Experimental investigations on Lehmer's conjecture for elliptic curves
Sven Cats, John Michael Clark, Charlotte Dombrowsky, Mar Curco Iranzo, Krystal Maughan, Eli Orvis
TL;DR
This work tackles Lehmer’s conjecture for elliptic curves by computing the smallest possible canonical height $\hat{h}(P)$ of a non-torsion point defined over extensions of fixed degree, formalized as $C_E = \inf \{ \hat{h}(P) [K(P):K] \}$. It also considers Lang’s refinement $C_{K,d}$ using a curve-invariant bound $M_E = \max\{ h(j_E), \log |N_{K/\mathbb{Q}}\Delta_E|, 1 \}$ and $C_{K,d} = \inf\{ \hat{h}(P)/M_E \}$ for degree-$d$ extensions. The authors develop a finite-search framework by reducing to a finite subfamily $\mathscr{F}'$ via a discriminant-bound $\Delta(D,E,F)$ and a refined height bound $h(P)-\hat{h}(P)$ (a modified CPS bound), enabling practical computation of minimal-height points over quadratic fields. Applying this to $17{,}834$ curves over $\mathbb{Q}$, they obtain 86 curves with provably smallest quadratic-field points and provide detailed field-defining data, observing limited correlation with classical invariants and offering empirical guidance for Lehmer–Lang-type conjectures and future computational directions.
Abstract
In this short note, we give a method for computing a non-torsion point of smallest canonical height on a given elliptic curve $E/\mathbb{Q}$ over all number fields of a fixed degree. We then describe data collected using this method, and investigate related conjectures of Lehmer and Lang using these data.
