---
title: Parabolic-Type Sub-VOAs in Lattice Theory
url: https://www.emergentmind.com/topics/parabolic-type-sub-vertex-operator-algebras-voas
type: topic
---

# Parabolic-Type Sub-VOAs in Lattice Theory

Parabolic-type sub-vertex operator algebras are a class of subVOAs of lattice VOAs defined from additive submonoids of a positive-definite even lattice and modeled on the role of parabolic subalgebras in Lie theory. In the lattice setting, if \(L\) is a positive-definite even lattice and \(P\le L\) is a parabolic-type submonoid, the associated subVOA is
\[
V_P=\bigoplus_{\alpha\in P} M_{\widehat{\mathfrak h}(1,\alpha)} \subset V_L,
\]
where \(V_L\) is the lattice VOA and \(M_{\widehat{\mathfrak h}(1,\alpha)}\) is the Heisenberg Fock space of momentum \(\alpha\). These VOAs are of CFT-type, are not necessarily strongly finitely generated, and exhibit a mix of Lie-theoretic, lattice-theoretic, and nonsemisimple features. The subject has developed along three closely related lines: the introduction of conic-, Borel-, and parabolic-type subVOAs of lattice VOAs [2402.02278]; the complete rank-two classification, module theory, and fusion theory of parabolic-type subVOAs [2508.21662]; and the quasi-triangular-decomposition and induction-theoretic perspective, especially for the embedding \(V_P\hookrightarrow V_{A_2}\) [2503.23632].

## 1. Definition through lattice submonoids

The ambient setting is a positive-definite even lattice \(L\) in a Euclidean space \(E\), with \(\mathfrak h=\mathbb C\otimes_{\mathbb Z}L\). The lattice VOA decomposes as
\[
V_L=\bigoplus_{\gamma\in L} M_{\widehat{\mathfrak h}(1,\gamma)},
\]
and the vertex-operator product respects momentum addition:
\[
Y\big(M_{\widehat{\mathfrak h}(1,\alpha)},z\big)M_{\widehat{\mathfrak h}(1,\beta)}
\subset M_{\widehat{\mathfrak h}(1,\alpha+\beta)}((z)).
\]
Because of this simple-current-type law, any additive submonoid \(M\le L\) determines a subVOA
\[
V_M=\bigoplus_{\gamma\in M}M_{\widehat{\mathfrak h}(1,\gamma)}.
\]
In particular, conic-type, Borel-type, and parabolic-type subVOAs are defined by choosing \(M\) of the corresponding monoid type [2402.02278].

The monoid-theoretic definitions are modeled on half-space geometry. For nonzero \(\gamma\in E\),
\[
P(\gamma)=\{v\in E:(\gamma\mid v)=0\},\qquad
P^+(\gamma)=\{v\in E:(\gamma\mid v)>0\},\qquad
P^{\ge 0}(\gamma)=\{v\in E:(\gamma\mid v)\ge 0\}.
\]
A submonoid \(B\le L\) is Borel-type if \(B\cup(-B)=L\), \(B\cap(-B)=\{0\}\), and \(B\subset P^{\ge 0}(\gamma)\) for some hyperplane \(P(\gamma)\). A proper submonoid \(P\le L\) is parabolic-type if it contains a Borel-type submonoid. Every Borel-type submonoid is parabolic-type, every Borel-type submonoid contains a conic-type submonoid, and the supporting hyperplane of a Borel-type monoid is unique [2402.02278].

This construction produces genuine subVOAs rather than merely distinguished subspaces. All such \(V_M\) inherit the vacuum \(\mathbf1\) and conformal vector \(\omega\) of \(V_L\), hence are of CFT-type. The terminology is not decorative: in the presence of lattice roots, the degree-one piece of a parabolic-type subVOA is literally a parabolic Lie subalgebra of \((V_L)_1\), and the normalizer property
\[
N_{V_L}(V_P)=V_P
\]
holds, paralleling the Lie-theoretic identity \(N_{\mathfrak g}(\mathfrak p)=\mathfrak p\) [2402.02278].

## 2. Rank-two classification

For a rank-two positive-definite even lattice
\[
L=\mathbb Z\alpha_1\oplus \mathbb Z\alpha_2,
\]
the classification of parabolic-type subVOAs reduces completely to the classification of parabolic-type submonoids. The decisive theorem states that every rank-two parabolic-type submonoid is of exactly one of two forms [2508.21662].

Before stating the structural consequences, it is useful to isolate the two normal forms.

| Type | Monoid normal form | Cartan-part quotient |
|---|---|---|
| Type I | \(\displaystyle P=(\mathbb Z_{\ge 0}\alpha)\cup(P^+(\gamma)\cap L)\), with \(P(\gamma)\cap L=\mathbb Z\alpha\) | \(\displaystyle V_H=M_{\widehat{\mathfrak h}(1,0)}\) |
| Type II | \(\displaystyle P=P^{\ge 0}(\gamma)\cap L\), with \(P(\gamma)\cap L=\mathbb Z\alpha\neq\{0\}\) | \(\displaystyle V_H\cong M_{\widehat{\mathbb C\beta}(1,0)}\otimes V_{\mathbb Z\alpha}\) |

In type I, the monoid is exactly a Borel-type monoid. In type II, the monoid contains the full lattice line on the supporting hyperplane. Geometrically, the distinction is between retaining only one semiaxis on the boundary line and retaining the entire boundary lattice. This is the rank-two form of the Lie-theoretic difference between a Borel-like object and a proper parabolic retaining Levi directions on the wall.

The corresponding VOA structure is uniform. Every rank-two parabolic-type VOA \(V_P\) admits a decomposition
\[
V_P=V_H\oplus V^+
\]
with \(V_H\) a simple subVOA, \(V^+\) the unique maximal proper ideal, and
\[
V_P/V^+\cong V_H
\]
as VOAs. In type I, \(V_H\) is the rank-two Heisenberg VOA. In type II, if \(P(\gamma)\cap L=\mathbb Z\alpha\neq 0\) and \(\mathbb C\beta=(\mathbb C\alpha)^\perp\), then
\[
V_H\cong M_{\widehat{\mathbb C\beta}(1,0)}\otimes V_{\mathbb Z\alpha}.
\]
The paper calls \(V_H\) the Cartan-part of \(V_P\) [2508.21662].

This decomposition isolates the simple quotient while preserving a nonsimple ambient VOA. The construction also clarifies an important point of terminology: in the rank-two theory the parabolic-type VOA itself is not simple, whereas the quotient \(V_H\) is simple and plays the role of a Levi- or Cartan-like core.

## 3. Irreducible modules, fusion, and finiteness properties

The rank-two representation theory is controlled entirely by the Cartan-part. The classification theorem states that irreducible admissible \(V_P\)-modules are exactly irreducible \(V_H\)-modules on which \(V^+\) acts trivially. Equivalently, the nontrivial ideal \(V^+\) contributes nilpotent structure to \(V_P\) but disappears on irreducibles [2508.21662].

For type I, with \(V_H=M_{\widehat{\mathfrak h}(1,0)}\), the irreducible admissible modules are
\[
\left\{\bigl(M_{\widehat{\mathfrak h}(1,\lambda)},\widetilde Y_M\bigr):\lambda\in\mathfrak h\right\}.
\]
For type II, if \((\alpha\mid\alpha)=2N\), \(\mathbb C\beta=(\mathbb C\alpha)^\perp\), and \(0\le i\le 2N-1\), the irreducible admissible modules are
\[
\left\{
\bigl(L^{(\mu,i)},\widetilde Y_M\bigr)
=
\bigl(M_{\widehat{\mathbb C\beta}(1,\mu)}\otimes V_{\mathbb Z\alpha+\frac{i}{2N}\alpha},\widetilde Y_M\bigr)
:\mu\in\mathbb C\beta,\ i=0,\dots,2N-1
\right\}.
\]
In both cases every irreducible admissible module is ordinary.

The fusion rules are equally explicit. In type I,
\[
N^{(M_{\widehat{\mathfrak h}(1,\theta)},\widetilde Y_M)}_{(M_{\widehat{\mathfrak h}(1,\gamma)},\widetilde Y_M)\,(M_{\widehat{\mathfrak h}(1,\eta)},\widetilde Y_M)}
=
\delta_{\gamma+\eta,\theta},
\]
while in type II,
\[
N^{L^{(\nu,k)}}_{L^{(\lambda,i)}\,L^{(\mu,j)}}=
\begin{cases}
1,& \lambda+\mu=\nu,\ \ i+j\equiv k\pmod{2N},\\
0,& \text{otherwise}.
\end{cases}
\]
Accordingly, all rank-two parabolic-type VOAs are simple current [2508.21662].

The finiteness properties separate the simple quotient from the ambient VOA. The Cartan-part \(V_H\) is always \(C_1\)-cofinite, irrational, and strongly unital. By contrast, \(V_P\) itself need not be \(C_1\)-cofinite: type-I \(V_P\) is not \(C_1\)-cofinite, while type-II \(V_P\) is \(C_1\)-cofinite under the explicit lattice inequality
\[
n^2+\ell^2-4\ell k\le 0
\]
when \((\alpha\mid\beta)=-n\), \((\beta\mid\beta)=2k\), and \((\alpha\mid\alpha)=2\ell\); in particular, if \((\alpha\mid\alpha)=2\) or \((\beta\mid\beta)=2\), then \(V_P\) is \(C_1\)-cofinite [2508.21662]. This dovetails with the earlier general picture: parabolic-type subVOAs are nonsimple and not rational, yet can still satisfy meaningful finiteness conditions in concrete families [2402.02278].

## 4. Lie-theoretic interpretation and basic examples

The analogy with parabolic Lie theory is built into both the lattice combinatorics and the weight-one Lie algebra. If
\[
\Phi(L)=\{\alpha\in L:(\alpha\mid\alpha)=2\}
\]
is nonempty, then \(\Phi(L)\) is an ADE root system. For a Borel-type monoid \(B\), the subset \(\Phi^+=B\cap\Phi(L)\) is a set of positive roots. On the VOA side,
\[
(V_L)_1=\mathfrak h\oplus \bigoplus_{\alpha\in\Phi(L)}\mathbb C e^\alpha,
\]
and for a parabolic-type subVOA \(V_P\),
\[
(V_P)_1=\mathfrak h(-1)\mathbf1\oplus \bigoplus_{\alpha\in P\cap\Phi(L)}\mathbb C e^\alpha
\]
is a parabolic subalgebra of \((V_L)_1\). Under a basis-extension hypothesis on the root lattice inside \(L\), Lie-theoretic parabolic subalgebras containing the Cartan part can be realized inside weight one by parabolic-type subVOAs [2402.02278].

The simplest example is rank one. For \(L=\mathbb Z\alpha\), the monoid \(\mathbb Z_{\ge 0}\alpha\) is simultaneously conic-type, Borel-type, and parabolic-type, so
\[
V_{\mathbb Z_{\ge 0}\alpha}=\bigoplus_{m\ge 0} M_{\widehat{\mathfrak h}(1,m\alpha)}
\subset V_{\mathbb Z\alpha}
\]
is the basic low-rank prototype. Its Zhu algebra is explicitly
\[
A(V_B)\cong \mathbb C\langle x,y\rangle /(y^2,\ yx+Ny,\ xy-Ny),
\]
equivalently
\[
A(V_B)=\mathbb C[x]+\mathbb C y,\qquad y^2=0,\quad yx=-Ny,\quad xy=Ny,
\]
for \((\alpha\mid\alpha)=2N\). The square-zero piece \(\mathbb C y\) is the clearest algebraic trace of the “positive” radical [2402.02278].

The model proper parabolic example is in \(A_2\). If
\[
A_2=\mathbb Za+\mathbb Z\beta,\qquad
(a\mid a)=(\beta\mid\beta)=2,\qquad
(a\mid\beta)=-1,
\]
the Borel-type monoid
\[
B=\{ma+n\beta:m,n\in\mathbb Z,\ n>0\}+\mathbb Z_{\ge 0}a
\]
is contained in
\[
P=\mathbb Za+\mathbb Z_{\ge 0}\beta.
\]
Since \(P\cap(-P)=\mathbb Za\neq\{0\}\), \(P\) is parabolic-type but not Borel-type. Its VOA
\[
V_P=\bigoplus_{\gamma\in P}M_{\widehat{\mathfrak h}(1,\gamma)}
\]
is the standard example of a proper parabolic-type subVOA; it is \(C_1\)-cofinite, whereas the associated \(A_2\) Borel-type example is not [2402.02278].

A structural theorem explains why these VOAs are generally nonsimple. If \(B\subset P\subset P^{\ge 0}(y)\) and
\[
S'=P\cap P^+(y)=B\cap P^+(y),
\]
then
\[
V_{S'}=\bigoplus_{\alpha\in S'} M_{\widehat{\mathfrak h}(1,\alpha)}
\]
is a proper ideal of \(V_P\). As a consequence, parabolic-type subVOAs are not simple and not rational [2402.02278].

## 5. Quasi-triangular decomposition and parabolic induction

A second formulation of parabolic-type subVOAs arises from quasi-triangular decomposition. A decomposition
\[
V=V_+\oplus V_H\oplus V_-
\]
is called quasi-triangular if \(V_H\) is a sub-vertex algebra containing \(\mathbf1\), \(V_\pm\) are sub-vertex algebras without vacuum, \(V_H\) and \(V_+\) are \(\mathfrak{sl}_2(\mathbb C)\)-invariant, and
\[
(V_\pm\mid V_\pm)=(V_H\mid V_+)=(V_H\mid V_-)=0.
\]
In this framework the associated parabolic-type subVOA is
\[
V_P=V_H\oplus V_+,
\]
the direct VOA analogue of a Lie-theoretic decomposition \(\mathfrak p=\mathfrak l\oplus\mathfrak n_+\) [2503.23632].

The principal worked example is again \(A_2\). Define
\[
N^+ := \mathbb Z\alpha+\mathbb Z_{>0}\beta,\qquad
T:=\mathbb Z\alpha,\qquad
N^-:=\mathbb Z\alpha+\mathbb Z_{<0}\beta.
\]
Then
\[
V_{A_2}=V_{N^+}\oplus V_T\oplus V_{N^-}
\]
is a quasi-triangular decomposition, and
\[
V_P=V_T\oplus V_{N^+}
\]
for
\[
P=\mathbb Z\alpha+\mathbb Z_{\ge 0}\beta.
\]
At degree one, \((V_P)_1\) is the standard parabolic subalgebra of \(\mathfrak{sl}_3(\mathbb C)\) consisting of block upper-triangular matrices. The paper describes \(V_P\hookrightarrow V_{A_2}\) as a VOA generalization of the inclusion \(\mathfrak p\hookrightarrow \mathfrak{sl}_3(\mathbb C)\) [2503.23632].

This formulation supports an induction theory. For an embedding \(U\hookrightarrow V\), if \(A(U)\twoheadrightarrow A_U\subseteq A(V)\) and \(W\) is an admissible \(U\)-module whose \(A(U)\)-action on \(W(0)\) factors through \(A_U\), the degree-zero induction is
\[
\operatorname{Ind}_U^V W := M\!\left(A(V)\otimes_{A_U} W(0)\right).
\]
A Frobenius reciprocity theorem gives a canonical injective map
\[
\Phi:\operatorname{Hom}_U(W,\operatorname{Res}_U^V M)\to
\operatorname{Hom}_V(\operatorname{Ind}_U^V W,M),
\]
and \(\Phi\) is an isomorphism under generalized-Verma or rationality hypotheses [2503.23632].

For the \(A_2\) parabolic-type VOA, the Zhu algebra admits a generators-and-relations description and is a nilpotent extension of a skew-polynomial algebra:
\[
A(V_P)=\bar A_P\oplus J,\qquad J^2=0,\qquad
\bar A_P\cong A(V_{A_1})[y;\operatorname{Id};\delta].
\]
The irreducible \(V_P\)-modules are precisely
\[
\mathcal E(P)=\{(L(0,\lambda),Y_M),(L(\tfrac{\alpha}{2},\lambda),Y_M):\lambda\in (\mathbb C\alpha)^\perp\}.
\]
Among these, the inducible irreducibles are
\[
L(0,0),\qquad L(0,\pm\lambda_2),\qquad L(\tfrac{\alpha}{2},\pm\lambda_2),
\]
and their inductions are
\[
\operatorname{Ind}_{V_P}^{V_{A_2}}L(0,0)\cong V_{A_2},
\]
\[
\operatorname{Ind}_{V_P}^{V_{A_2}}L(0,\lambda_2)\cong V_{A_2+\lambda_2},
\]
\[
\operatorname{Ind}_{V_P}^{V_{A_2}}L(\tfrac{\alpha}{2},\lambda_2)\cong V_{A_2+\lambda_1},
\]
while
\[
\operatorname{Ind}_{V_P}^{V_{A_2}}L(0,-\lambda_2)\cong 0,\qquad
\operatorname{Ind}_{V_P}^{V_{A_2}}L(\tfrac{\alpha}{2},-\lambda_2)\cong 0.
\]
The possibility of induction to the zero module is one of the distinctly VOA-specific features of this theory [2503.23632].

## 6. Related constructions, analogues, and scope of the term

The phrase “parabolic-type subVOA” is not used uniformly across the VOA literature. In the papers considered here, it is used explicitly in the lattice-VOA setting of submonoids and quasi-triangular decompositions, while several related works develop close analogues without adopting the same terminology.

For OZ-type VOAs generated by Ising vectors of \(\sigma\)-type, the paper on classification does not formulate a separate theory of parabolic-type subVOAs, but it identifies the nearest analogue as subVOAs generated by \(\sigma\)-closed subsets of Ising vectors, equivalently by subsets cut out by suitable 3-transposition subgroups or root-subsystem truncations. In that setting, if \(E'\subset E\) is \(\sigma\)-closed, then
\[
W=\langle E'\rangle
\]
has \(W_2\) spanned by \(E'\), its Griess algebra is a quotient of the Matsuo algebra attached to the subgroup generated by the corresponding Miyamoto involutions, and examples such as
\[
K(A_{n-1},2)\subset K(A_n,2),\qquad
V_{\sqrt2D_{n-1}}^+\subset V_{\sqrt2D_n}^+,\qquad
K(E_6,2)\subset K(E_7,2)\subset K(E_8,2)
\]
play the role of diagram-truncation or root-subsystem subVOAs [2312.09713].

A different but related line concerns \(\mathbb N\)-graded VOAs with nontrivial \(V_0\). There the key structural result is that if \(S\) is a Levi subalgebra of the canonical ideal \(F=\operatorname{Ann}_{V_1}(N)\), the fields \(Y(S,z)\) generate an affine Kac–Moody-type subVOA; in the simple, self-dual, \(\mathbb N\)-graded, \(C_2\)-cofinite case, the subVOA generated by a Levi factor is a tensor product of simple affine VOAs of positive integral levels. The same paper analyzes radicals of \(V_1\), but it does not construct a general VOA analogue of a Lie-theoretic parabolic \(\mathfrak l\oplus\mathfrak u\) [1310.0545].

A further generalization is provided by \(\mathbb C\)-graded vertex algebras and pseudo vertex operator algebras. These objects allow complex gradings, generalized \(L(0)\)-eigenspaces, and conformal deformations
\[
\omega_h=\omega+L(-1)h,\qquad L_h(0)=L(0)-h(0),
\]
while retaining Zhu theory and a bijection between simple admissible modules and simple Zhu-algebra modules. This framework does not define parabolic-type subVOAs, but it provides a natural ambient language when subalgebras are expected to carry nonintegral or nonsemisimple conformal structure [1308.0557].

A common misconception is therefore to treat “parabolic-type subVOA” as a single universal notion across VOA theory. The current literature instead supports a more precise statement: the term is explicit and fully developed for lattice-VOA subVOAs defined by parabolic-type submonoids and for the quasi-triangular constructions built from them, whereas Ising-generated, Levi-generated, and \(\mathbb C\)-graded theories supply analogies, background machinery, or adjacent constructions rather than the same definition.

Source: https://www.emergentmind.com/topics/parabolic-type-sub-vertex-operator-algebras-voas