Erdős·erdos:501
Erdős Problem 501
No independent check recorded yet. A formal artifact, declaration or published object is attached, but no rebuild of it is recorded here.
A Palomar registration names this problem. It is shown below but not counted as a check of the claim.
Fidelity F2: The correspondence is declared through a Comparator challenge, an alignment table and written divergences.
combinatorics, set theory·Source
AI activity
How grades workCandidate conditional partial result (conditional solution to first part)
Reasoning and sources
Autonomy
AI collaborating with humans
Sources
Does the formal statement say what was claimed?
F2 declared The correspondence is declared through a Comparator challenge, an alignment table and written divergences.
Computed from what the project declares and what the register holds, never from reading the mathematics. How fidelity is graded.
Checks
1-
verified Palomar
Registered by Palomar at 218d1c1e: Comparator confirmed 7 theorems prove the recorded statement within Palomar's axiom policy, replayed through Lean's kernel and the independent nanoda kernel. The project names this problem, which does not establish that it proves the result claimed here, so it is not counted as a check of it
Declared by the projects
1Read from each project's formalization.yaml. A declaration is what the authors say about their own work, recorded so that a check can confirm or contradict it.
Erdős Problem #501 in Lean 4: the closed case and the independence of the first question
- authors
- Elliot Glazer, Sol
- method
- agent — Claude (Anthropic) — Fable 5 (claude-fable-5) and Opus 4.8 (claude-opus-4-8); the final integration and the Mathlib ModelTheory bridge (Erdos501/FOL/, Challenge.lean, Solution.lean) were produced by Fable 5, GPT-5.6 ("Sol")
- review
- agent-reviewed (Claude (Anthropic) audit sessions: statements of the seven targets against erdosproblems.com/501 and formal-conjectures (docs/audits/2026-08-16-audit-formal-conjectures-501-statements.md), the paper (docs/audits/2026-08-16-audit-rev10-profile-certificate.md), the vendored port (third_party/flypitch4/), and the forcing development (docs/audits/2026-08-17-erdos501-forcing-audit-355bc1e.txt), Elliot Glazer (review of the trusted statements and of the mathematics))
- axioms
- Classical.choice, Quot.sound, propext
- sorry
- 0 unproved goals declared
- results
- 7 main results named, checked with Comparator, with an alignment table
- sources
- Some unsolved problems (Problem II.9) — background, authors n/a; Unsolved problems in set theory (Problem 38) — background, authors n/a; Erdős problem #501 (erdosproblems.com) — background, authors not-contacted; Erdős problem #501 — forum discussion (erdosproblems.com) — background, authors participated
- related
- flypitch/flypitch — builds-on; ianklatzco/flypitch — builds-on; google-deepmind/formal-conjectures — other; leanprover-community/mathlib4 — builds-on
- divergences
- (1) Independence is proved for the first-order sentence Erdos501 ("every complete ordered field has the Erdős property"), i.e. for the first question rendered inside ZFC; the target erdos501_sentence_faithful certifies that in Mathlib's ZFSet this sentence is equivalent to the Mathlib statement of the first question, so nothing weaker is being claimed. Inside the sentence, "outer measure < 1" is rendered as the existence of a cover by countably many open intervals of total length < 1 (the definition of Lebesgue outer measure), "bounded" as bounded above and below, "infinite" as "ω injects into X". (2) The theory ZFC is Flypitch's axiomatization (extensionality, empty set, ordered pairs, union, power set, infinity, regularity, Zorn's lemma, strong collection), which is equivalent to the usual ZFC. (3) Independence is stated semantically (Mathlib has no proof calculus): ¬ (ZFC ⊨ᵇ φ) means that some model of ZFC (with carrier in Type 0; by Löwenheim–Skolem this is no restriction) satisfies ¬φ. The underlying Flypitch results are the syntactic ¬ (ZFC ⊢ₛ' Erdos501_f) and ¬ (ZFC ⊢ₛ' ∼Erdos501_f) (comparator-flypitch.json). (4) The consistency of a positive answer is obtained from the ran
- checked by
- Palomar
Follow and discuss
All discussionGet an email when this problem moves
A new claim, a check, a bounty or a discussion. One link to confirm, one click to stop.
Discussion and bounties for this problem load here.
Formal material
Formal statements · 1
Recorded elsewhere
Compare the registries- palomar — elliotglazer/erdos501
- vibemathed — Erdős Problem #501: infinite independent sets for families of small outer measure
Also known as · 3
- https://www.erdosproblems.com/501
- https://github.com/google-deepmind/formal-conjectures/blob/main/FormalConjectures/ErdosProblems/501.lean
- FormalConjectures/ErdosProblems/501.lean