Joint equidistribution of newforms

Asbjørn Christian Nordentoft 1, Radu Toma 2 · 2025

Read Joint equidistribution of newforms by Asbjørn Christian Nordentoft 1, Radu Toma 2 (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-2509-01602.txt.

Introduction

Equidistribution asks whether a sequence of measures spreads evenly over a space in the long run, a deep question in number theory. Asbjorn Christian Nordentoft and Radu Toma study newforms, special functions on arithmetic hyperbolic surfaces built from quaternion orders. Covers of growing prime level carry these forms, and a Hecke correspondence embeds each cover into a product of the base surface with itself.

The pair conjecture that the measure built from a newform's squared size, pushed forward through that embedding, converges to the uniform measure as the level grows. They prove the conjecture, with an effective rate for the convergence, assuming the generalized Riemann hypothesis. That assumption is stated openly in the work. The result links the fine distribution of automorphic forms to uniform spread, a theme at the heart of modern analytic number theory.

Catalog record

CREATORAsbjørn Christian Nordentoft 1, Radu Toma 2
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE134 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-01602

Browse the full library · About the Tower of Babel

Full text

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

Joint equidistribution of newforms Asbjørn Christian Nordentoft 1, Radu Toma 2 Abstract . Let Y 1 Y_{1} be a compact arithmetic hyperbolic surface associated to a maximal quaternion order, let Y q Y_{q} be a cover associated to an Eichler suborder of prime level q q , and let ι q \iota_{q} be embedding of Y q Y_{q} as the Hecke correspondence into Y 1 × Y 1 Y_{1}\times Y_{1} . Let μ 1 \mu_{1} and μ q \mu_{q} be the invariant probability measures on Y 1 Y_{1} and Y q Y_{q} , respectively. If F F is a newform on Y q Y_{q} , we conjecture that the pushforward measure ( ι q ) ∗ ​ ( | F | 2 ​ μ q ) (\iota_{q})_{\ast}(\lvert{F}\rvert^{2}\mu_{q}) converges weakly to the uniform measure μ 1 × μ 1 \mu_{1}\times\mu_{1} , as q q tends to infinity. We prove this conjecture with an effective rate of equidistribution, assuming GRH. 1. Introduction Understanding the distribution of mass of Laplace eigenfunctions on hyperbolic manifolds is a profound subject with roots in quantum chaos (see [ Ber02 ] ). It has proven particularly fruitful to consider arithmetic manifolds and employ number theory to address this problem (see [ Sar95 ] ). This article proposes a novel and deeper study of the interplay between quantum chaos and number theory by investigating an equidistribution problem that combines both quantum chaos and the equidistribution of Hecke points. Let B be an indefinite quaternion algebra over \mathbb{Q} (in this introduction, the reader may simply consider the split case B=M_{2} ). Choose a maximal order and, for each integer q\geq 1 coprime to the discriminant of B , take an Eichler order of level q and define its associated hyperbolic surface Y_{q}=\Gamma_{q}\backslash\mathbb{H} (in the split case, these are the modular curves Y_{0}(q) of level q ). We endow Y_{1} and Y_{q} with invariant probability measures \mu_{1} and \mu_{q} , respectively. There are two natural covering maps \alpha_{q},\beta_{q}:Y_{q}\rightarrow Y_{1} , given by z\mapsto z and z\mapsto qz in the split case. Together, they define the Hecke correspondence \iota_{q}:Y_{q}\rightarrow Y_{1}\times Y_{1}, by setting \iota_{q}=(\alpha_{q},\beta_{q}) . If F is a cuspidal Hecke–Maaß newform on Y_{q} , L^{2} -normalized with respect to \mu_{q} , then \mu_{F}:=\lvert{F}\rvert^{2}\mu_{q} defines a new probability measure on Y_{q} . It was proposed by Kowalski–Michel–VanderKam [ KMV02 ] that the push-forward measures (\alpha_{q})_{\ast}\mu_{F} on Y_{1} should converge to the uniform measure \mu_{1} as q goes to infinity. This is an analogue of the original quantum unique ergodicity conjecture of Rudnick and Sarnak [ RS94 ] , which states that \lvert{\psi}\rvert^{2}\mu_{1} converges to \mu_{1} as the eigenvalue of a Maaß cuspform \psi on Y_{1} tends to infinity. Note that assuming F is a newform is important, although not quite necessary. 1 1 1 We recall that, at least for q prime, newforms are defined as those automorphic forms orthogonal to old forms of the shape \alpha_{q}^{\ast}(\psi) and \beta_{q}^{\ast}(\psi) for \psi defined on Y_{1} . Clearly, requiring that F be a newform in the conjecture of [ KMV02 ] is wise, since the measures attached to old forms \alpha_{q}^{\ast}(\psi) for some cuspform \psi on Y_{1} are always equal to \mu_{\psi} , which is independent of q . However, the newform assumption is not quite necessary, and it is an amusing observation that the measures attached to old forms \beta_{q}^{\ast}(\psi) do converge to the uniform measure. Indeed, this reduces to the equidistribution of Hecke points, as can be seen in the definition of the Hecke operator in ( 2.1 ) below. We propose the following stronger conjecture, which combines the two maps \alpha_{q} and \beta_{q} and takes the equidistribution of the Hecke correspondence into account. Conjecture A . Let (F_{q}) be a sequence of Hecke–Maaß newforms, where F_{q} is defined on Y_{q} . Then, as q goes to infinity, the measures (\iota_{q})_{\ast}\mu_{F_{q}} converge to the uniform measure \mu_{1}\otimes\mu_{1} on Y_{1}\times Y_{1} . A similar conjecture can be made for holomorphic modular forms. Notice that, since \alpha_{q} is equal to the composition of \iota_{q} and the projection to the first factor, this conjecture implies the one in [ KMV02 ] . In [ Nel11 ] , Nelson proves the Kowalski–Michel–VanderKam conjecture for holomorphic modular forms, generalizing the work of Holowinsky–Soundararajan [ HS10 ] and Watson [ Wat02 ] . Nelson’s work also shows that the case of Maaß forms follows from conjectured subconvexity results for certain L -functions, or the Generalized Lindelöf Hypothesis (GLH), more generally. The main result of the present paper is the following. Theorem 1.1 . For B a division algebra, along a sequence of prime levels q , Conjecture A holds under the Generalized Riemann Hypothesis (GRH) with an effective rate (\log q)^{-1/4+\varepsilon} for any \varepsilon>0 . 1.1. Comparison with the mixing conjecture Conjecture A and Theorem 1.1 share many features with the Mixing Conjecture of Michel–Venkatesh [ MV06 ] and the currently available partial results towards it. To state it, let D<0 be a fundamental discriminant and let \operatorname{Cl}(D) be the class group of \mathbb{Q}(\sqrt{D}) . Take a class \sigma\in\operatorname{Cl}(D) and denote by q be the minimal norm of an integral ideal in \sigma . Now consider a sequence of such data (D_{i},\sigma_{i},q_{i}) and let \mathcal{H}_{D_{i}} be the set of Heegner points in the modular curve Y_{0}(1) attached to \operatorname{Cl}(D_{i}) . Michel and Venkatesh conjecture that, as q_{i} tends to infinity (and D_{i}\to\infty implicitly, as discussed below), the sets \{(z,\sigma_{i}z)\mid z\in\mathcal{H}_{D_{i}}\}\subset Y_{0}(1)\times Y_{0}(1) equidistribute with respect to the uniform product measure \mu_{1}\otimes\mu_{1} . Note that, when projecting to one of the factors, equidistribution was already shown by Duke in [ Duk88 ] using subconvexity of L -functions (in the guise of bounds for Fourier coefficients of half-integral weight forms). We recall Minkowski’s theorem (see [ MV06 , Sec. 6.5] ), which implies that q=O(\sqrt{\lvert{D}\rvert}) . In particular, if q\to\infty , then D\to\infty , the latter being a condition that is commonly found in the statement of the Mixing Conjecture. We also observe that the set \{(z,\sigma_{i}z)\mid z\in\mathcal{H}_{D_{i}}\} lies in the Hecke correspondence \iota_{q}(Y_{0}(q)) , as explained in Section 2.2 of [ BBK22 ] . Thus, the Mixing Conjecture is about point masses inside a Hecke correspondence and their equidistribution inside the product space, given that the projections are already known to equidistribute. This is similar to Conjecture A , where we replace the points by the mass of an automorphic form. However, we note that the required subconvexity to prove the equidistribution in the projections, i.e. the conjecture of Kowalski–Michel–VanderKam, is not yet known in full generality. To deepen the analogy, we suggest comparing the quantity D/q^{2} to the Laplace eigenvalue \lambda_{F} . Minkowski’s theorem shows that D/q^{2}\gg 1 and the uniform spectral gap or Selberg’s conjecture imply that \lambda_{F}\gg 1 . Moreover, if D/q^{2}\gg q^{\eta} and \lambda_{F}\gg q^{\eta} for some \eta>0 , we expect both conjectures to be easier. This is the idea of equidistribution in stages explored by Ellenberg–Michel–Venkatesh [ EMV13 ] for the Mixing Conjecture (see [ BBK22 , Sec. 2.5] for a nice discussion). Namely, if measures on subsets Y_{i} of a space X converge fast enough to the uniform measure on Y_{i} and, in turn, the uniform measures on Y_{i} converge fast enough

Read the complete text