Randomized trial demonstrates conditional harmonic memory in prime gaps, suggesting new arithmetic insights.
Let (g_n=pₙ₊₁-p_n) denote consecutive prime gaps near a large height (X), with local scale (L=log X). This paper develops a unified framework for two complementary forms of arithmetic memory. First, in the coordinates (s=g_n+gₙ₊₁) and (t=g_n-gₙ₊₁), a model that exactly matches the empirical marginal distribution of (s) and weights each admissible split by the Hardy–Littlewood triple singular series remains systematically too broad in the conditional (t)-direction. Across seven numerical bands containing 489,720,096 adjacent-gap pairs, the conditional variance ratio follows (1-κ/L) to first order, with (κ=0.351±0.009). The identity (Cov(g_n,gₙ₊₁)=[Var(s)-Var(t)]/4) localizes the residual dependence inside the conditional splitting law. Second, under an explicitly stated quantitative Hardy–Littlewood/mod-Poisson count expansion and a square-tail uniform-integrability condition, an exact Palm inversion yields, for every fixed (k), the conditional asymptotic law[Cov(g_n,gₙ₊ₖ)=-L/2k+o(L),-ρ_kL1/2k.]The harmonic coefficient arises from the discrete identity (Δ^2(mH_m)=1/m), rather than from curve fitting. An exact centered-singular-series transform identifies the contribution of every factorial rank to the count law, while the same Palm calculation produces all fixed block moments and predicts the parameter-free third-cumulant ratio (R_m⁽³⁾(L)→3). A segmented-sieve computation over 1,825,728,368 prime gaps at six heights tests both predictions. At (10¹⁵), ranks (3,4,5) fit (q_k=C/k) with (C=0.49519±0.00222), and a joint fixed-limit fit of the third-cumulant statistic for (m=1,2) gives (χ^2=6.975) on ten degrees of freedom, with (p=0.728). Exact identities, conditional theorems, model consequences, and finite computations are explicitly distinguished throughout the manuscript.
No takes yet. Share an insight, caveat, or question.
Salem Eid (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: