Deformations of GR, Geometrodynamics and Reality Conditions

TL;DR

This paper investigates complexified deformations of four-dimensional General Relativity, focusing on their Hamiltonian structure, reality conditions, and the implications for geometrodynamics and metric reconstruction.

Contribution

It introduces a four-parameter class of deformations extending Ashtekar's formalism and analyzes their impact on the closure of evolution equations and reality conditions.

Findings

Modified theories generally do not close in metric variables.

Reality conditions are not preserved under dynamics in Lorentzian signature.

Only standard GR and Self-Dual Gravity maintain metric-only evolution equations.

Abstract

In four dimensions complexified General Relativity (GR) can be non-trivially deformed: There exists an (infinite-parameter) set of modifications all having the same count of degrees of freedom. It is trivial to impose reality conditions that give versions of the deformed theories corresponding to Riemannian and split metric signatures. We revisit the Lorentzian signature case. To make the problem tractable, we restrict our attention to a four-parameter set of deformations that are natural extensions of Ashtekar's Hamiltonian formalism for GR. The Hamiltonian of the later is a linear combination of $EEE$ and $EEB$. We consider theories for which the Hamiltonian constraint is a general linear combination of $EEE, EEB, EBB$ and $BBB$. Our main result is the computation of the evolution equations for the modified theories as geometrodynamics evolution equations for the 3-metric. We show that only for GR (and the related theory of Self-Dual Gravity) these equations close in the sense that they can be written in terms of only the metric and its first time derivative. Modified theories are therefore seen to be essentially non-metric in the sense that their dynamics cannot be reduced to geometrodynamics. We then show this to be related to the problem with Lorentzian reality conditions: the conditions of reality of the 3-metric and its time derivative are not acceptable because they are not preserved by the dynamics. Put differently, their conservation implies extra reality conditions on higher-order time derivatives, which then leaves no room for degrees of freedom.