Categoricity and non-arithmetic Fuchsian groups Thanks: J. Nagloo is partially supported by NSF grant DMS-2348885. Some of the work in this paper was completed while J. Nagloo was supported by an AMS Centennial Fellowship. John T. Baldwin, Joel Nagloo Abstract . Let Γ ⊂ P ​ S ​ L 2 ​ ( ℝ ) \Gamma\subset PSL_{2}(\mathbb{R}) be a non-arithmetic Fuchsian group of the first kind with finite covolume, and let j Γ j_{\Gamma} be a corresponding uniformizer. In this paper we introduce a natural L ω 1 , ω L_{\omega_{1},\omega} -axiomatization T S ​ F ∞ T^{\infty}_{SF} of the theory of j Γ j_{\Gamma} viewed as a covering map. We show that T S ​ F ∞ T^{\infty}_{SF} is categorical in all infinite cardinalities, extending to the non-arithmetic setting earlier results of Daw and Harris obtained in the arithmetic case. We also show that the associated first-order theory T j Γ T_{j_{\Gamma}} is complete, admits elimination of quantifiers, and is ω \omega -stable. 1. Introduction Let \Gamma\subset PSL_{2}(\mathbb{R}) be a Fuchsian group of the first kind and of finite covolume, and let j_{\Gamma}:\mathbb{H}_{\Gamma}\rightarrow S_{\Gamma}(\mathbb{C}) be a uniformizer, where \mathbb{H}_{\Gamma}:=\mathbb{H}\cup C_{\Gamma} and C_{\Gamma} is the set of cusps of \Gamma . Thus, j_{\Gamma} is a \Gamma -automorphic function: a covering map realizing the smooth projective curve S_{\Gamma}(\mathbb{C}) as the quotient \Gamma\setminus\mathbb{H}_{\Gamma} . This classical situation, originating in the work of Poincaré and Klein at the turn of the twentieth century, lies at the intersection of complex analysis, hyperbolic geometry, and the theory of automorphic functions. This paper draws on two streams of work studying the function j_{\Gamma} through the lens of model theory. On the one hand, there is the work of Blázquez-Sanz, Casale, Freitag and Nagloo (cf. [ CFN21 ] , [ BCFN21 ] ), which provides a full classification of the structure of the set \mathcal{X}_{\Gamma} defined — in a differentially closed field — by the Schwarzian differential equation satisfied by j_{\Gamma} . Their work shows that a key dividing line in geometric group theory, namely the arithmeticity of the group \Gamma , fully governs the complexity of algebraic relations among the solutions in \mathcal{X}_{\Gamma} . Roughly speaking, they show that the Fuchsian group \Gamma is arithmetic if and only if \mathcal{X}_{\Gamma} is not \omega -categorical, that is, if and only if the solution set admits infinitely many ‘algebraic symmetries’. It is a fundamental result of Margulis that the existence of infinitely many ‘Hecke’ correspondences (equivalently, a large commensurator \text{Comm}(\Gamma) ) characterizes the arithmeticity of the group. In particular, one of the main achievements of the work of Blázquez-Sanz, Casale, Freitag and Nagloo is to show that these ‘Hecke’ correspondences account for all algebraic relations in \mathcal{X}_{\Gamma} . The second line of work in model theory studying j_{\Gamma} is also concerned with questions of categoricity, but in uncountable cardinalities rather than \aleph_{0} . Namely, one may view j_{\Gamma}:\mathbb{H}_{\Gamma}\rightarrow S_{\Gamma}(\mathbb{C}) as a two-sorted structure (\mathbb{H}_{\Gamma};S_{\Gamma}(\mathbb{C})) in a natural language and find an L_{\omega_{1},\omega} -axiomatization of the theory T_{j_{\Gamma}} of that structure. This perspective is rooted in Zilber’s program to find axiomatizations of canonical structures that are categorical in all uncountable powers. This gives an ‘algebraic’ characterization of structures with inherent analytic properties. A natural question is whether T_{j_{\Gamma}} exemplifies this program — in particular, whether the unique model of cardinality continuum must coincide with the classical analytic structure. Daw and Harris [ DH17 ] showed that if \Gamma is arithmetic, then T_{j_{\Gamma}} is indeed categorical in all uncountable powers. This result was later generalized by Eterović [ Ete22 ] to include arithmetic uniformizations in higher dimensions. Their work makes essential use of deep results from arithmetic geometry, in particular open image theorems for Galois representations governing Hecke orbits on Shimura varieties, while also showing that categoricity itself can force arithmetic statements about the Galois representations associated with points on the Shimura variety via a theorem of Keisler. In higher dimensions, this relies on open image conditions that are presently conjectural and closely related to the Mumford–Tate conjecture. It is natural to ask whether the arithmeticity of the group \Gamma again provides a dividing line for categoricity in this new sense. In this paper, we show that, unlike in the context of differentially closed fields, the categoricity of T_{j_{\Gamma}} is not governed by the arithmetic nature of \Gamma . More precisely, we prove that when \Gamma is non-arithmetic, the theory T_{j_{\Gamma}} is categorical in all uncountable cardinalities. Moreover, the proof in the non-arithmetic case is substantially simpler than in the arithmetic setting, and its structure clarifies the respective roles played by model-theoretic and geometric considerations in the earlier results. Our argument follows the same pattern as [ DH17 ] : i) axiomatize the class of models in L_{\omega_{1},\omega} (which we named T^{\infty}_{SF} ) and prove by a back-and-forth that the class satisfies all conditions on a quasi-minimal excellent class except \aleph_{0} -homogeneity over models (see Definition 3.7 .2.c and Definition 5.1 .1 and 5.1 .2.a; ii) Apply further deep geometric results to obtain the needed homogeneity. Our first main theorem is Theorem A . All models of T^{\infty}_{SF} are (\infty,\omega) equivalent and satisfy the same Scott sentence. The associated first order theory T_{j_{\Gamma}} is complete, admits elimination of quantifiers, and is \omega -stable. Compared to the arithmetic setting, there are two important variants in the proof. The first variant is that the non-arithmeticity both vastly simplifies the argument for Theorem A and allows the conclusion of \omega -stability at this point. This improvement allows the second, namely, a completely different argument for our next main result: Theorem B . The class is \aleph_{0} -homogeneous over models, so almost quasiminimal- excellent and thus categorical in all infinite cardinalities. Rather than relying on open-image theorems of Serre et al ( [ DH17 , §5.1] ), we apply the characterization [ BHH+14 ] of full excellence ( \aleph_{0} -homogeneity over models) in terms of non-splitting, and thus we obtain the result via \omega -stability. The paper is organized as follows. After some background on Fuchsian groups § 2 , we describe in Section 3 the formalism for the current study. In Section 4 we show Theorem A (Theorem 4.11 ) and introduce a notion of determined type which clarifies the argument for both the arithmetic and non-arithmetic cases. In Section 5 , we use \omega -stability and the notion of determination to conclude Theorem B (Theorem 5.2 and Theorem 5.8 ). 2. Background on Fuchsian groups and the covering maps We begin by recalling some of the basic facts about Fuchsian groups that will play a role in our analysis. Standard references for this section are [ Kat92 ] and [ Shi71 ] . Throughout, \mathbb{H} will denote the upper half complex plane, while \mathbb{H}^{*}:=\mathbb{H}\cup{\bf P}^{1}(\mathbb{R}) . For F a subfield of \mathbb{C} , we will consider the action on \mathbb{H} (and \mathbb{H}^{*} ) of SL_{2}(F) and PSL_{2}(F) by linear fractional transformation: for g=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in SL_{2}(F) and \tau\in\mathbb{H} , we write g\cdot\tau=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\cdot\tau:=\frac{a\tau+b}{c\tau+d} . In the case when F=\mathbb{R} , this action gives all the orientation preserving isometries of \mathbb{H} . Let \Gamma\subset PSL_{2}(\mathbb{R}) be a Fuchsian group, namely, a discrete subgroup of PSL_{2}(\mathbb{R}) . We assume throughout that \Gamma is of the first kind and of finite covolume. This means, respectively, that its limit set is {\bf P}^{1}(\mathbb{R}) and that it has a fundamental domain with a finite area. Indeed, in this paper a “Fuchsian group” will always mean one where these two conditions hold. By a cusp of \Gamma , we mean a point \tau\in\mathbb{H}^{*} such that its stabilizer group \Gamma_{\tau}=\{g\in\Gamma\;:\;g\cdot\tau=\tau\} contains a parabolic element \gamma\in\Gamma , that is, an element \gamma=\begin{pmatrix}a&b\\ c&d\end{pmatrix}\in\Gamma with trace \text{tr}(\gamma):=a+d=\pm 2 . It is a fact (cf. [ Shi71 , §1.3] ) that \Gamma acts on the set C_{\Gamma} of its cusps and the action of \Gamma on \mathbb{H}_{\Gamma}:=\mathbb{H}\cup C_{\Gamma} , yields a compact Riemann surface \Gamma\setminus\mathbb{H}_{\Gamma} or equivalently a projective non-singular irreducible curve S_{\Gamma} . Notice that, if \Gamma\setminus\mathbb{H} is non-compact, then the non-compactness occurs exactly at finitely many cusps. By an automorphic function for \Gamma (or \Gamma -automorphic), we mean a meromorphic function f on \mathbb{H} which is meromorphic at every cusp of \Gamma and which is invariant under the action of \Gamma : f(g\cdot\tau)=f(\tau)\;\;\;\text{ for all }\;g\in\Gamma\text{ and }\tau\in\mathbb{H}. One usually denotes the field of \Gamma -automorphic functions by \mathcal{A}_{0}(\Gamma) and the latter is isomorphic to the function field of the curve S_{\Gamma} . In particular, we have a distinguished automorphic function for \Gamma , denoted j_{\Gamma}:\mathbb{H}_{\Gamma}\twoheadrightarrow S_{\Gamma}(\mathbb{C}) , called a uniformizer that is a generator of \mathcal{A}_{0}(\Gamma) . The function j_{\Gamma} is not unique since there are nontrivial automorphisms of the curve S_{\Gamma} . Moreover, uniqueness follows once the value of j_{\Gamma} at three points have been specified. The following notion is central to the study of Fuchsian groups. Definition 2.1 . Let \Gamma and \Gamma^{\prime} be two Fuchsian groups. (1) We say that \Gamma and \Gamma^{\prime} are commensurable (as subgroups of PSL_{2}(\mathbb{R}) ) if \Gamma\cap\Gamma^{\prime} is of finite index in both \Gamma and \Gamma^{\prime} , written \Gamma\sim\Gamma^{\prime} . (2) The commensurator of \Gamma is the subgroup of PSL_{2}(\mathbb{R}) defined as \text{Comm}(\Gamma)=\{g\in{\mathop{\rm PSL}}_{2}(\mbox{\msbm R})\colon g^{-1}\Gamma g\ \text{\ is commensurable with}\ \Gamma\}. In this paper we are mainly concerned with non-arithmetic Fuchsian groups. The definition of arithmeticity, while not difficult, is quite involved and will not be given here. We refer the reader to the references above. For the purposes of our work, it suffices and indeed essential to consider an equivalent characterization of arithmeticity due to Margulis. Fact 2.2 (Margulis) . Let \Gamma be a Fuchsian group. Then, \Gamma is arithmetic if and only if it has infinite index in \text{Comm}(\Gamma) . In particular, a characteristic feature of arithmetic Fuchsian groups is the existence of infinitely many “Hecke” correspondences in S_{\Gamma}(\mathbb{C})\times S_{\Gamma}(\mathbb{C}) . Let us make this notion of correspondences more precise; note that non-arithmetic Fuchsian groups will only have finitely many of those. If g\in\text{Comm}(\Gamma) is given, then by definition we have three pairwise commensurable Fuchsian groups: \Gamma , \Gamma^{g}:=g^{-1}\Gamma g and \Gamma_{g}:=\Gamma\cap\Gamma^{g} . It is not hard to prove that these groups have the same set of cusps (cf. [ Shi71 , Proposition 1.30] ). So the uniformizers j_{\Gamma} and j_{\Gamma}\circ g of \Gamma and g^{-1}\Gamma g respectively, are also automorphic functions for their common subgroup \Gamma_{g} . A classical result due to Poincaré gives that any two automorphic functions for a Fuchsian group are related by a complex polynomial: Fact 2.3 . For each g\in\text{Comm}(\Gamma)\setminus\Gamma , there is a non-zero polynomial \Phi_{g}\in\mathbb{C}[X,Y] such that for all z\in\mbox{\msbm H} , \Phi_{g}(j_{\Gamma}(z),j_{\Gamma}(g\cdot z))=0 and \Phi_{g}(j_{\Gamma}(z),y)=0 if and only y=j_{\Gamma}(h\cdot z) for some h\in\Gamma g\Gamma . The polynomial \Phi_{g} is called a \Gamma -special polynomial and does not depend on the choice of g but rather on the choice of the double coset \Gamma g\Gamma [ Shi71 , Pages 51-52, 76-77] . A proof of its existence can be found in [ Leh64 , Page 185] . Moreover, the full statement of Fact 2.3 is also obtained as a statement about finite branched covers of Riemann surfaces. Before giving the details of that approach, we fix some notation for non-arithmetic groups that we will use throughout the paper. Note however, that the argument will also hold for arithmetic groups since we will only use the properties of the commensurator of \Gamma . Notation 2.4 . Let G=\text{Comm}(\Gamma) . (1) [G:\Gamma]=k and G=\cup_{i