Skip to content

Erdős Problem #15

Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (not automatically the live result).
Subject areas: number theory, primes.

Read this status correctly

The status above is reproduced from GCL's immutable external-catalogue snapshot. It is not a new claim of current mathematical resolution, a source-semantic review, an independent proof check, or GCL certification. A published proof and a GCL replay are separate events.

The question in plain language

Let \(p_n\) denote the prime of index \(n\), starting with \(p_1=2\). Erdős Problem #15 asks whether the ordered real series

\[ \sum_{n=1}^{\infty}(-1)^n\frac{n}{p_n} \]

converges. Its first terms are \(-1/2+2/3-3/5+4/7-\cdots\). The signs alternate according to the prime's index, not whether the prime itself is odd or even.

There is a critical formalization trap: ordinary unconditional summability in Lean is not the correct predicate for convergence of an ordered, potentially conditionally convergent alternating series.

A small example

Pair the terms associated with consecutive primes \(a=p_{2m-1}\) and \(b=p_{2m}\), and write \(g=b-a>0\). Their exact sum is

\[ -\frac{2m-1}{a}+\frac{2m}{b} =\frac{a-(2m-1)g}{ab}. \]

For \(m=1\), the primes 2 and 3 give \(-1/2+2/3=1/6\). For \(m=2\), the primes 5 and 7 give \(-3/5+4/7=-1/35\). These pair identities are exact finite arithmetic, not evidence that infinitely many pairs sum to a limit.

What is known

Tao, The convergence of an alternating series of Erdős, assuming the Hardy–Littlewood prime tuples conjecture, Communications of the American Mathematical Society 4 (2024), 80–96, proves convergence conditionally on a sufficiently strong quantitative Hardy–Littlewood prime-tuples hypothesis. That is not an unconditional resolution.

The prime number theorem implies \(p_n\sim n\log n\), so the absolute magnitudes \(n/p_n\) are asymptotic to \(1/\log n\). Their positive series diverges. Therefore replacing the original question with unconditional or absolute summability produces a false problem statement even if the original ordered real series eventually proves convergent.

GCL's current state

Source audit #1284 checks the frozen provider and identifies two separate defects: it uses rational-valued rather than real-valued summability, and an unordered unconditional notion rather than the ordered partial-sum limit. The audit verifies a prime-pair identity with exact rational arithmetic and documents a public counterexample result against the defective formal statement. That result is a formalization defect, not a solution of the original Erdős question.

What remains

Prove convergence or divergence of the actual ordered real partial sums without assuming the required Hardy–Littlewood conjecture. Any future formalization must use convergence of sequential finite partial sums, not general unordered summability, and must distinguish finite term arithmetic from tail control.

Sources and claim boundaries

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1284 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
  • GCL triage: PROTECTED_READY_BACKLOG.
  • Formal-statement coverage: A provider statement was included in the pinned snapshot. Presence does not mean the statement or proof has been checked.
  • External mathematical status: consult current problem page and its dated bibliography.

Back to catalogue · Compendium introduction