Lemma 2.5: strict radical layers #
Lemma 2.5 (lem:strict-radical-layers, graded-category.tex). Let
J be the radical of the category of finite-length modules and let
a : U → Y be a radical monomorphism such that a + h is monic for every
h ∈ J²(U, Y). Then for every W and n ≥ 1,
a ∘ z ∈ J^n(W, Y) ⇔ z ∈ J^{n-1}(W, U). Every direct sum of left almost split
inclusions satisfies the hypothesis.
This file proves the lemma in an additive category with kernels and binary
biproducts, for the categorical radical ideal I (the vendored homIdeal) and
its powers, under two hypotheses which hold for mod A:
- every object
Yhas a right radical approximation: a radical mapp : E → Ythrough which every radical map intoYfactors, with a radical monic kerneli : T → Esuch that every map inI²(T, Z)isi ≫ qwithqradical (in the manuscript,pis the sum of the minimal right almost split maps of the summands ofY, withrad P → Pat projective summands, andiis a sum of left almost split inclusions); - the Krull–Schmidt normal form: every map becomes, after an isomorphism of its
source, a row of a split monomorphism and a radical map (the manuscript's
"split off the invertible blocks of
u").
The proof is the manuscript's: the identity
p ∘ J^n(W, E) = J^{n+1}(W, Y) (eq:right-radical-surjection), splitting off
the given monomorphism, the complementary monomorphism b with the same
perturbation property, and induction on the radical degree for all such
monomorphisms simultaneously. In Lean composition is written z ≫ a for the
manuscript's a ∘ z, and the exponents are shifted: z ≫ a ∈ I^{n+1} implies
z ∈ I^n.
The n-th power of the categorical radical.
Instances For
A radical map followed by a map of I^n lies in I^{n+1}.
A map of I^n followed by a radical map lies in I^{n+1}.
1 + q is invertible for a radical endomorphism q.
A right radical approximation of Y (see the module docstring).
- E : C
- p : self.E ⟶ Y
- T : C
- p_rad : MBDVendor.QSE.CategoricalRadical.IsRadicalMorphism self.p
- factors {W : C} (g : W ⟶ Y) : MBDVendor.QSE.CategoricalRadical.IsRadicalMorphism g → ∃ (g' : W ⟶ self.E), CategoryTheory.CategoryStruct.comp g' self.p = g
- i_mono : CategoryTheory.Mono self.i
- i_rad : MBDVendor.QSE.CategoricalRadical.IsRadicalMorphism self.i
Instances For
The Krull–Schmidt split–radical normal form.
Instances For
eq:right-radical-surjection: every map of I^{n+1}(W, Y) is w ≫ p with
w ∈ I^n(W, E).
Splitting off: if u ≫ p = a with a having the perturbation property and
p radical, then u is a split monomorphism.
Lemma 2.5 (lem:strict-radical-layers), abstract form. Let a : U ⟶ Y
be radical with a + h monic for every h ∈ I²(U, Y). Then
z ≫ a ∈ I^{n+1}(W, Y) implies z ∈ I^n(W, U).
The reverse implication of Lemma 2.5: z ∈ I^n implies z ≫ a ∈ I^{n+1}
for radical a.
A left almost-split-type map i (every map of I² out of its source factors
as i ≫ q with q radical) satisfies the perturbation hypothesis of Lemma 2.5:
i + h = i ≫ (1 + q) is monic for every h ∈ I².
The kernel map of a right radical approximation has the perturbation property of Lemma 2.5.
Injectivity on layers (Prop 2.6(1), first map): z ≫ i ∈ I^{n+2} implies
z ∈ I^{n+1}.
Exactness at the middle term on layers (Prop 2.6(1)): if w ∈ I^{n+1}
and w ≫ p ∈ I^{n+3}, then w ≡ z ≫ i modulo I^{n+2} for some z ∈ I^n.
The lowest degree (Prop 2.6(1)): w ≫ p ∈ I² implies w ∈ I.