Skip to content

Erdős Problem #352

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

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

Erdős Problem 352 asks whether there is one finite number that is large enough to force a unit-area triangle inside every sufficiently large measurable subset of the plane. In symbols, the set is a measurable \(A \subseteq \mathbb{R}^2\), its Lebesgue measure is at least some \(c>0\), and the desired conclusion is that three points of \(A\) are the vertices of a triangle whose area is exactly \(1\).

The quantifiers matter. The same constant \(c\) must work for every measurable set, including disconnected, unbounded, and highly irregular sets. The triangle may be chosen anywhere in the set; no convexity or boundedness is assumed. The problem asks whether some finite \(c\) exists, not whether the historically suggested value is already proved.

There is also an endpoint issue. Several sources formulate the forcing condition with measure strictly greater than \(c\), while the live problem page and the pinned formal statement use measure at least \(c\). Those formulations should not be silently identified. The open disk whose largest possible inscribed triangle has area \(1\) has area \(4\pi/\sqrt{27}\), but its boundary is absent, so no triangle in that open disk has area exactly \(1\). Thus this disk rules out the non-strict assertion at that particular endpoint; it does not rule out the existence of some larger finite threshold.

A small example

The discussion attached to the live problem introduces a useful discrete model. For a finite set \(S\subseteq\mathbb{Z}^2\), call it \(N\) avoiding when every triple from \(S\) has \(\lvert \mathrm{area}(\triangle)-N\rvert\) greater than the triangle's Euclidean diameter. This is a scaled finite model, not an equivalent theorem by itself.

Here is an exact check for \(N=2\). Take the four corners

\[ S=\{(0,0),(1,0),(0,1),(1,1)\}. \]

Among all \(4^3=64\) ordered triples, the ordinary triangle area is either \(0\) or \(1/2\). The largest possible diameter is \(\sqrt{2}\). In the first case, \(\lvert 0-2\rvert=2>\sqrt{2}\); in the second, \(\lvert 1/2-2\rvert=3/2>\sqrt{2}\), since \(9/4>2\). This verifies the predicate using only integer arithmetic after squaring the positive quantities. It is a reproducible finite witness for the toy model. It does not prove a bound for arbitrary measurable planar sets, and it says nothing about the asymptotic growth of the largest sets avoiding \(N\).

What is known

The historical sources cited by the catalogue record Erdős's observation that an unbounded measurable set of positive measure, and in particular a set of infinite measure, contains triangles of prescribed positive area. The difficult case is therefore a bounded finite-measure set. Erdős suggested the constant \(c_0=4\pi/\sqrt{27}\approx 2.418\) because an equilateral triangle of area \(1\) is inscribed in a disk of that area. The disk calculation gives a sharp obstruction to any smaller threshold, with the endpoint caveat described above.

R. Daniel Mauldin states the problem as Problem 5.1 in his 2013 paper Some Problems and Ideas of Erdős in Analysis and Geometry (PDF pp. 6–7, publication pp. 371–372). He records the reduction to unions of interiors of finitely many compact convex sets as Problem 5.2. The same discussion reports that the conjectured constant is optimal for one, two, and three convex bodies. These are substantial restricted results, but they do not settle the unrestricted measurable-set question. The live catalogue also attributes to Freiling and Mauldin a result saying that outer measure greater than \(c_0\) forces a triangle of area greater than \(1\); that statement is useful evidence, but “greater than \(1\)” is not automatically “equal to \(1\)” for an arbitrary disconnected measurable set.

The live Erdős Problems page currently labels the problem OPEN and says that no complete or partial solution is registered there. Its forum contains finite-model observations and computational claims, but the page explicitly warns that comments are not verified mathematical results. They are therefore context, not proof authority.

GCL's current state

The original issue is MATH-PROGRAMME issue #1622, assignment ERDOS-CATALOGUE-0352-S01. Its pinned catalogue source is teorth/erdosproblems commit 536408716724a2402bdfa7c1d764bf83d64a8c89, data/problems.yaml, entry 352, blob 4e60263a80af6207aca494d2e1388e44fa372889. That snapshot says open, last updated 2025-08-31, with geometry as its topic and statement-formalized metadata marked yes.

The protected formal source is FormalConjectures/ErdosProblems/352.lean at commit 85f863718beeec7b58a3a1926ee92e3472bc2020, blob 32df5e5342d9e3b4ed300bf904f509554c22865a. It captures the measurable-set hypotheses, the existential positive constant, and the unit-area triangle, but its theorem body is by sorry. It is a statement interface, not a proof or GCL certification. The issue has no existing comment and no existing RESULT/1.

What remains

The smallest useful next step is a source-locked semantic audit of the restricted convex-union theorem. It should transcribe the exact strictness of each measure hypothesis, the meaning of “contains,” the convexity and compactness assumptions, and whether each conclusion gives area exactly \(1\) or merely greater than \(1\). A bounded accompanying lemma can verify that, inside a convex set, a triangle of area greater than \(1\) can be continuously reduced along a segment to one of area exactly \(1\). Acceptance would require a line-by-line match to the cited pages and a proof of that interpolation lemma; a mismatch in endpoint, set class, or area conclusion would falsify the package. The task should stop after the source comparison and lemma check, without presenting the restricted result as a solution of the full problem.

For the unrestricted question, what remains is a genuinely global argument controlling arbitrary measurable sets, or a counterexample showing that every proposed finite threshold fails. Finite lattice searches can test models and produce intuition, but they cannot decide this continuum statement. No source inspected here supplies a complete resolution.

Sources and claim boundaries

The primary historical source is Erdős, Set-theoretic, measure-theoretic, combinatorial, and number-theoretic problems concerning point sets in Euclidean space, pp. 122–123; the retrieved PDF has SHA-256 dc61e56e53e4ed0ff5d86487a1459b1baaa139e9e7b50804744a482e8fa932. The primary expository source used for the restricted results is Mauldin, Some Problems and Ideas of Erdős in Analysis and Geometry, Problem 5.1–5.2; the retrieved PDF has SHA-256 12dde62f1ee9937612f9bdebffd93260295cba3970d4d5f69ffa9dcbd78fdce2.

The live status and discussion are Erdős Problems #352 and its discussion thread. The original GCL SOURCE RESULT/1 citation is authenticated original RESULT/1. This article reports source evidence and a finite check; it supplies no mathematical certification, publication decision, novelty claim, or proof of the parent problem. Accepted compendium coverage is decided by the V2 ledger and its protected evidence, not by this article alone.

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1622 — 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