SCIENCE CHINA Mathematics, https://doi.org/10.1007/s11425-018-1616-0

## A refined Poisson summation formula for certain Braverman-Kazhdan spaces

• AcceptedNov 5, 2019
Share
Rating

### Abstract

Braverman and Kazhdan (2000) introduced influential conjectures aimed at generalizing the Fourier transform and the Poisson summation formula. Their conjectures should imply that quite general Langlands $L$-functions have meromorphic continuations and functional equations as predicted by Langlands' functoriality conjecture. As evidence for their conjectures, Braverman and Kazhdan (2002) considered a setting related to the so-called doubling method in a later paper and proved the corresponding Poisson summation formula under restrictive assumptions on the functions involved. The connection between the two papers is made explicit in work of Li (2018). In this paper, we consider a special case of the setting in Braverman and Kazhdan's later paper and prove a refined Poisson summation formula that eliminates the restrictive assumptions of that paper. Along the way we provide analytic control on the Schwartz space we construct; this analytic control was conjectured to hold (in a slightly different setting) in the work of Braverman and Kazhdan (2002).

### Acknowledgment

This work was supported by the National Science Foundation of the USA (Grant Nos. DMS-1405708 and DMS-1901883). The second author was supported by the National Science Foundation of the USA (Grant Nos. DMS-1702218 and DMS-1848058), and by a start-up fund from the Department of Mathematics at Purdue University. The authors thank Herve Jacquet and Aaron Pollack for their interest in the results in this paper and for helpful comments and suggestions. The authors thank Wen-Wei Li, Yiannis Sakellaridis, and Freydoon Shahidi for useful conversations, and thank Freydoon Shahidi and Wen-Wei Li for sharing [24,25] with them. The authors also thank Heekyoung Hahn for the help with editing and for her constant encouragement. Thanks are due to the anonymous referees for several useful comments. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

### Supplement

Appendix

Proof of Proposition 4.7 in the Archimedean case

Assume that $F$ is an Archimedean local field. Recall that $C_c^\infty(X(F),K)~<~C_c^\infty(X(F))$ is the subset of functions that are right $K$-finite. Our goal here is to prove Proposition 4.7 in the current Archimedean setting. We recall that Proposition 4.7 simply states that $C_c^\infty(X(F),K)~\leq~\mathcal{S}(X(F),K)$.

It is convenient to begin with a few convergence lemmas.

Lemma 21. Let $\Phi~\in~C_c^\infty(X(F),K)$. Then for $\mathrm{Re}(s)~\geq~n^2/2$ one has $$|M_{w_0}\Phi_{\chi_s}(I_{2n})| \ll_{\Phi} 1.$$

Proof. Consider the map \begin{align} \mathrm{Pl}:X(F) \rightarrow \wedge^n F^{2n}-0. \tag{72} \end{align} Since $X$ is a homogeneous space for $\mathrm{Sp}_{2n}$ it is smooth (as a scheme over $F$). The map 72 is an injective diffeomorphism onto its image, a closed submanifold of $\wedge^n~F^{2n}-0$. In particular, there is a function $\Psi~\in~C_c^\infty(\wedge^n~F^{2n})$ such that $\Phi=\Psi~\circ~\mathrm{Pl}$. For $\mathrm{Re}(s)$ sufficiently large we have \begin{align*}M_{w_0} \Phi_{\chi_s}(I_{2n})&=\int_{N(F)} \Phi_{\chi_s}(w_0^{-1}n)dn \\ &=\int_{N(F)}\bigg(\int_{M^{\mathrm{ab}}(F)}\delta_P^{1/2}(m)\chi_s(\omega(m))\Psi( \mathrm{Pl}(m^{-1}w_0^{-1}n))dm^\times \bigg)dn. \end{align*} Temporarily denote \begin{align*}&\mathrm{Pl}_0:M_{n \times n}^{\oplus 2}(F)\rightarrow \wedge^{n}F^{2n}, \\ &(X,Y) \mapsto \mathrm{Pl}\left( \begin{matrix} * & * \\ X & Y \end{matrix}\right). \end{align*} This is just taking the wedge product of the $n$ rows of the $n~\times~2n$ matrix $(X\,~Y)$, going from top to bottom. Then the integral above can be written as \begin{align*}&\int_{\mathrm{Sym}^n(F)}\left(\int_{F^\times}\chi_s(a)|a|^{(n+1)/2}\Psi\left( \mathrm{Pl}_0\left(-\left(\begin{matrix} a & \\ & I_{n-1}\end{matrix} \right)J',-\left(\begin{matrix} a & \\ & I_{n-1}\end{matrix} \right)J'z \right)\right)da^\times \right)dz, \end{align*} where $\mathrm{Sym}^n(F)$ is the $F$-vector space of symmetric $n~\times~n$ matrices and $$J'=\left(\begin{smallmatrix} & & 1 \\ & \reflectbox{\ddots} & \\ 1 & & \end{smallmatrix}\right).$$ We note that $\mathrm{Pl}_0$ is invariant under multiplication by $\mathrm{SL}_n$ on the left to see that the above is equal to \begin{align*}&\int_{\mathrm{Sym}^n(F)}\left(\int_{F^\times}\chi_s(a)|a|^{(n+1)/2}\Psi\left( -\mathrm{Pl}_0\left(\left(\begin{matrix} I_{n-1}& \\ & a\end{matrix} \right),\left(\begin{matrix} I_{n-1}& \\ & a\end{matrix} \right)z \right)\right)da^\times \right)dz. \end{align*} Take a change of variables $z~\mapsto~a^{-1}z$ to arrive at \begin{align*}&\int_{\mathrm{Sym}^n(F)}\left(\int_{F^\times}\chi_{s+(1-n^2)/2}(a)\Psi\left(- \mathrm{Pl}_0\left(\left(\begin{matrix} I_{n-1} & \\ & a\end{matrix} \right),\left(\begin{matrix} I_{n-1} & \\ & a\end{matrix} \right)a^{-1}z \right)\right)da^\times \right)dz. \end{align*} By inspection the integrand is rapidly decreasing as a function of $z~\in~\mathrm{Sym}^n(F)$ and $a~\in~F$, so this integral converges absolutely and is bounded by a constant depending only on $\Psi$ for $\mathrm{Re}(s)\geq~n^2/2$.

Lemma 22. Let $\Phi~\in~C_c^\infty(X(F),K)$ and $A~\in~\RR$. Then for $\mathrm{Re}(s)~\geq~A$ one has \begin{align*}|M_{w_0}\Phi_{\chi_s}(I_{2n})| \ll_{\Phi,A} 1. \end{align*}

Proof. Let $\alpha~\in~\ZZ_{\geq~0}$. Notice that $$\widetilde{\Phi}:=(\omega\bar{\omega})^{\alpha}\Phi \in C_c^\infty(X(F),K),$$ where $\omega\bar{\omega}$ is defined as in 43. As in the proof of Lemma 5.5 we have \begin{align*}((\omega \bar{\omega})^{-\alpha}\widetilde{\Phi})_{\chi_s} &=\widetilde{\Phi}_{\chi_{s+2\alpha[F:\RR]^{-1}}}. \end{align*} By Lemma 21 we therefore have $$|M_{w_0}\Phi_{\chi_s}(I_{2n})|= |M_{w_0}((\omega \bar{\omega})^{-\alpha}\widetilde{\Phi})_{\chi_s}(I_{2n})|=|M_{w_0} \widetilde{\Phi}_{\chi_{s+2\alpha[F:\RR]^{-1}}}(I_{2n})| \ll_{\Phi,\alpha}\! 1,$$ for $\mathrm{Re}(s)+2\alpha~[F:\RR]^{-1}\geq~n^2/2.$ Taking $\alpha$ sufficiently large we deduce the lemma.

ProofProof of Proposition $\ref{prop:compact}$ in the Archimedean case. If $\Phi~\in~C_c^\infty(X(F),K)$ then it is easy to see that $\Phi_{\chi_s}$ is holomorphic for all $\chi$. Hence by [8] we deduce that $\Phi_{\chi_s}$ is a good section. We thus have to verify that for all $g~\in~\mathrm{Sp}_{2n}(F)$, all characters $\chi$, $A<B$, and all $P_w$ as in the definition of an excellent section that \begin{align} |\Phi_{\chi_s}(g)|_{A,B,P_{\mathrm{Id}}} \tag{73} \end{align} and \begin{align} |M_{w_0}\Phi_{\chi_s}(g)|_{A,B,P_{w_0}} \tag{74} \end{align} are finite. In fact this is enough to complete the proof since the space $C_c^\infty(X(F),K)$ is preserved under the differential operators $D$ and $\bar{D}$ of 32 and 33.

Write $\chi$ as in 28. By Lemma 5.9 for any $N,N'~\in~\ZZ_{~\geq~0}$ with $N'=0$ if $F$ is real, one has \begin{align*}|D^N\Phi_{\chi_s}|_{A,B,1}=\bigg|\bigg({\rm i}t+s+\frac{n+1}{2} \bigg)^N\Phi_{\chi_s}(g)\bigg|_{A,B,1}, \end{align*} if $F$ is real, and \begin{align*}|D^N\bar{D}^{N'}\Phi_{\chi_s}(g)|_{A,B,1}= \bigg|\bigg(\frac{\alpha}{2}+{\rm i}t+s+\frac{n+1}{2} \bigg)^N\bigg(-\frac{\alpha}{2}+{\rm i}t+s+\frac{n+1}{2} \bigg)^{N'}\Phi_{\chi_s}(g)\bigg|_{A,B,1}, \end{align*} if $F$ is complex. Since $\Phi~\in~C_c^\infty(X(F),K)$ the left-hand sides here are bounded by a constant depending on $A,B$ $\Phi,N$ and $N'$. We deduce that 73 is finite for all $P~\in~\CC[x]$. An analogous argument, using Lemma 22, allows us to deduce that 74 is finite for all $P~\in~\CC[x]$.

### References

[1] Blomer V, Brumley F. On the Ramanujan conjecture over number fields. Ann Math, 2011, 174: 581-605 CrossRef Google Scholar

[2] Bouthier A, Ng? B C, Sakellaridis Y. On the formal arc space of a reductive monoid. Am J Math, 2016, 138: 81-108 CrossRef Google Scholar

[3] Braverman A, Kazhdan D. Normalized Intertwining Operators. MMJ, 2002, 2: 533-553 CrossRef Google Scholar

[4] Braverman A, Kazhdan D. $\gamma$-functions of representations and lifting. In: Visions in Mathematics. Modern Birkhäuser Classics. Basel: Birkhäuser, 2010, 237--278. Google Scholar

[5] Brumley F. Effective multiplicity one on GL n and narrow zero-free regions for Rankin-Selberg L- functions. Am J Math, 2006, 128: 1455-1474 CrossRef Google Scholar

[6] Cheng S, Ngo B C. On a conjecture of Braverman and Kazhdan Int Math Res Not IMRN, 2018, 20: 6177--6200. Google Scholar

[7] Duke W, Rudnick Z, Sarnak P. Density of integer points on affine homogeneous varieties. Duke Math J, 1993, 71: 143-179 CrossRef Google Scholar

[8] Folland G B. A Course in Abstract Harmonic Analysis. Studies in Advanced Mathematics. Boca Raton: CRC Press, 1995. Google Scholar

[9] Franke J, Manin Y I, Tschinkel Y. Rational points of bounded height on Fano varieties. Invent Math, 1989, 95: 421-435 CrossRef Google Scholar

[10] Gelbart S S, Lapid E M. Lower bounds for L -functions at the edge of the critical strip. Am J Math, 2006, 128: 619-638 CrossRef Google Scholar

[11] Gelbart S, Piatetski-Shapiro I, Rallis S. Explicit Constructions of Automorphic $L$-Functions. Lecture Notes in Mathematics, vol. 1254. Berlin: Springer-Verlag, 1987. Google Scholar

[12] Gelbart S, Shahidi F. J Amer Math Soc, 2001, 14: 79-108 CrossRef Google Scholar

[13] Getz J. Nonabelian Fourier transforms for spherical representations. Pac J Math, 2018, 294: 351-373 CrossRef Google Scholar

[14] Getz J R, Liu B. A summation formula for triples of quadratic spaces. Adv Math, 2019, 347: 150-191 CrossRef Google Scholar

[15] Godement R, Jacquet H. Zeta Functions of Simple Algebras. Lecture Notes in Mathematics, vol. 260. Berlin-New York: Springer-Verlag, 1972. Google Scholar

[16] Ikeda T. On the location of poles of the triple $L$-functions. Compos Math, 1992, 83: 187--237. Google Scholar

[17] Jacquet H. Archimedean Rankin-Selberg integrals. In: Automorphic forms and $L$-functions II Local Aspects. Contemporary Mathematics, vol. 489. Providence: Amer Math Soc, 2009, 57--172. Google Scholar

[18] Kudla S S, Rallis S. Poles of Eisenstein series and $L$-functions. In: Festschrift in honor of II. Israel Mathematical Conference Proceedings, vol. 3. Ramat Gan: Bar-Ilan University, 1990, 81--110. Google Scholar

[19] Kudla S S, Rallis S. A regularized Siegel-Weil formula: The first term identity. Ann of Math (2), 1994, 140: 1--80. Google Scholar

[20] Lafforgue L. Noyaux du transfert automorphe de Langlands et formules de Poisson non linéaires. Jpn J Math, 2014, 9: 1--68. Google Scholar

[21] Li W-W. Basic functions and unramified local $L$-factors for split groups. Sci China Math, 2017, 60: 777--812. Google Scholar

[22] Li W-W. Appendix: A comparison of Basic Functions. Geometric Aspects of the Trace Formula. New York: Springer, 2018. Google Scholar

[23] Li W-W. Towards Generalized Prehomogeneous Zeta Integrals. Lecture Notes in Mathematics, vol. 2221. Berlin: Springer-Verlag, 2018. Google Scholar

[24] Li W-W. Zeta Integrals, Schwartz Spaces and Local Functional Equations. Lecture Notes in Mathematics, vol. 2228. Berlin: Springer-Verlag, 2018. Google Scholar

[25] Moreno C J. Advanced Analytic Number Theory: $L$-Functions. Mathematical Surveys and Monographs, vol. 115. Providence: Amer Math Soc, 2005. Google Scholar

[26] Müller W. On the singularities of residual intertwining operators. GAFA Geom funct anal, 2000, 10: 1118-1170 CrossRef Google Scholar

[27] Ngô B C. On a certain sum of automorphic $L$-functions. Contemp Math, 2014, 614: 337--343. Google Scholar

[28] Piatetski-Shapiro I, Rallis S. Rankin triple $L$ functions. Compos Math, 1987, 64: 31--115. Google Scholar

[29] Sakellaridis Y. ANT, 2012, 6: 611-667 CrossRef Google Scholar

[30] Sakellaridis Y. Inverse Satake Transforms. Geometric Aspects of the Trace Formula. New York: Springer, 2018. Google Scholar

[31] Serre J-P. Galois Cohomology. Berlin: Springer-Verlag, 1997. Google Scholar

[32] Shahidi F. Eisenstein Series and Automorphic $L$-Functions. American Mathematical Society Colloquium Publications, vol. 58. Providence: Amer Math Soc, 2010. Google Scholar

[33] Shahidi F. Local Factors, Reciprocity and Vinberg Monoids. Prime Numbers and Representation Theory. Lecture Series of Modern Number Theory, vol. 2. Beijing: Science Press, 2018. Google Scholar

[34] Shahidi F. On generalized Fourier Transforms for Standard ${L}$-Functions. Geometric Aspects of the Trace Formula. New York: Springer, 2018. Google Scholar

[35] Tate J T.Fourier analysis in number fields and Hecke's zeta-functions.PhD Thesis. Princeton: Princeton University, 1950. Google Scholar

•
 $F$ $~~~dx$ $\RR$ Lebesgue measure $\CC$ Twice Lebesgue measure Non-Archimedean $dx(\OO)=|\mathfrak{d}|^{1/2}$

Citations

• #### 0

Altmetric

Copyright 2020 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有