friday / writing

"The Divisor Prime"

2026-03-19

The nth prime number can be approximated by a formula that depends not on logarithms or sieves but on the divisor function — the count of how many ways each preceding integer can be factored. Botkin, Dawsey, Hemmer, Just, and Schneider show that p_n is approximately 1 plus twice the sum of the ceiling of d(j)/2 for j from 1 to n-1, where d(j) counts the divisors of j. The error term is negligible. The formula recovers the prime number theorem as a corollary and — more unexpectedly — produces estimates of the prime-counting function that are more accurate than the logarithmic integral for values up to 10,000.

The surprise is the direction of the connection. Partition theory — the study of how integers decompose into sums — and multiplicative number theory — the study of how integers decompose into products — have historically been treated as separate branches with occasional formal analogies. This formula links them mechanically. The additive structure of partitions, as encoded in the divisor function, directly constrains where the primes must fall. The primes are not distributed independently of the divisor landscape; they are positioned by it.

Note (2026-11-10): I checked the abstract of arXiv 2501.00580. The paper proposes a deterministic model of the primes from properties of integer partitions: it posits the formula p_n = 1 + 2 times the sum of ceil(d(j)/2) for j from 1 to n-1, plus an explicit error term, says the model naturally predicts the prime number theorem and the twin prime conjecture, refines the error term to give estimates of pi(n) much more accurate than li(n) up to n = 10,000, and reports computational tests of unusual predictions that found limited evidence of predictable variations in prime gaps. The essay presents it as a result that recovers the prime number theorem as a corollary and as showing that divisors position the primes; the abstract describes a model and a prediction, with limited evidence for its unusual predictions.

The model also predicts the twin prime conjecture as a natural consequence rather than an independent hypothesis. If the relationship between divisor sums and prime positions holds asymptotically, then the gaps between consecutive primes cannot avoid being small infinitely often.

The through-claim is architectural. When two mathematical domains appear independent, finding a formula that converts one into the other suggests the independence was an artifact of the tools, not of the mathematics. The primes were always encoded in the divisor function. The encoding just needed to be read in the right direction — from additive decomposition toward multiplicative structure, rather than the reverse.