---
title: Marked Multi-Colorings in Lie Superalgebras
url: https://www.emergentmind.com/topics/marked-multi-colorings
type: topic
---

# Marked Multi-Colorings in Lie Superalgebras

Marked multi-colorings are combinatorial colorings attached to a marked graph in which each vertex carries a prescribed color multiplicity, while the marking determines where repetitions of colors are permitted. They were introduced to encode root multiplicities of partially commutative Lie superalgebras and, through denominator identities, certain Borcherds–Kac–Moody superalgebras by graph-theoretic invariants [2503.11230]. In this framework, the ordinary chromatic polynomial is replaced by a family of marked chromatic polynomials indexed by multiplicity vectors, and the resulting theory links graph colorings, independence series, PBW-type combinatorics, and right-angled Coxeter groups. The construction was later extended from graphs to hypergraphs and subspace arrangements, where analogous polynomiality and positivity phenomena appear [2507.20847].

## 1. Marked graphs and the defining constraints

A marked multi-coloring is defined on a simple graph \(\mathcal G=(I,E)\) with countable vertex set \(I\). The marking data consist of a subset \(I_1\subseteq I\) of odd vertices, its complement \(I_0=I\setminus I_1\) of even vertices, and a distinguished subset \(I_1^{\mathrm{it}}\subseteq I_1\) of odd isotropic vertices. In the graphical conventions of the original work, odd isotropic vertices are drawn by \(\Circle\), while even and odd non-isotropic vertices are drawn by \(\CIRCLE\) [2503.11230].

Fix a multiplicity vector \(\mathbf m=(m_i:i\in I)\in\mathbb Z_+^I\) with finite support and an integer \(q\in\mathbb N\). A marked multi-coloring associated to \(\mathbf m\) using at most \(q\) colors is a map from \(I\) to multisets of \(\{1,\dots,q\}\) such that three conditions hold. First, if \(i\notin I_1^{\mathrm{it}}\), then the image at \(i\) is an ordinary subset, so no color may repeat there. Second, the cardinality of the multiset at \(i\), counted with multiplicity, is exactly \(m_i\). Third, if \((i,j)\in E\), then the color multisets at \(i\) and \(j\) are disjoint. Repetition is therefore permitted only at odd isotropic vertices, and even there it is constrained by adjacency.

This definition simultaneously generalizes several familiar notions. When \(\mathbf m\) is the all-ones vector and \(I_1^{\mathrm{it}}=\varnothing\), it reduces to ordinary proper colorings. For general \(\mathbf m\), it becomes a vertex multi-coloring problem with prescribed multiplicities. The marking by \(I_1^{\mathrm{it}}\) is not an auxiliary decoration: it is the combinatorial reflection of the parity and isotropy data carried by the associated Lie superalgebra.

A common simplification is to regard marked multi-colorings as a routine weighted variant of ordinary graph coloring. That is inaccurate. The essential novelty is not merely that vertices may carry several colors, but that multiplicities are permitted only on the isotropic part of the marking, and properness is enforced by disjointness of entire multisets rather than inequality of single color labels.

## 2. Marked chromatic polynomials and their relation to ordinary colorings

The number of marked multi-colorings of \(\mathcal G\) associated to \(\mathbf m\) using at most \(q\) colors is denoted
\[
{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q).
\]
For fixed \(\mathbf m\), this counting function is a polynomial in \(q\), called the marked chromatic polynomial associated to \(\mathbf m\) [2503.11230].

Its first combinatorial expansion is indexed by ordered decompositions of the vertex multiset into nonempty independent multisets. If \(P_k(\mathbf m,\mathcal G)\) denotes the set of ordered \(k\)-tuples \((P_1,\dots,P_k)\) of nonempty multisets such that each \(P_r\) is independent and their disjoint union is exactly the multiset \(\{i^{m_i}:i\in\mathrm{supp}(\mathbf m)\}\), then
\[
{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)
=
\sum_{k\ge 1}|P_k(\mathbf m,\mathcal G)|\binom{q}{k}.
\]
The linear coefficient in \(q\) is therefore an alternating sum built from the cardinalities \(|P_k(\mathbf m,\mathcal G)|\), exactly paralleling classical chromatic-polynomial expansions.

A second formula relates the marked theory to ordinary chromatic polynomials of blow-up graphs. For \(\mathbf m\in\mathbb Z_+^I\), one forms \(\mathcal G(\mathbf m)\) by replacing each vertex \(i\) with a clique of size \(m_i\) and connecting distinct cliques according to \(E(\mathcal G)\). Let \(S(\mathbf m)\) be the set of families of partitions \(\boldsymbol\lambda=(\boldsymbol\lambda_i)_{i\in I}\) such that \(\boldsymbol\lambda_i\vdash m_i\) for all \(i\), with \(\boldsymbol\lambda_i=(1^{m_i})\) whenever \(i\notin I_1^{\mathrm{it}}\). Writing \(\boldsymbol s(\boldsymbol\lambda)=(\ell(\boldsymbol\lambda_i):i\in I)\), the exact relation is
\[
{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)
=
\sum_{\boldsymbol\lambda\in S(\mathbf m)}
\frac{\pi_{\mathcal G(\boldsymbol s(\boldsymbol\lambda))}(q)}
{\displaystyle\prod_{i\in\mathrm{supp}(\mathbf m)}\prod_{k=1}^\infty d_k^{\boldsymbol\lambda_i}!}.
\]
Thus marked multi-colorings are a refined version of ordinary colorings of a blown-up graph, with factorial denominators correcting for indistinguishable permutations of equal parts in the partition data.

This formula shows precisely how the marked theory extends classical coloring. When no isotropic vertices are present, every partition is forced to be trivial, and the marked polynomial reduces to the ordinary chromatic polynomial of the corresponding blow-up graph. When isotropic vertices are present, nontrivial partitions record the admissible repetitions of colors at those vertices.

## 3. Independence series and explicit evaluation formulas

The generating object behind the entire family of marked chromatic polynomials is the marked multivariate independence series. A multiset \(U\) on \(I\) is called independent if its underlying set spans no edges and every vertex in \(I\setminus I_1^{\mathrm{it}}\) appears with multiplicity at most one. Denoting by \(\mathcal I(\mathcal G)\) the set of all finite such multisets and by
\[
x(U)=\prod_{i\in I}x_i^{m_i}
\]
the associated monomial, one defines
\[
\mathcal I(\mathcal G,\mathbf x)=\sum_{U\in\mathcal I(\mathcal G)}x(U)\in\mathbb C[[x_i:i\in I]].
\]
Its \(q\)-th powers encode all marked chromatic polynomials through the coefficient identity
\[
\mathcal I(\mathcal G,\mathbf x)^q
=
\sum_{\mathbf m\in\mathbb Z_+^I}{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)\,x^{\mathbf m},
\qquad q\in\mathbb Z
\]
[2503.11230]. In other words, the coefficient of \(x^{\mathbf m}\) in \(\mathcal I(\mathcal G,\mathbf x)^q\) is exactly the marked chromatic polynomial associated to \(\mathbf m\).

For graphs admitting a perfect elimination ordering, the marked chromatic polynomial has an explicit product formula. If \(\mathcal G\) is a PEO-graph and \(b_j^{\boldsymbol\lambda}\) denotes the sum of partition lengths over the clique determined by the elimination ordering at vertex \(j\), then
\[
{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)
=
\sum_{\boldsymbol\lambda\in S(\mathbf m)}
\prod_{j\in\mathrm{supp}(\mathbf m)}
\binom{q-b_j^{\boldsymbol\lambda}}{\ell(\boldsymbol\lambda_j)}
\frac{\ell(\boldsymbol\lambda_j)!}{\displaystyle\prod_{k=1}^\infty d_k^{\boldsymbol\lambda_j}!}.
\]
This gives a closed combinatorial expression for a large class of graphs, particularly relevant when multiplicity formulas for Lie superalgebra roots are applied afterward.

The worked star-graph example makes the mechanism concrete. If vertex \(1\) is the center of a star and \(\boldsymbol\lambda\in S(\mathbf m)\), then the center contributes \(\binom{q}{m_1}\), while each leaf contributes a factor of the form
\[
\binom{q-m_1}{\ell(\boldsymbol\lambda_i)}
\frac{\ell(\boldsymbol\lambda_i)!}{\prod_k d_k^{\boldsymbol\lambda_i}!}.
\]
The distinction between isotropic and non-isotropic leaves appears entirely through the admissible partition families \(S(\mathbf m)\).

## 4. Partially commutative Lie superalgebras and the denominator identity

Given marked graph data \((I,E;I_1,I_1^{\mathrm{it}})\), the associated partially commutative Lie superalgebra is
\[
\mathcal P(\mathcal G)=\mathcal F(X_I)/\mathcal J,
\]
where \(\mathcal F(X_I)\) is the free Lie superalgebra on generators \(e_i\) of parity determined by \(I_0\) and \(I_1\), and \(\mathcal J\) is generated by
\[
\{[e_i,e_j]: i\neq j,\ (i,j)\notin E\}\cup\{[e_i,e_i]: i\in I_1^{\mathrm{it}}\}.
\]
The \(\mathbb Z_+^I\)-grading is given by \(\deg(e_i)=\alpha_i\), and the root multiplicity of \(\beta\in\mathbb Z_+^I\) is
\[
\mathrm{mult}_{\mathcal G}(\beta)=\dim \mathcal P(\mathcal G)_\beta
\]
[2503.11230].

The universal enveloping algebra \(\mathbf U_{\mathcal G}\) has Hilbert series
\[
\mathcal H_{\mathbf U_{\mathcal G}}(\mathbf x)
=
\frac{\displaystyle\prod_{\beta\in\Delta_1^{\mathcal G}}(1+x^\beta)^{\mathrm{mult}_{\mathcal G}(\beta)}}
{\displaystyle\prod_{\beta\in\Delta_0^{\mathcal G}}(1-x^\beta)^{\mathrm{mult}_{\mathcal G}(\beta)}}.
\]
The key combinatorial statement is the denominator identity
\[
\mathcal I(\mathcal G,-\mathbf x)^{-1}
=
\frac{\displaystyle\prod_{\beta\in\Delta_1^{\mathcal G}}(1+x^\beta)^{\mathrm{mult}_{\mathcal G}(\beta)}}
{\displaystyle\prod_{\beta\in\Delta_0^{\mathcal G}}(1-x^\beta)^{\mathrm{mult}_{\mathcal G}(\beta)}}.
\]
This is the direct analogue, for partially commutative Lie superalgebras, of the denominator identity familiar from Borcherds–Kac–Moody theory.

Its proof is combinatorial rather than purely representation-theoretic. The underlying mechanism is a modified Cartier–Foata monoid \(\mathcal M'(\mathcal G)\), or equivalently a super heaps monoid, in which words ending and beginning with the same isotropic letter produce zero, reflecting the relation \([e_i,e_i]=0\). The inversion lemma identifies the generating series of this monoid with \(\mathcal I(\mathcal G,-\mathbf x)^{-1}\), while a PBW basis identifies the same series with the Hilbert series of \(\mathbf U_{\mathcal G}\). Marked multi-colorings then enter through the expansion of \(\mathcal I(\mathcal G,\mathbf x)^q\).

This placement of marked multi-colorings inside the proof architecture is essential. They are not merely a post hoc interpretation of multiplicity formulas; they are the graph-side invariants that mediate between independence series and the Lie-theoretic denominator identity.

## 5. Root multiplicities, root structure, and Coxeter-theoretic consequences

Combining the denominator identity with the coefficient formula for marked chromatic polynomials yields an explicit root-multiplicity formula. Define
\[
\epsilon(\mathbf m)=
\begin{cases}
1,& \text{if }\sum_{i\in I_1}m_i\text{ is even},\\
-1,& \text{if }\sum_{i\in I_1}m_i\text{ is odd}.
\end{cases}
\]
If \(\ell\mid \mathbf m\) means \(\ell\mid m_i\) for every coordinate and \(\mu\) is the Möbius function, then
\[
\mathrm{mult}_{\mathcal G}(\mathbf m)
=
\epsilon(\mathbf m)
\sum_{\ell\mid \mathbf m}
\frac{\mu(\ell)}{\ell}\,
\epsilon(\mathbf m/\ell)\,
(-1)^{|\mathbf m|/\ell-1}\,
{}_{\mathbf m/\ell}\Pi^{\mathrm{mark}}_{\mathcal G}(q)[q].
\]
When the coordinates of \(\mathbf m\) are relatively prime, only \(\ell=1\) contributes, and one obtains the simpler identity
\[
\mathrm{mult}_{\mathcal G}(\mathbf m)
=
\bigl|{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)[q]\bigr|.
\]
For primitive degree vectors, root multiplicities are therefore exactly the absolute values of linear coefficients of marked chromatic polynomials [2503.11230].

The same framework also detects non-roots. A combinatorial description of the root set \(\Delta^{\mathcal G}\) is given in terms of an admissible subset \(P\subset\mathbb Z_+^I\) together with forbidden star configurations \(K=K_2^0\cup K_3^1\), yielding
\[
\Delta^{\mathcal G}
=
(P\setminus K)\ \cup\ \{\alpha_i:i\in I\}\ \cup\ \{2\alpha_i:i\in I_1\setminus I_1^{\mathrm{it}}\}.
\]
The forbidden sets are precisely families where the multiplicity formula forces vanishing.

The paper further identifies partially commutative Lie superalgebras as positive parts of suitable Borcherds–Kac–Moody superalgebras. In the classes considered there, the corresponding BKM root multiplicities coincide with \(\mathrm{mult}_{\mathcal G}(\mathbf m)\), so marked multi-colorings become explicit combinatorial formulas for BKM multiplicities as well.

A separate specialization occurs when every vertex is odd isotropic, i.e. \(I=I_1^{\mathrm{it}}\). Then the universal enveloping algebra satisfies \(e_i^2=0\) and super-commutation relations matching those of the right-angled Coxeter group
\[
W_{\mathcal G}=\langle s_i\ (i\in I)\mid s_i^2=1,\ s_is_j=s_js_i\text{ if }(i,j)\notin E\rangle.
\]
In this case
\[
\mathbf U_{\mathcal G}\cong \mathbb C[W_{\mathcal G}]
\]
as graded vector spaces, and therefore
\[
\mathcal H_{\mathcal G}(\mathbf x)
=
\sum_{w\in W_{\mathcal G}}\mathbf x(w)
=
\frac{1}{\mathcal I(\mathcal G,-\mathbf x)}.
\]
After the specialization \(x_i=t\), one recovers the Poincaré series formula
\[
P_{\mathcal G}(t)=\frac{1}{I_{\mathcal G}\!\left(\frac{-t}{1+t}\right)}.
\]
A plausible implication is that marked multi-colorings provide a common language for multiplicity formulas on the Lie side and growth series on the Coxeter side, rather than two unrelated enumerative constructions.

## 6. Extensions to hypergraphs and subspace arrangements

The later extension to hypergraphs replaces the marked graph by a simple hypergraph \(\mathcal G=(\mathcal V,\mathcal E)\) together with a subset \(\mathcal V^{\mathrm{sp}}\subseteq\mathcal V\) of special vertices. A marked-independent multiset may use arbitrary multiplicities on \(\mathcal V^{\mathrm{sp}}\) and multiplicity at most one elsewhere, subject to the condition that its underlying set contain no edge. The marked independence series is then
\[
I^{\mathrm{mark}}(\mathcal G,x)
=
\sum_{U\in I^{\mathrm{mark}}(\mathcal G)}x(U).
\]
For a multiplicity vector \(\mathbf m\) and \(q\in\mathbb N\), a marked multi-coloring assigns a multiset of colors to each vertex, forbids repetitions at non-special vertices, prescribes \(|\Gamma(v)|=m_v\), and imposes the hypergraph properness condition
\[
\bigcap_{v\in e}\Gamma(v)=\varnothing
\qquad\text{for every }e\in\mathcal E.
\]
The resulting counting function \({}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)\) is again polynomial, and the basic generating identity survives unchanged:
\[
I^{\mathrm{mark}}(\mathcal G,x)^q
=
\sum_{\mathbf m\in\mathbb Z_{\ge 0}^{\mathcal V}}
{}_{\mathbf m}\Pi^{\mathrm{mark}}_{\mathcal G}(q)\,x^{\mathbf m}
\]
[2507.20847].

The same paper extends the theory to subspace arrangements. If \(\mathcal A\) is an arrangement in \(\mathbb R^n\), \(\mathbf m=(m_1,\dots,m_n)\), and \(S\subseteq \mathrm{supp}(\mathbf m)\), an \(S\)-marked multi \(q\)-coloring is a map \(f:[n]\to P^{\mathrm{mult}}(\mathbb F_q)\), \(f(i)=C_i\), such that non-special coordinates use sets rather than multisets, \(|C_i|=m_i\), and there is no vector \((x_1,\dots,x_n)\in\mathbb F_q^n\) with \(x_i\in C_i\) for all \(i\) lying in any subspace of the induced arrangement over \(\mathbb F_q\). The number of such colorings is polynomial in \(q\), with an explicit formula as a sum of characteristic polynomials of derived arrangements divided by factorial symmetry factors.

For hyperplane arrangements, the associated independence series has a sign-positivity property:
\[
I(\mathcal A,-x)^{-q}\ge 0
\qquad\text{for all }q\in\mathbb Z_{\ge 0},
\]
meaning that every coefficient is nonnegative. The hypergraph analogue remains conjectural: the proposed criterion is that \(I(\mathcal G,-x)^{-q}\) has nonnegative coefficients for all \(q\ge 0\) if and only if every edge of \(\mathcal G\) has even cardinality. The necessity of evenness is proved; the sufficiency is open.

These extensions show that marked multi-colorings are not confined to graph-theoretic Lie combinatorics. They persist at the level of hypergraph independence theory and arrangement complements, where they interact with characteristic polynomials, Möbius inversion on intersection lattices, and additive-combinatorial avoidance problems.

Source: https://www.emergentmind.com/topics/marked-multi-colorings