Feit–Thompson theorem: every finite group of odd order is solvable
ProvedCayleyErdosGyarfas.feit_thompsonLet be a finite group of odd order. Then is solvable: its derived series , defined by , reaches the trivial subgroup after finitely many steps:
This is the odd order theorem of Feit and Thompson (1963). The proof of Theorem 1.1 of the source paper (§2, p. 4) invokes it to conclude that a minimal counterexample to the claim made there (every Cayley graph of a noncyclic group of odd order contains a symmetric figure-eight) is solvable, so that its minimal normal subgroups are elementary abelian. It is a supporting statement of this mission, cited rather than proved by the paper, and Mathlib does not contain it.
Formalization Note Solvability is Mathlib's Group.IsSolvable G: some term derivedSeries G n of the derived series equals ⊥. The type is that of OddOrder.feitThompson in Yawara Ishida's Lean 4 development (https://github.com/yawara/odd-order), a port of which is vendored and proved in Palomar entry PALOMAR-2026-11-04-000015.
import Mathlib.GroupTheory.Solvable import Mathlib.SetTheory.Cardinal.Finite
namespace CayleyErdosGyarfas
/-- **Feit–Thompson theorem** (Feit–Thompson 1963): every finite group of odd order is
solvable. Supporting statement: the proof of Theorem 1.1 of arXiv:2611.05060v1 invokes it
(§2, p. 4). The type is that of `OddOrder.feitThompson` in Yawara Ishida's Lean 4
development. -/
theorem feit_thompson {G : Type*} [Group G] [Finite G] (hodd : Odd (Nat.card G)) :
Group.IsSolvable G := by sorry
end CayleyErdosGyarfasRead-back
What the Lean code literally says, in plain math · claude-opus-5-5
For every group , satisfying the two hypotheses below, the statement asserts that is solvable. Here is an arbitrary type, in any universe of types, with a group structure written multiplicatively with identity . The hypotheses are:
- is finite, i.e. there is a bijection between and for some ;
- the order is odd, i.e. for some .
Solvable has this exact meaning: there exists , with allowed, such that
Here is the derived series. It is defined by recursion and starts at index with the whole group:
In this formula is the subgroup of generated by the set , and the commutator of two elements is . So is the commutator subgroup of . The series is decreasing (), and once a term equals , every later term also equals . "Some term is trivial" therefore means the same as "the series is eventually trivial". The statement gives no bound on (the derived length) and asserts nothing beyond the existence of such an .
The order is the library's natural-number cardinality of the underlying set of . It is the number of elements when is finite, and by convention when is infinite. Here is assumed finite and is nonempty because it contains . So is the actual number of elements, , and the junk value cannot occur. The oddness hypothesis alone also rules out the value , because is even. Oddness therefore already forces to be finite, which means the finiteness hypothesis follows from the oddness hypothesis.
Edge cases: the trivial group satisfies both hypotheses (). For it the conclusion holds with , because . The hypotheses can be satisfied, for example by every cyclic group of odd order, so the statement is not vacuous. Every notion used is the standard library definition, and no custom definitions are involved. These notions are group, finiteness, cardinality, oddness of a natural number, subgroup generated by a set, commutator, derived series and solvability.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.