Elliptic Gauge Complexes and Gribov Obstructions

An elliptic gauge complex separates three infinitesimal questions: stabilizers, true deformations, and obstructions. For an anti-self-dual connection these are the cohomology groups $H_A^0,H_A^1,H_A^2$. Ellipticity makes them finite-dimensional and gives a Fredholm index; it does not prove that all infinitesimal deformations integrate, nor that a local gauge slice is globally unique.

Required background. Gauge configuration groupoids supplies the gauge action. Constraints and reduction supplies tangent reduction. Gribov copies and limits of local gauge fixing distinguishes local zero modes from global copies. Helpful background. Hilbert spaces, completion, and Riesz representation supplies adjoints and orthogonal decompositions.

The anti-self-dual deformation complex

Let $X$ be a closed oriented Riemannian four-manifold, $P\to X$ a principal compact-$G$ bundle, and $A$ a smooth anti-self-dual connection:

$$F_A^+=0.$$

The site’s Lorentzian metric convention is (+---); this page concerns the Riemannian elliptic problem after continuation, where $*^2=1$ on two-forms and $\Omega^2=\Omega^{2,+}\oplus\Omega^{2,-}$. Linearizing $F_{A+a}^+=0$ gives $d_A^+a=0$, while an infinitesimal gauge transformation gives $a=d_A\xi$. Because

$$d_A^+d_A\xi=[F_A^+,\xi]=0,$$

these maps form the elliptic complex

$$0\longrightarrow\Omega^0(X,\operatorname{ad}P) \xrightarrow{d_A}\Omega^1(X,\operatorname{ad}P) \xrightarrow{d_A^+}\Omega^{2,+}(X,\operatorname{ad}P) \longrightarrow0.$$

Its cohomology has distinct meanings:

$$H_A^0=\ker d_A, \qquad H_A^1=\frac{\ker d_A^+}{\operatorname{im}d_A}, \qquad H_A^2=\frac{\Omega^{2,+}}{\operatorname{im}d_A^+}.$$

$H_A^0$ is the infinitesimal stabilizer, $H_A^1$ is the formal tangent space to the moduli problem, and $H_A^2$ is the obstruction space. These roles must not be exchanged: a vanishing Faddeev–Popov eigenvalue detects $H_A^0$, whereas a Kuranishi obstruction lies in $H_A^2$.

The single gauge-fixed operator

$$D_A=d_A^*\oplus d_A^+: \Omega^1(\operatorname{ad}P)\longrightarrow \Omega^0(\operatorname{ad}P)\oplus\Omega^{2,+}(\operatorname{ad}P)$$

has the same elliptic symbol data. Its kernel represents harmonic $H_A^1$, and its cokernel is $H_A^0\oplus H_A^2$. Atiyah, Hitchin, and Singer derive the complex, its ellipticity, and the Kuranishi construction in Atiyah, Hitchin, and Singer 1978, §6, pp. 444–448.

The gauge component has a complementary analytic interpretation. On an irreducible stratum, the orbit direction $\operatorname{im}d_A$ is closed in the Sobolev one-forms and its $L^2$-orthogonal complement is $\ker d_A^*$. Thus the same Hodge decomposition that produces the cohomology also produces the local Coulomb slice. Kondracki and Rogulski establish the relevant Sobolev orbit and slice structure in Kondracki and Rogulski 1986, §3.1, pp. 30–35. This is still stratum-local: at a reducible connection the stabilizer changes, so the quotient chart changes type.

Index, regularity, and the nonlinear equation

Elliptic regularity makes every Sobolev cohomology representative smooth and each $H_A^i$ finite-dimensional. The Fredholm index is

$$\operatorname{ind}D_A =\dim H_A^1-\dim H_A^0-\dim H_A^2.$$

For an $SU(2)$ bundle with $c_2(P)=k$, the standard orientation convention gives

$$\operatorname{ind}D_A =8k-3\bigl(1-b_1(X)+b_2^+(X)\bigr).$$

This is the expected dimension, not automatically the actual local dimension. If $A$ is irreducible modulo the center, then $H_A^0=0$. If it is also regular, meaning $H_A^2=0$, the implicit-function theorem makes the moduli space a smooth manifold near $[A]$ with tangent $H_A^1$ and dimension $\operatorname{ind}D_A$. If $H_A^2\neq0$, nearby solutions are instead described by a Kuranishi map

$$\kappa:H_A^1\longrightarrow H_A^2, \qquad \kappa(0)=0, \qquad \mathrm d\kappa_0=0,$$

and the local moduli space is modeled on $\kappa^{-1}(0)$ modulo the finite stabilizer. The nonlinear term is obtained by projecting $(a\wedge a)^+$ to $H_A^2$ after solving its orthogonal complement with the Green operator. Thus the proof mechanism is elliptic decomposition, bounded Green operator, then a finite-dimensional obstruction equation—not index counting alone.

Application and the global boundary

The first application is the gauge-fixed functional framework in Gauge Fixing, BRST, and Gribov Structure. At an irreducible ASD background, $D_A$ combines the linearized field equation with the Coulomb condition. A nonzero $H_A^1$ gives genuine zero modes of the quadratic action after gauge directions have been removed. A nonzero $H_A^0$ instead says the gauge action was not locally free. A nonzero $H_A^2$ says a formal tangent may fail to extend to a neighboring solution.

None of these finite-dimensional cohomology statements excludes a second, distant intersection of the gauge orbit with the Coulomb slice. Ellipticity controls the principal symbol and local Fredholm problem. A Gribov copy is global information about the nonlinear gauge action. Likewise, global uniqueness cannot be inferred from the positivity of $d_A^*d_A$ at one connection.

An independent symbol check makes the local claim precise. At $p\neq0$, the symbol sequence

$$0\longrightarrow\mathfrak g \xrightarrow{p\wedge}\Lambda^1\otimes\mathfrak g \xrightarrow{\pi_+(p\wedge\cdot)}\Lambda^{2,+}\otimes\mathfrak g \longrightarrow0$$

is exact. Hence the complex is elliptic. This calculation contains no information about how widely separated points of an orbit meet a slice, so it cannot support the global converse.

Exactness can be checked without invoking the index theorem. Choose an oriented orthonormal frame with $p=e^1$. The first symbol maps $\xi$ to $\xi e^1$. If $\pi_+(e^1\wedge a)=0$, the three components of $a$ transverse to $e^1$ vanish because the self-dual forms $e^1\wedge e^j+*(e^1\wedge e^j)$ are independent. Hence $a$ is proportional to $e^1$ and lies in the first image. The dimensions $1,4,3$ then also prove surjectivity of the second symbol. Tensoring with $\mathfrak g$ preserves exactness.

Failure test

The adversarial move is to argue: “$D_A$ is elliptic, therefore Coulomb gauge is unique on the entire orbit.” Insert a second connection $A^u$ satisfying the same gauge condition with $u$ outside the small transformation neighborhood. The principal symbol and Fredholm index at $A$ remain unchanged, yet uniqueness fails. The missing step is a global section theorem, not another local elliptic estimate.

Exercises

Show that the expected dimension of the charge-one $SU(2)$ ASD moduli space on $S^4$ is five.

Solution

For $S^4$, $b_1=0$ and $b_2^+=0$. With $k=1$, the index formula gives $8-3(1)=5$. This is the local dimension only at irreducible regular connections; the arithmetic alone does not prove either hypothesis or compactness of the moduli space.

References

• Atiyah, Michael F., Nigel J. Hitchin, and Isadore M. Singer. “Self-Duality in Four-Dimensional Riemannian Geometry.” Proceedings of the Royal Society of London A 362 (1978): 425–461. DOI; Open PDF.
• Kondracki, Witold, and Jan S. Rogulski. On the Stratification of the Orbit Space for the Action of Automorphisms on Connections. Dissertationes Mathematicae 250. Warsaw: Polish Scientific Publishers, 1986. Repository record and PDF.