On a divisor modular form and a theta lift
Andreas Mono, Larry Rolen, Johann Stumpenhusen · 2025
Read On a divisor modular form and a theta lift by Andreas Mono, Larry Rolen, Johann Stumpenhusen (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-2509-01378.txt.
Introduction
Modular forms are highly symmetric functions central to number theory, and in 1975 Zagier introduced a family of hyperbolic Poincare series that became very influential. Andreas Mono, Larry Rolen, and Johann Stumpenhusen connect the divisor modular form built from those series to a new object, a weak Maass form with shifted weight. Maass forms extend modular forms to a broader, less rigid class.
The team shows the generating function of their new form shares the same modularity traits as the theta kernel of Kohnen and Zagier, the tool that produces Zagier's series. That match yields a new theta lift, a construction that moves forms from one setting to another. Such lifts are prized because they turn hard questions in one world into workable ones in another, linking arithmetic objects across theories.
Catalog record
| CREATOR | Andreas Mono, Larry Rolen, Johann Stumpenhusen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 41 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2509-01378
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On a divisor modular form and a theta lift Thanks: The first author received support for this research by an AMS–Simons Travel Grant. The second author’s work was supported by a grant from the Simons Foundation (853830, LR) Andreas Mono, Larry Rolen, Johann Stumpenhusen Abstract . In 1975 1975 , Zagier introduced the highly influential hyperbolic Poincaré series f k , D f_{k,D} . We connect the divisor modular form of f k , D f_{k,D} to a new weak Maaß form ω k + 1 , D \omega_{k+1,D} . Furthermore, we show that the generating function of ω k + 1 , D \omega_{k+1,D} has the same modularity properties as Kohnen and Zagier’s fruitful theta kernel generating the f k , D f_{k,D} ’s. This yields a new theta lift. 2020 Mathematics Subject Classification 11F11 (Primary); 11E16, 11F27 (Secondary) 1. Introduction and Statement of Results Throughout, we let 2<k\in 2\mathbb{N} , D\in\mathbb{N} be a discriminant, and \displaystyle\mathcal{Q}_{D}\coloneqq\left\{Q(X,Y)=aX^{2}+bXY+cY^{2}\colon a,b,c\in\mathbb{Z},\ b^{2}-4ac=D\right\}. In 1975, Zagier [ zagier75 ] *Appendix 2 introduced the function \displaystyle f_{k,D}(z)\coloneqq\sum_{Q\in\mathcal{Q}_{D}}\frac{1}{Q(z,1)^{k}},\qquad z\in\mathbb{H}\coloneqq\left\{z=x+iy\in\mathbb{C}\colon y>0\right\}, which turns out to be a weight 2k cusp form for {\text{\rm SL}}_{2}(\mathbb{Z}) (see Definition 2.3 ). Soon after that, Katok [ katok ] showed that f_{k,D} can be written as a hyperbolic Poincaré series by averaging over z^{-k} , which is the constant term in the hyperbolic expansion of a modular form. Such expansions go back to Petersson [ pet43 ] , and an excellent survey on such expansions can be found in [ IOS ] . For this reason, the f_{k,D} -functions generate the space of weight 2k cusp forms for varying D , paralleling the parabolic situation of the usual holomorphic Poincaré series of exponential type (see ( 2.4 )). Bengoechea [ beng13 ] studied the f_{k,D} functions for negative discriminants D , which have poles at the corresponding CM points. Let \displaystyle j(z)\coloneqq q^{-1}+744+O(q),\qquad q\coloneqq e^{2\pi iz}, be Klein’s modular invariant for {\text{\rm SL}}_{2}(\mathbb{Z}) and, for every n\in\mathbb{N}_{0} , let j_{n} be the unique modular function for {\text{\rm SL}}_{2}(\mathbb{Z}) having a Fourier expansion of the form q^{-n}+O(q) . The first few are explicitly given by \displaystyle j_{0}(z)=1,\qquad j_{1}(z)=j(z)-744,\qquad j_{2}(z)=j(z)^{2}-1488j(z)+159768, and the j_{n} generate the space of weakly holomorphic modular functions for {\text{\rm SL}}_{2}(\mathbb{Z}) (see Definition 2.3 ). For example, the functions j_{n} were central to Zagier’s [ zagier02 ] work on traces of singular moduli. It is natural to consider the generating function (1.1) \displaystyle H_{z}(\tau) \displaystyle\coloneqq\sum_{n\geq 0}j_{n}(z)e^{2\pi in\tau},\qquad\tau=u+iv\in\mathbb{H}. If v>y , Asai, Kaneko, and Ninomiya [ askani ] established the expansion \displaystyle H_{z}(\tau)=\frac{\frac{1}{2\pi i}\frac{\partial}{\partial\tau}\left(j(\tau)\right)}{j(z)-j(\tau)}, which shows that H_{z}(\tau) is a meromorphic modular form of weight 2 for {\text{\rm SL}}_{2}(\mathbb{Z}) (see Definition 2.3 ). Their identity is equivalent to the denominator formula for the Monster Lie algebra. Let f be a weight k meromorphic modular form (see Definition 2.3 ). The divisor modular form of f is defined as \displaystyle f^{\mathrm{div}}(\tau)\coloneqq\sum_{z\in{\text{\rm SL}}_{2}(\mathbb{Z})\backslash\mathbb{H}}\frac{2}{\left\lvert\mathrm{Stab}_{z}({\text{\rm SL}}_{2}(\mathbb{Z}))\right\rvert}\ \mathrm{ord}_{z}(f)H_{z}(\tau). More explicitly, we have (using ( 2.1 )) \displaystyle\frac{2}{\left\lvert\mathrm{Stab}_{z}({\text{\rm SL}}_{2}(\mathbb{Z}))\right\rvert}=\begin{cases}\frac{1}{2}&\text{if }z\sim i,\\ \frac{1}{3}&\text{if }z\sim e^{\frac{2\pi i}{3}},\\ 1&\text{else},\end{cases} where z\sim z^{\prime} indicates that z and z^{\prime} lie in the same orbit of \mathbb{H} under the action of {\text{\rm SL}}_{2}(\mathbb{Z}) . By definition, the divisor modular form f^{\mathrm{div}} is a weight 2 meromorphic modular form detecting the zeros and poles of f . If f is normalized such that its first non-zero Fourier coefficient is equal to 1 , then Bruinier, Kohnen, and Ono [ brukoon ] *Theorem 1.1 showed that (1.2) \displaystyle\frac{1}{2\pi i}\frac{\partial f}{\partial z}=\frac{k}{12}E_{2}f-ff^{\mathrm{div}},\qquad E_{2}(z)\coloneqq 1-24\sum_{n=1}^{\infty}\sum_{d\mid n}dq^{n}. Recently, Bringmann, Kane, Löbrich, Ono and the second author [ brikaloeonro ] generalized the j_{n} to \Gamma_{0}(N) and showed that their generalizations are also related to divisor modular forms. Consequently, it is natural to investigate f_{k,D}^{\mathrm{div}} . To this end, we introduce the quantity \displaystyle Q_{z}\coloneqq\frac{1}{y}\left(a\left\lvert z\right\rvert^{2}+bx+c\right)\in\mathbb{R},\qquad Q=[a,b,c]\in\mathcal{Q}_{D}, which encodes the hyperbolic geodesic S_{Q}\coloneqq\left\{z\in\mathbb{H}\colon Q_{z}=0\right\}. The endpoints of the geodesic S_{Q} are the two roots of Q(z,1) . Since D>0 throughout, these roots are contained in \mathbb{R}\cup\{\infty\} (if a=0 then b\neq 0 and the two roots are -\frac{c}{b} and \infty ). Hence, the function \displaystyle\omega_{k+1,D}(z)\coloneqq\sum_{Q\in\mathcal{Q}_{D}}\frac{Q_{z}}{Q(z,1)^{k+1}} has no poles on \mathbb{H} and converges absolutely for k>2 (see Proposition 3.1 ). Roughly speaking, weak Maaß forms are certain real-analytic generalizations of (complex-analytic) modular forms. In particular, they satisfy the same modular transformation laws as modular forms. Our definition of weak Maaß forms follows Bruinier and Funke [ brufu02 ] , see Definition 2.4 below. Our first main result collects the properties of \omega_{k+1,D} . Theorem 1.1 . Let 2<k\in 2\mathbb{N} and D\in\mathbb{N} be a discriminant. (i) The function \omega_{k+1,D} is a weak Maaß form of weight 2k+2 for {\text{\rm SL}}_{2}(\mathbb{Z}) and eigenvalue 2k under the hyperbolic Laplace operator \Delta_{2k+2} (defined in ( 2.6 )). Furthermore, it satisfies \displaystyle\lim_{z\to i\infty}\omega_{k+1,D}(z)=0. (ii) A splitting of \omega_{k+1,D} into a holomorphic and a non-holomorphic part is given by \displaystyle\omega_{k+1,D}(z)=-i\sum_{Q\in\mathcal{Q}_{D}}\frac{Q^{\prime}(z,1)}{Q(z,1)^{k+1}}+\frac{1}{y}f_{k,D}(z),\qquad Q^{\prime}(z,1)\coloneqq\frac{\mathrm{d}}{\mathrm{d}z}Q(z,1). (iii) Let c_{f_{k,D}} be the first non-vanishing Fourier coefficient of f_{k,D} in its Fourier expansion about i\infty (see Zagier [ zagier75 ] *p. 45). Let \displaystyle E_{2}^{*}(z)\coloneqq E_{2}(z)-\frac{3}{\pi y} be the non-holomorphic Eisenstein series of weight 2 . Then, we have \displaystyle\frac{1}{c_{f_{k,D}}}f_{k,D}^{\mathrm{div}}(z)=\frac{k}{2\pi}\frac{\omega_{k+1,D}(z)}{f_{k,D}(z)}+\frac{k}{6}E_{2}^{*}(z). Remarks . (1) Further motivation for \omega_{k+1,D} arises from other functions of Zagier [ zagier77 ] *eq. (48), work of the first author [ mon22 , mon24 ] , and work of Bringmann and the first author [ brimo ] . (2) Following Zagier, [ zagier75 ] *p. 44, the function \displaystyle\omega_{k+1,0}(z)=\frac{1}{\zeta(k+1)}\sum_{\gamma\in{\text{\rm SL}}_{2}(\mathbb{Z})_{\infty}\backslash{\text{\rm SL}}_{2}(\mathbb{Z})}y^{-1}\big|_{k+1}\gamma is a non-holomorphic Eisenstein series. Here, \zeta denotes the Riemann \zeta -function, the Petersson slash operator is defined in ( 2.2 ), and {\text{\rm SL}}_{2}(\mathbb{Z})_{\infty} is defined in ( 2.3 ). In the 1980 ’s, Kohnen and Zagier [ koza81 , koza84 ] introduced the two-variable generating function \displaystyle\Omega_{k}(\tau,z)\coloneqq\sum_{D>0}D^{k-\frac{1}{2}}f_{k,D}(z)e^{2\pi iD\tau} of the f_{k,D} . This can be viewed as a “hyperbolic analog” of the functions H_{z}(\tau) from ( 1.1 ). Employing a result by Vignéras [ vign1 , vign2 ] (see Theorem 2.5 and the subsequent discussion), one can show that \Omega_{k}(\cdot,z) is a cusp form of weight k+\frac{1}{2} for \Gamma_{0}(4) in Kohnen’s plus space (see Section 2.2 for the definition) for fixed z\in\mathbb{H} . Alternatively, Kohnen [ koh85 ] *Theorem 1 proved that \displaystyle\Omega_{k}(\tau,z)=\frac{(-1)^{\frac{k}{2}}3(2\pi)^{k}}{(k-1)!}\sum_{m\geq 1}m^{k-1}\sum_{d\mid m}d^{k}P_{k+\frac{1}{2},d^{2}}^{+}(\tau)q^{m}, where P_{k+\frac{1}{2},d^{2}}^{+} is the weight k+\frac{1}{2} holomorphic Poincaré series of exponential type in the plus space (see ( 2.5 )). Kohnen and Zagier proved that \Omega_{k} , utilized as a kernel function of a theta lift, realizes the famous Shimura [ shim ] and Shintani [ shin ] lifts. From this, they established that the even periods of f_{k,D} are rational and deduced deep results on (twisted) central L -values of cusp forms. More recent work on such values can be found in work of Ehlen, Guerzhoy, Kane and the second author [ egkr ] as well as of Males, Wagner and the first two authors [ mmrw ] . Imitating the construction of \Omega_{k} , we define \displaystyle\Lambda_{k}(\tau,z)\coloneqq\sum_{D>0}D^{k-\frac{1}{2}}\omega_{k+1,D}(z)e^{2\pi iD\tau}. Our second result shows that \Lambda_{k} shares the aforementioned modularity properties with \Omega_{k} . Theorem 1.2 . The function \Lambda_{k}(\cdot,z) is modular of weight k+\frac{1}{2} for \Gamma_{0}(4) in the plus space. Remark . In contrast to \Omega_{k} , our function \Lambda_{k} is not holomorphic as a function of z . Consequently, \Lambda_{k} gives