Herzog-Schönheim conjecture for subnormal uniform covers
ProvedSubnormalCover.herzog_schonheim_subnormalThe Herzog--Sch"onheim conjecture for subnormal uniform covers. In 1974 Herzog and Sch"onheim conjectured that if the left cosets () partition a group then the finite indices cannot be pairwise distinct. Sun proves the conjecture, in the stronger setting of uniform covers (each covered exactly times, arbitrary), whenever the are subnormal of finite index and not all equal to : two of the indices must coincide.
Formalization note. The conclusion asserts distinct indices into the family, in , so it says two distinct parts of the cover have equal index; the subgroups and themselves may well be equal, which is the usual reading of the conjecture. Nontriviality is ∃ i, K i ≠ ⊤, which together with the uniform-cover hypothesis already forces , so no separate assumption appears.
import Mathlib import Definitions.Def_SubnormalCover_uniformCover
namespace SubnormalCover
theorem herzog_schonheim_subnormal {G : Type*} [Group G] {k : ℕ}
(K : Fin k → Subgroup G) (a : Fin k → G) (w : ℕ) (hcov : IsUniformCover K a w)
(hfin : ∀ i, (K i).FiniteIndex) (hsn : ∀ i, (K i).IsSubnormal) (hnt : ∃ i, K i ≠ ⊤) :
∃ i j : Fin k, i ≠ j ∧ (K i).index = (K j).index := by
sorry
end SubnormalCoverRead-back
What the Lean code literally says, in plain math · claude-opus-5
Setting. Let be an arbitrary type equipped with a group structure, and let be an implicit natural number. Assume given a family of subgroups indexed by (no distinctness assumption: the same subgroup may occur at several indices), a family of elements , a natural number , and the four hypotheses:
- uniform cover of weight : for every , the number of indices such that — that is, such that belongs to the left coset — is exactly . This is an exact count for every single element of ; disjointness of the cosets is not assumed, and is an arbitrary given natural number, not required to be .
- each has finite index: , where denotes the cardinality of the left-coset space , with Mathlib's convention that this value is precisely when the coset space is infinite. So each is a positive integer.
- each is subnormal in : it can be joined to by a finite chain with each subgroup normal in its successor (Mathlib's inductive
IsSubnormal, generated by " is subnormal" and "if , subnormal and normal in , then subnormal"). - nontriviality: there exists at least one index with . This implicitly forces ; together with it also forces , since lies in .
Conclusion. There exist indices with and
That is, two distinct positions in the indexing carry subgroups of equal index; nothing asserts that as subgroups (the two positions may hold the very same subgroup), nothing asserts that this common index is or that are among the indices witnessing , and no uniqueness is claimed. Since the conclusion requires two distinct elements of , it is only satisfiable when ; the case is already excluded by , and the case would make the conclusion false, so the hypotheses are being asserted to preclude it. The parameters and do not appear in the conclusion.
Confirmed by the mission captain (proposal self-audit).