The stalk of a closed subscheme is the quotient of the stalk by the stalk ideal #
For an ideal sheaf Z on a scheme X with closed subscheme ι : V(Z) ⟶ X and a point
z ∈ V(Z) over x = ι z, the stalk map ι^♯_z : 𝒪_{X,x} → 𝒪_{V(Z),z} is surjective (a closed
immersion) with kernel the stalk Z_x = stalkIdeal Z x of the ideal sheaf, so
𝒪_{V(Z),z} ≅ 𝒪_{X,x}/Z_x (stalkQuotientEquiv) and the residue fields agree
(residueFieldStalkEquiv). This is the identification 𝒪_{Z,x} = 𝒪_{X,x}/I_x, Kollár's "the
local ring of Z at x", used without comment throughout the sources.
Proof of the kernel formula (the local form of [TSPA, Tags 01HP, 01QN]): on an affine open U ∋ x
the stalk 𝒪_{X,x} is the localization Γ(X, U)_𝔭 and 𝒪_{V(Z),z} is the localization of
Γ(V(Z), ι⁻¹U) = Γ(X, U)/Z(U) at the corresponding prime 𝔮; the stalk map is the localization
of the surjection Γ(X, U) → Γ(X, U)/Z(U) (Mathlib's ker_subschemeι_app,
subschemeι_app_surjective, comap_primeIdealOf_appLE), and the kernel of a localized map is the
localization of the kernel (IsLocalization.ker_map), which is Z(U)·𝒪_{X,x} = Z_x
(stalkIdeal_eq_map_germ).
Used wherever the local ring of a closed subscheme is computed from that of the ambient scheme
(Hironaka/Scheme/Smooth/AdaptedCoordinates.lean, Hironaka/Snc/, Hironaka/Sequence/,
Hironaka/Resolution/).
A point of the support of Z is the image of a point of the closed subscheme V(Z).
The kernel of the stalk map of the closed immersion V(Z) ⟶ X at z is the stalk of Z
at ι z.
𝒪_{V(Z),z} ≅ 𝒪_{X,ι z} / Z_{ι z}: the stalk of the closed subscheme is the quotient of the
stalk of X by the stalk of the ideal sheaf.
Equations
Instances For
The residue fields of 𝒪_{X,ι z} and 𝒪_{V(Z),z} agree (the stalk map is a surjective local
homomorphism).
Equations
Instances For
A closed point of X in the support of Z is a closed point of V(Z).