---
title: Hermitian Modular Operators Overview
url: https://www.emergentmind.com/topics/hermitian-modular-operators
type: topic
---

# Hermitian Modular Operators Overview

Searching arXiv for recent papers on Hermitian modular operators, differential operators, and related Hecke/theta operator frameworks.
arxiv_search(query="Hermitian modular operators differential operators Hermitian modular forms U(n,n) pullback formula", max_results=10)
arxiv_search(query="Hermitian modular forms differential operators pullback formula U(n,n) Hecke theta operator", max_results=10)
Hermitian modular operators are operator constructions acting on Hermitian modular forms on unitary groups such as \(U(n,n)\). In the arithmetic literature, the term encompasses several distinct but related families: Hecke operators defined by double cosets, differential operators in matrix variables, Rankin–Cohen type bilinear brackets, Jacobian-based vector-valued operators, and Fourier-coefficient operators such as the Hermitian theta and heat operators. These operators are used to preserve or alter automorphic transformation laws, to pass from scalar-valued to vector-valued forms, to construct Eisenstein and Klingen-type series by pullback, and to study congruences and \(L\)-functions [2412.06532].

## 1. Arithmetic setting and automorphy

The standard setting is a CM field \(K\), often a quadratic imaginary extension of a totally real field \(K^+\), together with the algebraic unitary group \(U_n\) over \(K^+\). In the \(U(n,n)\) realization, the Hermitian upper half-space of degree \(n\) is
\[
\mathfrak{H}_n=\{Z\in M_n(\mathbb{C})\mid \operatorname{Im}(Z)>0\},
\]
and an element \(g=\begin{pmatrix}A&B\\ C&D\end{pmatrix}\) acts by
\[
g\langle Z\rangle=(AZ+B)(CZ+D)^{-1}.
\]
For vector-valued theory one uses the pair of automorphy factors
\[
\lambda(g,Z)=CZ+D,\qquad \mu(g,Z)=\overline{C}^{\,t}Z+\overline{D},
\]
or, equivalently, \(M(g,Z)=(\lambda(g,Z),\mu(g,Z))\), and a representation \((\rho,V)\) of \(\prod_{\tau}(GL_n(\mathbb{C})\times GL_n(\mathbb{C}))\). The slash operator is
\[
F|_{\rho}[g](Z)=\rho(M(g,Z))^{-1}F(g\langle Z\rangle),
\]
and a Hermitian modular form is a holomorphic \(V\)-valued function fixed by the relevant congruence subgroup [2412.06532].

Scalar-valued formulations appear in parallel. For a congruence subgroup \(\Gamma\subset U(n,n)(\mathcal{O}_L)\), a scalar-valued Hermitian modular form \(F\) of weight \(k\) satisfies
\[
F(Z)=\det(CZ+D)^{-k}F(\gamma Z),
\]
or equivalently \(F(\gamma Z)=\det(CZ+D)^kF(Z)\) [2404.04371]. In degree \(2\), one frequently works with \(\Gamma_2(\mathcal{O}_K)\) or \(U(2,2;\mathcal{O}_K)\), and the Fourier expansion is indexed by positive semidefinite Hermitian matrices. A typical expansion is
\[
F(Z)=\sum_{T\ge 0} a(F,T)e^{2\pi i \operatorname{Tr}(TZ)},
\]
with cusp forms characterized by support on positive definite indices [2412.06532].

The underlying representation theory is intrinsic to the operator theory. Dominant integral weights \(k\) and \(l\) parametrize irreducible algebraic representations \(\rho_{n,k}\) and \(\rho_{n,l}\) of \(GL_n(\mathbb{C})\), and tensor products \(\rho_{n,(k,l)}\) encode vector-valued weights. In low-dimensional structure theorems, the standard tensor product representation \(St\otimes St\) appears as the natural target for first-order differential constructions [1402.1929].

## 2. Differential-operator criteria on \(U(n,n)\)

A central problem is to determine when a differential operator on scalar Hermitian modular forms preserves automorphy after restriction and produces a vector-valued form of prescribed weight. In the pullback framework, one considers operators \(D=(P_\tau(\partial_{Z_\tau}))_\tau\), where \(P_\tau\) is a polynomial in matrix derivatives. The operator-preservation criterion, formulated as Condition (A), states that the restriction of \(D(F)\) must commute with the slash action in the appropriate vector-valued weight. The characterization is exact: \(D\) satisfies Condition (A) if and only if the associated polynomial symbols are pluriharmonic and covariant under blockwise \(GL_{n_i}(\mathbb{C})\times GL_{n_i}(\mathbb{C})\)-action [2412.06532].

The pluriharmonicity condition is expressed after rewriting a polynomial \(P_\tau(T)\) as
\[
\widetilde{P_\tau}(X_1,\dots,X_d,Y_1,\dots,Y_d)
=
P_\tau([X_s^tY_t]_{1\le s,t\le d}),
\]
and requiring \(\widetilde{P_\tau}\) to be pluriharmonic in each pair \((X_i,Y_i)\). The covariance condition is
\[
P_\tau(\operatorname{diag}(A_1,\dots,A_d)\,T\,\operatorname{diag}({}^tB_1,\dots,{}^tB_d))
=
(\rho_{n_1,k,l}(A_1,B_1)\otimes\cdots\otimes \rho_{n_d,k,l}(A_d,B_d))P_\tau(T).
\]
This gives an exact intertwining criterion for an operator to map scalar forms of weight \(\det^\kappa\) to vector-valued forms of weight \(\det^\kappa\rho_{n_1,k,l}\otimes\cdots\otimes\det^\kappa\rho_{n_d,k,l}\) [2412.06532].

The recent explicit theory on \(U(n,n)\) refines this criterion by introducing two-variable spherical pluriharmonic polynomials and explicit bases for their spaces. For \(X,Y\in M_{n,\kappa}(\mathbb{C})\), the mixed Laplacians
\[
\Delta_{ij}=\sum_{\nu=1}^{\kappa}\frac{\partial^2}{\partial X_{i,\nu}\partial Y_{j,\nu}}
\]
define pluriharmonicity, while the induced Hermitian mixed Laplacians on \(C[T]\) are
\[
D_{ij}^{(\kappa)}=\kappa \partial_{ij}+\sum_{k,l=1}^n t_{k,l}\,\partial_{i,l}\partial_{k,j}.
\]
The spaces \(P^n(\kappa)\) of higher spherical pluriharmonic polynomials are then characterized by the equations \(D_{ij}^{(\kappa)}P=0\) on specified block indices, and explicit monomial and descending bases are constructed. The symbol calculus is encoded by
\[
\phi_\kappa(P)(T)=\left.(-1)^d\bigl(P(\partial_W)\det(I_n-W\,{}^tT)^{-\kappa}\bigr)\right|_{W=0},
\]
for homogeneous \(P\) of degree \(d\) [2506.18236].

A complementary bilinear theory constructs Rankin–Cohen type differential operators on Hermitian modular forms of signature \((n,n)\). For polynomials \(Q\in\mathcal{H}_{n,v}(k_1,\dots,k_r)\), the commutation relation
\[
Q\!\left(\frac{\partial}{\partial Z}\right)\left(F(\gamma Z_{1},\dots,\gamma Z_{r})\prod_{i=1}^{r}\det(CZ_{i}+D)^{-k_{i}}\right)\Big|_{Z_{1}=\cdots=Z_{r}=Z}
=
\det(CZ+D)^{-(k+2v)}Q\!\left(\frac{\partial}{\partial (\gamma Z)}\right)\big(F(\gamma Z,\dots,\gamma Z)\big)
\]
holds if and only if \(Q\) lies in the appropriate pluriharmonic space. In the bilinear case \(r=2\), uniqueness up to scale is proved for \(n>1\) and \(v>0\) [2404.04371].

## 3. Pullback formulas and Eisenstein series

Pullback formulas provide one of the principal applications of Hermitian modular differential operators. Given \(n_1\ge n_2\) and \(n=n_1+n_2\), one embeds \(G_{n_1}\times G_{n_2}\) into \(G_n\) by block diagonal insertion. The operator under study is a differential operator \(D\) applied to a Hermitian Eisenstein series \(E_{n,\kappa}(g,s;\mathfrak{n},\chi)\), then restricted along the embedding and paired against a cusp form on \(G_{n_2}\) [2412.06532].

For a Hecke character \(\chi\) of \(K\) with prescribed infinity type, the Eisenstein series is defined from local sections \(\epsilon_{n,\kappa,v}(g,s;\mathfrak{n},\chi)\), and convergence for \(\operatorname{Re}(s)>n\) is standard. The pullback theorem states that, after applying \(D\), the inner product with a Hecke eigen cusp form \(f\) factors into explicit local contributions. In the equal-rank case \(n_1=n_2\), the result is
\[
(f,(D E_{n,\kappa}^\theta)(\iota(g_1,*),\overline{s};\mathfrak{n},\chi))
=
\Bigl[\prod_{v|\infty}\omega_K\cdot 2^{n_2(\kappa_v-2s-n_2^2)-|\rho_{n_2,v}|}\cdot c(s,\rho_{n_2,v})\Bigr]
\cdot
\Bigl[\prod_{v|\mathfrak{n}}[K_{n,v}:K_{n,v}(\mathfrak{n})]\Bigr]
\cdot
D_S(s,f;\overline{\chi})\cdot f^\natural(g_1),
\]
while in level \(1\) the same pullback produces a Klingen-type Eisenstein series \([f^\natural]_{n_2}^{n_1}(g_1,s;\overline{\chi})\) with coefficient \(D(s,f;\overline{\chi})\) [2412.06532].

The later differential-operator treatment makes the archimedean part explicit in terms of the symbol \(\phi_\kappa(P)\) of the operator. The constant \(c(s,\rho_{n_2})\) is given by an integral over the matrix ball
\[
\mathfrak{S}_{n_2}=\{S\in M_{n_2}(\mathbb{C})\mid I_{n_2}-\overline{S}S>0\},
\]
and for \(\mu\ge n\) a closed product formula is obtained in terms of factorials and Pochhammer symbols. This yields an exact pullback formula for Hermitian Eisenstein series, analogous to the Siegel case but with the Hermitian representation parameters \((k,l)\) and the symbol \(\phi_\kappa(P)\) built into the archimedean factor [2506.18236].

On Fourier expansions, the operator \(P(\partial_Z)\) acts termwise:
\[
P(\partial_Z)e^{2\pi i \operatorname{Tr}(TZ)}
=
(2\pi i)^{\deg P}({}^tP(T))\,e^{2\pi i \operatorname{Tr}(TZ)},
\]
so pullback and differentiation transform coefficients by explicit polynomial symbols in the Fourier index \(T\). This is the mechanism behind the appearance of operator symbols in the pullback formula and the restriction to compatible block indices [2412.06532].

## 4. Hecke algebras, theta and heat operators, and congruence operators

Hecke operators form another major class of Hermitian modular operators. In the Hel Braun setting, the pair \((\Gamma_n,\Delta_n)\) is a Hecke pair, the Hecke algebra \(H(\Gamma_n,\Delta_n)\) is commutative, and double cosets \([\Gamma_n M\Gamma_n]\) act on Hermitian modular forms by summing slash actions over right coset representatives. For inert primes, the local Hecke algebra is generated by the analogues of the Siegel generators \(T_n(p)\) and \(T_{n,j}(p^2)\), and the inert part of the Hecke algebra decomposes as a restricted tensor product over inert primes [1911.03157].

In degree \(2\) over imaginary quadratic fields, explicit good-prime formulas are available. For \(p\nmid \Delta_K\), inert primes admit generators \(T_p\) and \(T_{p^2}\), while split primes admit \(T_p\), \(T_{\mathfrak p}\), and \(T_{\overline{\mathfrak p}}\), each given by explicit double-coset representatives and explicit Fourier-coefficient transformations. The corresponding degree-\(6\) Euler factors are then written directly in terms of the Hecke eigenvalues \(\lambda_p\), \(\lambda_{p^2}\), \(\lambda_{\mathfrak p}\), and \(\lambda_{\overline{\mathfrak p}}\) [2505.23497].

A distinct Fourier-coefficient operator is the Hermitian theta operator. In degree \(2\), it is defined by
\[
\Theta\!\left(\sum_H a(F;H)q_H\right)=\sum_H \det(H)\,a(F;H)\,q_H.
\]
Over \(\mathbb{C}\), \(\Theta(F)\) need not be modular. Over \(\mathbb{Z}_{(p)}\), however, a modularity-lifting result holds: for \(p\ge 5\), \(\Theta(F)\) is congruent modulo \(p\) to a cusp form of weight \(k+p+1\). This is used to define the mod \(p\) kernel of \(\Theta\) and to prove explicit congruences such as
\[
\Theta(E^{(2)}_{12,K})\equiv \Theta(\Theta^{(2)}(Z;H_4))\equiv \Theta(\Theta^{(2)}(Z;H_5))\equiv 0 \pmod{11}
\]
over the Eisenstein field, and
\[
\Theta(E_{p+1,K}^{(2)})\equiv 0 \pmod p
\]
for class number \(1\) when \(\chi_K(p)=-1\) in the Gaussian-field setting [1806.10326].

The heat operators are Jacobi- and modular-form analogues of \(\Theta\). For Hermitian Jacobi forms over \(\mathbb{Q}(i)\),
\[
L_m=-\frac{1}{2\pi i}\left(\frac{\partial^2}{\partial z_1\partial z_2}-2\pi i\,m\,\frac{\partial}{\partial \tau}\right),
\]
and for degree-\(2\) Hermitian modular forms,
\[
D=-\frac{1}{2\pi i}\left(\frac{\partial^2}{\partial \tau\,\partial \tau'}-\frac{\partial^2}{\partial z_1\,\partial z_2}\right).
\]
On Fourier coefficients, \(D\) multiplies \(A_F(n,r,m)\) by \(4(nm-N(r))\). These operators control \(U(p)\)-congruences and Ramanujan-type congruences. For example, \(F\) has a Ramanujan-type congruence at \(b\) if and only if
\[
D^2(F)\equiv -\left(\frac{b}{p}\right)D(F)\pmod p,
\]
and there are precise filtration criteria distinguishing the cases \(F|U(p)\equiv 0\) and \(F|U(p)\not\equiv 0\) [1908.05980].

## 5. Vector-valued Jacobian and Rankin–Cohen constructions

In low-dimensional Hermitian theory, vector-valued operators are often built directly from the Jacobian of the modular action. For \(U(n,n)\), the Jacobian of \(Z\mapsto MZ\) on \(C^{n\times n}\) is
\[
\operatorname{Jac}(M,Z)(W)=(CZ'+D)'^{-1}W(CZ+D)^{-1},
\]
with determinant
\[
\det \operatorname{Jac}(M,Z)=\det(M)^{-n}\det(CZ+D)^{-2n}.
\]
This leads to vector-valued transformation laws of \(St\otimes St\)-type. In particular, for \(Q=St\otimes St\), a \(C^{n\times n}\)-valued form transforms by
\[
f(MZ)=\chi(M)\det(CZ+D)^r (CZ+D)f(Z)(CZ'+\overline{D})'.
\]
The same paper develops an analogous Jacobian automorphy factor for the quaternionic case, but the Hermitian case is the relevant template for Hermitian modular operators in the sense of vector-valued differential constructions [1402.1929].

The basic first-order operator is the Rankin–Cohen bracket
\[
\{f,g\}=g^2\,d(f/g),
\]
which takes scalar-valued generators to \(St\otimes St\)-valued forms. In the Eisenstein and Gaussian cases, the graded modules of vector-valued Hermitian modular forms are generated over the scalar ring by such brackets among finitely many theta-constant generators. For the Eisenstein field,
\[
M=\sum_{1\le i<j\le 5} A\cdot \{\Theta_i,\Theta_j\},
\]
and for the Gaussian field,
\[
N=\sum_{1\le i<j\le 5} A\cdot \{\Theta(i)^2,\Theta(j)^2\},
\]
with defining skew-symmetry and Plücker-type relations. Determinant identities involving matrices of brackets and explicit cusp forms control holomorphy and denominators [1402.1929].

The broader Rankin–Cohen theory on Hermitian modular forms of signature \((n,n)\) extends these constructions from first-order brackets to bilinear differential operators of arbitrary order \(v\). For \(r=2\), the bilinear bracket is
\[
\{f,g\}^{(v)}(Z)=Q\!\left(\frac{\partial}{\partial Z_1},\frac{\partial}{\partial Z_2}\right)\bigl(f(Z_1)g(Z_2)\bigr)\Big|_{Z_1=Z_2=Z},
\]
and has weight \(k_1+k_2+2v\). For \(n>1\) and \(v>0\), the operator is unique up to rescaling, and for \(v>0\) its image consists of cusp forms. When \(n=1\), the construction specializes to the classical Rankin–Cohen brackets [2404.04371].

A common misconception is that vector-valued Hermitian modular operators are merely ad hoc derivatives. The recent literature instead ties them to explicit representation theory: covariance is governed by \(GL(n,\mathbb{C})\times GL(n,\mathbb{C})\)-types, pluriharmonicity is enforced by mixed Laplacians, and determinant identities or symbol maps ensure precise automorphic behavior [1402.1929].

## 6. Broader operator landscapes and terminological ambiguity

The phrase “Hermitian modular operator” is not uniform across the literature. In arithmetic geometry and automorphic forms it refers to operators acting on Hermitian modular forms, such as the Hecke, theta, heat, Jacobian, and differential operators described above. In operator algebra and quantum field theory, by contrast, “modular operator” refers to the Tomita–Takesaki modular operator \(\Delta\) associated with a von Neumann algebra \(A(O)\) and cyclic separating vector \(\Omega\), with polar decomposition
\[
S=J\Delta^{1/2},
\]
modular automorphism group
\[
\sigma_t(A)=\Delta^{it}A\Delta^{-it},
\]
and modular Hamiltonian
\[
K=-\log \Delta.
\]
That usage is conceptually distinct from Hermitian modular operators in the arithmetic theory of modular forms [2506.00504].

The distinction matters because the same word “modular” labels very different structures. In the QFT setting, the bounded Hermitian observables used in Bell–CHSH analysis are functions of smeared fields, for example
\[
X(f)=\frac{W(f)+W(f)^\dagger}{2}=\cos(\phi(f)),\qquad
Y(f)=\frac{W(f)-W(f)^\dagger}{2i}=\sin(\phi(f)),
\]
while \(\Delta\) and \(K\) belong to modular theory in the von Neumann algebraic sense. This suggests a terminological separation between arithmetic Hermitian modular operators and Tomita–Takesaki modular operators, even though both are built from highly structured transformation theories [2506.00504].

Within arithmetic Hermitian theory itself, the operator landscape is still expanding. A recent conjectural correspondence relates Hermitian modular forms of degree \(2\) to algebraic modular forms on \(SO(6)\), with Hecke operators, Atkin–Lehner involutions, and a theta map entering the proposed dictionary. Under that conjecture, unramified Hecke eigenvalues on the Hermitian side match Kneser-neighbor Hecke data on the \(SO(6)\) side, and the theta-map criterion is expected to detect Sugano Maass space forms [2505.23497]. This suggests that the study of Hermitian modular operators is increasingly representation-theoretic, linking explicit operator formulas to conjectural functorial correspondences.

Overall, the modern theory presents Hermitian modular operators not as a single construction but as a family of exact mechanisms for controlling automorphy, vector-valued structure, Fourier expansions, congruences, and \(L\)-functions. Differential criteria based on pluriharmonicity, Hecke-theoretic commutativity and local generators, mod \(p\) theta and heat operations, and Jacobian/Rankin–Cohen constructions together form the operative toolkit of current Hermitian modular-form theory [2506.18236].

Source: https://www.emergentmind.com/topics/hermitian-modular-operators