Square a7 · Mathematics · Queenside

Prime number theorem

Prime-number theorem: π(x) grows like x / ln x.

Meaning

Let π(x)\pi(x) denote the number of primes less than or equal to xx. The prime number theorem asserts that π(x)∼xln⁡x\pi(x) \sim \frac{x}{\ln x} as x→∞x\to\infty. The symbol ∼\sim means that the ratio of the two sides tends to 11: lim⁡x→∞π(x)x/ln⁡x=1.\lim_{x\to\infty}\frac{\pi(x)}{x/\ln x}=1. Thus primes become, on average, about ln⁡x\ln x apart near xx: scarcer among larger integers, but in a precisely quantified way. Equivalent formulations replace x/ln⁡xx/\ln x by the logarithmic integral li(x)\mathrm{li}(x) or involve the Chebyshev function ψ(x)∼x\psi(x)\sim x.

Context

Gauss, examining prime tables as a teenager in the 1790s, conjectured that π(x)\pi(x) grows like x/ln⁡xx/\ln x (or like li(x)\mathrm{li}(x)). Legendre proposed a related empirical formula. Chebyshev (1850s) proved bounds of the right order of magnitude and showed that if the limit of π(x)ln⁡x/x\pi(x)\ln x/x exists, it must be 11, but existence remained open.

The theorem was proved independently in 1896 by Jacques Hadamard and by Charles Jean de la Vallée Poussin, using complex analysis applied to Riemann’s zeta function ζ(s)\zeta(s). The key analytic input is that ζ(s)\zeta(s) has no zeros on the line Re(s)=1\mathrm{Re}(s)=1. Riemann’s 1859 memoir had already linked the distribution of primes to the zeros of ζ(s)\zeta(s); the Riemann hypothesis—that all nontrivial zeros satisfy Re(s)=1/2\mathrm{Re}(s)=1/2—would yield much sharper error terms for π(x)\pi(x), but remains unproved.

Elementary proofs avoiding complex analysis were given by Erdős and Selberg in 1948, answering a challenge implicit in earlier work. Those proofs are “elementary” in method, not in ease.

The prime number theorem earns a square because it transforms the qualitative infinitude of primes into a quantitative law of density. It is a flagship theorem of analytic number theory and a principal reason the Riemann zeta function occupies a central place in modern mathematics.

Error terms matter as much as the main term: the prime number theorem with a strong remainder estimate is essentially equivalent to a wide zero-free region for ζ(s)\zeta(s). In that sense the square’s asymptotic is the first chapter of a still-open story about how regularly primes appear.

Sources