The moduli space of representations of the modular group into G 2

Angelica Babei, Andrew Fiori, Cameron Franc · 2025

Read The moduli space of representations of the modular group into G 2 by Angelica Babei, Andrew Fiori, Cameron Franc (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-2501-17276.txt.

Introduction

Angelica Babei, Andrew Fiori, and Cameron Franc study representations of the modular group in this 2025 math paper. A representation maps group elements to matrices, here inside the exceptional group G2. The space of all such maps, the moduli space, has two main parts.

One part is one-dimensional and holds rigid examples that Katz studied. The authors focus on the larger four-dimensional part, whose members are not rigid. They build a big family there as an etale cover of degree ninety-six over an open set. They also give algebra tests that ensure the maps cover G2 over a finite field for primes at least five. New congruence subgroups result.

Catalog record

CREATORAngelica Babei, Andrew Fiori, Cameron Franc
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE116 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2501-17276

Browse the full library · About the Tower of Babel

Full text

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

The moduli space of representations of the modular group into G 2 G_{2} Thanks: The authors gratefully acknowledge financial support received from NSERC through their respective Discovery Grants, the financial support of the University of Lethbridge, and the use of computational resources made available through WestGrid and Compute Canada. Angelica Babei, Andrew Fiori, Cameron Franc Abstract . In this paper we construct a large four-dimensional family of representations of the modular group into G 2 G_{2} . Precisely, this family is an etale cover of degree 96 96 of an open subset of the moduli space of such representations. This moduli space has two main components, of dimensions one and four. The one-dimensional component consists of well-studied rigid representations, in the sense of Katz. We focus on the four-dimensional component which consists of representations that are not rigid. We also provide algebraic conditions to ensure that the specializations surject onto G 2 ​ ( 𝐅 p ) G_{2}(\mathbf{F}_{p}) for primes p ≥ 5 p\geq 5 . These representations give new examples of ϕ \phi -congruence subgroups of the modular group as introduced in [ 1 ] . 1. Introduction In this paper we construct a large family of representations, \phi , of the modular group \Gamma=\PSL_{2}(\mathbf{Z}) into G_{2} . Such a representation is determined by choosing elements of G_{2} of order two and three, and mapping respective free-generators \displaystyle S \displaystyle=\left(\begin{matrix}0&-1\\ 1&0\end{matrix}\right), \displaystyle R \displaystyle=\left(\begin{matrix}0&-1\\ 1&1\end{matrix}\right), for \Gamma to these elements. We are particularly interested in representations that surject onto G_{2}(\mathbf{F}_{p}) for primes p . This is motivated by our earlier paper [ 1 ] where we introduced \phi -families of subgroups of \Gamma by reducing linear representations \phi of \Gamma modulo primes. In [ 1 ] we considered two main examples: a family arising from an indecomposable but not irreducible representation of rank two, as well as families arising from maps into the symplectic group of genus two. The symplectic case is interesting because by results of [ 13 ] , most choices of order two and three elements in classical groups lead to surjective representations of \Gamma , whereas for the g=2 symplectic group, if p is large enough then only approximately half the choices yield surjective representations. Thus, the symplectic group is somewhat exceptional and it is harder to construct representations onto it. In [ 1 ] we were able to write down a large supply of such symplectic representations by using the description of the character variety of rank four irreducible representations of \Gamma described by Tuba-Wenzl [ 15 ] . By [ 15 ] , character varieties of \Gamma are well-understood up to rank five, and Le Bruyn has written a number of interesting papers on higher rank cases, see for example [ 12 ] . Thus, the case of representations of G_{2} of rank seven is an interesting exceptional case of examples that lies just outside of the known literature of representations of the modular group, but which intersects the active area of study of G_{2} -local systems. As for the growing literature on G_{2} -local systems, much of it has focused on rigid representations of the full free group F_{2} on two generators, a subject introduced by Katz [ 9 ] . See for example the papers [ 4 , 6 , 7 ] for some work on this subject. Boalch and Paluba [ 3 ] have studied G_{2} -character varieties obtained from the fundamental group of a four-punctured sphere, where three of the conjugacy classes of local monodromies are fixed to consist of elements of order 3 in G_{2} , whereas the fourth local monodromy can be a more general semisimple conjugacy class. As this fourth conjugacy class is allowed to vary, Boalch-Paluba show that the resulting two-parameter family of complex surfaces is related to a subfamily of the Fricke family of cubic surfaces xyz+x^{2}+y^{2}+z^{2}+b_{1}x+b_{2}y+b_{3}z=0. These examples of G_{2} -character varieties parameterize local systems that are minimally nonrigid in the sense of Katz, and so provide a suitably rich class of examples for studying local systems beyond the rigid case. In a similar vein, our examples below correspond to character varieties of a three-punctured sphere where two local monodromies are fixed to lie in conjugacy classes of order 2 and 3 , and the third class is allowed to vary. More precisely, as we will see below, there are two disconnected components of irreducible representations \Gamma\to G_{2} , one of which is one-dimensional while the other is four-dimensional. It transpires that the one-dimensional component consists of rigid representations, and none of these lead to surjections onto G_{2}(\mathbf{F}_{p}) . On othe other hand, the larger component X consists of non-rigid representations and to the best of our knowledge, it has not received as much study as the rigid component. Therefore, in our analysis below, we will ignore the more well-studied rigid component and we focus our attention on the non-rigid component X . 1 1 1 Technically, our variety X of representations will contain many representations that are not irreducible, though the irreducibles are dense in X . We use the adjective “irreducible” in this introduction mainly to rule out trivial components such as the trivial representation, et cetera. Our approach to studying X is to realize it concretely as a double-quotient of G_{2} of the following form: X=\GL_{2}^{\short}\backslash G_{2}/\SO_{\mathbf{H}}. In this optic, we are considering simultaneous conjugacy classes of the generating elements of \Gamma of orders 2 and 3 . Ultimately we shall prove the following: Theorem 1.1 . There exists a nonempty open subset Z\subseteq X in the large component and a cover \widetilde{X}_{0}\to Z described concretely by the family of representations: \displaystyle\phi(S) \displaystyle=\left(\begin{smallmatrix}a_{1}^{2}-a_{2}^{2}&0&0&0&2a_{1}a_{2}&0&0\\ 0&b_{1}^{2}-b_{2}^{2}&0&0&0&2b_{1}b_{2}&0\\ 0&0&c_{1}^{2}-c_{2}^{2}&0&0&0&2c_{1}c_{2}\\ 0&0&0&-1&0&0&0\\ 2a_{1}a_{2}&0&0&0&-a_{1}^{2}+a_{2}^{2}&0&0\\ 0&2b_{1}b_{2}&0&0&0&-b_{1}^{2}+b_{2}^{2}&0\\ 0&0&2c_{1}c_{2}&0&0&0&-c_{1}^{2}+c_{2}^{2}\end{smallmatrix}\right), \displaystyle\phi(R) \displaystyle=\left(\begin{smallmatrix}1&0&0&0&0&0&0\\ 0&1&0&0&0&0&0\\ 0&0&1&0&0&0&0\\ 0&0&0&-\frac{1}{2}&-\frac{3}{2}(1-2y_{1})&-3x_{1}&3w_{1}\\ 0&0&0&\frac{1}{2}(1-2y_{1})&-\frac{1}{2}&w_{1}&3x_{1}\\ 0&0&0&3x_{1}&-3w_{1}&-\frac{1}{2}&-\frac{3}{2}(1-2y_{1})\\ 0&0&0&-w_{1}&-3x_{1}&\frac{1}{2}(1-2y_{1})&-\frac{1}{2}\end{smallmatrix}\right). where (1) a_{1}^{2}+a_{2}^{2}=b_{1}^{2}+b_{2}^{2}=c_{1}^{2}+c_{2}^{2}=1 , (2) w_{1}^{2}+3x_{1}^{2}+y_{1}^{2}=y_{1} , and (3) \left(\begin{smallmatrix}a_{1}&-a_{2}\\ a_{2}&a_{1}\end{smallmatrix}\right)\left(\begin{smallmatrix}b_{1}&-b_{2}\\ b_{2}&b_{1}\end{smallmatrix}\right)\left(\begin{smallmatrix}c_{1}&-c_{2}\\ c_{2}&c_{1}\end{smallmatrix}\right)=I . This map \widetilde{X}_{0}\to Z is generically finite of degree 48 . The Galois group of this cover is generated by the permutations of the pairs (a_{1},a_{2}) , (b_{1},b_{2}) , (c_{1},c_{2}) in addition to sign changes of the pairs that respect equation (3). If we restrict to specializations where a_{1}b_{1}c_{1}\neq 0 and set \displaystyle t(\phi) \displaystyle=-3\frac{(a_{2}b_{1}c_{1})^{2}+(a_{1}b_{2}c_{1})^{2}+(a_{1}b_{1}c_{2})^{2}}{(a_{1}b_{1}c_{1})^{2}}, \displaystyle u(\phi) \displaystyle=-3\frac{a_{2}b_{2}c_{1}+a_{2}b_{1}c_{2}+a_{1}b_{2}c_{2}}{a_{1}b_{1}c_{1}}, and let \phi_{1} and \phi_{2} be two specializations of this family, then the corresponding representations are equivalent if and only if we have t(\phi_{1})=t(\phi_{2}) , u(\phi_{1})=u(\phi_{2}) , and further the y_{1} coordinates of the specializations are equal, whereas the x_{1} and w_{1} coordinates of the two specializations are equal up to sign. Theorem 1.1 is a consequence of Theorem 4.14 and a slight reparameterization discussed in Section 5 . See the proof of Theorem 1.1 in Section 5.1 . Note that the map \widetilde{X}_{0}\to Z is in particular finite of degree 48 over the locus of points discussed in Theorem 1.2 below. The invariants t(\phi) and u(\phi) appearing in Theorem 1.1 above are defined in Section 4.1.2 below, and we have retained the factors of -3 to be consistent with that later material. Before outlining the layout of the paper, we introduce some notation that will both allow us to connect this work to other results in the literature, and which will allow us to list algebraic conditions that will ensure

Read the complete text