormalizable and unitary in four spacetime dimensions.