Siegel modular forms associated to Weil representations

Chun-Hui Wang · 2025

Read Siegel modular forms associated to Weil representations by Chun-Hui Wang (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2501-12140.txt.

Introduction

Chun-Hui Wang builds special symmetry functions from a classic representation in a 2025 math paper. Siegel modular forms are functions with rich symmetry on a space of matrices, prized in number theory. Weil representations give one factory for them, through the old theta group.

A snag: eighth roots of unity cling to these forms, noted by many past authors. Wang attacks them with a tool called a 2-cocycle, drawn from a line of past work. The study then widens to the full Siegel group, yielding two matrix-valued forms. A quotient group of Igusa's ties the pieces together.

Catalog record

CREATORChun-Hui Wang
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE185 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-12140

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Siegel modular forms associated to Weil representations Chun-Hui Wang Abstract . We study some explicit Siegel modular forms from Weil representations. For the classical theta group Γ m ​ ( 1 , 2 ) \Gamma_{m}(1,2) with m > 1 m>1 , there are some eighth roots of unity associated with these modular forms, as noted in the works of Andrianov, Friedberg, Maloletkin, Stark, Styer, Richter, and others. We apply 2 2 -cocycles introduced by Rao, Kudla, Perrin, Lion-Vergne, Satake-Takase to investigate these unities. We extend our study to the full Siegel group Sp 2 ​ m ⁡ ( ℤ ) \operatorname{Sp}_{2m}(\mathbb{Z}) and obtain two matrix-valued Siegel modular forms from Weil representations; these forms arise from a finite-dimensional representation Ind Γ ~ m ′ ​ ( 1 , 2 ) Sp ~ 2 ​ m ′ ​ ( ℤ ) ​ ( 1 Γ m ​ ( 1 , 2 ) ⋅ Id μ 8 ) − 1 \operatorname{Ind}_{\widetilde{\Gamma}^{\prime}_{m}(1,2)}^{\widetilde{\operatorname{Sp}}^{\prime}_{2m}(\mathbb{Z})}(1_{\Gamma_{m}(1,2)}\cdot\operatorname{Id}_{\mu_{8}})^{-1} , which is related to Igusa’s quotient group Sp 2 ​ m ⁡ ( ℤ ) Γ m ​ ( 4 , 8 ) \tfrac{\operatorname{Sp}_{2m}(\mathbb{Z})}{\Gamma_{m}(4,8)} . 2010 Mathematics Subject Classification 11F27, 20C25 1. Introduction 1.1. Notations and conventions Let \mathbb{R} and \mathbb{C} denote the usual real numbers and complex numbers. Let (W,\langle,\rangle) be a symplectic vector space of dimension 2m over \mathbb{R} . Let \operatorname{H}(W)=W\oplus\mathbb{R} denote the usual Heisenbeg group, defined by (w,t)\cdot(w^{\prime},t^{\prime})=(w+w^{\prime},t+t^{\prime}+\tfrac{\langle w,w^{\prime}\rangle}{2}),\quad\quad w,w^{\prime}\in W,t,t^{\prime}\in\mathbb{R}. Let \operatorname{Sp}(W) denote the corresponding symplectic group. Let \{e_{1},\cdots,e_{m};e_{1}^{\ast},\cdots,e_{m}^{\ast}\} be a symplectic basis of W . Let X=\operatorname{Span}_{\mathbb{R}}\{e_{1},\cdots,e_{m}\} , X^{\ast}=\operatorname{Span}_{\mathbb{R}}\{e_{1}^{\ast},\cdots,e_{m}^{\ast}\} . In \operatorname{Sp}(W) , let P_{X^{\ast}}(W)=\{g\in\operatorname{Sp}(W)\mid X^{\ast}g=X^{\ast}\}=\{\begin{pmatrix}a&b\\ 0&(a^{\ast})^{-1}\end{pmatrix}\in\operatorname{Sp}(W)\} , N_{X^{\ast}}(W)=\{\begin{pmatrix}1_{m}&b\\ 0&1_{m}\end{pmatrix}\in\operatorname{Sp}(W)\} , M_{X^{\ast}}(W)=\{\begin{pmatrix}a&0\\ 0&(a^{\ast})^{-1}\end{pmatrix}\in\operatorname{Sp}(W)\} , N^{-}_{X^{\ast}}(W)=\{\begin{pmatrix}1_{m}&0\\ c&1_{m}\end{pmatrix}\in\operatorname{Sp}(W)\} . For a subset S\subseteq\{1,\cdots,m\} , let us define an element \omega_{S} of \operatorname{Sp}(W) as follows: (e_{i})\omega_{S}=\left\{\begin{array}[]{lr}-e_{i}^{\ast}&i\in S,\\ e_{i}&i\notin S,\end{array}\right. and (e_{i}^{\ast})\omega_{S}=\left\{\begin{array}[]{lr}e_{i}^{\ast}&i\notin S,\\ e_{i}&i\in S.\end{array}\right. If S=\{1,\cdots,m\} , we write \omega for simplicity. For elements within \operatorname{Sp}(W) , we denote by u(b)=\begin{pmatrix}1_{m}&b\\ 0&1_{m}\end{pmatrix} , u_{-}(c)=\begin{pmatrix}1_{m}&0\\ c&1_{m}\end{pmatrix} , h(a)=\begin{pmatrix}a&0\\ 0&(a^{\ast})^{-1}\end{pmatrix} . Let S_{m} denote the permutation group on m elements. For s\in S_{m} , let \omega_{s} be an element of \operatorname{Sp}(W) defined as (e_{i})\omega_{s}=e_{s(i)} and (e_{j}^{\ast})\omega_{s}=e_{s(j)}^{\ast} . Let \mathfrak{W}_{P_{X^{\ast}}}=\{\omega_{S_{i}}\mid S_{0}=\emptyset\textrm{ or }S_{i}=\{1,\cdots,i\}\} and \mathfrak{W}=\{\omega_{s}\omega_{S}\mid S\subseteq\{1,\cdots,m\},s\in S_{m}\} . Given the basis \{e_{1},\cdots,e_{m};e_{1}^{\ast},\cdots,e_{m}^{\ast}\} of W , we identity \operatorname{Sp}(W) with \operatorname{Sp}_{2m}(\mathbb{R}) . We use the same notations for elements of these two groups and replace W by \mathbb{R} for the corresponding subgroups. Let N_{0}(\mathbb{R})=\{h(a)\in\operatorname{Sp}_{2m}(\mathbb{R})\mid a=\begin{pmatrix}1&*&*\\ &\ddots&*\\ &&1\end{pmatrix}\in\operatorname{GL}_{m}(\mathbb{R})\} . Let N^{-}_{0}(\mathbb{R})=\omega^{-1}N_{0}(\mathbb{R})\omega , N(\mathbb{R})=N_{X^{\ast}}(\mathbb{R})N_{0}(\mathbb{R}) and N^{-}(\mathbb{R})=N_{X^{\ast}}^{-}(\mathbb{R})N^{-}_{0}(\mathbb{R}) . For the elements in \operatorname{M}_{m}(\mathbb{R}) or \operatorname{M}_{2m}(\mathbb{R}) , we denote by \epsilon_{ij}(t) the m\times m -matrix where the entry in the i -th row and j -th column set to t and all other entries set to 0 . We then define the matrices u_{ii}(t),u^{-}_{ii}(t),u_{ij}(t),u_{ij}^{-}(t) , v_{ij}(t) ( i\neq j ) as follows: u_{ii}(t)=\begin{pmatrix}1_{m}&\epsilon_{ii}(t)\\ 0&1_{m}\end{pmatrix} , u^{-}_{ii}(t)=\begin{pmatrix}1_{m}&0\\ \epsilon_{ii}(-t)&1_{m}\end{pmatrix} , u_{ij}(t)=\begin{pmatrix}1_{m}&\epsilon_{ij}(t)+\epsilon_{ji}(t)\\ 0&1_{m}\end{pmatrix} ,and u_{ij}^{-}(t)=\begin{pmatrix}1_{m}&0\\ \epsilon_{ij}(-t)+\epsilon_{ji}(-t)&1_{m}\end{pmatrix} , v_{ij}(t)=\begin{pmatrix}1_{m}+\epsilon_{ij}(t)&0\\ 0&1_{m}+\epsilon_{ji}(-t)\end{pmatrix} . If t=1 , we simply write them as u_{ii} , u_{ii}^{-} , u_{ij} , u_{ij}^{-} , v_{ij} . If \psi is a non-trivial character of \mathbb{R} and a\in\mathbb{R} , we will write \psi^{a} for the character: t\longrightarrow\psi(at) . Let i=\sqrt{-1} . We will let \psi_{0} denote the fixed character of \mathbb{R} defined as: t\longmapsto e^{2\pi it} , for t\in\mathbb{R} . Let \mu_{n}=\langle e^{\tfrac{2\pi i}{n}}\rangle , e^{\tfrac{2\pi i}{n}}\in\mathbb{C}^{\times} . Let T=\{e^{i\theta}\mid\theta\in\mathbb{R}\} . For a finite set G , let |G| denote the cardinality of G . On \mathbb{R}^{m} , let du denote the Lebsegue measure. On \mathbb{C}^{m} , let dz=du+idv , and d(z)=dudv . 1.2. Weil representations Let \psi be a non-trivial continuous unitary character of \mathbb{R} . According to the Stone-von Neumann’s theorem, there exists only one unitary irreducible complex representation of \operatorname{H}(W) with central character \psi , known as the Heisenberg representation. Details are in Section 3 . According to Weil’s work, the Heisenberg representation can be extended to a projective representation of \operatorname{Sp}(W) , and then to an actual representation of a \mathbb{C}^{\times} -covering group over \operatorname{Sp}(W) . This group is called the Metaplectic group, and the representation is called the Weil representation. In the real field case, this special central cover can descend to an 8 -degree or 2 -degree cover over \operatorname{Sp}(W) . There are three common models for realizing the Weil representation: the Schödinger model, the Lattice model and the Fock model. Details are in Section 4 . 1.3. Siegel modular forms Recall that the complex domain \mathbb{H}_{m}=\{x+iy\in\operatorname{M}_{m}(\mathbb{C})\mid x^{T}=x,y^{T}=y,y>0\} is called the Siegel upper half space. There exists an action of \operatorname{Sp}_{2m}(\mathbb{R}) on \mathbb{H}_{m} , given by z\longrightarrow g(z)=(az+b)(cz+d)^{-1}, for g=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\operatorname{Sp}_{2m}(\mathbb{R}) , z\in\mathbb{H}_{m} . Let \operatorname{U}_{m}(\mathbb{C})=\{g\in\operatorname{Sp}_{2m}(\mathbb{R})\mid g(i1_{m})=i1_{m}\}=\{\begin{pmatrix}a&b\\ -b&a\end{pmatrix}\in\operatorname{Sp}_{2m}(\mathbb{R})\} . It is clear that \mathbb{H}_{m}\simeq\operatorname{Sp}_{2m}(\mathbb{R})/\operatorname{U}_{m}(\mathbb{C}) . Throughout the text, we will fix an element z_{0}=i1_{m}\in\mathbb{H}_{m} . Assuming m>1 , following [ Fr2 , Ge , Kl , Sc ] , let \rho:\operatorname{GL}(m,\mathbb{C})\to\operatorname{GL}(V_{\rho}) be a rational representation on a finite dimensional complex vector space V_{\rho} . Let \Gamma be a discrete group of \operatorname{Sp}_{2m}(\mathbb{R}) . Let \lambda(-) be a multiplier system of weight 1/2 , with respect to \Gamma . Let \gamma be a representations of \Gamma . Let M_{\rho,\gamma} be a vector space over \mathbb{C} that is a left \rho -module and a right \gamma -module, comprised of matrices. Definition 1.1 . A matrix-valued function M:\mathbb{H}_{m}\to M_{\rho,\gamma} is called a M_{\rho,\gamma} -valued Siegel modular form with a multiplier system \lambda(-) of weight 1/2 , with respect to \Gamma if: (1) M is holomorphic, (2) M(gz)=\lambda(g)\sqrt{\det(cz+d)}\rho(cz+d)M(z)\gamma(g^{-1}) , for any g=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\Gamma . This definition is mainly for number-theoretic research and giving clear examples. In Langlands programs, it relates to automorphic forms and automorphic representations. For deeper understanding, see Borel and Jacquet’s paper [ BoJa ] , or Takase’s [ Ta0 ] . As shown in [ BeCaRaRo ] , some central coverings over discrete groups don’t split. In such cases, the definition can be extended to a covering group \widetilde{\operatorname{Sp}}_{2m}(\mathbb{R}) by replacing \operatorname{U}_{m}(\mathbb{C}) and \Gamma with their covering subgroups \widetilde{\operatorname{U}}_{m}(C) and \widetilde{\Gamma} . 1.4. The main results 1.4.1. For z\in\mathbb{H}_{m} , the theta function and the theta vector function associated with the Weil representation are defined as: \displaystyle\theta_{1/2}(z) \displaystyle=\sum_{n\in\mathbb{Z}^{m}}e^{i\pi nzn^{T}}, \displaystyle\theta_{3/2}(z) \displaystyle=\sum_{n\in\mathbb{Z}^{m}}n^{T}e^{i\pi nzn^{T}}. Let \Gamma_{m}(1,2)=\{g=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\operatorname{Sp}_{2m}(\mathbb{Z})\mid\textrm{ diagronal }ac^{T}\textrm{ even and }\textrm{ diagronal }bd^{T}\textrm{ even}\} . For g=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\operatorname{Sp}_{2m}(\mathbb{R}) , z\in\mathbb{H}_{m} , r\in\Gamma_{m}(1,2) , we define: (1) J(g,z)=cz+d . (2) J_{1/2}(g,z)=\epsilon(g;z,z_{0})\cdot|\det J(g,z)|^{1/2} . (3) \sqrt{\det(cz+d)}=J_{1/2}(g,z)m_{X^{\ast}}(g)^{-1} . (4) \lambda(r)=m_{X^{\ast}}(r)\widetilde{\beta}^{-1}(r) . Theorem 1.2 (Thm. 8.15 ) . For r=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\Gamma_{m}(1,2) , we have: (1) \theta_{1/2}(rz)=\lambda(r)\sqrt{\det(cz+d)}\theta_{1/2}(z) . (2) \theta_{3/2}(rz)=\lambda(r)\sqrt{\det(cz+d)}(cz+d)\theta_{3/2}(z) . It is well-known that theta functions have a long and rich history. For examples, one can see [ Be ] , [ Fr1 ] , [ Fr2 ] , [ Ig ] , [ Kl ] , [ Mu ] , [ Na ] , [ Ri1 , Ri2 ] , [ Sc ] , [ Ta2 ] , [ Vi ] , [ Ya1 ] . Here, we approach them in a more personal way, using the theory of Weil representations. The Weil representation \pi_{\psi_{0}} has two irreducible components, and the above theta functions correspond to each component in a one-to-one manner. Here, we derive Theorem 1.2 (1) using a single irreducible Weil representation of the Metaplectic group, without the assistance of Jacobi forms and Jacobi groups. To obtain the above theorem, we mainly examined and compared three

Read the complete text