Erdős Problem #325
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (not automatically the live result).
Subject areas: number theory, powers.
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 \(k\geq 3\), let \(f_{k,3}(x)\) count the integers \(n\leq x\) that can be written as a sum of three nonnegative kth powers:
where \(a,b,c\) are nonnegative integers. Erdős Problem #325 asks whether, for every such fixed \(k\), the number of represented integers grows at least on the natural scale \(x^{3/k}\). In standard asymptotic notation, the first proposed form is
The notation means that a positive constant may depend on \(k\), but not on sufficiently large \(x\). The problem also asks the weaker alternative \(f_{k,3}(x)\gg_\epsilon x^{3/k-\epsilon}\) for every \(\epsilon>0\). These are asymptotic statements: checking many small values of \(x\) cannot settle them, and the quantifier “for every \(k\geq 3\)” is essential.
The exponent \(3/k\) is the heuristic scale suggested by choosing three variables of size about \(x^{1/k}\). Collisions are unavoidable, however: different triples can give the same sum. The question is whether enough distinct integers survive those collisions for a power-sized lower bound.
A small example
Take \(k=3\) and \(x=100\). The cubes of nonnegative integers that do not exceed \(100\) are
An exact finite calculation forms every ordered triple from this five-element list, keeps sums at most \(100\), and removes duplicates. It gives
For a check on the arithmetic, \(91\) occurs because
This is a reproducible exact-arithmetic finite check: the only operations are integer powers, integer addition, comparison with \(100\), and set deduplication. It illustrates what \(f_{3,3}(100)\) means under the nonnegative convention used by the problem statement and the pinned Lean definition. It does not establish a lower bound for all large \(x\), does not test every \(k\), and does not distinguish the two asymptotic claims.
What is known
The maintained Erdős Problems entry for #325 gives the question in the same form and records earlier binary-power work of Erdős and Mahler. Its bibliography identifies Erdős and Mahler, “On the number of integers which can be represented by a binary form,” Documenta Mathematica (2019), 475–481, as [ErMa38], and Trevor D. Wooley, “Sums of three cubes, II,” Acta Arithmetica 170 (2015), 73–100, as [Wo15].
Wooley’s primary paper is the relevant unconditional result for \(k=3\). Theorem 1.1 on printed p. 73 states that if \(N(X)\) counts integers not exceeding \(X\) represented as sums of three cubes of natural numbers, then
The paper explains that this follows from a sixth-moment estimate for smooth cubic Weyl sums, stated as Theorem 1.2 and developed through the later sections. It is a lower bound for the cubic case, not a proof of the all-k assertion in #325. The wording “natural numbers” also deserves a convention check against the problem’s explicit “nonnegative” variables; the nonnegative formulation is the one used in the Lean predicate, and this audit does not silently identify conventions that the source does not define.
The live entry says the problem has been formalised, but neither it nor Wooley’s paper supplies a solution for every \(k\geq 3\). The available primary evidence therefore supports a strong solved special case while leaving the general catalogue question open.
GCL's current state
The frozen catalogue packet is teorth/erdosproblems data/problems.yaml, at commit 536408716724a2402bdfa7c1d764bf83d64a8c89; the entry is Git blob 4e60263a80af6207aca494d2e1388e44fa372889, with source SHA-256 be65b0db56fe13becec3178f2937685ee7e6a2f5df9325712474722c279c4182. At that snapshot the informal status is open, last updated 2025-08-31; the reported formal-solution status is unformalized, while statement-formalized metadata is yes. The programme registry classifies it PROTECTED_READY_BACKLOG.
The pinned formal source is FormalConjectures/ErdosProblems/325.lean, at commit 85f863718beeec7b58a3a1926ee92e3472bc2020; its Git blob is 00f87d674b69be7fde60d49976ff1c49c33c1999, and the checked raw-file SHA-256 is 46e4eefbacce27428007ed817d9f3718d22b5468808d6df6780dd87fd97314f2. It defines the counting predicate over natural numbers and formalises the main statement, a weaker epsilon-variant, and Wooley’s cubic bound. Each theorem body ends in by sorry. Thus the file is a formal statement interface, not a Lean proof or GCL mathematical certification.
The original GCL issue #1595 had the required source-audit packet and zero comments at readback, including no existing RESULT/1. Its source return is therefore not represented here as an accepted result. V2 ledger state alone decides accepted editorial coverage.
What remains
The smallest useful successor is a source-semantic bridge: compare the exact quantifiers in the catalogue wording with the formal theorem, settle whether “integers” includes \(0\), and record the convention for “natural numbers” in Wooley’s \(N(X)\). The bridge should then isolate the fixed k = 3 theorem from the unresolved all-k claim and from the weaker epsilon-variant.
A bounded formalisation task could next prove the finite counting lemma for a chosen \(k\) and \(x\) in the exact Lean definitions, then state the cubic Wooley interface without filling its proof. Such work would improve reproducibility and expose coercion or quantifier mismatches, but it would not replay Wooley’s analytic argument. Any broader claim that the original all-k problem is solved would require a primary theorem covering those quantifiers, or a separately verified reduction to one.
Sources and claim boundaries
- Erdős Problems #325, maintained statement and bibliography; live page read for this audit.
- Frozen catalogue commit and file, commit and blob recorded above.
- Wooley, “Sums of three cubes, II,” publisher PDF, Acta Arithmetica 170 (2015), 73–100, Theorem 1.1 p. 73; the read primary PDF has SHA-256
4b34807d2b80cbc9273b1762964dcbb7000a63be2ea36f2a9ffa6cb56d5df32e. - Wooley’s arXiv record and HTML, arXiv:1502.01944, submitted 2015-02-06; it exposes the same theorem and definitions.
- Pinned FormalConjectures file, a statement source whose theorem proofs are
sorry. - Original issue #1595; original SOURCE RESULT/1 citation: authenticated original RESULT/1.
These sources support the stated question, the dated open snapshot, the exact finite example, and Wooley’s cubic theorem. They do not certify the general conjecture, establish semantic equivalence across every convention, replay a Lean proof, or supply GCL mathematical certification. Accepted coverage remains a separate V2-ledger decision.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1595 — 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.