Gelfand-Fuchs 2-Cocycles

• Gelfand-Fuchs 2-cocycles are Lie algebra cocycles defined on formal vector fields that capture central extension phenomena, notably underpinning the classical Virasoro algebra.
• They appear in multiple realizations—including classical residue formulas, relative cohomology, and Hamiltonian reduction—demonstrating versatile applications across formal geometry and representation theory.
• Advanced methods such as crystal basis theory and factorization algebras are employed to compute these cocycles, linking concepts from algebraic geometry to characteristic classes of foliations.

Gelfand-Fuchs 2-cocycles are Lie algebra cocycles associated with vector fields and Hamiltonian vector fields, most classically as the cocycle defining the central extension of the Witt algebra that underlies the Virasoro algebra. In the literature represented here, they occur in several interconnected forms: as relative cohomology classes of formal vector fields and formal Hamiltonian vector fields, as explicit cocycles on multipoint Witt algebras, as cocycles produced by non-equivariant momentum maps in Hamiltonian reduction, and as algebraic-geometric classes recovered through factorization algebras and Gelfand-Fuchs fibrations [(Khoroshkin, 2013); (Mikami, 2014); (Jurisich et al., 2014); (Hennion et al., 2018); (Goncharov, 19 Sep 2025)].

1. Classical realizations

In the one-dimensional formal setting, the classical Gelfand-Fuchs $2$-cocycle appears on $W_1$, the Lie algebra of formal vector fields on the line. One formula given for $W_1$ is

$$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$

with $X,Y\in W_1$ (Khoroshkin, 2013). For the classical $1$-point Witt algebra with basis $L_k$, the cocycle is presented as

$\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$

which already exhibits the cubic term characteristic of the Virasoro central extension (Jurisich et al., 2014).

The same classical object is also written in analytic and Hamiltonian language. For vector fields on the circle arising in semidirect-product reduction, the relevant cocycle is

$$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$

where $f,g\in C^\infty(S^1)$ (Goncharov, 19 Sep 2025). For Laurent vector fields on the punctured disk, the cocycle is expressed as

$W_1$0

and is described as giving the central extension underlying the Virasoro algebra (Hennion et al., 2018).

A common reduction of the subject to the Virasoro extension alone is incomplete. The cited works show that the classical $W_1$1-cocycle is the lowest-dimensional instance of a broader cohomological framework involving higher-dimensional formal vector fields, Hamiltonian vector fields, and multi-point analogues [(Khoroshkin, 2013); (Jurisich et al., 2014)].

2. Relative cohomology, formal vector fields, and characteristic classes

For $W_1$2, the Lie algebra of formal vector fields on $W_1$3, Anton Khoroshkin proves the conjecture of Feigin, Fuchs and Gelfand on the relative cohomology with coefficients in symmetric powers of the coadjoint representation. The principal statement is

$W_1$4

with

$W_1$5

The absolute cohomology is likewise described explicitly: $W_1$6 for $W_1$7, and vanishes otherwise (Khoroshkin, 2013).

These results place Gelfand-Fuchs cocycles inside a larger family of cohomology classes. The paper describes the higher-$W_1$8, higher-$W_1$9 representatives in terms of “wheels” graph-cocycles and identifies them with invariants in

$W_1$0

For $W_1$1 and any $W_1$2, a representative is

$W_1$3

(Khoroshkin, 2013).

The geometric significance is explicit. The relative cohomology classes correspond to characteristic classes of foliations, and the extension to Lie algebras preserving flags at the origin encodes characteristic classes of flags of foliations. Khoroshkin’s computation is also tied to the local Riemann-Roch theorem through earlier work of Feigin and Tsygan (Khoroshkin, 2013). This suggests that Gelfand-Fuchs $W_1$4-cocycles are best viewed not only as central-extension data, but also as part of a universal characteristic-class mechanism for formal geometry.

3. Formal Hamiltonian vector fields and low-weight Gel'fand-Kalinin-Fuks classes in dimension six

For the Lie algebra $W_1$5 of formal Hamiltonian vector fields, the relevant object is the relative Gel'fand-Kalinin-Fuks cohomology

$W_1$6

in the $W_1$7-dimensional case $W_1$8. The cohomology is graded by weight, where for a wedge product $W_1$9, with $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$0 dual to $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$1, the weight is

$$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$2

The coboundary on a $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$3-cochain is

$$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$4

with $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$5 the Poisson bracket (Mikami, 2014).

The low-weight computations at weights $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$6 are especially relevant for Gelfand-Fuchs $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$7-cocycles. At weight $$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$8, the only non-zero cochain space is

$$\gamma(X,Y)=\operatorname{Res}_0 \operatorname{tr}(X'Y''-X''Y'),$$9

with $X,Y\in W_1$0, and the only non-zero cohomology group is

$X,Y\in W_1$1

The paper identifies this unique class as the classical first GF $X,Y\in W_1$2-cocycle in this relative Hamiltonian setting (Mikami, 2014).

For weights $X,Y\in W_1$3 and $X,Y\in W_1$4, the nontrivial classes move to higher degrees:

Weight	Nontrivial cohomology groups	Dimension(s)
$X,Y\in W_1$5	$X,Y\in W_1$6	$X,Y\in W_1$7
$X,Y\in W_1$8	$X,Y\in W_1$9	$1$0
$1$1	$1$2, $1$3, $1$4	$1$5

At weight $1$6, the trivial $1$7-summands occur in $1$8, $1$9, and $L_k$0, with Betti numbers $L_k$1 and all others zero. At weight $L_k$2, the complex includes components such as $L_k$3, $L_k$4, and $L_k$5, and the Betti numbers are $L_k$6, $L_k$7, $L_k$8 (Mikami, 2014).

Methodologically, the paper uses crystal basis theory because classical decomposition rules become unmanageable for $L_k$9. The procedure is to decompose exterior and tensor powers of $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$0-modules, isolate trivial subrepresentations, and then compute the coboundary operator explicitly in multi-index bases dual to monomials. Some explicit bases in weight $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$1 are described as sums with more than $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$2 terms (Mikami, 2014). In this setting, the classical GF $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$3-cocycle survives as a uniquely determined low-weight invariant, while higher weights produce genuinely new cohomology classes.

4. Multipoint Witt algebras and explicit cocycle formulas

The three-point Witt algebra considered in the literature is $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$4, where

$\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$5

with basis

$\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$6

Its structure is richer than the $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$7-point Witt algebra because it involves two families of derivations and admits two independent $\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$8-cocycles rather than one (Jurisich et al., 2014).

The computation proceeds by transferring cocycles from a more tractable algebra. Writing

$\psi_{\mathrm{GF}}(L_k,L_l)=(k^3-k)\delta_{k+l,0},$9

a ring isomorphism $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$0 induces a Lie algebra isomorphism $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$1. If $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$2 is a $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$3-cocycle on $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$4, then

$$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$5

defines a $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$6-cocycle on $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$7 (Jurisich et al., 2014).

The paper gives explicit formulas for two cocycles $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$8 and $$c_{\mathrm{GF}}(f\partial,g\partial)=\int_{S^1} f''g'\,d\theta,$$9. For $f,g\in C^\infty(S^1)$0,

$f,g\in C^\infty(S^1)$1

and the mixed terms are

$f,g\in C^\infty(S^1)$2

The second cocycle satisfies

$f,g\in C^\infty(S^1)$3

together with

$f,g\in C^\infty(S^1)$4

(Jurisich et al., 2014).

The relation to the classical Gelfand-Fuchs cocycle is structural rather than merely formal. The coefficients $f,g\in C^\infty(S^1)$5 and the $f,g\in C^\infty(S^1)$6-supported terms mirror the one-point Witt case, but the support shifts, the dependence on two basis families, and the presence of mixed terms reflect the three-puncture geometry. These cocycles determine the universal central extension and thereby the $f,g\in C^\infty(S^1)$7-point Virasoro algebra, which the paper describes as important for the representation theory of $f,g\in C^\infty(S^1)$8-point current algebras, including free field and Fock representations (Jurisich et al., 2014).

5. Hamiltonian reduction and cocycles from semidirect products

A different but complementary construction of Gelfand-Fuchs $f,g\in C^\infty(S^1)$9-cocycles arises from Hamiltonian actions and momentum maps. If a Lie group $W_1$00 acts in a Hamiltonian way on a symplectic manifold $W_1$01, the non-equivariance of a momentum map $W_1$02 yields a central extension of the Lie algebra $W_1$03, measured by the cocycle

$W_1$04

for $W_1$05 (Goncharov, 19 Sep 2025).

The semidirect-product situation $W_1$06 is especially useful. After Hamiltonian reduction by $W_1$07,

$W_1$08

one obtains a residual Hamiltonian action of $W_1$09. Directly computing the resulting residual $W_1$10-cocycle can be difficult, especially in infinite-dimensional settings. The key simplification identified in the paper is the splitting condition

$W_1$11

Under this assumption, the residual cocycle for the $W_1$12-action coincides with the restriction of the original cocycle to the stabilizer $W_1$13, viewed as naturally isomorphic to $W_1$14: $W_1$15 The significance is explicit: the cocycle can be computed without reconstructing the reduced symplectic or Poisson structures (Goncharov, 19 Sep 2025).

The motivating application is to Teichmüller spaces for surfaces with boundary. There, one takes

$W_1$16

reduces by $W_1$17, and obtains the residual symmetry

$W_1$18

whose Lie algebra is the Witt algebra. The resulting Hamiltonian action of the Virasoro algebra on Teichmüller space carries a central extension measured by the Gelfand-Fuchs cocycle, and the restriction method recovers the explicit formula

$W_1$19

(Goncharov, 19 Sep 2025). Relative to the more elaborate computation in Alekseev–Meinrenken’s 2024 work, this isolates the cocycle as a stabilizer restriction problem.

6. Algebraic geometry, factorization algebras, and global GF classes

For a smooth affine variety $W_1$20 over a field $W_1$21 of characteristic $W_1$22, with

$W_1$23

the Lie algebra cohomology $W_1$24 is computed in topological terms after choosing an embedding $W_1$25. The main theorem identifies it with the cohomology of the space of sections of the holomorphic Gelfand-Fuchs fibration: $W_1$26 where the fiber is the classical skeleton $W_1$27 acted on by $W_1$28 through the tangent bundle (Hennion et al., 2018).

The proof uses factorization-algebra techniques associated with Beilinson-Drinfeld, Gaitsgory, and Lurie. One assigns to an open subset $W_1$29 the Chevalley-Eilenberg complex

$W_1$30

regarded on the Ran diagram. Because this factorization algebra is generally not locally constant in the algebraic setting, covariant Verdier duality is applied to obtain a locally constant object. The diagonal filtration then organizes the Chevalley-Eilenberg complex by arity, and factorization homology identifies the global cohomology with the cohomology of the section space (Hennion et al., 2018).

In this framework, Gelfand-Fuchs $W_1$31-cocycles become global algebraic objects. For $W_1$32, the group $W_1$33 yields a collection of $W_1$34-cocycles corresponding to central extensions,

$W_1$35

and these classes are described as symmetric polynomials of degree $W_1$36 in the Chern classes: $W_1$37 For curves, the resulting cohomology recovers the classical Virasoro extension; in higher dimensions, the classes are algebraic analogues of the classical Gelfand-Fuchs cocycles, constructed by factorization-homological gluing of local formal-geometry data (Hennion et al., 2018).

Across these settings, Gelfand-Fuchs $W_1$38-cocycles retain a common function—measuring central extension phenomena in Lie algebras of vector fields—while their realizations range from explicit residue and integral formulas to relative cohomology classes, graph-cocycle representatives, stabilizer restrictions in Hamiltonian reduction, and factorization-algebra constructions in algebraic geometry.