---
title: Quasi-lisse Vertex Algebras
url: https://www.emergentmind.com/topics/quasi-lisse-vertex-algebras
type: topic
---

# Quasi-lisse Vertex Algebras

Quasi-lisse vertex algebras are finitely strongly generated vertex algebras whose associated Poisson variety has finitely many symplectic leaves. Introduced by Arakawa as a geometric weakening of lisse, or \(C_2\)-cofinite, behavior, the notion has become a central organizing principle for nonrational but still highly constrained vertex-algebraic systems. It links Zhu’s \(C_2\)-theory, Poisson geometry, modular linear differential equations, arc spaces, BRST constructions, and several forms of gauge-theoretic correspondence, including the 4d/2d and 3d/2d interfaces. In contemporary work, quasi-lisse examples arise from admissible affine vertex algebras, affine and finite \(W\)-algebras, Feigin–Tipunin-type extensions, small \(\mathcal N=4\) systems on Hilbert schemes, and boundary vertex operator superalgebras attached to hypertoric varieties [1610.05865] [2309.17308] [2606.24708].

## 1. Definition through Zhu’s \(C_2\)-geometry

For a vertex algebra or vertex superalgebra \(V\), the \(C_2\)-subspace is
\[
C_2(V)=\operatorname{Span}_{\mathbb C}\{a_{(-2)}b\mid a,b\in V\}.
\]
The Zhu \(C_2\)-algebra is
\[
R_V=V/C_2(V),
\]
with commutative product and Poisson bracket induced by the \((-1)\)- and \(0\)-products:
\[
\bar u\cdot \bar v=\overline{u_{(-1)}v},\qquad \{\bar u,\bar v\}=\overline{u_{(0)}v}.
\]
Papers in the subject use either the reduced affine Poisson scheme
\[
X_V=\operatorname{Spec}((R_V)_{\mathrm{red}})
\]
or the maximal-spectrum convention
\[
X_V=\operatorname{Specm}(R_V),
\]
and both serve as the associated variety controlling the Poisson geometry of \(V\) [2606.24708] [2511.02209].

The lisse condition means that \(R_V\) is finite-dimensional; equivalently, the associated variety is zero-dimensional. Quasi-lisse replaces this with the weaker requirement that \(X_V\) have finitely many symplectic leaves. Arakawa and Kawasetsu showed that this geometric finiteness is strong enough to force modular-type constraints on ordinary characters, while allowing many examples that are not \(C_2\)-cofinite [1610.05865].

The associated variety is part of a larger filtration package. For any vertex (super)algebra \(V\), Li’s filtration produces an associated graded Poisson vertex algebra \(\operatorname{gr}^F(V)\), and there is a canonical surjective morphism
\[
\psi:J_\infty(R_V)\to \operatorname{gr}^F(V),
\]
from the arc algebra of \(R_V\). The corresponding singular support
\[
SS(V)=\operatorname{Spec}(\operatorname{gr}^F(V))\subset J_\infty(\widetilde X_V)
\]
provides the jet-level refinement of \(X_V\). In the quasi-lisse setting, the natural topological map \(SS(V)\to J_\infty(\widetilde X_V)\) is a homeomorphism onto its image, which is one reason arc-space methods enter the theory so naturally [2308.04993].

## 2. Finiteness, modularity, and representation theory

The foundational consequence of quasi-lisse is Poisson-theoretic finiteness. Etingof–Schedler’s theorem implies that if an affine Poisson variety has finitely many symplectic leaves, then its zeroth Poisson homology is finite-dimensional. In the VOA setting this becomes the key input for Arakawa’s modularity theorem: for a quasi-lisse vertex operator algebra, the normalized character of an ordinary module satisfies a modular linear differential equation. In Serre-derivative form, an MLDE has the shape
\[
\sum_{j=0}^{N}P_j(\tau)D_k^{(j)}f(\tau)=0,
\]
with coefficients \(P_j(\tau)\) in classical modular forms and
\[
D_kf=q\frac{d}{dq}f-\frac{k}{12}E_2(q)f
\]
the Serre derivative [1610.05865].

This modular control is accompanied by categorical finiteness. A quasi-lisse VOA has only finitely many simple ordinary modules, and the same geometric mechanism constrains conformal blocks. Recent work sharpened the module geometry further: if \(V\) is conical, simple, self-dual, and quasi-lisse, and \(M\) is a simple ordinary module, then
\[
\dim X_M=\dim X_V,
\]
where \(X_M\) is the support of \(M/F^1M\) over \(R_V\); if \(X_V\) is irreducible, then \(X_M=X_V\) [2511.02209]. This rules out the possibility that ordinary modules generically live on strictly smaller Poisson supports.

The same paradigm extends beyond untwisted ordinary characters. Li, Li, and Yan proved that ordinary \(\mathbb Z_2\)-twisted modules of quasi-lisse VOAs also satisfy MLDEs, with coefficients adapted to congruence subgroups such as \(\Gamma^0(2)\). Their examples include affine \(A_1^{(1)}\) and \(A_2^{(1)}\) at admissible or boundary admissible levels, the Bershadsky–Polyakov algebra, and non-admissible \(D_4\) at level \(-2\) [2304.09681]. In parallel, Li developed twisted arc spaces and established new finiteness conditions for the convergence of genus-0 and genus-1 \(n\)-point functions of quasi-lisse vertex (super)algebras; in particular, quasi-lisse implies finiteness of simple ordinary \(g\)-twisted modules for every finite-order automorphism \(g\) [2308.04993].

A recurring theme is that quasi-lisse is strictly weaker than lisse but still rigid enough to force even-dimensionality phenomena on Poisson supports, finite ordinary representation theory, and modular or quasimodular behavior of characters. This suggests that quasi-lisse is the natural geometric finiteness condition for nonrational vertex algebras that remain analytically tractable.

## 3. Geometric constructions: sheaves, BRST reduction, and symplectic singularities

A major development is the realization of quasi-lisse vertex algebras as global sections of sheaves over symplectic resolutions. In the affine Feigin–Tipunin setting, \(\mathrm{FT}_p(\mathfrak{sl}_2)\) is constructed as the global sections of a sheaf of vertex algebras on the flag variety \(SL_2/B\), and also realized as a screening kernel via inverse Hamiltonian reduction. Its associated variety is the nilpotent cone \(\mathcal N\subset \mathfrak{sl}_2\), so \(\mathrm{FT}_p(\mathfrak{sl}_2)\) is quasi-lisse [2306.13568].

Hypertoric geometry provides a different but closely related mechanism. For a unimodular weight matrix \(\Delta\) and generic stability parameter \(\delta\), one has the projective hypertoric variety
\[
Y_\delta(\Delta)=\mu^{-1}(0)/_\delta G
\]
and affine quotient
\[
Y_0(\Delta)=\mu^{-1}(0)/G=\operatorname{Spec}(\mathbb C[\mu^{-1}(0)]^G).
\]
Arakawa’s associated variety machinery is brought into a BRST–microlocal framework by constructing a sheaf of \(\hbar\)-adic vertex superalgebras \(\mathcal V^{\hbar}_{\delta,\Delta}\) on \(Y_\delta(\Delta)\). Its global sections yield the boundary hypertoric vertex operator superalgebra \(V(\Delta)\), and the main theorem identifies
\[
X_{V(\Delta)}\cong Y_0(\Delta).
\]
Since \(Y_0(\Delta)\) is a symplectic singularity with finitely many symplectic leaves, \(V(\Delta)\) is quasi-lisse [2606.24708].

The Hilbert-scheme construction for the symmetric group \(S_N\) is analogous in spirit. A sheaf of \(\hbar\)-adic vertex operator superalgebras is constructed on \(\operatorname{Hilb}^N(\mathbb C^2)\), with global sections \(V_{S_N}\) satisfying
\[
X_{V_{S_N}}\cong (\mathbb C^2)^N/S_N.
\]
After splitting off a rank-one \(\beta\gamma\)-system and rank-one symplectic fermion factor, one obtains \(W_{S_N}\) with
\[
X_{W_{S_N}}\cong M_{S_N},
\]
where \(M_{S_N}\) is the canonical symplectic singularity attached to \(S_N\). Hence \(W_{S_N}\) is quasi-lisse, and in this case the construction also produces an explicit small \(\mathcal N=4\) structure [2309.17308].

These sheaf-theoretic realizations share a common pattern: one starts from free fields, imposes a BRST differential, proves concentration of cohomology in degree \(0\), identifies the global sections, and then compares the reduced \(C_2\)-algebra with the coordinate ring of a symplectic singularity. In each case, quasi-lisse is not an auxiliary property but the geometric output of the identification \(X_V\cong X\).

## 4. Affine, \(W\)-algebraic, and extension-type examples

The original motivating class was affine. For an affine VOA \(V_k(\mathfrak g)\), quasi-lisse is equivalent to the inclusion of the associated variety in the nilpotent cone \(\mathcal N_{\mathfrak g}\). This covers admissible affine VOAs and, by Drinfeld–Sokolov reduction, many quasi-lisse \(W\)-algebras [1610.05865].

A particularly influential family consists of the Deligne exceptional series at
\[
k=-\frac{h^\vee}{6}-1.
\]
For \(\mathfrak g\in\{A_1,A_2,G_2,D_4,F_4,E_6,E_7,E_8\}\), the affine vertex algebra \(V_k(\mathfrak g)\) is quasi-lisse and has associated variety equal to the closure of the minimal nilpotent orbit. Its vacuum character satisfies a second-order MLDE, and the resulting solutions reproduce explicit modular or quasimodular forms; these are the homogeneous Schur indices of the corresponding 4d \(\mathcal N=2\) SCFTs [1610.05865].

More recent affine examples show that quasi-lisse behavior extends beyond standard admissible regimes. Adamović and Vukorepa proved that \(L_{-14/3}(D_4)\) is quasi-lisse by showing
\[
X_{L_{-14/3}(D_4)}\subset \mathcal N_{D_4}.
\]
Their argument uses explicit singular vectors of conformal weight six, Zhu-theoretic polynomial constraints, and a Chevalley-projection computation excluding nonzero semisimple points. The same paper shows that \(L_{-14/3}(D_4)\) has \(405\) irreducible modules in category \(\mathcal O\), but a unique irreducible ordinary module [2504.13783].

Extension constructions supply a second major source. In the Feigin–Tipunin direction, \(\mathrm{FT}_p(\mathfrak{sl}_2)\) is an affine analogue of the triplet algebra with associated variety the nilpotent cone of \(\mathfrak{sl}_2\); its strong generators include \(e,h,f\) and additional nilpotent fields \(x_{ij}\), and it supports infinitely many simple modules despite being quasi-lisse [2306.13568]. In a different construction, Adamović and Milas defined a family \(\mathcal C_p\), \(p\ge 2\), related to the chiral differential operator algebra on \(SL_2\) at level \(-2+\frac1p\). For \(p=3,4,5\), they established isomorphisms with known quasi-lisse affine or \(W\)-algebras:
\[
\mathcal C_3\cong L_{-5/3}(\mathfrak g_2),\qquad
\mathcal C_4\cong W^{-23/4}(F_4,\widetilde A_1+A_1),\qquad
\mathcal C_5\cong W^{-119/5}(E_8,A_4+A_2),
\]
and used these identifications to support the conjectural quasi-lisse property of the whole family [2502.01766].

There are also instructive contrasts. Kuwabara’s purely even hypertoric algebras \(V_{\min}(\Delta)\), constructed on the universal family of Poisson deformations, are generally not quasi-lisse because the universal deformation typically has infinitely many symplectic leaves. The boundary hypertoric superalgebras \(V(\Delta)\) of the newer construction are fermionic simple-current extensions of \(V_{\min}(\Delta)\otimes V_{J^\perp}\), and the extension eliminates central parameters so that the associated variety collapses from the universal deformation to the singular hypertoric variety \(Y_0(\Delta)\), restoring quasi-lisse behavior [2606.24708].

## 5. Twisted sectors, spectral flow, and charged refinements

Quasi-lisse geometry interacts strongly with twisting and spectral flow. In admissible \(\widehat{\mathfrak{sl}}_2\), all irreducible \(\mathbb Z_2\)-twisted modules in category \(\mathcal O\) are obtained by half-integer spectral flow from untwisted admissible modules, and at boundary admissible levels the category of ordinary \(\mathbb Z_2\)-twisted modules is semisimple [2304.09681]. This yields explicit twisted MLDEs, twisted Zhu-algebra calculations, and fusion rules between untwisted and twisted sectors. The same paper also derives ordinary characters of some non-vacuum modules for affine \(D_4\) at non-admissible level \(-2\) from spectral flow automorphisms [2304.09681].

The module-level geometry is now understood much more sharply. For a conical, simple, self-dual, quasi-lisse vertex algebra \(V\), the support variety \(X_M\) of any simple ordinary module has the same dimension as \(X_V\), and equals \(X_V\) whenever the latter is irreducible [2511.02209]. In affine and affine \(W\)-algebra examples, this recovers the phenomenon that ordinary modules do not shrink the associated variety.

A further refinement is the charged or relative theory. If \(h\in V_1\) acts semisimply with integer eigenvalues, one forms the \(h\)-relative Zhu algebra \(R_V^h\) by quotienting \(R_V\) by the ideal generated by nonzero \(h_0\)-charge pieces. The triple \((V,\omega,h)\) is stably quasi-lisse if the relative associated variety
\[
X_V^h=\operatorname{Specm}(R_V^h)
\]
has finitely many symplectic leaves. This framework leads to stable modules, a stable Zhu algebra
\[
A^{\mathrm{stab}}(V)=A(V)^{(0)}/(A(V)A(V)^{(>0)})^{(0)},
\]
and charged trace functions
\[
S_M(u;\alpha,\tau)=\operatorname{Tr}_M\,u_0\,y^{h_0}q^{L_0-c/24},
\qquad y=e^{2\pi i\alpha}.
\]
Under stable quasi-lissity and stable rationality, these traces span the flat sections of a Jacobi-invariant connection on the moduli of degree-zero line bundles over elliptic curves, substantially generalizing Zhu’s modular-invariance theorem to the quasi-lisse setting [2605.29921].

This charged theory reorganizes several older modular phenomena. For admissible affine algebras with \(h=\rho\), and for admissible \(W\)-algebras of standard Levi type with \(h=\rho_f\), the relevant stable categories satisfy the hypotheses of the theorem, and the resulting charged characters transform as vector-valued Jacobi objects with MLDE constraints [2605.29921].

## 6. Modularity, physical correspondences, and current directions

Character theory is one of the most visible outputs of quasi-lisse structure. In the Deligne-series affine examples, the MLDE is rigid enough to determine the vacuum character explicitly as a modular or quasimodular form [1610.05865]. In \(\mathrm{FT}_p(\mathfrak{sl}_2)\), quasi-lisse coexists with logarithmic features and infinitely many simple modules, while the associated variety remains the nilpotent cone of \(\mathfrak{sl}_2\) [2306.13568]. In the Hilbert-scheme family \(W_{S_N}\), the supercharacter is an explicit quasimodular form of mixed weight, and the construction matches the Schur index of 4d \(\mathcal N=4\) \(SL_N\) super-Yang–Mills theory [2309.17308]. In hypertoric boundary superalgebras, the supercharacter is especially simple:
\[
\operatorname{sch}_{V(\Delta)}(q)=\eta(q)^{2M},
\]
while the ordinary characters admit Lambert-type series expressions and are conjectured to be quasimodular for congruence subgroups of level \(2\) [2606.24708].

Physical applications are not incidental. The 4d/2d correspondence predicts that the associated variety of a chiral algebra should recover the Higgs branch of the 4d theory; the \(W_{S_N}\) construction proves this when the Higgs branch is \((\mathbb C^2)^N/S_N\) [2309.17308]. The 3d Higgs branch conjecture asserts
\[
X_{V^A(\mathcal T)}\cong \mathcal M_H(\mathcal T),
\]
and the hypertoric construction verifies it for all 3d \(\mathcal N=4\) abelian gauge theories with unimodular \(\Delta\) [2606.24708]. Mirror symmetry and Gale duality then transport these statements to dual Coulomb-branch constructions.

Several current directions remain open. One is irreducibility: Arakawa–Moreau’s conjecture, cited in later work, predicts that the associated variety of a simple conical quasi-lisse vertex algebra is irreducible; if true, it would force \(X_M=X_V\) for all simple ordinary modules [2511.02209]. Another is extension beyond presently verified families: \(\mathcal C_p\) is conjecturally quasi-lisse for all \(p\ge 2\), and the simple-current extension
\[
\mathcal U(n)\cong \operatorname{Com}_{V_1(\mathfrak{psl}(n|n))}(L_{\mathfrak{sl}(n),1}(\Lambda_0))
\]
is conjectured to be quasi-lisse on the basis of its finite ordinary representation theory and quasimodular character behavior [2502.01766] [1805.09771]. In the \(D_4\) family, the sharper conjecture is that
\[
X_{L_{-14/3}(D_4)}=\overline{\mathcal O_{\mathrm{sreg}}},
\]
the closure of the subregular nilpotent orbit, together with a collapsing statement for the subregular \(W\)-reduction [2504.13783].

Across these examples, quasi-lisse vertex algebras occupy a precise intermediate regime: broader than lisse and often nonrational, yet still governed by a rigid Poisson geometry that propagates into characters, conformal blocks, twisted sectors, and gauge-theoretic correspondences. The subject has therefore become a meeting point of Poisson algebraic geometry, modular analysis, BRST localization, and nonsemisimple representation theory.

Source: https://www.emergentmind.com/topics/quasi-lisse-vertex-algebras