---
title: 'Diagonal Coinvariant Ring: Structure and Theory'
url: https://www.emergentmind.com/topics/diagonal-coinvariant-ring
type: topic
---

# Diagonal Coinvariant Ring: Structure and Theory

The diagonal coinvariant ring is a quotient attached to a diagonal group action on several copies of a reflection representation. For a finite complex reflection group \(W\subseteq \GL(V)\), one formulation is
\[
DCR_W=\mathbb{C}[V\oplus V]/I_W,\qquad
I_W=\langle f\in \mathbb{C}[V\oplus V]^W:\deg f>0\rangle
\]
[2112.01419]. In the symmetric-group literature, closely related quotients coexist: one writes
\[
DR_n=\mathbb{C}[x_1,\dots,x_n,y_1,\dots,y_n]/I_n
\]
with \(I_n\) generated by all \(S_n\)-invariants of positive total degree [2402.05221], while another recent source defines
\[
R_n=\mathbb{C}[\mathbf{x},\mathbf{y}]/I
\]
with \(I\) generated by symmetric polynomials in the \(x\)-variables and in the \(y\)-variables having zero constant term [2508.12521]. This suggests that the term is convention-dependent, but across these formulations the recurring structures are a natural grading or bigrading, a diagonal \(S_n\)-action, and strong links with Catalan combinatorics, parking functions, diagonal harmonics, and Cherednik-theoretic representation theory.

## 1. Definitions, gradings, and model realizations

Several standard realizations appear in the recent literature.

| Setting | Quotient | Grading convention |
|---|---|---|
| Type \(A\), diagonal invariants [2402.05221] | \(DR_n=\mathbb{C}[\mathbf{x},\mathbf{y}]/I_n\) | Bigraded by the \(x\)- and \(y\)-degrees |
| Type \(A\), separate symmetric ideals [2508.12521] | \(R_n=\mathbb{C}[\mathbf{x},\mathbf{y}]/I\) | \(\deg(x_i)=(1,0)\), \(\deg(y_i)=(0,1)\) |
| Irreducible complex reflection group [2112.01419] | \(DCR_W=\mathbb{C}[V\oplus V]/I_W\) | A monomial has bidegree \((a,b)\) from total \(x\)-degree \(a\) and total \(y\)-degree minus \(x\)-degree \(b\) |
| \(r\) diagonal copies [1105.4358] | \(R_W^{(r)}=\mathbb{C}[x^{(1)},\dots,x^{(r)}]/I_W^{(\mathrm{dir})}\) | Multigraded by \(q_1,\dots,q_r\) |

For \(W\subseteq \GL_n(\mathbb{C})\), the diagonal action sends
\[
w\cdot f\bigl(x^{(1)},\dots,x^{(r)}\bigr)=
f\bigl(w\cdot x^{(1)},\dots,w\cdot x^{(r)}\bigr),
\]
and the ideal is generated by \(W\)-invariants with zero constant term [1105.4358]. In the two-copy case, the diagonal action of \(S_n\) simultaneously permutes the indices in the \(x\)- and \(y\)-variables [2402.05221].

A second standard model is the diagonal harmonic space
\[
H_W^{(r)}=\{\,f : p(\partial)f=0 \text{ for all } p\in \mathbb{C}[x^{(1)},\dots,x^{(r)}]^W_+\,\},
\]
which is the orthogonal complement of the invariant ideal for the constant-term pairing and is canonically isomorphic to the diagonal coinvariant quotient [1105.4358]. This harmonic realization is one reason the subject is often studied through Hilbert series and graded Frobenius characteristics rather than only through quotient presentations.

## 2. Classical type \(A\): Hilbert series, Frobenius character, and Catalan structures

In the symmetric-group case with two sets of commuting variables, the quotient by diagonal invariants has total dimension
\[
\dim_{\mathbb{C}} DR_n=(n+1)^{\,n-1},
\]
and its bigraded Frobenius characteristic is
\[
\mathrm{Frob}_{q,t}(DR_n)=\nabla e_n
\]
[2402.05221]. Another source records the same dimension as Haiman’s theorem and emphasizes that the bigraded Hilbert series is symmetric in \(q\) and \(t\), while the \(S_n\)-invariant piece is the \((q,t)\)-Catalan polynomial [2607.04999].

In the multivariate framework, the case \(W=S_n\) and \(r=2\) recovers the original diagonal coinvariant ring \(R_n\) and its harmonic model \(H_n\). The same source states that its bigraded Hilbert series is the \((q,t)\)-Catalan polynomial
\[
C_{n+1}(q,t)=\sum_{P\in \mathrm{Dyck}_{n+1}} q^{\mathrm{area}(P)}t^{\mathrm{dinv}(P)},
\]
and that \(\dim R_n=(n+1)^{\,n-1}\) [1105.4358]. Read together with the source-specific definitions above, these statements show that Catalan series enter the theory in more than one way: as a full Hilbert series in one presentation, as an invariant or isotypic contribution in another, and as a combinatorial shadow of the full Frobenius characteristic.

Representation-theoretically, \(DR_n\) decomposes as
\[
DR_n\cong \bigoplus_{\lambda\vdash n} V_\lambda\otimes (\text{multiplicity space}),
\]
with multiplicities encoded by the Schur expansion
\[
\nabla e_n=\sum_{\lambda\vdash n}\tilde K_\lambda(q,t)\,s_\lambda
\]
[2402.05221]. This formulation places the ring at the intersection of Macdonald-theoretic operators, tableaux combinatorics, and parking-function models.

## 3. Alternating diagonal coinvariants and the Vandermonde basis

A recent explicit basis theorem concerns the alternating subspace in the quotient
\[
R_n=\mathbb{C}[\mathbf{x},\mathbf{y}]/I
\]
with \(I\) generated by symmetric polynomials in the \(x\)- and \(y\)-variables of zero constant term. Under the diagonal \(S_n\)-action, one has the isotypic decomposition
\[
R_n=\bigoplus_{\lambda\vdash n} (R_n)^\lambda,
\]
and the sign-isotypic component is
\[
\mathcal{A}_n=R_n^{\mathrm{sign}}=e_{\mathrm{sign}}(R_n),
\qquad
e_{\mathrm{sign}}=\frac1{n!}\sum_{\sigma\in S_n}\mathrm{sgn}(\sigma)\,\sigma
\]
[2508.12521].

The indexing set is the set of Dyck paths \(\pi\) of semilength \(n\). For such a path, one defines the area sequence
\[
\mathrm{area}(\pi)=(a_1(\pi),\dots,a_n(\pi))
\]
and the dinv sequence
\[
\mathrm{dinv}(\pi)=(d_1(\pi),\dots,d_n(\pi)),
\qquad
d_i(\pi)=\bigl|\{\,j>i: a_j(\pi)\in\{a_i(\pi),a_i(\pi)-1\}\}\bigr|.
\]
From the exponent-pair sequence
\[
X(\pi)=\bigl((d_1(\pi),a_1(\pi)),\dots,(d_n(\pi),a_n(\pi))\bigr),
\]
one forms the bivariate Vandermonde determinant
\[
\Delta_{X(\pi)}(\mathbf{x},\mathbf{y})
=
\det\!\bigl[x_i^{\,d_j(\pi)}y_i^{\,a_j(\pi)}\bigr]_{i,j=1}^{n}.
\]

Jiang’s theorem states that the collection
\[
\left\{\Delta_{X(\pi)}(\mathbf{x},\mathbf{y})\right\}_{\pi\in L_{n,n}^+}
\]
is a \(\mathbb{C}\)-basis of \(\mathcal{A}_n\). Each determinant is nonzero and alternating, the number of basis elements is the Catalan number
\[
C_n=\frac1{n+1}\binom{2n}{n}=\dim \mathcal{A}_n,
\]
and the bigrading is
\[
\deg_x\bigl(\Delta_{X(\pi)}\bigr)=\mathrm{dinv}(\pi),
\qquad
\deg_y\bigl(\Delta_{X(\pi)}\bigr)=\mathrm{area}(\pi)
\]
[2508.12521]. Consequently,
\[
\mathrm{Hilb}(\mathcal{A}_n;q,t)
=
\sum_{\pi\in L_{n,n}^+} q^{\mathrm{dinv}(\pi)}t^{\mathrm{area}(\pi)}
=:C_n(q,t).
\]

The proof uses a monomial basis of \(R_n\) indexed by parking functions, due to Carlsson–Oblomkov, of the form \(\mathbf{y}^{\mathrm{area}(P)}\mathbf{x}^{\mathrm{dinv}(P)}\). Antisymmetrizing these monomials produces linear combinations of bivariate Vandermonde determinants; a triangular-shape argument based on distinct leading monomials then yields linear independence [2508.12521]. The result answers a question of Stump and gives an explicit basis for a distinguished isotypic component rather than only a Hilbert-series description.

## 4. Reflection groups, Cherednik algebras, and lower bounds

For an irreducible complex reflection group \(W\) of rank \(n\) with \(N\) reflections, Griffeth sets
\[
g=\frac{2N}{n}
\]
and constructs a finite-dimensional irreducible representation \(L\) of the rational Cherednik algebra \(H_c(W)\) of dimension \((g+1)^n\), together with a \(W\)-equivariant surjection
\[
DCR_W\twoheadrightarrow L.
\]
This yields the general lower bound
\[
\dim_{\mathbb{C}} DCR_W\ge (g+1)^n
\]
[2112.01419].

The relevant representation-theoretic notion is that of a module of coinvariant type: an irreducible finite-dimensional \(H_c(W)\)-module \(L\) is of coinvariant type if the determinant character \(\det\) of \(W\) occurs in \(L\) with multiplicity one. In that situation, the associated graded module receives a surjection
\[
DCR_W\otimes_{\mathbb{C}}\det \twoheadrightarrow \mathrm{gr}\,L,
\]
so lower bounds for \(DCR_W\) reduce to constructing large coinvariant-type modules [2112.01419]. The same work also proves the exponents-duality relation
\[
e_i(V)+e_{n-i+1}\bigl(K(V^*)\bigr)=g,
\]
hence \(g\in \mathbb{Z}\), and establishes that both the finite Hecke algebra and the spherical subalgebra \(eH_c(W)e\) are invariant under dot-action by the Namikawa Weyl group [2112.01419].

For the type \(B_n\) Weyl group, Ajila and Griffeth obtain a refined lower bound improving the standard estimate \((2n+1)^n\). Writing
\[
R_W=\mathbb{C}[x_1,\dots,x_n,y_1,\dots,y_n]/I_W,
\]
they prove
\[
\dim R_W\ge (2n+1)^n+\Delta_n,
\]
where
\[
\Delta_n=
\begin{cases}
n(n-4)/4,& n\equiv 0\pmod 4,\\
n(n-6)/4,& n\equiv 2\pmod 4,\\
(n-1)(n-3)/4,& n\text{ odd}.
\end{cases}
\]
In particular, \(\Delta_4=1\) and \(\Delta_6=3\) [2110.03077]. The proof combines the Gordon module, hook bipartitions, charged-content combinatorics, and the inequality
\[
\mathrm{ch}\,R_W\ge \sum_{L\ \mathrm{coinv\text{-}type}}\mathrm{ch}(L\otimes \det^{-1}),
\]
showing that the error term beyond the principal contribution \((2n+1)^n\) grows quadratically in \(n\) [2110.03077].

## 5. Multivariate, supersymmetric, fermionic, and dihedral extensions

Bergeron studies \(r\) diagonal copies of the reflection representation and proves that the multigraded Hilbert series admits a universal Schur-positive expansion
\[
H_W(q_1,\dots,q_r)=\sum_{\mu\subseteq \Lambda_W} c_\mu\, s_\mu(q_1,\dots,q_r),
\]
where the finite index set of partitions and the coefficients \(c_\mu\) depend only on \(W\), not on \(r\) [1105.4358]. The same source records a parallel universal expansion for the graded Frobenius characteristic and observes a partial stability phenomenon: low-degree components become independent of \(r\) once \(r\) exceeds the degree. For \(r=1\) one recovers ordinary coinvariants, while \(r=2\) gives the diagonal theory [1105.4358].

The bosonic–fermionic extension replaces several sets of commuting variables by a mixture of commuting and anticommuting sets. For the symmetric group with \(k\) bosonic sets and \(j\) fermionic sets,
\[
R_n^{(k,j)}
=
\frac{\mathbb{C}[\mathbf{x}^{(1)},\dots,\mathbf{x}^{(k)},\mathbf{\theta}^{(1)},\dots,\mathbf{\theta}^{(j)}]}
{\langle (\mathbb{C}[\mathbf{x},\mathbf{\theta}])^{S_n}_+\rangle}
\]
inherits a \(U(\mathfrak{gl}(k|j))\otimes \mathbb{C}[S_n]\)-module structure, and its Frobenius series has the form
\[
\Frob\bigl(R_n^{(k,j)};\mathbf q;\mathbf u\bigr)
=
\sum_{\lambda,\mu} c_{\lambda\mu}\,
s_\lambda(\mathbf q/\mathbf u)\,s_\mu(\mathbf z),
\]
with nonnegative integers \(c_{\lambda\mu}\) independent of \(k,j\). In the symmetric-group case this proves Bergeron’s “Diagonal Supersymmetry” conjecture [2505.14885].

The purely fermionic analogue,
\[
FDR_n=\wedge\{\Theta_n,\Xi_n\}/I,
\]
where \(I\) is generated by the \(S_n\)-invariants with vanishing constant term, admits a basis indexed by a family \(\mathcal{Y}(n)\) of noncrossing set partitions. Its total dimension is
\[
\dim FDR_n=\sum_{k=0}^{\lfloor n/2\rfloor}2^{\,n-1-2k}C_k,
\]
and the bigraded Frobenius series is given by an explicit coefficient extraction formula in \(z\) [2204.06059]. This model is combinatorial only after restriction to \(S_{n-1}\), but it gives a fully explicit basis for the entire ring [2204.06059].

For the dihedral group \(\mathfrak I_2(n)\), the bosonic–fermionic coinvariant rings
\[
R_{\mathfrak I_2(n)}^{(k,j)}
\]
also exhibit universal multigraded character and Hilbert series, together with an explicit monomial basis obtained from invariant polarization and straightening. The numerical Hilbert series is
\[
\Hilb\,R^{(k,j)}_{\mathfrak I_2(n)}(\mathbf q;\mathbf u)
=
1+s_{(1,1)}(\mathbf q/\mathbf u)+s_{(n)}(\mathbf q/\mathbf u)+2\sum_{i=1}^{n-1}s_{(i)}(\mathbf q/\mathbf u)
\]
[2509.15054]. This is the dihedral counterpart of the general supersymmetric perspective.

## 6. Related constructions, non-relations, and current directions

A recent comparison with the projective coinvariant algebra \(P_n\) shows that formal similarity need not imply structural identification. The natural \(S_n\)-equivariant map
\[
\phi_n:P_n\to DR_n
\]
factors through the one-dimensional quotient \(P_n/(P_n)_+=\mathbb{C}\), so every positive bidegree piece of \(P_n\) maps to zero. In particular, no nontrivial subquotient or homogeneous component of \(P_n\) recovers \(DR_n\) [2602.15017]. This resolves a potentially misleading analogy between a bigraded Segre-type degeneration and the Garsia–Haiman diagonal coinvariant algebra.

The same work places coinvariant constructions in a cohomological context. Borel’s theorem identifies the classical coinvariant algebra \(R_n\) with \(H^*(Fl_n;\mathbb{C})\), and for a Young subgroup \(S_\alpha\) one has
\[
R_\alpha=(R_n)^{S_\alpha}\cong H^*(F_\alpha;\mathbb{C}).
\]
The deformations \(P_\alpha\) interpolate between flag-variety cohomology and Segre coordinate rings; for two-step flags they coincide with small quantum cohomology, while for longer flags they give an explicit flat deformation that does not coincide with the standard quantum ring [2602.15017].

On the combinatorial side, a uniform higher Specht basis for the full two-variable diagonal coinvariant ring remains open. Gillespie introduces higher Specht polynomials under the diagonal action and proves a higher Specht basis theorem for the hook-shape Garsia–Haiman modules \(DR_\mu\), together with a Frobenius formula in terms of new generalized cocharge statistics on tableaux [2402.05221]. A plausible implication is that the missing uniform basis problem may require exponent data adapted to parking functions or diagonal tableaux rather than only ordinary tableau combinatorics.

Another current direction connects diagonal coinvariants to cluster parking functions and dihedral symmetry. The reduced homology of the cluster-parking complex is conjecturally isomorphic, as an \(S_n\times D_{2(n+2)}\)-module up to tensoring by a sign character of the dihedral group, to diagonal coinvariants. The conjecture is formulated as a \(q,t\)-dihedral sieving phenomenon, with rotations evaluated at \((q,t)=(\zeta,\zeta^{-1})\) and reflections at \((1,-1)\) [2607.04999]. This line of work does not supply a coordinate-level basis for the diagonal coinvariant ring, but it proposes a symmetry-rich homological avatar for its bigraded character.

Source: https://www.emergentmind.com/topics/diagonal-coinvariant-ring