Skip to content

Erdős Problem #365

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

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

A positive integer is powerful when every prime factor occurs at least twice in its prime factorization. If \(P(n)\) denotes that property, Erdős Problem 365 asks two related questions about consecutive integers. First, if \(P(n)\) and \(P(n+1)\) both hold, must one of the two numbers be a perfect square? Second, if

\[ N(x)=\#\{n:0<n\leq x,\ P(n),\ P(n+1)\}, \]

is the number of such pairs up to \(x\), is \(N(x)\) bounded by a fixed power of \(\log x\) for all sufficiently large \(x\)? The first question has a negative answer. The second remains open in the sources checked here. These are different assertions: an infinite supply of square-containing pairs says nothing by itself about the growth of the pairs in which neither member is a square.

The relevant ambient convention matters. The maintained page speaks about positive integers, while the current formal file notes that Lean’s Nat.Powerful also includes \(0\) and therefore adds the hypothesis \(0<n\). The source-normalized first statement is therefore: for every \(n\in\mathbb N\), if \(0<n\), \(P(n)\), and \(P(n+1)\) hold, then either \(n\) or \(n+1\) is a square. The answer is false.

A small example

The decisive finite example is

\[ 12167=23^3,\qquad 12168=2^3\,3^2\,13^2, \]

and \(12168=12167+1\). Every prime divisor in these displayed factorizations has exponent at least two, so both numbers are powerful. They are not squares: exact arithmetic gives

\[ 110^2=12100<12167<12168<12321=111^2. \]

Thus this one pair refutes the universal square assertion. The calculation is reproducible directly from integer arithmetic: \(23^3=12167\), \(2^3\cdot3^2\cdot13^2=12168\), and the two strict square bounds above all evaluate exactly. It establishes a counterexample, not a classification of all consecutive powerful pairs, not an asymptotic estimate for \(N(x)\), and not a proof that the displayed pair belongs to any particular infinite family.

The first part of the current external Lean source records this same example, including explicit checks of prime divisors and the square bounds. Its theorem is stated with answer(False). That source is useful formalization evidence, but this article does not claim a Lean build or GCL certification.

What is known

The original references distinguish two types. Type I pairs have a square member and arise from Pell equations. The maintained problem page gives the basic example \(x^2=2^3y^2+1\): the pair \(8y^2,x^2\) is consecutive and powerful whenever the Pell equation is solved, so there are infinitely many pairs of this kind.

Golomb’s primary paper, “Powerful Numbers,” The American Mathematical Monthly 77 (1970), 848–852, is the historical source associated with the definition and the initial consecutive-pair discussion. The journal record is available through the DOI record. The maintained page attributes the explicit \(12167,12168\) observation to Golomb.

Walker’s primary paper, “Consecutive Integer Pairs of Powerful Numbers and Related Diophantine Equations,” The Fibonacci Quarterly 14 (1976), 111–116, is available as the scanned article. Section 2 treats the Pell-equation, square-containing type. Section 3 treats the non-square type through equations of the form \(mX^2-nY^2=\pm1\) together with a divisibility condition ensuring that \(mX^2\) and \(nY^2\) are powerful. Theorem 3.5 states the corresponding odd-power parametrization, and the page-116 example computes a consecutive pair from \(7X^2-3Y^2=1\).

There is a normalization point to retain. The maintained page displays the related equation \(7^3x^2=3^3y^2+1\) and attributes infinitely many counterexamples to Walker. The page-116 example in the scanned paper is printed in the squarefree-coefficient form \(7X^2-3Y^2=1\). These may be related by the paper’s general framework, but the exact change of variables and the claim that the displayed cubic-coefficient equation itself is the one proved were not reconstructed here. The safe statement is that Walker gives a primary-source Type II mechanism and a concrete infinite-family framework; the exact displayed equation needs source-semantic reconciliation before stronger attribution.

GCL's current state

The frozen catalogue evidence is teorth/erdosproblems at commit 536408716724a2402bdfa7c1d764bf83d64a8c89, file data/problems.yaml, blob 4e60263a80af6207aca494d2e1388e44fa372889. Its entry for 365 was last updated 2025-09-20: informal status open, formal status unformalized, statement-formalized no. Its note already records the \(12167,12168\) counterexample and says the counting question remains open. The protected Programme triage therefore remains STATEMENT_CAPTURE_REQUIRED; the admitted FC-GDM-85F86371 snapshot has no protected statement file for this problem.

The maintained live problem page was last edited 31 October 2025 and now says that the first answer is no and links a formalized statement. The current upstream file is FormalConjectures 365.lean, exact commit dc6995d9248a8e5bde2587b1f219b981cbf48acb, blob c5ca30c6237a1ccd1b37eb1385133a47e2a92a11, dated 2026-10-09. It contains a proved-looking first theorem in source text and a separate counting theorem whose proposition is answer(sorry) and whose proof is by sorry. This is a dated live-source change, not a protected GCL admission and not a certification.

What remains

The smallest useful successor is a source-normalization task for the counting question. Its input should be the positive-integer definition of powerful numbers, the exact set counted by \(N(x)\), and the original Golomb and Walker references. Its target should be one of two bounded outcomes: either a primary-source theorem with exact hypotheses that implies a polylogarithmic upper bound for \(N(x)\), or a rigorously documented reduction showing the first missing lemma needed for such a bound.

The task should first reconcile the \(7^3\) equation with Walker’s squarefree-coefficient notation, then isolate what Walker proves about Type II families versus what he does not say about their total counting function. Acceptance requires a page-located citation, normalized quantifiers, and a proof or reduction whose variables and positivity conditions match \(N(x)\). A falsification is either a counterexample to the proposed reduction or a source passage showing that the claimed hypothesis is absent. Stop after the source reconciliation and one bounded reduction attempt if the primary text does not yield a counting theorem; do not turn a citation or finite computation into an asymptotic claim.

Sources and claim boundaries

This article is tied to original issue #1635, assignment ERDOS-CATALOGUE-0365-S01. The original SOURCE RESULT/1 citation is authenticated original RESULT/1. The frozen catalogue commit and blob above are the source lock used for the dated baseline; the live page and later FormalConjectures commit are reported as drift, not silently substituted into that baseline. The Walker PDF was read as a primary source; the Golomb journal record was verified, while no claim is made here to have independently reconstructed every argument in Golomb’s article.

External proof text, an upstream Lean theorem containing sorry, finite exact arithmetic, and GCL mathematical certification are separate categories. No GCL certification, publication decision, accepted-coverage decision, or claim of a completed solution to the counting question is supplied. V2 ledger state alone decides accepted coverage.

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1635 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
  • GCL triage: STATEMENT_CAPTURE_REQUIRED.
  • Formal-statement coverage: No statement was confirmed in GCL's pinned formal-provider 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