Documentation

TauCeti.NumberTheory.HeckeRing.GL2.Gamma1.WithCenter

The Hecke triple of Γ₁(N)·{±I} #

Gamma1/Basic.lean puts Γ₁(N) into a Hecke triple Γ₁(N) ≤ Δ₀(N) ≤ commensurator(Γ₁(N)). This file does the same for Γ₁(N)·{±I}, the subgroup (Gamma1 N).withCenter obtained by adjoining the centre of SL₂(ℤ).

Why the enlarged group is the one some statements need #

Γ₁(N) does not contain -I once N ≥ 3 — -I has diagonal (-1, -1) while Γ₁(N) asks for ≡ (1, 1), so CongruenceSubgroup.neg_one_mem_Gamma1_iff holds exactly when N ∣ 2. Any statement whose hypothesis is "this subgroup contains -I" is therefore simply false at Γ₁(N) for almost every level, and a fundamental-domain or Petersson argument that needs it has to run over a group that does contain it.

Γ.withCenter = Γ ⊔ Z(G) contains -I for every Γ, and it is the group TauCeti's Petersson layer already works with: CuspForm.peterssonInnerCosets sums over SL(2, ℤ) ⧸ Γ.withCenter. That this enlarged group is itself a Hecke triple with the same Δ₀(N) is what lets a Hecke coset be formed over it at all.

Nothing here is deep — the point is that the passage from Γ₁(N) to Γ₁(N)·{±I} costs nothing on either side of the triple. Containment in Δ₀(N) survives because Γ₀(N) already contains -I, so it absorbs the central factor and Γ₁(N)·{±I} ≤ Γ₀(N) still holds; commensurability survives because the enlarged group still has finite index in SL₂(ℤ), containing Γ₁(N).

This file declares only the instance. Neither half of the triple mentions Γ₁(N): the Δ₀(N) containment needs only H ≤ Γ₀(N) and is map_withCenter_le_Delta0 in Gamma0/Basic.lean, while the commensurator half needs only finite index — nothing about withCenter at all — and is Delta0_le_commensurator_map in Delta0.lean.

Main results #

References #

The Hecke triple of Γ₁(N)·{±I}: Γ₁(N)·{±I} ≤ Δ₀(N) ≤ commensurator(Γ₁(N)·{±I}) inside GL₂(ℚ), with the same monoid Δ₀(N) as the triple of Γ₁(N) itself.