← All papers
First page of No regular icosahedron has a rational point on every face: a proof checked in Lean

No regular icosahedron has a rational point on every face: a proof checked in Lean

Mikhail Levit

math.MG Nov 8, 2026 · v1
Full proof, including 48 kernel-checked exact-arithmetic certificates over Z[φ], is formalized in Lean 4 with Mathlib; formalization developed by AI under author direction.
We prove that no regular icosahedron in $\mathbb{R}^3$ has a point with rational coordinates on each of its twenty closed faces, in every position, size and orientation, reflections included. This answers a question of I. Bogdanov on MathOverflow. A field automorphism of $\mathbb{C}$ that sends $\sqrt5$ to $-\sqrt5$ reduces the problem to the following statement: for every complex orthogonal $3\times3$ matrix $T$ some real linear combination of $\mathrm{Re}\,T$ and $\mathrm{Im}\,T$ separates ten intervals determined by the faces. We prove this statement by a computer-assisted case analysis over rational parametrizations of rotation frames, in which every inequality is an exact polynomial bound over $\mathbb{Z}[\varphi]$, $\varphi=(1+\sqrt5)/2$; no floating-point arithmetic and no Lipschitz estimates are used. The whole argument, including the $48$ certificates, is formalized in the Lean 4 proof assistant with Mathlib and checked by the Lean kernel; the main theorem depends only on the standard axioms. The mathematical argument and the formalization were produced by AI systems (OpenAI Codex and Anthropic Claude) under the direction of the author; see the statement at the end of the paper.

Bogdanov asked on MathOverflow whether a regular icosahedron in R^3 can have a rational point on each of its twenty closed faces. The other Platonic solids admit such configurations, and Rouse showed the analogous question for face planes has a positive answer.

A field automorphism of C sending √5 to −√5 reduces the problem to showing that, for every complex orthogonal 3×3 matrix T, some real combination of Re T and Im T separates ten intervals determined by the faces. This is proved by a computer-assisted case analysis over rational quaternion-based chart parametrizations of rotation frames. Every inequality is an exact polynomial bound over Z[φ], with no floating-point arithmetic. The certificates are checked by `decide +kernel` using natural-number arithmetic. The mathematical argument was developed by OpenAI Codex and the Lean formalization by Anthropic Claude.

No regular icosahedron, in any position, size or orientation (reflections included), has a rational point on every closed face. The Lean 4/Mathlib proof depends only on propext, Classical.choice and Quot.sound, and the trusted base is the kernel plus 58 lines of definitions.