Kernels of commuting endomorphisms #
A linear map f intertwining endomorphisms d and e, that is with f ∘ d = e ∘ f, sends the
kernel of d into the kernel of e (LinearMap.map_mem_ker_of_comp_eq); this is how a chain map
acts on cycles.
Let r₁ and r₂ be endomorphisms of a module M with ker r₁ ⊓ ker r₂ = ⊥, let l₁ and l₂
be endomorphisms commuting with both, and let Y ≤ ker r₁ and Z ≤ ker r₂ be submodules. If
ξ ∈ Y ⊔ Z satisfies r₁ ξ ∈ ker l₂ and r₂ ξ ∈ ker l₁, then ξ ∈ (Y ⊓ ker l₁) ⊔ (Z ⊓ ker l₂).
This is the exactness step in the proof of §3 Theorem 2 of Popa and Zagier, with r₁, r₂ the
right multiplications by π_S, π_U on their ℛ (whose kernels meet trivially by the
right-action form of their Lemma 2), l₁, l₂ the left multiplications by 1 - π_S, 1 - π_U,
Y = ℛ (1 - π_S) and Z = ℛ (1 - π_U): an element of Y + Z in their ℬ (6) is in their 𝒥 (7).
Main results #
LinearMap.map_mem_ker_of_comp_eq: an intertwining map sends kernel to kernel.TauCeti.End.mem_inf_ker_sup_inf_ker_of_mem_sup: the splitting statement above.
References #
- A. Popa and D. Zagier, An elementary proof of the Eichler–Selberg trace formula, J. Reine Angew. Math. 762 (2020), 105–122, arXiv:1711.00327, Section 3, proof of Theorem 2.
A linear map f with f ∘ d = e ∘ f sends the kernel of d into the kernel of e.
Splitting along disjoint kernels of commuting endomorphisms: let r₁ and r₂ have
disjoint kernels and let l₁, l₂ commute with both. If ξ ∈ Y ⊔ Z for submodules Y ≤ ker r₁
and Z ≤ ker r₂, and r₁ ξ ∈ ker l₂, r₂ ξ ∈ ker l₁, then ξ ∈ Y ⊓ ker l₁ ⊔ Z ⊓ ker l₂.