Abstract
In this paper, we prove certain multiplicity one theorems and define twisted gamma factors for irreducible generic cuspidal representations of split \(\mathrm {G}_2\) over finite fields k of odd characteristic. Then we prove the first converse theorem for exceptional groups, namely, \({\mathrm {GL}}_1\) and \({\mathrm {GL}}_2\)-twisted gamma factors will uniquely determine an irreducible generic cuspidal representation of \({\mathrm {G}}_2(k)\).
Similar content being viewed by others
Notes
This relation is not explicitly given in [15]. Due to its importance for our calculation, we give some details in this footnote. Let \(\varphi _\alpha : {\mathrm {SL}}_2(k)\rightarrow {\mathrm {G}}_2(k)\) be the embedding such that \(\varphi _\alpha \left( \begin{pmatrix} 1&{}x \\ &{}1 \end{pmatrix} \right) =\mathbf{{x}}_\alpha (x)\) and \(\varphi _\alpha ({\mathrm {diag}}(a,a^{-1}))=h_\alpha (a,a^{-1})\). For \(g,h\in {\mathrm {G}}_2(k)\), denote the conjugation \(g^{-1}hg\) by h.g. The conjugation of \(\varphi _\alpha (g)\) for \(g\in {\mathrm {SL}}_2(k)\) on \(\mathbf{{x}}_\beta (r_0)(0,r_2,r_3,r_4,0)\) is given in [14, p. 196, (3.5)]. From that description, one can check the following relations \(\mathbf{{x}}_\beta (1)(0,0,-1,2,0).\varphi _\alpha \begin{pmatrix} 1&{} 1/2\\ -1 &{}1/2 \end{pmatrix}=(0,0,1,1/2,0),\) \((0,0,1,1/2,0).\varphi _\alpha \begin{pmatrix} &{}1\\ -1 &{} \end{pmatrix}=\mathbf{{x}}_\beta (1/2)(0,1,0,0,0),\) and \(\mathbf{{x}}_\beta (1/2)(0,1,0,0,0).\varphi _\alpha \begin{pmatrix} 1&{} 0\\ -1/6 &{} 1 \end{pmatrix}=\mathbf{{x}}_{\alpha +\beta }(1)\). This shows that \(\mathbf{{x}}_\beta (1)(0,0,-1,2,0)\sim _{{\mathrm {G}}_2(k)} \mathbf{{x}}_{\alpha +\beta }(1)\).
Recall that an irreducible character \(\theta \) of a reductive group H over a finite field is cuspidal if and only if for any proper parabolic subgroup \(Q=M_QU_Q\) with Levi \(M_Q\) and unipotent \(U_Q\), one has \(\sum _{u \in U_Q}\theta (uh)=0\) for all \(h\in H\), see [12, Corollary 9.1.2] for example. In fact, from the character Table 8, one can check that
$$\begin{aligned} \langle {\theta _5,I(\chi )\otimes \omega _\psi } \rangle =\left\{ \begin{array}{lll} 1, &{} \text { if } \epsilon \chi \ne 1, \\ 2, &{} \text { if } \epsilon \chi =1. \end{array}\right. \end{aligned}$$Thus \(\theta _5\) indeed does not satisfy the conclusion of Theorem 2.1 if \(\chi =\epsilon ^{-1}=\epsilon \).
References
Adrian, M., Liu, B.: Some results on simple supercuspidal representations of \({\rm GL}_n(F)\). J. Number Theory 160, 117–147 (2016)
Adrian, M., Liu, B., Shaun, S., Xu, P.: On the Jacquet Conjecture on the local converse problem for p-adic \({\rm GL}_N\). Represent. Theory 20, 1–13 (2016)
Adrian, M., Liu, B., Shaun, S., Tam, K.-F.: On sharpness of the bound for the local converse theorem of \(p\)-adic \({\rm GL}_{prime}\), tame case. Proc. Am. Math. Soc. Ser. B 5, 6–17 (2018)
Adrian. M., Takeda, S.: A local converse theorem for Archimedean \({\rm GL}(n)\). Preprint. (2018)
Baruch, E. M.: Local factors attached to representations of \(p\)-adic groups and strong multiplicity one. PhD thesis, Yale University (1995)
Baruch, E.M.: On the gamma factors attached to representations of \({{\rm U}}(2,1)\) over a p-adic field. Israel J. Math. 102, 317–345 (1997)
Baruch, E.M., Rallis, S.: A uniqueness theorem of Fourier Jacobi models for representations of \({\rm Sp}(4)\). J. Lond. Math. Soc. 62, 183–197 (2000)
Bernstein, J., Zelevinski, A.V.: Representations of the group \({\rm GL}(n, F)\), where \(F\) is a non-archimedean local field. Rus. Math. Surv. 31, 1–68 (1976)
Bump, D.: Automorphic forms and Representations, Cambridge Studies in Advanced Mathematics, vol. 55. Cambridge University Press, Cambridge (1997)
Bushnell, C., Henniart, G.: Langlands parameters for epipelagic representations of \({\rm GL}_n\). Math. Ann. 358(1–2), 433–463 (2014)
Bushnell, C., Kutzko, P.: The Admissible Dual of \({{\rm GL}}_N\) via Restriction to Compact Open Subgroups. Annals of Mathematics Studies 129. Princeton University Press, Princeton (1993)
Carter, R.: Finite Groups of Lie Type: Conjugacy Classes and Complex Characters, vol. xii+554, p. John Wiley and Sons, Chichester, New York, Brisbane, Toronto and Singapore (1985)
Chai, J.: Bessel functions and local converse conjecture of Jacquet. J. Eur. Math. Soc. (JEMS) 21(6), 1703–1728 (2019)
Chang, B.: The conjugate classes of Chevalley groups of type \(({\rm G}_2)\). J. Algebra 9, 190–211 (1968)
Chang, B., Ree, R.: The characters of \(G_2(q)\), pp. 395–413. Instituto Nazionale di Alta Matematica, Symposia Mathematica XIII (1974)
Chen, J.-P.: Local Factors, Central Characters, and Representations of the General Linear Group Over Non-archimedean Local Fields. Yale University, Thesis (1996)
Chen, J.-P.: The \(n \times (n-2)\) local converse theorem for \({\rm GL}(n)\) over a \(p\)-adic field. J. Number Theory 120(2), 193–205 (2006)
Cogdell, J., Kim, H., Piatetski-Shapiro, I., Shahidi, F.: On lifting from classical groups to \({ m GL}_N\). Publ. Math. Inst. Hautes Études Sci. No. 93, 5–30 (2001)
Cogdell, J., Kim, H., Piatetski-Shapiro, I., Shahidi, F.: Functoriality for the classical groups. Publ. Math. Inst. Hautes Études Sci. No. 99, 163–233 (2004)
Cogdell, J., Piatetski-Shapiro, I.: Converse theorems for \({{\rm GL}}_n\). Inst. Hautes. Études. Sci. Publ. Math. 79, 157–214 (1994)
Cogdell, J., Piatetski-Shapiro, I.: Converse theorems for \({{ m GL}}_n\). II. J. Reine Angew. Math. 507, 165–188 (1999)
Cogdell, J., Piatetski-Shapiro, I., Shahidi, F.: Functoriality for the quasisplit classical groups. On certain L-functions. In: Clay Math. Proc., 13, pp 117–140. Am. Math. Soc., Providence, (2011)
Deligne, P., Lusztig, G.: Representations of reductive groups over finite fields. Ann. Math. 103, 103–161 (1976)
Enomoto, H.: The characters of the finite Chevalley group \(G_2(q), q=3^f\). Jpn. J. Math. 2, 191–248 (1976)
Fulton, W., Harris, J.: Representation Theory, A First Course, Graduate Texts in Mathematics, 129. Springer (1991)
Gan, W.T., Gross, B.H., Prasad, D.: Symplectic local root numbers, central critical L-values and restriction problems in the representation theory of classical groups. Astérisque 346, 1–109 (2012)
Geck, M.: Character sheaves and generalized Gelfand–Graev characters. Proc. Lond. Math. Soc. (3) 78, 139–166 (1999)
Geck, M., Hézard, D.: On the unipotent support of character sheaves. Osaka J. Math. 45, 819–831 (2008)
Gérardin, P.: Weil representations associated to finite fields. J. Algebra 46, 54–101 (1977)
Ginzburg, D.: On the standard \(L\)-function for \(G_2\). Duke Math. J. 69, 315–333 (1993)
Gurevich, S., Howe, R.: Small representations of finite classical groups. In: Representation Theory, Number Theory, and Invariant Theory. Honor of Roger Howe on the occasion of his 70th Birthday, edited by Jim Cogdell, Ju-Lee Kim and Chen-bo Zhu, Progress in Math, 323, pp 209–234. Birkhäuser (2017)
Hakim, J., Offen, O.: Distinguished representations of \({\rm GL}(n)\) and local converse theorems. Manuscripta Math. 148, 1–27 (2015)
Henniart, G.: Caractérisation de la correspondance de Langlands locale par les facteurs \(\epsilon \) de paires. Invent. Math. 113(2), 339–350 (1993)
Jacquet, H., Piatetski-Shapiro, I., Shalika, J.: Automorphic forms on \({\rm GL }(3)\). Ann. Math. (2) 109(1–2), 169–258 (1979)
Jacquet, H., Piatetski-Shapiro, I., Shalika, J.: Rankin–Selberg convolutions. Am. J. Math. 105, 367–464 (1983)
Jacquet, H., Langlands, R.: Automorphic forms on \({\rm GL}(2)\). Lecture Notes in Mathematics, vol. 114. Springer, Berlin-New York (1970)
Jacquet, H., Liu, B.: On the local converse theorem for \(p\)-adic \({ m GL}_n\). Am. J. Math. 140, 1399–1422 (2018)
Jiang, D., Nien, C., Stevens, S.: Towards the Jacquet conjecture on the local converse problem for \(p\)-adic \({\rm GL}_n\). J. Eur. Math. Soc. 17(4), 991–1007 (2015)
Jiang, D., Soudry, D.: The local converse theorem for \({\rm SO }(2n+1)\) and applications. Ann. Math. (2) 157(3), 743–806 (2003)
Jiang, D., Soudry, D.: On local descent from GL(n) to classical groups. Am. J. Math. 134(3), 767–772 (2012). (appendix to a paper by D. Prasad and D. Ramakrishnan)
Kawanaka, N.: Generalized Gelfand–Graev representations and Ennola duality. In: Algebraic Groups and Related Topics. Advanced Studies in Pure Math., Vol. 6, pp. 175–206. Kinokuniya and North-Holland, Tokyo and Amsterdam (1985)
Kawanaka, N.: Generalized Gelfand Graev representations of exceptional algebraic groups over a finite field. Invent. Math. 84, 575–616 (1986)
Kawanaka, N.: Shintani lifting and Gelfand-Graev representations. In: Proc. Sympos. Pure Math., Vol. 47, pp. 147-163. Amer. Math. Sot., Providence, (1987)
Kudla, S.: Notes on the local theta correspondence, Lecture notes from the European School of. Group Theory (1996). http://www.math.toronto.edu/~skudla/ssk.research.html
Lapid, E., Rallis, S.: On the local factors of representations of classical groups. Automorphic representations, L-functions and applications: progress and prospects, 309–359, Ohio State Univ. Math. Res. Inst. Publ., 11, de Gruyter, Berlin (2005)
Liu, B., Moss, G.: On the local converse theorem and the descent theorem in families. Math. Z. 295(1–2), 463–483 (2020)
Liu, B., Zhang, Q.: Uniqueness of certain Fourier–Jacobi models over finite fields. Finite Fields Appl. 58, 70–123 (2019)
Liu, B., Zhang, Q.: Gamma factors and converse theorems for classical groups over finite fields. J. Number Theory. https://doi.org/10.1016/j.jnt.2021.06.024
Lusztig, G.: Characters of Reductive Groups over a Finite Field. Annals of Math Studies, Volume 107. Princeton University Press (1984)
Lusztig, G.: A unipotent support for irreducible representations. Adv. Math. 94, 139–179 (1992)
Morimoto, K.: On the irreducibility of global descents for even unitary groups and its applications. Trans. Am. Math. Soc. 370, 6245–6295 (2018)
Moss, G.: Gamma factors of pairs and a local converse theorem in families. Int. Math. Res. Not. IMRN 16, 4903–4936 (2016)
Nien, C.: A proof of the finite field analogue of Jacquet’s conjecture. Am. J. Math. 136(3), 653–674 (2014)
Nien, C.: Gamma factors and quadratic extension over finite fields. Manuscripta Math. 158(1–2), 31–54 (2019)
Nien, C., Zhang, L.: Converse theorem of Gauss sums. (with an appendix by Zhiwei Yun). J. Number Theory 221, 365–388 (2021)
Rallis, S., Schiffmann, G.: Theta correspondence associated to \({\rm G}_2\). Am. J. Math. 111, 801–849 (1989)
Paskunas, V., Stevens, S.: On the realization of maximal simple types and epsilon factors of pairs. Am. J. Math. 130(5), 1211–1261 (2008)
Piatetski-Shapiro, I.I., Rallis, S., Schiffmann, G.: Rakin–Selberg integral for the group \({\rm G}_2\). Am. J. Math. 114, 1269–1315 (1992)
Roditty, E.A.: On Gamma factors and Bessel functions for representations of general linear groups over finite fields. Master Thesis, Tel Aviv University. (2010)
Shahidi, F.: Fourier transforms of intertwining operators and Plancherel measures for \({\rm GL}(n)\). Am. J. Math. 106(1), 67–111 (1984)
Steiberg, R.: Lectures on Chevalley groups. Yale University, (1967)
Sun, B.: Multiplicity one theorems for Fourier–Jacobi models. Am. J. Math. 134, 1655–1678 (2012)
Taylor, J.: On unipotent supports of reductive groups with a disconnected centre. J. Algebra 391, 41–61 (2013)
Xu, P.: A remark on the simple cuspidal representations of \({\rm GL}_n\). Preprint (2013). arXiv:1310.3519
Zhang, Q.: A local converse theorem for \({{\rm U}}(1,1)\). Int. J. Number Theory 13, 1931–1981 (2017)
Zhang, Q.: A local converse theorem for \({{\rm U}}(2,2)\). Forum Math. 29, 1471–1497 (2017)
Zhang, Q.: A local converse theorem for \({\rm Sp}_{2r}\). Math. Ann. 372, 451–488 (2018)
Zhang, Q.: A local converse theorem for \({\rm U}_{2r+1}\). Trans. Am. Math. Soc. 371, 5631–5654 (2019)
Acknowledgements
The authors would like to thank James Cogdell, Clifton Cunningham, Dihua Jiang and Freydoon Shahidi for their interest, constant support and encouragement. We thank Hikoe Enomoto, Meinolf Geck and Jay Taylor for helpful communications. This project was initiated when the second-named author was a student at the Ohio State University. The collaboration of the two authors started from the 2016 Paul J. Sally, Jr. Midwest Representation Theory Conference in University of Iowa. Part of the work was done when the second-named author worked at University of Calgary, Canada. We would like to express our gratitude to the above mentioned institutes. We also would like to thank the referees for their careful reading and many useful suggestions.
Author information
Authors and Affiliations
Corresponding author
Additional information
Communicated by Wei Zhang.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
The first-named author is partially supported by NSF Grants DMS-1702218, DMS-1848058, and start-up funds from the Department of Mathematics at Purdue University. The second-named author is partially supported by NSFC Grant 11801577.
Appendices
Appendix A. Computation of certain Gauss sums
1.1 Basic Gauss sum
Let \(\psi \) be a nontrivial additive character of \(k={\mathbb {F}}_q\). Recall that we have fixed a square root \(\sqrt{\epsilon _0}\) of \(\epsilon _0\) such that
For \(a\in k^\times \), let
We then have
and
Thus we get the following
Lemma A.1
We have \(A_1(a)-A_\kappa (a)=\epsilon (a)\sqrt{\epsilon _0 q}\).
We write \(A_r(1)\) as \(A_r\) for simplicity, for \(r=1,\kappa \).
1.2 Computation of \(B_r^i\)
We now compute the sums \(B_r^i\) for \(r=1,\kappa \) and \(i=0,1,2,3\) in (3.6) used in Sect. 3. We assume \(q\equiv 1\ \mathrm {mod}\ 3\). Given \(r\in \left\{ {1,\kappa }\right\} , r_3\in k^\times , r_4\in k^\times /{\pm 1}\), let \(z(r,r_3,r_4)=-2-\frac{rr_4^2}{r_3^3}\in k\). Note that for any \(a\in k\), the equation \(t+t^{-1}=a\) for t is solvable over \(k_2\). Given \(r,r_3,r_4\) as above, and recall that \(t=t(r,r_3,r_4)\) denotes a solution of the equation \(t+t^{-1}=z(r,r_3,r_4)\). Note that \(rr_3r_4\ne 0\) implies that \(t\ne -1\). Although there are two choices of \(t(r,r_3,r_4)\) in general, one can check that the condition \(t(r,r_3,r_4)\in \left\{ {\pm 1}\right\} \) (resp. \(t(r,r_3,r_4)\in k^{\times ,3}-\left\{ {\pm 1}\right\} \), \(t(r,r_3,r_4)\in k^{\times }-k^{\times ,3}\), \(t(r,r_3,r_4)\in k_2- k^{\times }\)) is independent on the choice of \(t(r,r_3,r_4)\).
Lemma A.2
We have
Proof
Notice that the condition \(-2-rr_4^2/r_3^3=t+t^{-1}\) implies that \(t\ne -1\) and
We first compute \(B_r^0\). We first assume that \(r=1\). When \(t=1\), (8.1) becomes \((-r_3)^3=(r_4/2)^2\). Since \(k^{\times }\) is a cyclic group generated by \(\kappa \), the condition \((-r_3)^3=(r_4/2)^2\) implies that \(-r_3\in k^{\times ,2}\). Moreover, for each \(-r_3\in k^{\times ,2}\), there exists a unique \(r_4\in k^{\times }/\left\{ {\pm 1}\right\} \) such that the equation \((-r_3)^3=(r_4/2)^2\) holds. Thus we get
Similarly, we have \(B_\kappa ^0=A_\kappa (-1)\). Thus we have \(B_1^0-B_\kappa ^1=A_1(-1)-A_\kappa (-1)=\epsilon _0\sqrt{\epsilon _0 q}\) by Lemma A.1.
We next compute \(B_{r}^1\), \(r=1,\kappa \). Let \(t=t(r,r_3,r_4)\in k^{\times ,3}-\left\{ {\pm 1}\right\} \). Let \(a\in k^\times \) with \(t=a^3\). We first assume that \(r=1\). From (8.1), we have \(-a^{-1}r_3\in k^{\times ,2}\). Thus the contribution of each fixed \(t=t(1,r_3,r_4)\) to the sum \(B_1^1\) is
where \(t^{1/3}\) is any cubic root of t in \(k^{\times }\). Because t and \(t^{-1}\) contributes the same to the sum \(B_1^1\), we have
Similarly, we have
Thus by Lemma A.1, we have
We have \(k^{\times ,3}=\left\{ {\kappa ^{3i}: 1\le i\le \frac{q-1}{3}}\right\} \). Thus we get
where the last equality follows from the fact that \(\epsilon (\kappa )=-1\) and \(\frac{q-1}{3}\) must be even. Thus we get
We next consider \(B_r^2\). Note that \( k^\times -k^{\times ,3}=\kappa k^{\times ,3}\coprod \kappa ^2 k^{\times ,3}\). For \(j=1,2\), we define
We have \(B_{r}^2=B_r^{2,1}+B_{r}^{2,2}\). Take an element \(t\in k^\times -k^{\times ,3}\) with \(t(r,r_3,r_4)=t\). Then the condition \(-2-\frac{rr_4^2}{r_3^3}=t+t^{-1}\) implies (8.1). Note that if \(r=1\) and \(t\in \kappa k^{\times ,3}\), equation (8.1) implies that \(-r_3\in \kappa k^{\times ,2}\), and for such an \(r_3\), there is a unique \(r_4\) satisfying that equation. Thus we get
For \(t\in \kappa ^2 k^{\times ,2}\) and \(r=\kappa \), we also have that \(-r_3\in \kappa k^{\times ,2}\) and a unique \(r_4\) determined by these datum. This shows that
Similarly, we have \(B_\kappa ^{2,1}=B_1^{2,2}\). Thus we get \(B_1^2=B_{\kappa }^2\).
Finally, we consider \(B_r^3\). We have
Thus, from the previous results, we get
This concludes the proof of the lemma. \(\square \)
1.3 Computation of \(C_r^i\)
In this subsection, we compute the sums \(C_r^i\) for \(r=1,\kappa ,\) and \(i=0,1,2,3\) defined in (3.8). Note that in this case, \(q\equiv -1\ \mathrm {mod}\ 3\). Recall that \(k_2\) is the unique quadratic extension of \(k={\mathbb {F}}_q\). We can realize \(k_2\) as \(k[\sqrt{\kappa }]\). Let \({\mathrm {Nm}}:k_2\rightarrow k\) be the norm map. We have \({\mathrm {Nm}}(x+y\sqrt{\kappa })=x^2-y^2\kappa \). Recall that \(k_2^1\) is the norm 1 subgroup of \(k_2^\times \).
Lemma A.3
-
1.
If an element \(u\in k_2^1\) has a cubic root \(v\in k_2^\times \), then we must have \(v\in k_2^1\).
-
2.
Let \(t\in k_2^1\) and \(t\ne -1\). Then \(t+t^{-1}+2\) is a square in \(k^\times \) if and only if t is a square in \(k_2^1\).
Proof
-
(1)
Since \(u=v^3\in k_2^{1}\), we have \(v^{3q+3}=1\). On the other hand, we have \(v^{q^2-1}=1\) since \(v\in k_2^\times \). Since \(q\equiv -1 \ \mathrm {mod}\ 3\), the greatest common divisor of \(q^2-1\) and \(3q+3\) is \(q+1\). Thus \(v^{q+1}=1\), which means that \(v\in k_2^1\).
-
(2)
Suppose that \(t=\beta ^2\) with \(\beta \in k_2^1\). We write \(\beta =a+b\sqrt{\kappa }\) with \(a,b\in k\). Then \(\beta \in k_2^1\) means that \(a^2-b^2\kappa =1\), which implies that \(b^2\kappa =a^2-1\). We have \(t=\beta ^2=a^2+b^2\kappa +2ab\sqrt{\kappa }\). Thus
$$\begin{aligned} t+t^{-1}+2=2(a^2+b^2\kappa )+2=4a^2\in k^{\times ,2}. \end{aligned}$$Conversely, suppose that \(t+t^{-1}+2\in k^{\times ,2}\). Suppose that \(t=x+y\sqrt{\kappa }\) with \(x,y\in k\) and \(t+t^{-1}+2=a^2\) with \(a\in k^{\times }\). Note that \(t+t^{-1}+2=2x+2\). Thus \(a^2=2x+2\). On the other hand, we have
$$\begin{aligned} a^2=t+t^{-1}+2=t^{-1}(t+1)^2. \end{aligned}$$Thus, we have \(t=(a^{-1}(t+1))^2\). It suffices to show that \(a^{-1}(t+1)\in k_2^1\). We have
$$\begin{aligned} {\mathrm {Nm}}(t+1)=(x+1)^2-y^2\kappa =2+2x=a^2, \end{aligned}$$where we used \(x^2-y^2\kappa =1\). Thus \({\mathrm {Nm}}(a^{-1}(t+1))=1\).
\(\square \)
Lemma A.4
We have
Proof
Note that \(C_r^0=B_r^0\) and thus \(C_1^0-C_2^0=\epsilon _0\sqrt{\epsilon _0 q}\) follows from Lemma A.2. To compute \(C_r^2\), we take an element \(t\in k^\times -\left\{ {\pm 1}\right\} \) and let \(t(r,r_3,r_4)=t\), which implies
see (8.1). Note that any \(t\in k^{\times }\) is has a cubic root in \(k^{\times }\). Let \(t^{1/3}\in k^\times \) be one cubic root of t. Then the above equation implies that
If \(r=1\), this implies that \(r_3\in -t^{1/3}k^{\times ,2}\), and for such an \(r_3\) (and a fixed t), there is a unique \(r_4\in k^\times /\left\{ {\pm 1}\right\} \) such that \((-r_3/t^{1/3})^3=r\left( \frac{r_4}{t+1}\right) ^2 \). Thus the contribution of a single t with \(t(1,r_3,r_4)\) to the sum \(C_1^1\) is
Since t and \(t^{-1}\) have the same contribution, we have
Since \(t\mapsto t^{3}\) is a bijection from \(k^\times -\left\{ {\pm 1}\right\} \) to itself, we get
Similarly, we have
Thus by Lemma A.1, we have
Since \(\epsilon \) is a nontrivial character on \(k^\times \), we have \(\sum _{t\in k^\times }\epsilon (t)=0\). Thus we have
We next consider \(C_r^3\). Let \(\alpha \) be a generator of \(k_2^1\). Note that \(\alpha \) has no cubic root in \(k_2^1\). By Lemma A.3 (1), we have
Consider the subsets \(S_1,S_2\) of \(k_2^1-k_2^{\times ,3}:\)
Note that \(|S_1|=|S_2|=\frac{q+1}{3}\). For \(i=1,2\), let
We have \(C_r^{3}=C_{r}^{3,1}+C_{r}^{3,2}\). Take \(t\in S_i\), the condition \(t(r,r_3,r_4)=t\) implies that
If \(t\in S_1\), by Lemma A.3, we have \(t+t^{-1}+2\in \kappa k^{\times ,2}\). Thus for \(r=1,t\in S_1\), we have \(-r_3\in \kappa k^{\times ,2}\), and for each \(-r_3\in \kappa k^{\times ,2}\), there is a unique \(r_4\in k^{\times }/\left\{ {\pm 1}\right\} \) such that \(t(1,r_3,r_4)=t\) (for fixed t). Thus, we get
where the 1/2 was appeared since t and \(t^{-1}\) have the same contribution to the above sum. Similarly, we have
In particular, we have \(C_{1}^{3,1}=C_{\kappa }^{3,2}\). Similarly, we have \(C_1^{3,2}=C_\kappa ^{3,1}\). Thus we have \(C_1^3-C_\kappa ^3=0\).
Finally, to compute \(C_1^2-C_{\kappa }^2\), it suffices to notice that
and thus
One can also compute \(C_1^2-C_\kappa ^2\) directly from Lemma A.3. \(\square \)
1.4 Computation of \(D_r^i\)
In this subsection, let \(q=3^f\) and \(k={\mathbb {F}}_q\). We compute the Gauss sums in (4.2).
Lemma A.5
We have
Proof
Note that we have \(D_r^0=B_r^0\). Thus the first identity follows from Lemma A.2. The second identity can be computed similarly as the computation of \(C_1^1-C_\kappa ^1\). Since \(D_1^0+D_1^1+D_1^2=D_\kappa ^0+D_\kappa ^1+D_\kappa ^2\), the last identity follows from the first one. \(\square \)
Appendix B. Embedding of \({\mathrm {G}}_2\) into \({\mathrm {SO}}_7\)
In this appendix, based on [56], we give an explicit matrix realization of \(\mathbf{{x}}_\gamma (r)\) for each root \(\gamma \) of \({\mathrm {G}}_2\), which gives an explicit embedding of \({\mathrm {G}}_2(k)\) into \({\mathrm {SO}}_7(k)\). Here \({\mathrm {SO}}_7(k)=\left\{ {g\in {\mathrm {GL}}_7(k): {}^t g Qg=Q}\right\} \), with \(Q=\begin{pmatrix} &{}&{}s_3 \\ &{}2&{} \\ {}^t \!s_3 \end{pmatrix},\) where \(s_3=\begin{pmatrix} &{}&{}1 \\ &{}1&{} \\ -1&{}&{}\end{pmatrix}.\) The explicit realization of \(\mathbf{{x}}_\gamma (r)\) is given as follows.
and
Rights and permissions
About this article
Cite this article
Liu, B., Zhang, Q. On a converse theorem for \({\mathrm {G}}_2\) over finite fields. Math. Ann. 383, 1217–1283 (2022). https://doi.org/10.1007/s00208-021-02250-2
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00208-021-02250-2