Generalized Redshift Invariant in Relativity

• Generalized redshift invariant is defined as the proper-time average of dt/dτ over a radial libration in extreme-mass-ratio inspirals.
• It provides a robust gauge-invariant observable that bridges black-hole perturbation theory with post-Newtonian and effective-one-body approaches.
• The invariant’s formulations extend to differential geometry, modified dispersion relations, and Finsler spacetimes, highlighting its versatile applications.

Searching arXiv for recent and foundational papers on “generalized redshift invariant” and closely related formulations. The term generalized redshift invariant has a precise meaning in gravitational self-force theory and a broader family of related meanings in differential geometry, modified dispersion relations, and generalized kinematics. In the eccentric-orbit extreme-mass-ratio-inspiral literature, it denotes the proper-time average of $u^t=dt/d\tau$ over one radial libration,

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$

a central conservative-sector gauge invariant used for cross-checks between black-hole perturbation theory and post-Newtonian theory, for the first law of binary mechanics, and for calibrating effective-one-body potentials (Munna et al., 2022). In other settings, the phrase is used more loosely for generalized redshift laws, conserved quantities controlling redshift, or redshift potentials; several papers state explicitly that they do not define a new invariant in the strict sense (Hasse et al., 2019).

1. Terminological scope

The phrase does not have a single uniform meaning across the literature. In some works it denotes a specific gauge-invariant scalar attached to orbital dynamics; in others it denotes a redshift potential, a conserved quantity along null rays, or only a generalized redshift law.

Context	Quantity	Status
Eccentric GSF/EMRI theory	$\langle u^t\rangle_\tau=T_r/\mathcal T_r$	Explicit conservative gauge invariant
Contact geometry of null rays	$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$	Exact geometric reformulation
Finsler spacetime	$\omega=-p_\mu u^\mu$, and when applicable $K^\mu p_\mu$ with potential $f$	Generalized formula / redshift potential
Homogeneous-isotropic MDR cosmology	$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$	Inferred first-order generalized invariant
Generalized Lorentz transformations	$(ct)^2-r^2$	Interval invariant; modified redshift law

This multiplicity is not accidental. The self-force literature isolates a conservative observable of bound motion, whereas the geometric and cosmological literatures ask whether redshift can be written as an endpoint ratio, a contact-form scaling, or a deformed conservation law. A recurring distinction is therefore between an invariant quantity and a generalized redshift formula.

2. Conservative invariant in gravitational self-force theory

For circular orbits, Detweiler’s invariant is the instantaneous quantity

$u^t=\frac{dt}{d\tau}.$

For eccentric motion, $$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$0 is not constant along the orbit, so Barack and Sago generalized it to the proper-time average over one radial libration,

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$1

Here $$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$2 is the radial period measured in Schwarzschild coordinate time $$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$3, and $$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$4 is the radial period measured in proper time $$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$5. The first-order decomposition is

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$6

with

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$7

The regularized invariant is assembled mode by mode from

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$8

with leading regularization parameter

$$\left\langle u^t \right\rangle_\tau=\frac{1}{\mathcal T_r}\int_0^{\mathcal T_r}u^t\,d\tau=\frac{T_r}{\mathcal T_r},$$9

The gauge-invariant prescription is to hold the observable radial libration frequency fixed when comparing background and perturbed motion; then the needed first-order information is entirely contained in $\langle u^t\rangle_\tau=T_r/\mathcal T_r$0 (Munna et al., 2022).

The orbital parametrization is by Darwin variables $\langle u^t\rangle_\tau=T_r/\mathcal T_r$1, with

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$2

specific energy and angular momentum

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$3

and gauge-invariant frequency variable

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$4

The same invariant is important because it is a clean conservative gauge invariant computable in both black-hole perturbation theory and post-Newtonian theory; it enters the first law of binary mechanics for eccentric orbits, maps directly to the EOB nongeodesic potential $\langle u^t\rangle_\tau=T_r/\mathcal T_r$5, and helps calibrate analytic models of extreme-mass-ratio inspirals and strong-field conservative dynamics (Munna et al., 2022).

3. High-order post-Newtonian structure

For eccentric nonspinning motion on Schwarzschild, the first-order self-force correction admits dual expansions in PN order and eccentricity. One explicit form is

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$6

and equivalently

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$7

The calculation reaches 10PN relative order and $\langle u^t\rangle_\tau=T_r/\mathcal T_r$8. A central structural result is that the first logarithm is directly proportional to the Peters-Mathews quadrupole flux enhancement,

$\langle u^t\rangle_\tau=T_r/\mathcal T_r$9

with

$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$0

The 5.5PN conservative term inherits the same eccentricity dependence as the 1.5PN tail enhancement in the dissipative energy flux,

$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$1

and the 7PN $\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$2 term is tied to the 3PN log energy-flux function,

$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$3

The paper emphasizes that identifying the correct eccentricity singular factor at each PN order is crucial for recognizing closed forms, identifying inherited flux functions, and improving convergence of truncated $\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$4 series (Munna et al., 2022).

For eccentric equatorial motion about a Kerr black hole, the same invariant was extended with the dependence on the primary spin parameter $\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$5 kept exact. The first-order correction is expanded as

$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$6

through 6PN relative order with terms to $\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$7, and separately through 8PN relative order with terms to $\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$8. The computation uses the Teukolsky formalism, CCK reconstruction in ingoing radiation gauge, MST for low multipoles, and a direct general-$\alpha_{M'}/\alpha_M=1+z(M,M',\gamma)$9 expansion for the tail of the mode sum. The resulting coefficient functions again exhibit closed forms in eccentricity, but now organized by spin sectors and, at higher order, involving $\omega=-p_\mu u^\mu$0 and polygamma structures (Munna, 2023).

A different but related extension concerns a spinning small body on a circular equatorial orbit in Schwarzschild, treated to linear order in the particle’s spin and first order in the mass ratio. In that setting the gauge-invariant spin-linear first-order self-force correction is not just a regularized $\omega=-p_\mu u^\mu$1, but the specific combination

$\omega=-p_\mu u^\mu$2

where the extra term $\omega=-p_\mu u^\mu$3 is crucial for gauge invariance because the spinning particle follows an accelerated, nongeodesic circular orbit. The result was computed through 8.5PN and shown to agree between Regge-Wheeler gauge and radiation gauge (Bini et al., 2018).

4. Geometric reformulations

In globally hyperbolic Lorentzian geometry, redshift can be reformulated on the space of light rays $\omega=-p_\mu u^\mu$4. For each spacelike Cauchy surface $\omega=-p_\mu u^\mu$5, the identification of a light ray with a unit covector on $\omega=-p_\mu u^\mu$6 induces a contact form

$\omega=-p_\mu u^\mu$7

and for two Cauchy surfaces $\omega=-p_\mu u^\mu$8 and $\omega=-p_\mu u^\mu$9 the main theorem is

$K^\mu p_\mu$0

Equivalently,

$K^\mu p_\mu$1

In this setting the geometric invariant underlying redshift is the canonical co-oriented contact structure on $K^\mu p_\mu$2, while the redshift factor is the multiplicative transition factor between contact forms induced by different Cauchy surfaces (Chernov et al., 2017).

A different geometric reformulation concerns the comparison between static black-hole and cosmological redshift. For an observer with four-velocity $K^\mu p_\mu$3, the measured frequency is

$K^\mu p_\mu$4

and this is the invariant core of the observable. In static spherically symmetric spacetimes, one may pass to a synchronous free-fall frame, where the transport laws become

$K^\mu p_\mu$5

with corresponding endpoint formulas

$K^\mu p_\mu$6

Returning to the static frame by a local Lorentz boost restores the usual lapse-based formula. The paper’s central claim is therefore that the observable redshift is invariant, whereas its decomposition into gravitational, cosmological, and Doppler pieces is frame-dependent (Toporensky et al., 2017).

In Finsler spacetime, the general redshift formula keeps the same pairing structure,

$K^\mu p_\mu$7

so that

$K^\mu p_\mu$8

When a suitable symmetry exists, a conserved quantity

$K^\mu p_\mu$9

is associated with a Finsler conformal Killing field $f$0. If

$f$1

for the observer congruence $f$2, then

$f$3

and $f$4 is a redshift potential. In the cosmological example this role is played by

$f$5

so that $f$6. The paper is explicit that the generalized redshift invariant is not a single universal scalar valid in every Finsler spacetime; rather, the invariant content is the frequency pairing $f$7, and, when symmetry permits, the conserved quantity $f$8 and its associated potential (Hasse et al., 2019).

5. Generalized redshift laws beyond standard relativistic kinematics

For homogeneous and isotropic modified dispersion relations, the Hamiltonian

$f$9

deforms the standard FLRW photon relation. The first-order redshift formula becomes

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$0

The nearest equivalent to a generalized redshift invariant is the conserved combination

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$1

for massless particles, which is directly implied by the perturbed mass-shell relation (Pfeifer, 2018).

In generalized Lorentz transformations with a modified boost parameter $$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$2, the explicit invariant is not a redshift scalar but the ordinary Minkowski interval,

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$3

The Doppler relation is modified by the substitutions $$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$4 and $$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$5, leading to

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$6

The paper states explicitly that it does not derive a new invariant involving redshift. The invariant is the spacetime interval; the redshift is a modified observable, and in the one-sided mass-dependent proposal it may depend on the observer’s mass (Gupta, 2013).

A more radical proposal arises from spinning-photon dynamics in Robertson-Walker spacetime. Using the Souriau-Saturnini equations, the paper defines redshift through the spin-precession or helix period rather than the atomic period, obtaining

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$7

The authors are explicit that this is not a new exact invariant in the usual geometric sense, but an invariant-like ratio of a spin-induced intrinsic period (Duval et al., 2018).

A still milder generalization is Baldry’s logarithmic wavelength-shift variable

$$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$8

Its importance is compositional rather than invariant-theoretic: $$a(t)(-p_t)-\epsilon \frac{a(t)^2}{2w}h=\text{const.}=w$$9 For pure radial motion,

$(ct)^2-r^2$0

so $(ct)^2-r^2$1 is rapidity-like and, for one-dimensional motion, exactly the rapidity itself. The paper nevertheless states that $(ct)^2-r^2$2 is not a true invariant in the strict relativistic sense; it is a more natural generalized redshift variable because it linearizes the physically correct multiplicative structure of redshift effects (Baldry, 2018).

6. Nonexistence results, redshift drift, and controversies

In generic $(ct)^2-r^2$3 Szekeres cosmologies, there is no simple generalized redshift invariant analogous to the FLRW law $(ct)^2-r^2$4. The paper derives instead a coupled propagation system for the time-lag variable $(ct)^2-r^2$5 and transverse separation variables $(ct)^2-r^2$6, together with the constraint

$(ct)^2-r^2$7

Its principal conclusion is negative: in a general Szekeres model generic light rays do not have repeatable paths, and the only Szekeres spacetimes in which all rays are repeatable light paths are the Friedmann models. In proper Szekeres models, repeatable paths exist only in axially symmetric subcases, and in $(ct)^2-r^2$8 models, including Lemaître-Tolman, only the radial geodesics are repeatable (Krasiński et al., 2010).

For redshift drift in arbitrary spacetime, the exact general formula is

$(ct)^2-r^2$9

The line-of-sight contribution admits an exact finite multipole expansion and is invariant under affine-parameter rescalings of the null congruence, but the paper’s main conclusion is again restrictive: redshift drift cannot in general be thought of as a direct probe of the average expansion rate of the Universe, because structure along the beam contributes monopole, dipole, quadrupole, and octupole terms. It also predicts the general presence of a dipolar and a quadrupolar offset for observers placed in locally anisotropic environments (Heinesen, 2020).

The literature also contains explicit methodological disputes. One 2025 paper argues against a purported generalized redshift formula based only on energy conservation and insists that a correct treatment of Doppler, gravitational, and cosmological redshift requires both energy and momentum conservation. Its own conclusion is that it does not provide a single invariant formula such as $u^t=\frac{dt}{d\tau}.$0, but only a quasi-unifying conservation principle built from

$u^t=\frac{dt}{d\tau}.$1

The same paper states that it does not derive a rigorous covariant generalized redshift invariant in the modern relativistic sense (Wilhelm et al., 31 Jan 2025).

Observationally, the most explicit test among these generalized proposals concerns the exotic spinning-photon cosmological redshift. When fitted to the supernova Hubble diagram, the paper finds that supernova standardization is worse, that intrinsic magnitude dispersion increases from about $u^t=\frac{dt}{d\tau}.$2 to $u^t=\frac{dt}{d\tau}.$3, that the best exotic fit prefers

$u^t=\frac{dt}{d\tau}.$4

with

$u^t=\frac{dt}{d\tau}.$5

and concludes that the exotic redshift is disfavored at at least $u^t=\frac{dt}{d\tau}.$6 confidence level (Duval et al., 2018).

Across these literatures, a stable pattern emerges. In the strictest sense, a generalized redshift invariant is an observer-adapted scalar or conserved quantity with a clear geometric definition, such as $u^t=\frac{dt}{d\tau}.$7 in GSF theory, $u^t=\frac{dt}{d\tau}.$8 on the space of light rays, or $u^t=\frac{dt}{d\tau}.$9 in Finsler spacetime. In looser usage, the phrase refers to generalized redshift laws, deformed conservation laws, or reparameterizations of redshift. The distinction between those two usages is itself one of the topic’s central technical issues.