Non-Coincident Gauge in Gravity and Gauge Theories

• Non-coincident gauge refers to approaches that maintain non-zero connection components in coordinate systems while enforcing flatness and torsion-free conditions, thereby highlighting nonmetricity or non-local holonomy.
• In symmetric teleparallel gravity, these gauges preserve the integrability of the affine connection without defaulting to the coincident gauge, enabling the explicit tracking of inertial and geometric contributions.
• In gauge-Higgs systems and QCD, non-aligned and contour gauges respectively avoid preset global symmetry directions and trivial Wilson line contributions, which is crucial for correctly interpreting physical observables.

“Non-coincident gauge” is not a single universal construction. In the literature it denotes several distinct ways of avoiding a preferred trivialized or aligned representation of gauge or connection data. In symmetric teleparallel gravity, it means working in arbitrary coordinates with a non-vanishing affine connection that is nevertheless flat and torsion-free, so that nonmetricity remains the only nontrivial geometric object. In lattice gauge theory and gauge-Higgs systems, it denotes a non-aligned supplementary gauge condition, typically enforcing $\langle \phi_i\rangle=0$, so that the gauge choice does not coincide with a particular global symmetry-breaking orientation. In QCD, it also refers to the contour gauge, a non-local path-dependent gauge in which a Wilson line between a reference point and the observation point is set to the identity (Jiménez et al., 2022, Maas, 2012, Anikin, 2021).

1. Non-coincident gauge in symmetric teleparallel geometry

In symmetric teleparallel gravity (STG), the affine connection is constrained by

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$

Flatness makes the connection integrable and of pure-gauge form, while torsionlessness makes it symmetric in its lower indices. A general teleparallel connection can be written as

$$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$

with $\Lambda(x)\in GL(4,\mathbb{R})$. Imposing the torsion-free restriction gives $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$, so that the symmetric teleparallel connection becomes

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$

The corresponding global $GL$ symmetry reduces to affine invariance $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$ (Jiménez et al., 2022).

The coincident gauge is any holonomic coordinate system in which

$\Gamma^{\alpha}{}_{\mu\nu}=0.$

For a flat, torsion-free connection, such coordinates exist at least locally; this follows from the complete integrability of the covariant derivative, $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$. A non-coincident gauge is simply any coordinate system in which $$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$0, while the connection still satisfies $$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$1 and $$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$2. The two descriptions are related by the standard coordinate-transformation law

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$3

and, when the starting chart is coincident, this reduces to the Hessian form

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$4

Global trivialization can fail because of topology or coordinate singularities, but the local existence of coincident coordinates is guaranteed (Jiménez et al., 2022).

In STG the nonmetricity tensor is

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$5

with traces

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$6

and disformation

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$7

A standard STG choice for the nonmetricity scalar is

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$8

which satisfies

$$R^{\alpha}{}_{\beta\mu\nu}=0,\qquad T^{\alpha}{}_{\mu\nu}=0.$$9

In coincident gauge, $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$0. In a non-coincident gauge, the explicit connection dependence reappears through

$$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$1

This is the basic covariant object through which non-coincident formulations are organized (Jiménez et al., 2022).

2. Dynamical interpretation and cosmological realizations in $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$2 gravity

A central subtlety is the status of autoparallels. For the affine connection $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$3, they satisfy

$$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$4

In coincident gauge they are straight lines in that coordinate system; in a non-coincident gauge they explicitly involve $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$5. However, for any symmetric connection, minimal coupling in the action leads to metrical coupling in the equations of motion, so standard matter propagates along Levi-Civita trajectories rather than autoparallels of the symmetric teleparallel connection. The coincident gauge is therefore a coordinate choice with no physical significance per se. The same analysis emphasizes that one may either interpret the $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$6 as Stückelberg fields, making the coincident gauge analogous to unitary gauge, or treat them as non-canonical coordinates parametrizing inertial frames; the distinction matters for whether one attributes independent field equations to $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$7. In STEGR, where the action differs from Einstein–Hilbert by a boundary term, non-coincident calculations must track explicit $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$8-dependent terms and total divergences carefully. The same work also stresses that the coincident gauge is compatible with rotational symmetry and FRW-type isometries, and that it must be distinguished from the Weitzenböck gauge (Jiménez et al., 2022).

Recent $$\Gamma^{\alpha}{}_{\mu\beta}=(\Lambda^{-1})^{\alpha}{}_{\lambda}\,\partial_{\mu}\Lambda^{\lambda}{}_{\beta},$$9 cosmology has used explicitly non-trivial flat, torsion-free FLRW connections. One realization adopts

$\Lambda(x)\in GL(4,\mathbb{R})$0

together with a spatially flat FLRW metric

$\Lambda(x)\in GL(4,\mathbb{R})$1

In this gauge,

$\Lambda(x)\in GL(4,\mathbb{R})$2

The resulting FLRW equations can be rewritten through

$\Lambda(x)\in GL(4,\mathbb{R})$3

leading to a minisuperspace Lagrangian

$\Lambda(x)\in GL(4,\mathbb{R})$4

whose cross-kinetic term yields a two-scalar-field, quintom-like representation. For the power-law model $\Lambda(x)\in GL(4,\mathbb{R})$5, the paper derives an analytic equation-of-state parameter $\Lambda(x)\in GL(4,\mathbb{R})$6, finds that $\Lambda(x)\in GL(4,\mathbb{R})$7 produces phantom-divide crossing, and reports $\Lambda(x)\in GL(4,\mathbb{R})$8 from DESI DR2 BAO and gamma-ray bursts. In the combinations including GRBs, the model gives a smaller $\Lambda(x)\in GL(4,\mathbb{R})$9 than $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$0CDM, while AIC indicates weak evidence in favor of the $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$1 theory and BIC still favors $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$2CDM because of the extra parameter (Paliathanasis, 15 Apr 2025).

A different non-coincident formalism uses a torsionless, flat, non-metric-compatible FLRW connection with a time-dependent function $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$3, and specifically chooses

$$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$4

In that setting,

$$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$5

so the background equations become

$$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$6

$$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$7

with $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$8. These equations are reported to be entirely distinct from those of $$\Lambda^{\alpha}{}_{\beta}=\partial_{\beta}\xi^{\alpha}$$9 theory. The same work introduces the phenomenological parameterization

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$0

and, from CC, Pantheon+SH0ES, and BAO, reports

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$1

with $$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$2. Around STEGR, the paper further studies exponential, logarithmic, and power-law corrections, finding NEC-compatible and late-time SEC-violating parameter windows for several models, while the quadratic correction $$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$3 is reported as thermodynamically unstable in the explored range (Pradhan et al., 2024).

A later dynamical-systems treatment formulates spatially flat FLRW $$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$4 cosmology in a connection-agnostic way by introducing the Hubble-normalized variables

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$5

together with

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$6

For the non-coincident branches $$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$7 and $$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$8, the analysis finds generic de Sitter attractors and matter-dominated points, identifies the invariant submanifold

$$\Gamma^{\alpha}{}_{\mu\beta}=\frac{\partial x^{\alpha}}{\partial \xi^{\lambda}}\,\partial_{\mu}\partial_{\beta}\xi^{\lambda}.$$9

and shows that on this submanifold the background kinematics are exactly $GL$0CDM-like,

$GL$1

even though the dynamics differ from GR. The same framework reconstructs the dynamical connection through a first integral and emphasizes the observational role of

$GL$2

For $GL$3, the non-coincident branches are reported to admit late-time acceleration and $GL$4CDM-like behavior without vacuum energy, while the coincident branch can fail to provide a viable matter fixed point for the same model (Dutta et al., 13 Aug 2025).

3. Non-aligned gauge in gauge-Higgs theory, color superconductivity, and BRST

In gauge-Higgs systems the term “non-coincident gauge” is used for a non-aligned supplementary gauge condition imposed after a local gauge choice that is invariant under global gauge transformations. The aligned prescription fixes a global direction in internal space,

$GL$5

whereas the non-aligned prescription enforces

$GL$6

The rationale is explicitly tied to the statement that local gauge symmetry cannot spontaneously break, that apparent symmetry breaking in non-gauge-invariant correlators is a gauge artifact, and that gauges selecting a global direction are themselves explicit hidings of the global gauge symmetry. The non-aligned construction therefore “hide[s] as little symmetry as possible” (Maas, 2012).

On the lattice, the procedure is two-step. One first fixes the local gauge degrees of freedom with a condition that does not refer to the global gauge freedom, such as minimal Landau gauge. One then omits any global alignment step, or averages observables over many global gauge transformations, or picks a single random global representative. Because no global weight factor is present, a random representative suffices and yields $GL$7 in the infinite-volume limit. By contrast, aligned gauges rotate the configuration-wise spatial average of the Higgs field into a fixed internal direction, reproducing the global content of the Landau-gauge limit of ’t Hooft gauge (Maas, 2012).

This choice affects gauge-dependent correlators but not gauge-invariant physics. In aligned gauges, the connected Higgs two-point function decomposes into longitudinal and transverse parts relative to the chosen direction,

$GL$8

In a non-aligned gauge, the correlator is diagonal with a single dressing function,

$GL$9

and vertices simplify even more. The same work emphasizes that in a non-aligned gauge “the conventional perturbative expansion around the classical Higgs mean field in a semi-classical way will not be possible,” because $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$0. Gauge-invariant diagnostics remain meaningful; in particular,

$$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$1

continues to distinguish the would-be Higgs and confinement regions (Maas, 2012).

The same framework is extended to high-density color superconductors and BRST symmetry. In full QCD, colored condensates are gauge-dependent; in a suitable non-aligned gauge such colored condensates vanish. For BRST beyond perturbation theory, the non-aligned choice is the Landau–Hirschfeld–Neuberger–von Smekal construction, which averages over all Gribov copies and restores global BRST symmetry. Selecting only a subset of copies, such as in minimal Landau gauge, can instead produce apparent symmetry breaking in non-invariant correlators (Maas, 2012).

4. Contour gauge as a non-local non-coincident gauge

In QCD and related gauge-theory applications, the contour gauge is presented as a non-local, path-dependent, generalized axial gauge, and also as a non-coincident gauge. For a fixed contour $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$2, the Wilson line is

$$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$3

Choosing a reference point $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$4 and a family of contours $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$5, the contour gauge is defined by

$$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$6

for all $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$7 along the chosen contour family. The constraint is non-local because it fixes the holonomy from $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$8 to $$\xi^{\alpha}\to M^{\alpha}{}_{\beta}\xi^{\beta}+\xi^{\alpha}_{0}$$9, rather than imposing a local condition on $\Gamma^{\alpha}{}_{\mu\nu}=0.$0 at a single point. Geometrically, the gauge picks out the horizontal plane $\Gamma^{\alpha}{}_{\mu\nu}=0.$1 on the principal fiber bundle (Anikin, 2021).

The contour gauge generalizes local axial gauges while removing their residual gauge freedom. In a local light-cone gauge $\Gamma^{\alpha}{}_{\mu\nu}=0.$2, residual transformations remain because the gauge function is only determined up to an arbitrary factor depending on the transverse and complementary light-cone coordinates. The contour condition instead trivializes the full path-ordered exponential and reduces the remaining ambiguity to a single global group element fixed at $\Gamma^{\alpha}{}_{\mu\nu}=0.$3. In this gauge the potential can be expressed directly through the field strength: $\Gamma^{\alpha}{}_{\mu\nu}=0.$4 The Fock–Schwinger gauge $\Gamma^{\alpha}{}_{\mu\nu}=0.$5 is a special straight-line realization, with

$\Gamma^{\alpha}{}_{\mu\nu}=0.$6

in the Abelian case (Anikin, 2021).

A key point is that different contours correspond to inequivalent non-local gauge fixings even when they project to the same local axial condition. Along $\Gamma^{\alpha}{}_{\mu\nu}=0.$7, the two formal representations

$\Gamma^{\alpha}{}_{\mu\nu}=0.$8

$\Gamma^{\alpha}{}_{\mu\nu}=0.$9

are not equivalent in the contour-gauge framework. This difference feeds directly into the $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$0 prescriptions for Drell–Yan soft functions,

$[\,\nabla_{\mu},\nabla_{\nu}\,]=0$1

and thereby into the treatment of gluonic poles, single-spin asymmetries, and gauge invariance of hadron tensors. The contour gauge is therefore used to eliminate unphysical longitudinal Wilson lines in DVCS and Drell–Yan operator definitions while retaining the process-dependent information carried by contour direction and endpoint choice (Anikin, 2021).

5. Conceptual distinctions and recurrent misconceptions

The same label thus covers three structurally different notions. In symmetric teleparallel gravity, a non-coincident gauge is a coordinate description with $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$2 but $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$3 and $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$4. In gauge-Higgs theory, it is a non-aligned global prescription imposed after local gauge fixing, typically enforcing $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$5. In contour-gauge QCD, it is a non-local holonomy condition $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$6. A plausible implication is that comparisons across these literatures must first determine whether “gauge” refers to a spacetime connection, a choice of global representative on a gauge orbit, or a path-dependent condition on Wilson lines (Jiménez et al., 2022, Maas, 2012, Anikin, 2021).

Several misconceptions recur across these contexts. In STG, the coincident gauge is not a physical separation of gravity from inertia, and the autoparallels of the symmetric teleparallel connection are not the physical worldlines of minimally coupled standard-model matter. In gauge-Higgs systems, a non-zero $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$7 in an aligned gauge does not imply gauge-invariant spontaneous breaking of local gauge symmetry; the symmetry is more precisely hidden than broken. In contour-gauge applications, the local condition $[\,\nabla_{\mu},\nabla_{\nu}\,]=0$8 does not by itself remove residual gauge freedom; the path choice and boundary data are part of the gauge fixing. Across all three uses, the non-coincident construction is chiefly a device for keeping otherwise hidden structure explicit: inertial contributions in nonmetricity formulations, unaligned global symmetry content in gauge-Higgs systems, or contour dependence in Wilson-line-based QCD factorization (Jiménez et al., 2022, Maas, 2012, Anikin, 2021).