Perturbative General Relativity and Gauge Invariance

Perturbative general relativity turns a nonlinear geometric theory into a controlled hierarchy without mistaking coordinate freedom for dynamics. This chapter follows one chain: vary the geometric equations, derive the gauge action, choose a representative without losing constraints, construct invariants, recover the variational flux structure, and test whether a weakly nonlinear approximation remains controlled in its stated norm and time interval.

Helpful background. Black-Hole Geometry and Mechanics fixes the background and horizon data used in applications; PDEs and Green Functions for Gravitation supplies the differential-operator viewpoint; Formulations of General Relativity separates covariant equations, constraints, and evolution; Presymplectic Potentials and Currents prepares the field-space geometry behind canonical energy.

From geometric variation to weak nonlinearity

Write a family of metrics as

$$g_{\mu\nu}(\varepsilon) =\bar g_{\mu\nu}+\varepsilon h_{\mu\nu} +\frac12\varepsilon^2k_{\mu\nu}+O(\varepsilon^3).$$

This innocent-looking series contains three separate structures. The field equation determines $h$ through a linear differential operator and determines $k$ through the same operator with a quadratic source. Diffeomorphism covariance makes some changes of $h$ and $k$ changes of identification rather than new geometry. Boundary and initial conditions select a physical solution and decide whether a gauge vector is redundant or carries a boundary charge.

The durable problem of the chapter is therefore not merely “expand Einstein’s equation.” It is to preserve the relationships among equations, identities, constraints, gauge transformations, observables, and flux as the expansion is truncated. The linearized Bianchi and matter identities check the equation hierarchy; gauge-invariant quantities check the identification; symplectic flux checks the boundary; and field-norm ordering, constraints, and invariant diagnostics test the approximation. An observable that first appears at second order must not be compared with an identically zero first-order coefficient.

Six steps through the chapter

The pages form one cumulative calculation and are best read in this order.

1. Linearizing Geometry and the Einstein Tensor derives the connection difference, curvature variation, Einstein tensor, a complete scalar matter example, and the coupled Noether identity without assuming a flat or vacuum background.
2. Infinitesimal Diffeomorphisms and Gauge Transformations derives first- and second-order transformation laws from identification maps, computes the ordered composition bracket, and translates active, passive, and inverse-map conventions.
3. Gauge Conditions, Residual Freedom, and Constraints compares Lorenz, Regge–Wheeler, light-cone, horizon-adapted, and radiation gauges by domain, residual equation, and compatibility with constraints.
4. Gauge-Invariant Observables and Master Variables builds invariants by cancellation, relates them to relational measurements, and states the harmonic and reconstruction domains of master variables.
5. Quadratic Actions, Symplectic Currents, and Canonical Energy derives linear dynamics from the action Hessian and follows boundary flux, gauge degeneracy, and the hypotheses behind a canonical-energy criterion.
6. Second-Order Perturbations and Mode Coupling constructs the sourced equation, harmonic selection rules, secular diagnostic, and the evidence boundary for nonlinear ringdown components.

The first four pages establish what the perturbation is. The fifth supplies its phase-space pairing and energy diagnostic. The sixth reuses every earlier distinction: a quadratic source is meaningful only after the first-order gauge, constraints, boundary conditions, and observable have been controlled.

Two reader routes

For a covariant foundations route, read all six pages in order. This is the right path if you need to translate source conventions, derive a perturbation operator on a new background, or decide whether a claimed scalar is truly invariant. Pause after the fourth page and verify that you can turn a coordinate-labeled curvature change into a relational one; then use the fifth page to understand which gauge directions survive at a boundary.

For a black-hole calculation route, read linearization and gauge transformations, move directly to gauge invariants and master variables, then return to gauge conditions when choosing a reconstruction. Finish with the quadratic-action page if flux or stability matters, or the second-order page if mode coupling matters. This route is efficient for readers proceeding to Schwarzschild master equations or Kerr curvature methods, but it should not skip the low-multipole and boundary qualifications.

A third useful path is diagnostic rather than sequential. When two papers seem to disagree, consult the diffeomorphism page for map signs, the gauge page for residual freedom, the invariant page for normalization and domain, and the second-order page for background drift and fitting choices. Many apparent contradictions disappear before any new calculation is required.

Recurring examples and checks

Four examples recur because each exposes a different failure mode.

Pure gauge on a curved solution. Set $h_{\mu\nu}=\mathcal L_\xi\bar g_{\mu\nu}$ and, when matter is present, $\delta\psi=\mathcal L_\xi\bar\psi$. Covariance requires the linearized equation to equal the Lie derivative of the background equation and hence vanish on shell. This tests curvature commutators more strongly than a flat-space limit. For arbitrary, not necessarily on-shell scalar perturbations, the separate divergence identity is $\bar\nabla^\mu E^{(1)}_{\mu\nu}=-8\pi G K^{(1)} \bar\nabla_\nu\bar\phi$, with $K^{(1)}$ the linearized scalar equation. For vacuum with a cosmological constant it is the divergence of $G^{(1)}+\Lambda h$, not necessarily of $G^{(1)}$ alone, that vanishes.

Schwarzschild parameter and radiative sectors. The mass shift lives in the monopole, an angular-momentum shift in an odd dipole, while the familiar radiative master fields begin at $\ell=2$. A formula that divides by an $\ell=0$ or $\ell=1$ factor has not proved those sectors absent; it has left its domain.

Horizon-crossing perturbations. A component singular in Schwarzschild coordinates may be regular in an ingoing chart. Gauge vectors, symplectic flux, and quadratic sources must therefore be checked in a regular frame at the future horizon, with its orientation and generator normalization fixed. The residual Lorenz example $\chi=A\log(1-2GM/r)$ solves the exterior scalar wave equation but generates a vector with divergent invariant norm at the horizon; a local residual equation is not enough.

A damped parent mode. At second order its self-product supplies a sum frequency, while its product with its complex conjugate decays without oscillating rather than remaining constant. Absorption can shift the background mass. The daughter response may contain forced frequencies, homogeneous resonances, and secular or drift terms. Separating them requires a gauge-invariant observable and a declared projection and time window, not just a fit to one component.

The identification-map calculus and gauge-invariance criterion are developed in Bruni et al. 1997, arXiv v2, p. 6, Eq. (2.13), and pp. 12–14, § 4, PDF; the individual pages above work through the geometric and boundary checks.

Distinctions that prevent wrong inferences

A gauge transformation changes the identification; a physical perturbation changes a relational quantity or charge. A gauge condition selects a representative; an Einstein constraint restricts admissible data. Constraint propagation does not finish gauge fixing, and a residual gauge equation does not replace the constraints.

A gauge-invariant variable is not automatically a detector observable. It may depend on a tetrad, harmonic normalization, asymptotic frame, or relational marker. A master variable is a compressed dynamical coordinate whose physical meaning is established by reconstruction or flux.

The presymplectic form is antisymmetric and may be degenerate; it is not a positive norm. Canonical energy pairs a perturbation with its stationary time translate and becomes a stability diagnostic only on a specified constrained phase space with boundary terms controlled. The relevant construction and ambiguities are given in Iyer and Wald 1994, §§II–III, while the black-hole criterion has the narrower hypotheses stated by Hollands and Wald 2013, §§2–4.

Finally, a quadratic forcing frequency is not automatically a new free mode, and a secular term is not automatically an instability. Both conclusions depend on the operator domain, initial data, boundary conditions, gauge, observable, and the time interval over which the asymptotic ordering holds.

Scope and onward use

After completing either main route, you should be able to linearize a metric–matter equation with the site’s curvature signs, derive its pure-gauge solutions, test a gauge condition for residual freedom, construct a simple relational invariant, state when a symplectic integral is conserved, and diagnose the breakdown of a second-order expansion. Those are the inputs for the later chapters on test fields, spherical perturbations, master equations, Kerr curvature methods, response, spectra, and stability.

Exact background construction remains in Exact Solutions of Einstein’s Equations; this chapter assumes a smooth background with its coordinate domain and horizon structure already fixed. Full finite-amplitude evolution and nonlinear endpoints remain in Black-Hole Dynamics. The chapter develops controlled linear and weakly nonlinear response, not a substitute for solving the complete merger or stability problem.

Chapter review

Consider a Ricci-flat stationary black hole and a first-order vacuum perturbation $h$ in Lorenz gauge. A second calculation differs by $\mathcal L_\xi\bar g$, where $\bar\Box\xi_\mu=0$. Both calculations produce the same radiative master field, but their second-order metric components and surface integrals differ. Explain what can be concluded and what must still be checked.

Solution

The homogeneous vector equation shows that both representatives remain in Lorenz gauge; it does not show that $\xi$ is admissible. One must compare its initial data, regularity at the future horizon, falloff at infinity, and action on every boundary. Agreement of the radiative master field is strong evidence that the two first-order metrics describe the same radiative degree of freedom, provided the master definition covers the harmonic sector. It says nothing by itself about monopole or dipole parameter shifts.

At second order the transformation is not just $k\mapsto k+\mathcal L_{\xi_2}\bar g$. It also contains $2\mathcal L_\xi h+\mathcal L_\xi^2\bar g$, so comparing components without these terms is meaningless. The background mass and spin convention must also be matched. For the surface integrals, evaluate the symplectic or charge boundary term. If $\xi$ vanishes in the required sense, the difference is a gauge degeneracy; if it approaches a boundary symmetry, it may change a physical charge. Finally, compare a second-order invariant flux and verify field-norm ordering, using each observable’s first nonzero order or a nonzero physical scale. Flux need not have a nonzero first-order term. Only then can the calculations be called physically equivalent to the claimed order.

References

• Bruni, Marco, Sabino Matarrese, Silvia Mollerach, and Sebastiano Sonego. “Perturbations of Spacetime: Gauge Transformations and Gauge Invariance at Second Order and Beyond.” Classical and Quantum Gravity 14 (1997): 2585–2606. DOI. Open PDF, arXiv:gr-qc/9609040v2.
• Hollands, Stefan, and Robert M. Wald. “Stability of Black Holes and Black Branes.” Communications in Mathematical Physics 321 (2013): 629–680. DOI. Open PDF.
• Iyer, Vivek, and Robert M. Wald. “Some Properties of Noether Charge and a Proposal for Dynamical Black Hole Entropy.” Physical Review D 50 (1994): 846–864. DOI. Open PDF.