Erdős Problem #672
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: verifiable (not automatically the live result).
Subject areas: number theory.
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
Can the product of four or more positive integers in an arithmetic progression be a perfect square, cube or higher integer power, when the initial term and the common difference are coprime?
Erdős Problem #672 asks whether there exist integers \(n,d,k,q,\ell\) with
such that
The maintained catalogue calls this VERIFIABLE: one genuine counterexample with all of the above hypotheses would answer yes. In the other direction, proving that none exist for any length \(k\geq4\) requires a general theorem, not a finite computation.
A small example
The bound \(k\geq4\) is essential. The three-term positive progression \(1,25,49\) has common difference \(24\) and coprime initial term, yet
So replacing \(k\geq4\) by \(k\geq3\) immediately changes the answer. Another three-term example is \(18,25,32\), whose product is \(14400=120^2\) and whose first term and difference \(7\) are coprime.
Similarly, allowing negative terms changes the question. The four integers \(-6,-1,4,9\) form an arithmetic progression of difference \(5\), and their product is \(216=6^3\). This does not satisfy the requirement that all progression terms be positive. A correct source audit must preserve both the length threshold and the positivity condition.
What is known
The classical theorem of Erdős and Selfridge says the product of consecutive positive integers is never a perfect power; that corresponds to common difference \(d=1\). Arbitrary positive \(d\) with \(\gcd(n,d)=1\) is a strictly broader problem.
The primary 2006 paper of Bennett, Bruin, Győry and Hajdu, Powers from Products of Consecutive Terms in Arithmetic Progression, Proceedings of the London Mathematical Society 92, 273–306, states in its publisher abstract that there are no such perfect-power products for \(4\leq k\leq11\).
The later Győry, Hajdu and Pintér article, Perfect powers from products of consecutive terms in arithmetic progression, Compositio Mathematica 145 (2009), 845–864, extends the excluded range to \(4\leq k\leq34\) in its author abstract. These are substantial published partial negative results, but they do not settle all arbitrarily large \(k\). The maintained catalogue reports additional results for sufficiently large \(k\) under restrictions on prime divisors of \(d\), and for large prime exponents subject to explicit bounds. Those extra hypotheses must not be dropped.
Formal-source boundary. The exact GCL-pinned 672.lean file has been directly inspected. It includes a named predicate for the arithmetic-progression product, an open parent theorem and historically proved Euler and Obláth special-case labels. However, its parent statement includes answer(sorry), its theorem proof is by sorry, and the two formal special-case proof bodies also contain sorry. Therefore this snapshot is not a complete Lean proof of the parent assertion or even a replayed GCL certificate for those special cases.
Checking the formal translation must cover the meaning of the arithmetic progression, the positive starting term and difference, the coprimality condition, the length of the finite set, the exponent and whether a purported answer asserts existence or nonexistence. A Lean source file's presence does not resolve its theorem's proof obligations.
GCL's current state
The original #1943 authenticated source audit records the reviewed 2006 and 2009 primary publisher abstracts, the signed and three-term finite fixtures, and the limitations of the pinned formal file. Its disposition is BOUNDED_TASK_DESIGNED. The original historical [Er97c] and [Va99] passages and the complete published proof pages were not independently checked.
This is an unaccepted editorial draft. It does not announce a new counterexample, an unconditional theorem for all \(k\), or independent Lean certification.
What remains
A bounded successor would compare the exact \(4\leq k\leq34\) publisher theorem with the pinned formal predicate and verify its hypotheses one by one. A further task could independently formalize one Euler or Obláth special case without a proof hole, but only if the precise source and dependencies are available.
Finite enumerations may discover and certify a positive example, if one exists, but failure to find an example cannot establish nonexistence for all integers. Any subsequent mathematical claim must distinguish the published finite range from the unrestricted infinite question.
The source-fidelity audit, accountable critical review, protected production merge, browser MathJax verification and V2 exact-hash editorial admission remain separate requirements.
Sources and claim boundaries
-
Maintained Erdős Problem #672, VERIFIABLE status and progressive partial results.
-
M. A. Bennett, N. Bruin, K. Győry and L. Hajdu, 2006 primary journal article, publisher abstract directly checked.
-
K. Győry, L. Hajdu and Á. Pintér, 2009 primary journal article, author abstract and bibliographic metadata checked.
-
Pinned Formal Conjectures 672.lean, statement and proof holes directly inspected.
-
Original GCL source return, math/source verification boundaries.
-
Historical Erdős [Er97c], [Va99, 1.32] primary pages not independently inspected.
Boundary: no original discovery, full primary-paper proof replay, Lean certificate, protected publication, deployed math rendering or editorial admission.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1943 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
- 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.