Metric Number Theory

Metric number theory studies the approximation properties of almost all real numbers, shifting the focus from individual numbers to the measure-theoretic and geometric structure of exceptional sets. Rather than asking whether a specific α admits good rational approximations, metric theory asks: what is the Lebesgue measure of the set of numbers that do? What is their Hausdorff dimension? How do these properties vary with the approximation function?

The founding result is the Khinchin theorem (1924): for a monotonic function ψ(q), the inequality |α − p/q| < ψ(q)/q has infinitely many solutions for almost all α if and only if Σ ψ(q) diverges. This zero-one law is one of the earliest instances of a probabilistic dichotomy in number theory, and it established that the approximation type of a typical real number is not a subtle property but a generic one governed by a simple convergence criterion.

The metric perspective reveals paradoxical structures. The set of badly approximable numbers has measure zero yet full Hausdorff dimension — they are negligible in volume but geometrically thick. This tension between measure and dimension is characteristic of the fractal geometry of number-theoretic sets, and it connects Diophantine approximation to dynamical systems through the ergodic theory of the continued fraction map.

Metric number theory is the systems-level complement to the individual results of Thue, Siegel, and Roth. Where those theorems classify specific numbers, metric theory classifies the space itself.