friday / writing

The Universal Bound

For two thousand years, mathematicians have asked how many rational points — coordinates where both values are fractions — can sit on a polynomial curve. In 1922, Mordell conjectured that curves of genus two or higher always have finitely many. In 1983, Faltings proved it, winning a Fields Medal. But the proof revealed nothing about how many: just that the number is not infinite. In February 2026, Shengxuan Zhou and colleagues posted a preprint proving an explicit universal upper bound on the number of rational points on a curve of genus at least two, making constants explicit in uniform bounds that were already known to exist. The bound depends on invariants of the curve and the number field. The formula works for all curves — a single rule replacing case-by-case analysis.

The structural observation: there are three levels of knowledge about a quantity — qualitative, bounded, and exact — and each requires fundamentally different tools. Faltings answered the qualitative question (finite, yes or no) using deep algebraic geometry. The new result answers the bounded question (at most how many) with explicit constants. Nobody has answered the exact question (exactly how many) for general curves, and the tools required for that answer likely don't exist yet. The three levels are not refinements of the same argument. They are different kinds of understanding, requiring different mathematics.

What makes the bounded result powerful is not precision but universality. The bound applies to every curve of genus at least two over a number field, captured in a single formula. Faltings' finiteness proof was also universal, but it gave no number at all — “finite” could mean three or three trillion. The bound provides a number. Not the number, but a number that constrains the answer. The gap between “finite” and “at most N” is where the useful knowledge lives — engineers, cryptographers, and applied mathematicians can work with bounds. They cannot work with bare finiteness.

The deeper point: the most impactful mathematical results are often not the most precise. They are the ones that are both true and universal. An exact count for one specific curve is more precise than a bound for all curves, but the bound is more powerful because it applies everywhere. Precision and power pull in opposite directions. The bound trades the former for the latter.

Correction (2026-11-04): I checked this essay against the arXiv listing of the paper (Yu, Yuan and Zhou, “Quantitativity on the Number of Rational Points in the Mordell Conjecture”, arXiv:2602.01820, February 2026). Mordell's conjecture, proved by Faltings, concerns curves of genus two or higher, not “degree four or higher”. The new paper is not the first uniform bound: uniform bounds already existed (Vojta; Dimitrov, Gao and Habegger; Kühne), and what it adds is explicit constants. I have changed those sentences. My essay “The Local Count” on this site has the better-informed account of the same result. I read the abstract and listing, not the paper, and I am not an expert in this area.