Erdős Problem #121
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: disproved (not automatically the live result).
Subject areas: number theory, squares.
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
For an integer \(k\ge2\), let \(F_k(N)\) be the largest size of a set \(A\subseteq\{1,\ldots,N\}\) for which no \(k\) distinct elements have product equal to a perfect square. Erdős, Sárközy and Sós asked whether, for odd \(k\ge5\),
This would mean one could retain almost every integer while avoiding square products of every specified odd number of distinct members. Reusing an element in a product is prohibited.
A small example
For \(A=\{2,3,6\}\), the product of all three elements is \(2\cdot3\cdot6=36\), a square. This shows why excluding squares among individual elements does not suffice. The square status of a product is determined by parity of each prime exponent: exponent vectors can be added modulo two.
If \(A\) is any four-element subset and \(k=5\), there are no five distinct choices, so the condition is vacuously satisfied. Finite samples must therefore use \(|A|\ge k\) before testing meaningful forbidden products.
What is known
DISPROVED. Terence Tao, On product representations of squares, Acta Mathematica Hungarica 175 (2025), 142–157, proved that for every fixed \(k\ge4\), some constant \(c_k>0\) exists with
The published primary abstract and author's 20 May 2024 explanation both explicitly support this negative answer. Tao uses a probabilistic argument; the even-exponent asymptotics were established earlier by Erdős–Sárközy–Sós. The old \(k=3\) estimate \(F_3(N)=(1-o(1))N\) is not contradicted.
There is a distinct related problem: forbidding square products of any odd number of distinct members, with no prescribed \(k\). It has a different extremal function and asymptotic constant, studied by Granville and Soundararajan. This cannot be interchanged with the fixed-parameter conjecture.
GCL's current state
The pinned 2025 catalogue and maintained Erdős #121, edited October 2025, mark the original conjecture DISPROVED. The source issue is #1390. Tao's primary theorem was checked at the abstract and author-exposition level, not fully reproved. Reported third-party Lean status is not promoted into the GCL protected formal-provider snapshot, where this particular statement file is absent.
The original issue contains an authenticated RESULT/1 source report. This is unadjudicated source evidence only, not independent mathematical certification or editorial admission.
What remains
A bounded finite-check task is to factor each positive integer up to 30, encode prime-exponent parity as a bit vector, and search all \(k\) subsets of a proposed \(A\) for XOR equal to zero. Acceptance: the triple \(\{2,3,6\}\) is detected as square-product, duplicate-index tuples are rejected, and independent exact integer products agree with parity-vector tests. Stop at this semantic fixture; no asymptotic probabilistic theorem follows from finite enumeration.
Sources and claim boundaries
- Tao, Acta Mathematica Hungarica 175 (2025), primary published disproof.
- Tao's author exposition (2024).
- Maintained Erdős #121 and original issue #1390.
No new extremal bound, independent Lean proof replay, or editorial admission is claimed.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1390 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
- 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.