Erdős Problem #638 — Let be a family of finite graphs such that for every there is some such that if the edges of are coloured with colours then there is a monochromatic triangle.
Let be a family of finite graphs such that for every there is some such that if the edges of are coloured with colours then there is a monochromatic triangle. Is it true that for every infinite cardinal there is a graph of which every finite subgraph is in and if the edges of are coloured with many colours then there is a monochromatic triangle.
References
Primary source
Additional references
UnsolvedMath, Erdős Problems set, ULAM AI, licensed CC BY 4.0.
Progress summary
The original wording is false, while the intended stronger hereditary version has a reported construction but no complete public proof.
Problem asks whether finite triangle-Ramsey richness of a family forces, for every infinite cardinal , a graph whose finite subgraphs lie in and which is triangle-Ramsey for edge-colours. The page records no proposer or date.
Known results
- Closure under arbitrary subgraphs is insufficient: a sparse family of complete graphs gives a counterexample.
- The intended formulation appears to require closure under finite subgraphs; substantial related literature is cited, but no named theorem is supplied in the retrieved material.
- A Lean project formalizes only the compactness and diagonal reduction; its finite-avoidance and block-sequence lemmas remain unproved.
Reported construction
A newer discussion says that a construction addresses the hereditary finite-subgraph interpretation, while treating the induced-subgraph analogue separately. No author, date, construction details, or complete proof are provided, so this is unverified progress rather than a settled solution.
Current status (as of March 2026): the original formulation has a counterexample; the intended hereditary version has only an undocumented reported construction and remains unverified.
Solutions 0
No solutions have been posted yet.