---
title: Solomon–Terao Bi-polynomial Overview
url: https://www.emergentmind.com/topics/solomon-terao-bi-polynomial
type: topic
---

# Solomon–Terao Bi-polynomial Overview

The Solomon–Terao bi-polynomial is a two-variable invariant attached to a hyperplane arrangement through the graded Hilbert series of its higher logarithmic derivation modules. In the literature summarized here, the notation \(\Psi(\mathcal A;x,t)\) is standard, but the naming varies: the 2018 paper of Abe–Maeno–Murai–Numata treats \(\Psi(\mathcal A;x,t)\) as the Solomon–Terao polynomial, whereas the 2025 regularity paper explicitly calls it the Solomon–Terao bi-polynomial [1802.04056], [2509.10047]. Its importance lies in the fact that distinguished specializations recover classical arrangement invariants, and in tame or free settings they admit algebraic, geometric, and recursive interpretations.

## 1. Algebraic setting and logarithmic modules

Let \(K\) be a field, \(V=K^\ell\), and
\[
S=\operatorname{Sym}(V^*)\cong K[x_1,\dots,x_\ell].
\]
For a finite central hyperplane arrangement \(\mathcal A\) in \(V\), choose defining linear forms \(\alpha_H\in V^*\) for \(H\in\mathcal A\) and set
\[
Q(\mathcal A)=\prod_{H\in\mathcal A}\alpha_H.
\]
The \(S\)-module of derivations is
\[
\operatorname{Der}S=\bigoplus_{i=1}^\ell S\,\partial_{x_i},
\]
and its exterior powers are
\[
\operatorname{Der}^pS=\bigwedge^p\operatorname{Der}S.
\]

The higher logarithmic derivation modules are defined by
\[
D^p(\mathcal A)=\left\{\theta\in \operatorname{Der}^pS\ \middle|\ \theta(\alpha_H,f_2,\dots,f_p)\in S\alpha_H\ \ (\forall H\in\mathcal A)\right\}.
\]
For \(p=1\), this is the usual logarithmic derivation module \(D(\mathcal A)\). The graded Hilbert series of a graded \(S\)-module \(M\) is denoted \(\operatorname{Hilb}(M;x)\).

A parallel differential-form theory is also used. For hyperplane arrangements, the logarithmic form modules \(\Omega^p(\mathcal A)\) are dual to the logarithmic derivation modules:
\[
D^p(\mathcal A)^*\simeq \Omega^p(\mathcal A),\qquad \Omega^p(\mathcal A)^*\simeq D^p(\mathcal A).
\]
This reflexive duality is one of the structural inputs behind addition–deletion arguments and regularity estimates [2305.10283].

## 2. Definitions, normalizations, and the role of the variables

The literature summarized here uses two displayed normalizations for \(\Psi(\mathcal A;x,t)\). In Abe–Maeno–Murai–Numata,
\[
\Psi(\mathcal A;x,t)=t^\ell\sum_{p=0}^{\ell}\operatorname{Hilb}(D^p(\mathcal A);x)\left(\frac{1-x}{t}-1\right)^p,
\]
while in Abe’s addition–deletion paper and in the regularity paper,
\[
\Psi(\mathcal A;x,t)=\sum_{p=0}^{\ell}\operatorname{Hilb}(D^p(\mathcal A);x)\bigl(t(x-1)-1\bigr)^p.
\]
In both conventions, \(\Psi(\mathcal A;x,t)\) is a polynomial rather than merely a formal series [1802.04056], [2305.10283].

The two variables have distinct formal roles. The variable \(x\) records the internal grading coming from the Hilbert series of the modules \(D^p(\mathcal A)\), whereas the parameter \(t\) records the exterior or homological degree \(p\). For this reason, “bi-polynomial” is a natural description of \(\Psi(\mathcal A;x,t)\). The subspace-arrangement generalization makes this point explicitly: \(\Psi(M^\bullet,x,t)\) is a two-parameter generating function built from Hilbert–Poincaré series, not a bigraded Hilbert series in two independent grading variables [1803.08672].

| Source | Displayed definition of \(\Psi\) | Distinguished specialization |
|---|---|---|
| [1802.04056] | \(t^\ell\sum_p \operatorname{Hilb}(D^p;x)\left(\frac{1-x}{t}-1\right)^p\) | \(x=1\), \(t=1\) |
| [2305.10283] | \(\sum_p \operatorname{Hilb}(D^p;x)(t(x-1)-1)^p\) | \(x=1\), \(t=-1\) |
| [2509.10047] | \(\sum_p \operatorname{Hilb}(D^p;x)(t(x-1)-1)^p\) | \(x=1\), \(t=-1\) |

## 3. Specializations and their interpretations

In the normalization used by Abe–Maeno–Murai–Numata, the specialization
\[
\Psi(\mathcal A;1,t)=\pi(\mathcal A;t)
\]
recovers the Poincaré polynomial of the arrangement complement, and the Orlik–Solomon algebra satisfies
\[
A(\mathcal A)\cong H^*(M(\mathcal A),\mathbb Z),
\]
so that this specialization is identified with the Hilbert series of \(A(\mathcal A)\otimes \mathbb Q\). In the tame case and for \(\eta\in U_2(\mathcal A)\),
\[
\operatorname{Hilb}(\operatorname{ST}(\mathcal A,\eta);x)=\Psi(\mathcal A;x,1),
\]
which interprets the \(t=1\) specialization as the Hilbert series of the Solomon–Terao algebra [1802.04056].

In the normalization used in the addition–deletion and regularity papers, the specialization at \(x=1\) yields the characteristic polynomial up to sign:
\[
(-1)^\ell\Psi(\mathcal A;1,t)=\chi(\mathcal A;t).
\]
The other distinguished specialization is
\[
ST(\mathcal A;x):=\Psi(\mathcal A;x,-1)=\sum_{p=0}^\ell \operatorname{Hilb}(D^p(\mathcal A);x)(-x)^p,
\]
which is called the Solomon–Terao polynomial in the 2025 paper [2305.10283], [2509.10047].

For ideal arrangements, \(\Psi(\mathcal A;x,-1)\) has a geometric meaning: it is essentially the same as the topological Poincaré polynomial of the regular nilpotent Hessenberg variety \(\operatorname{Hess}(I)\), and the 2025 paper states that \(ST(\mathcal A_I;x)\) coincides with the topological Poincaré polynomial of the regular nilpotent Hessenberg variety defined by the same lower ideal \(I\) [2305.10283], [2509.10047].

The Solomon–Terao algebra provides a further algebraic incarnation of these specializations. For a homogeneous polynomial \(\eta\), define the Solomon–Terao ideal
\[
\mathfrak a(\mathcal A,\eta)=\{\theta(\eta)\mid \theta\in D(\mathcal A)\},
\]
and
\[
\operatorname{ST}(\mathcal A,\eta)=S/\mathfrak a(\mathcal A,\eta).
\]
Solomon and Terao showed that for generic \(\eta\), this algebra is finite-dimensional, hence Artinian. In the multiarrangement setting, the 2025 paper states that for tame \((\mathcal A,m)\) and generic \(\eta\),
\[
\Psi\!\left(\mathcal A,m;x,\frac{1-x^d}{x-1}\right)=\operatorname{Hilb}(ST(\mathcal A,m,\eta);x),
\]
and in particular, for \(d=1\),
\[
ST(\mathcal A,m;x)=\operatorname{Hilb}(ST(\mathcal A,m,\eta);x).
\]
For ordinary tame arrangements this yields
\[
ST(\mathcal A;x)=\operatorname{Hilb}(ST(\mathcal A,\eta);x).
\]
The same paper also recalls the divisibility
\[
ST(\mathcal A;-1)=0,\qquad ST(\mathcal A;x)=(1+x)ST^+(\mathcal A;x)
\]
for the reduced Solomon–Terao polynomial [1802.04056], [2509.10047].

## 4. Free arrangements, factorization, and complete intersections

For free arrangements, the bi-polynomial admits explicit product formulas. If \(\mathcal A\) is free with exponents \(\exp(\mathcal A)=(d_1,\dots,d_\ell)\), Abe–Maeno–Murai–Numata prove
\[
\Psi(\mathcal A;x,t)=\prod_{i=1}^{\ell}\bigl(t(1+x+\cdots+x^{d_i-1})+x^{d_i}\bigr),
\]
and consequently
\[
\Psi(\mathcal A;x,1)=\prod_{i=1}^{\ell}(1+x+\cdots+x^{d_i}).
\]
In the later normalization, the corresponding product formula is
\[
\Psi(\mathcal A;x,t)=\prod_{i=1}^{\ell}\bigl(-tx^{d_i}+1+x+\cdots+x^{d_i-1}\bigr),
\]
hence
\[
\Psi(\mathcal A;x,-1)=\prod_{i=1}^{\ell}(1+x+\cdots+x^{d_i}).
\]
Thus, in either convention, freeness forces complete factorization of the distinguished one-variable specialization [1802.04056], [2305.10283].

The Solomon–Terao algebra sharpens this relation. If \(\mathcal A\) is free and \(\eta\in U_d(\mathcal A)\), then
\[
\operatorname{ST}(\mathcal A,\eta)
\]
is a complete intersection and
\[
\operatorname{Hilb}(\operatorname{ST}(\mathcal A,\eta);x)=\prod_{i=1}^{\ell}(1+x+\cdots+x^{d_i+d-2}),
\]
with
\[
\operatorname{socdeg}\operatorname{ST}(\mathcal A,\eta)=|\mathcal A|+\ell(d-2).
\]
Conversely, if \(\operatorname{ST}(\mathcal A,\eta)\) is a complete intersection and
\[
\operatorname{Hilb}(\operatorname{ST}(\mathcal A,\eta);x)=\prod_{i=1}^{\ell}(1+x+\cdots+x^{e_i}),
\]
then
\[
\exp(\mathcal A)=(e_1-d+2,\dots,e_\ell-d+2).
\]
This yields the criterion that \(\operatorname{ST}(\mathcal A,\eta)\) is a complete intersection if and only if \(\mathcal A\) is free [1802.04056].

Several geometric realizations lie inside this free picture. For an irreducible crystallographic Weyl group \(W\) with reflection arrangement \(\mathcal A_W\),
\[
\Psi(\mathcal A_W;x,1)=\operatorname{Poin}(G/B;\sqrt{x})=\prod_{i=1}^{\ell}(1+x+\cdots+x^{d_i^W}),
\]
and if \(P_1\) is the lowest degree basic invariant of \(S^W\), then
\[
\operatorname{ST}(\mathcal A_W,P_1)\cong \operatorname{coinv}(W)\cong H^*(G/B,\mathbb C).
\]
For a lower ideal \(I\subset \Phi^+\), the ideal arrangement \(\mathcal A_I\) is free and
\[
\operatorname{ST}(\mathcal A_I,P_1)\cong H^*(X(N,I)).
\]
These results place the Solomon–Terao specialization alongside coinvariant algebras and regular nilpotent Hessenberg varieties [1802.04056].

## 5. Addition–deletion theory and higher-order \(B\)-sequences

A major advance in the theory is the establishment of addition–deletion formulas for \(\Psi(\mathcal A;x,t)\). For a fixed \(H\in\mathcal A\), write
\[
\mathcal A'=\mathcal A\setminus\{H\},\qquad \mathcal A^H=\{L\cap H\mid L\in \mathcal A\setminus\{H\}\}.
\]
The basic “Euler exact sequence”
\[
0\to D^p(\mathcal A')\xrightarrow{\alpha_H}D^p(\mathcal A)\xrightarrow{\rho_H}D^p(\mathcal A^H)
\]
is not right exact in general. Abe proves that, under codimension-surjectivity and projective-dimension hypotheses,
\[
\Psi(\mathcal A;x,t)=x\,\Psi(\mathcal A';x,t)+\Psi(\mathcal A^H;x,t).
\]
Under analogous hypotheses, the deletion formula is
\[
\Psi(\mathcal A';x,t)=\Psi(\mathcal A;x,t)+x^d\bigl(t(x-1)-1\bigr)\Psi(\mathcal A^H;x,t),
\]
where
\[
d=|\mathcal A'|-|\mathcal A^H|.
\]
After specializing \(x=1\), these formulas recover the classical deletion–restriction relation for \(\chi(\mathcal A;t)\) [2305.10283].

The central new exact sequence is the higher-order \(B\)-sequence
\[
0\to D^p(\mathcal A)\to D^p(\mathcal A')\xrightarrow{\partial^p}D^{p-1}(\mathcal A^H)B,
\]
with
\[
\partial^p(\theta)(f_2,\dots,f_p)=\theta(\alpha_H,f_2,\dots,f_p).
\]
For \(p=1\), this reduces to Terao’s classical polynomial \(B\)-theory
\[
0\to D(\mathcal A)\to D(\mathcal A')\to SB.
\]
The paper identifies this as the correct higher-order extension of \(B\)-theory and derives the deletion theorem for \(\Psi\) from the induced Hilbert-series identities [2305.10283].

The free surjection theorem supplies a mechanism for right exactness. In particular, for logarithmic derivations, if \(\mathcal A'\) is free, then \(\rho_H\) is surjective. This makes the addition theorem available in many cases where a deletion is free. Explicit computations in the paper show that these recursions give nonfree examples that were previously difficult to access [2305.10283].

## 6. Generalizations, regularity results, and open questions

The theory extends beyond ordinary hyperplane arrangements in two different directions. First, for equidimensional subspace arrangements, Pol introduces a generalized Solomon–Terao function
\[
\Psi(M^\bullet,x,t)=\sum_q \operatorname{Poin}(M^q,x)(t(1-x)-1)^q
\]
for finite sequences of graded modules. The relevant modules are multi-logarithmic forms \(\Omega^\bullet(\log \mathcal X/\mathcal C)\) and logarithmic multi-residues \(\mathcal R_{\mathcal X}^\bullet\). The paper proves
\[
\Psi(\Omega^\bullet(\log \mathcal X/\mathcal C),x,t)\in \mathbb Z[x,x^{-1},t],
\]
and under the condition
\[
\Psi(\mathcal R_{\mathcal X_Y}^\bullet,1,1)=1
\quad\text{for all }Y\in L(\mathcal X)\setminus\{V\},
\]
obtains the generalized Solomon–Terao formula
\[
\chi(\mathcal X_Y,t)=t^\ell-\Psi(\mathcal R_{\mathcal X_Y}^\bullet,1,t).
\]
This condition holds for all line arrangements of arbitrary codimension, but the paper also gives an explicit codimension-\(2\) counterexample in \(\mathbb K^4\) where
\[
\Psi(\mathcal R_{\mathcal X}^\bullet,1,1)=2
\]
and the generalized formula fails [1803.08672].

Second, the 2025 regularity paper studies the top degree of the specialization
\[
ST(\mathcal A;x)=\Psi(\mathcal A;x,-1).
\]
For tame arrangements it proves the conjecture that \(ST(\mathcal A;x)\) is monic of degree \(|\mathcal A|\). More generally, for a tame multiarrangement \((\mathcal A,m)\),
\[
ST_{d+1}(\mathcal A,m;x):=\Psi\!\left(\mathcal A,m;x,\frac{1-x^d}{x-1}\right)
\]
is monic of degree
\[
|m|+\ell(d-1).
\]
The key technical input is the regularity estimate
\[
\operatorname{reg}(D^p(\mathcal A,m))\le |m|-\ell+p,
\qquad
\operatorname{reg}(\Omega^p(\mathcal A,m))\le -p.
\]
The same paper also identifies the second-highest coefficient of \(ST(\mathcal A,m;x)\) with the number of degree-\(|m|\) relations among a minimal generating set of \(D^{\ell-1}(\mathcal A,m)\), equivalently with the number of degree-\(0\) relations for \(\Omega^1(\mathcal A,m)\) [2509.10047].

Several open problems remain central. Abe–Maeno–Murai–Numata ask for a topological meaning of the full bivariate polynomial \(\Psi(\mathcal A;x,t)\) and of \(\operatorname{Hilb}(\operatorname{ST}(\mathcal A,\eta);x)\), ask whether deletion–restriction type relations exist in general, and formulate conjectures that freeness is characterized by product decomposition or by palindromicity of the relevant specialization. They also ask for criteria for Gorensteinness and ST-finiteness, formulate a conjecture on the top degree and one-dimensional socle of \(\operatorname{ST}(\mathcal A,\eta)\), and pose a Macaulay dual generator problem for Gorenstein Solomon–Terao algebras [1802.04056]. The addition–deletion paper proves recursive formulas in substantial cases, but explicitly does not claim full combinatorial invariance of \(\Psi(\mathcal A;x,t)\) in general [2305.10283].

Taken together, these results describe the Solomon–Terao bi-polynomial as a logarithmic-module invariant whose principal specializations recover either classical lattice-theoretic data or graded algebras with geometric realizations. Free arrangements yield closed product formulas, tame arrangements admit Hilbert-series interpretations and top-degree control, and higher-order \(B\)-theory provides a recursive calculus for nonfree cases. The main unresolved issue is that the full two-variable structure of \(\Psi(\mathcal A;x,t)\) remains less understood than either of its distinguished one-variable specializations.

Source: https://www.emergentmind.com/topics/solomon-terao-bi-polynomial