Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Index of an intersection of subnormal subgroups

Proved
SubnormalCover.index_iInf_dvd_prod

by burkh4rt · Sep 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

combinatoricsgroup-theory

Let GGG be a group and let K1,…,KkK_1,\dots,K_kK1​,…,Kk​ be subnormal subgroups of GGG, each of finite index. Then the index of their intersection divides the product of the individual indices: [G:⋂i=1kKi]  ∣  ∏i=1k[G:Ki]\bigl[G:\bigcap_{i=1}^{k}K_i\bigr] \;\Bigm|\; \prod_{i=1}^{k}[G:K_i][G:⋂i=1k​Ki​]​∏i=1k​[G:Ki​], and consequently P([G:⋂iKi])=⋃iP([G:Ki])P\bigl([G:\bigcap_i K_i]\bigr)=\bigcup_i P([G:K_i])P([G:⋂i​Ki​])=⋃i​P([G:Ki​]) for the sets of prime divisors. Subnormality is essential: for arbitrary finite-index subgroups Poincaré's theorem gives only the inequality [G:⋂iKi]≤∏i[G:Ki][G:\bigcap_i K_i]\le\prod_i[G:K_i][G:⋂i​Ki​]≤∏i​[G:Ki​], and divisibility fails already for the two subgroups ⟨(12)⟩,⟨(13)⟩\langle(12)\rangle,\langle(13)\rangle⟨(12)⟩,⟨(13)⟩ of S3S_3S3​, where the intersection is trivial of index 666 while 3⋅3=93\cdot 3=93⋅3=9. This arithmetic property of subnormality is the reason covers by subnormal subgroups are far better behaved than general coset covers. Formalization note: the family is indexed by Fin k, the intersection is the lattice infimum ⨅ i, K i, and "finite index" is the class Subgroup.FiniteIndex supplied as a plain hypothesis ∀ i, (K i).FiniteIndex (equivalently (K i).index ≠ 0), so a prover must haveI it before instance search can use it.

Preamble
import Mathlib
Formal statement
namespace SubnormalCover

/-- **Sun, Lemma 2.1 / equation (2.1).**

If `K 1, …, K k` are *subnormal* subgroups of finite index of a group `G`, then the index of
their intersection divides the product of their indices:
`[G : ⋂ᵢ Kᵢ] ∣ ∏ᵢ [G : Kᵢ]`.

For arbitrary (non-subnormal) finite-index subgroups, Poincaré's theorem only gives the
inequality `[G : ⋂ᵢ Kᵢ] ≤ ∏ᵢ [G : Kᵢ]` (`Subgroup.index_iInf_le`); divisibility genuinely
requires subnormality. -/
theorem index_iInf_dvd_prod {G : Type*} [Group G] {k : ℕ} (K : Fin k → Subgroup G)
    (hsn : ∀ i, (K i).IsSubnormal) (hfin : ∀ i, (K i).FiniteIndex) :
    (⨅ i, K i).index ∣ ∏ i, (K i).index := by
  sorry

end SubnormalCover
Source
Z.-W. Sun, On the Herzog-Schönheim conjecture for uniform covers of groups, J. Algebra 273 (2004) 153-175, https://doi.org/10.1016/S0021-8693(03)00526-X, Lemma 2.1, eq. (2.1) (restated from Lemma 3.1(ii) of [Su8])
Read-back

What the Lean code literally says, in plain math · claude-opus-5

Setting. Let GGG be a type in an arbitrary universe carrying a group structure (no finiteness, commutativity, or other assumption on GGG), let kkk be a natural number (kkk is implicit, so it is inferred, and k=0k = 0k=0 is allowed), and let K:{0,1,…,k−1}→Sub(G)K : \{0,1,\dots,k-1\} \to \mathrm{Sub}(G)K:{0,1,…,k−1}→Sub(G) be an arbitrary family of kkk subgroups of GGG indexed by the finite type Fin k\mathrm{Fin}\,kFink. Nothing requires the KiK_iKi​ to be distinct, proper, or nontrivial; repetitions and Ki=⊤K_i = \topKi​=⊤ or Ki=⊥K_i = \botKi​=⊥ are permitted.

Hypotheses. Two hypotheses are assumed, both ordinary (non-instance) hypotheses quantified over all indices iii:

  • KiK_iKi​ is subnormal in GGG for every iii. Mathlib's Subgroup.IsSubnormal is the inductive predicate generated by: ⊤\top⊤ is subnormal; and if H≤KH \le KH≤K, KKK is subnormal, and the image of HHH inside KKK is a normal subgroup of KKK, then HHH is subnormal. Equivalently, there is a finite chain Ki=H0⊴H1⊴⋯⊴Hn=GK_i = H_0 \trianglelefteq H_1 \trianglelefteq \cdots \trianglelefteq H_n = GKi​=H0​⊴H1​⊴⋯⊴Hn​=G with each term normal in the next (the chain length may depend on iii and is not bounded).
  • KiK_iKi​ has finite index for every iii: Mathlib's Subgroup.FiniteIndex is the one-field class asserting [G:Ki]≠0[G : K_i] \neq 0[G:Ki​]=0. Here [G:Ki][G : K_i][G:Ki​] is Subgroup.index, defined as Nat.card (G/Ki)\mathrm{Nat.card}\,(G / K_i)Nat.card(G/Ki​), the number of left cosets, with the convention that it is 000 when the coset space is infinite. So the hypothesis says exactly that each coset space G/KiG/K_iG/Ki​ is finite.

Conclusion. In the natural numbers,

[ G:⋂iKi ]  ∣  ∏i[G:Ki],\Big[\,G : \bigcap_{i} K_i\,\Big] \;\Big|\; \prod_{i} [G : K_i],[G:i⋂​Ki​]​i∏​[G:Ki​],

where ⋂iKi\bigcap_i K_i⋂i​Ki​ is the lattice infimum ⋀i∈Fin kKi\bigwedge_{i \in \mathrm{Fin}\,k} K_i⋀i∈Fink​Ki​ of the family (the subgroup whose underlying set is the intersection of the KiK_iKi​), its index again means Nat.card\mathrm{Nat.card}Nat.card of its coset space with value 000 if that space is infinite, and the product is the finite product over all i∈Fin ki \in \mathrm{Fin}\,ki∈Fink. The divisibility is stated in this direction only: the index of the intersection divides the product of the indices; no inequality, no reverse divisibility, and no statement about the quotient being an integer of any particular form.

Degenerate cases. If k=0k = 0k=0 the index type is empty, both hypotheses are vacuously true, the empty infimum is ⊤\top⊤ so the left side is [G:⊤]=1[G : \top] = 1[G:⊤]=1, the empty product is 111, and the claim reduces to 1∣11 \mid 11∣1. Because each [G:Ki]≠0[G : K_i] \neq 0[G:Ki​]=0, the right-hand product is nonzero; since 0∣n0 \mid n0∣n holds only for n=0n = 0n=0, the assertion implicitly entails that ⋂iKi\bigcap_i K_i⋂i​Ki​ also has finite index (its index cannot be the junk value 000). The statement uses only standard Mathlib notions; no definition from the accompanying platform definition files occurs in it.

Human review
  • Endorsed by Shuze Chen · Sep 14, 2026

  • Endorsed by burkh4rt · Sep 14, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me