Documentation

TauCeti.CategoryTheory.Monoidal.Tor

Tor from a projective presentation of length one #

Mathlib's CategoryTheory.Tor C n derives the tensor product in its second variable: ((Tor C n).obj X).obj M is Lₙ(X ⊗ -)(M). Let S : 0 ⟶ P₁ ⟶ P₀ ⟶ M ⟶ 0 be a short exact sequence with P₁ and P₀ projective. Then

Over a principal ideal domain the boundaries and cycles of a complex of free modules are free, so the homology Hⱼ has the presentation 0 ⟶ Bⱼ ⟶ Zⱼ ⟶ Hⱼ ⟶ 0. This identifies the Tor₁ terms of the Künneth short exact sequence 0 ⟶ ⨁ Hᵢ ⊗ Hⱼ ⟶ Hₙ(K ⊗ L) ⟶ ⨁ Tor₁(Hᵢ, Hⱼ) ⟶ 0 with the kernels of Hᵢ(K) ⊗ Bⱼ(L) ⟶ Hᵢ(K) ⊗ Zⱼ(L).

These are the specializations to X ⊗ - of the computation of left derived functors in TauCeti.CategoryTheory.Abelian.LeftDerived.

If e • 𝟙 X = 0 for a natural number e, then e kills every Torₙ(X, Y) (CategoryTheory.nsmul_id_Tor_eq_zero): the objects killed by e (CategoryTheory.ObjectProperty.isKilledBy) are closed under subobjects and quotients and contain every X ⊗ P.

For Mathlib's CategoryTheory.Tor' C n, which derives the first variable instead, ((Tor' C n).obj X).obj Y is Lₙ(- ⊗ Y)(X). The file records two facts about it in the underived variable Y: it is an additive functor of Y, and Tor'ₙ₊₁(X, 𝟙_ C) vanishes (CategoryTheory.isZero_Tor'_succ_tensorUnit), since - ⊗ 𝟙_ C is isomorphic to the identity functor and hence exact.

References #

For a projective presentation 0 ⟶ P₁ ⟶ P₀ ⟶ M ⟶ 0, Tor₁(X, M) is the kernel of X ◁ S.f : X ⊗ P₁ ⟶ X ⊗ P₀.

Instances For

    Torₙ(X, M) vanishes for n ≥ 2 when M has a projective presentation of length one.

    If a natural number e kills X, in the sense that e • 𝟙 X = 0, then it kills every Torₙ(X, Y).

    Tor' is additive in its second variable, the one that is not derived.

    The higher Tor' groups with the tensor unit as second variable vanish: - ⊗ 𝟙_ C is isomorphic to the identity functor, so it is exact.