André’s theorem and weakly bounded height

Guy Fowler · 2026

Read André’s theorem and weakly bounded height by Guy Fowler (2026) 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-2606-29369.txt.

Introduction

Fowler proves an effective form of a case of the André-Oort conjecture, a statement about special points on algebraic curves. His setting is a curve in the complex plane where two coordinate functions have different degrees. The points of interest have both coordinates equal to CM j-invariants, values tied to elliptic curves with extra symmetry, called complex multiplication. Earlier effective results depended badly on the height, a measure of a curve's arithmetic size.

Fowler shows that an effectively computable constant bounds the height of every such point on the curve. Crucially, that constant depends only linearly on the height of the curve itself. Linear dependence is much better than what previous methods gave. The result is an instance of what he calls weakly bounded height. It sharpens number theory tools for finding all special points on a given curve, and it does so with explicit, usable bounds.

Catalog record

CREATORGuy Fowler
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE64 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.29369

Browse the full library · About the Tower of Babel

Full text

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

André’s theorem and weakly bounded height Guy Fowler Abstract . Let V ⊂ 𝔸 2 ​ ( ℂ ) V\subset\mathbb{A}^{2}(\mathbb{C}) be an algebraic curve such that deg ⁡ X ≠ deg ⁡ Y \deg X\neq\deg Y , where X , Y X,Y denote the coordinate functions on 𝔸 2 ​ ( ℂ ) \mathbb{A}^{2}(\mathbb{C}) restricted to V V . We prove there exists an effectively computable constant c c , that depends linearly on the height of V V , such that max ⁡ { h ⁡ ( x ) , h ⁡ ( y ) } ≤ c \max\{h(x),h(y)\}\leq c for every ( x , y ) ∈ V (x,y)\in V with x x and y y both CM j j -invariants. This establishes, for such curves, an effective version of the André–Oort conjecture that has a better dependence on the height of V V than previous effective results. 2020 Mathematics Subject Classification 11G18, 14G35 1. Introduction A point (x,y)\in\mathbb{A}^{2}(\mathbb{C}) is a special point if x and y are both singular moduli , i.e. j -invariants of elliptic curves with complex multiplication. Denote by \Delta_{x} (respectively \Delta_{y} ) the discriminant of the endomorphism order of an elliptic curve with j -invariant x (respectively y ). An irreducible algebraic curve V\subset\mathbb{A}^{2}(\mathbb{C}) is a special curve if V is one of: (1) a horizontal line \mathbb{A}^{1}(\mathbb{C})\times\{y\} for some singular modulus y , (2) a vertical line \{x\}\times\mathbb{A}^{1}(\mathbb{C}) for some singular modulus x , (3) a modular curve \mathbb{V}(\Phi_{N}(X,Y)) for some N\in\mathbb{Z}_{>0} , where \Phi_{N}\in\mathbb{Z}[X,Y] is the modular polynomial of level N , see [ Lan87 , p. 55] . The curve V is non-special if it is not a special curve. André [ And98 ] proved that every non-special algebraic curve V\subset\mathbb{A}^{2}(\mathbb{C}) contains only finitely many special points. This is the simplest non-trivial case of the André–Oort conjecture for \mathbb{A}^{n}(\mathbb{C}) ; Pila [ Pil11 ] subsequently proved the full André–Oort conjecture for \mathbb{A}^{n}(\mathbb{C}) . These proofs were ineffective: in particular, they do not give an effectively computable bound on the height (or even the number) of special points on a non-special curve V . The main unconditional result of this paper is an effective version of André’s theorem for a certain class of non-special curves. In particular, we will prove that the height of special points on such curves V may be bounded linearly in terms of the height of the curve. See § 2.1 for the definitions of the different notions of heights that appear in this paper. Theorem 1.1 . Let V\subset\mathbb{A}^{2}(\mathbb{C}) be a geometrically irreducible, non-special algebraic curve defined over a number field L . Write X,Y for the coordinate functions on \mathbb{A}^{2}(\mathbb{C}) restricted to V . Let d_{1}=\deg X and d_{2}=\deg Y . Suppose that d_{1}\neq d_{2} . Then there exist an effectively computable constant c_{1}([L:\mathbb{Q}],\max\{d_{1},d_{2}\}) and, for every \delta>0 , an effectively computable constant c_{2}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},\delta) with the following properties: for every special point (x,y)\in V , (1.1) \displaystyle\max\{h(x),h(y)\}\leq c_{1}\left([L:\mathbb{Q}],\max\{d_{1},d_{2}\}\right)+26\max\{d_{1},d_{2}\}^{3}h(V) and (1.2) \displaystyle\max\{\lvert\Delta_{x}\rvert,\lvert\Delta_{y}\rvert\}\leq c_{2}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},\delta)\max\{1,h(V)\}^{2+\delta}. The condition in Theorem 1.1 that d_{1}\neq d_{2} implies, but is strictly stronger than, the condition that V is not a modular curve \mathbb{V}(\Phi_{N}(X,Y)) . The condition that V is a non-special curve is necessary, because every special curve contains infinitely many special points. For non-special curves in general, Kühne [ Küh12 ] and, independently, Bilu, Masser, and Zannier [ BMZ13 ] proved effective versions of André’s theorem. Compared to these previous effective versions of André’s theorem, the novel aspect of Theorem 1.1 is that, when applicable, it gives a significantly better dependence on the height of the curve V . Kühne’s result [ Küh12 , Theorem 2] gives a bound of the form of ( 1.2 ), but with exponent 8+\delta in place of 2+\delta . Wüstholz [ Wüs14 , Theorem 1.1] subsequently improved this exponent to 8 . The result of Bilu, Masser, and Zannier [ BMZ13 ] does not give an explicit dependence on the height of the curve. Kühne [ Küh12 , p. 654] notes that a bound as strong as ( 1.2 ) for general non-special curves would follow via his method from a conjectural elliptic analogue [ DHK09 , Conjecture 1.7] of the Lang–Waldschmidt conjecture. Such conjectures seem out of reach at present however. More recently, Binyamini [ Bin20 , Corollary 3] and Papas [ Pap26 , Corollary 1.7] have also given effective proofs of André’s theorem for certain classes of curves, but again without obtaining an explicit dependence on the height of the curve. Their proofs both follow the Pila–Zannier strategy, which was originally applied to André–Oort for \mathbb{A}^{n}(\mathbb{C}) by Pila [ Pil09b , Pil11 ] . Instead of the ineffective Pila–Wilkie theorem [ PW06 , Pil09a ] , Binyamini and Papas apply effective point counting results also due to Binyamini [ Bin19 , Bin24 ] . 1.1. A conditional result and uniformity in André’s theorem Our proof of Theorem 1.1 will use a result of Habegger [ Hab10 , Theorem 1.1] , that gives a height bound for points lying in the intersection of a modular curve \mathbb{V}(\Phi_{N}(X,Y)) and a curve V satisfying the hypotheses of Theorem 1.1 . This is the only part of the proof of Theorem 1.1 that requires the hypothesis that d_{1}\neq d_{2} , rather than just the condition that V is non-special. Habegger conjectures [ Hab10 , p. 44] that an analogous height bound to [ Hab10 , Theorem 1.1] should hold solely under the hypothesis that V is a non-special curve. Assuming such a conjecture, we can extend the proof of Theorem 1.1 to all non-special curves. In fact, a qualitatively weaker height bound than the one in [ Hab10 , Theorem 1.1] would suffice. Definition 1.2 . Let \eta\in(0,1/2) and \kappa\in\mathbb{R}_{\geq 0} . We say that the property \mathrm{WBH}(\eta,\kappa) holds if the following property is satisfied. Let V\subset\mathbb{A}^{2}(\mathbb{C}) be a geometrically irreducible, non-special algebraic curve defined over a number field L . Write X,Y for the coordinate functions on \mathbb{A}^{2}(\mathbb{C}) restricted to V . Let d_{1}=\deg X and d_{2}=\deg Y . There exists a constant c_{*}=c_{*}(\eta,\kappa,[L:\mathbb{Q}],\max\{d_{1},d_{2}\}) such that, for every special point (x,y)\in V and every N\in\mathbb{Z}_{>0} , if \Phi_{N}(x,y)=0 , then \max\{h(x),h(y)\}\leq c_{*}\max\{1,h(V)\}^{\kappa}N^{1/2-\eta}. Similar height bounds have been formulated in [ Hab17a , Appendix B] and [ Pil22 , Conjecture 21.24] . The key aspect to note in the formulation of \mathrm{WBH}(\eta,\kappa) is that it records the dependence on h(V) explicitly. The second main result of this paper is the following conditional result. Theorem 1.3 . Let \eta\in(0,1/2) and \kappa\in\mathbb{R}_{\geq 0} . Suppose that \mathrm{WBH}(\eta,\kappa) holds with a given c_{*} . Let V\subset\mathbb{A}^{2}(\mathbb{C}) be a geometrically irreducible, non-special algebraic curve defined over a number field L . Write X,Y for the coordinate functions on \mathbb{A}^{2}(\mathbb{C}) restricted to V . Let d_{1}=\deg X and d_{2}=\deg Y . (1) For every \delta>0 , there exists an effectively computable constant c_{1}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},c_{*},\delta) such that \quad\quad\quad\max\{h(x),h(y)\}\leq c_{1}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},c_{*},\delta)\max\{1,h(V)\}^{\kappa/2\eta+\delta} for every special point (x,y)\in V . (2) If \kappa<2\eta , then there exists an effectively computable constant c_{2}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},c_{*}) such that \max\{\lvert\Delta_{x}\rvert,\lvert\Delta_{y}\rvert\}\leq c_{2}([L:\mathbb{Q}],\max\{d_{1},d_{2}\},c_{*}) for every special point (x,y)\in V . The motivation for part (2) of this theorem is the following ineffective result of Pila [ Pil09b ] . This is a uniform version of

Read the complete text