Functoriality and decompositions of the universal coefficient space
OpenQAlgorithms.Papers.Independent.Intakenext100.P67.universalForm_functor_and_decompositionsFor a finite-dimensional real Lie algebra , write for its universal coefficient space. Every Lie homomorphism induces the map determined by
- If and is perfect, then , with the corresponding component formula on .
- If the quotient has a Lie section , then , under . The action is induced by . The map is injective and . For coordinates , the bracket is .
- A reductive decomposition into the center and simple ideals induces .
- A Levi decomposition induces an injection .
These statements explain how the universal coefficient space respects the given algebraic decompositions. Lean specifies each reductive decomposition by its actual summands, inclusions and Lie isomorphism; a Levi complement is a semisimple complement to the solvable radical.
import Definitions.Def_QAlgorithms_Papers_Independent_Intakenext100_P67
namespace QAlgorithms.Papers.Independent.Intakenext100.P67
open scoped TensorProduct
universe u v w
attribute [local instance] QAlgorithms.Papers.Independent.Intakenext100.P67.dependentLieRing
attribute [local instance] QAlgorithms.Papers.Independent.Intakenext100.P67.dependentLieAlgebra
section Functor
variable {L : Type u} {M : Type v}
variable [LieRing L] [LieAlgebra ℝ L] [LieRing M] [LieAlgebra ℝ M]
/-- Remark B.2, PDF p. 41, all five parts. Perfectness in (b) is the span of
brackets being the whole algebra. In (c), a split quotient is supplied by its
actual quotient map and Lie section; the displayed bracket identifies the
result with the semidirect product by the section's adjoint action. In (d),
the chosen reductive center/simple-ideal decomposition is supplied as an
actual Lie isomorphism, and the conclusion specifies the corresponding map
on κu. In (e), a Levi complement means a semisimple complement of the solvable
radical, not an arbitrary subalgebra. -/
theorem universalForm_functor_and_decompositions
(L : Type u) (M : Type v) (N : Type w)
[LieRing L] [LieAlgebra ℝ L] [FiniteDimensional ℝ L]
[LieRing M] [LieAlgebra ℝ M] [FiniteDimensional ℝ M]
[LieRing N] [LieAlgebra ℝ N] [FiniteDimensional ℝ N] :
(universalMap (LieHom.id : L →ₗ⁅ℝ⁆ L) = LinearMap.id) ∧
(∀ (f : L →ₗ⁅ℝ⁆ M) (g : M →ₗ⁅ℝ⁆ N),
universalMap (g.comp f) = (universalMap g).comp (universalMap f)) ∧
(Submodule.span ℝ {z : L | ∃ x y : L, z = ⁅x, y⁆} = ⊤ →
∃ e : UniversalFormSpace (L × M) ≃ₗ[ℝ]
UniversalFormSpace L × UniversalFormSpace M,
∀ (x y : L × M), e (universalForm (L × M) x y) =
(universalForm L x.1 y.1, universalForm M x.2 y.2)) ∧
(∀ (q : L →ₗ⁅ℝ⁆ M) (s : M →ₗ⁅ℝ⁆ L), q.comp s = LieHom.id →
Function.Injective (universalMap s) ∧
(universalMap q).comp (universalMap s) = LinearMap.id ∧
∃ e : L ≃ₗ[ℝ] (q.ker × M),
(∀ x : L, (e x).2 = q x ∧ ((e x).1 : L) = x - s (q x)) ∧
(∀ x y : L, ((e ⁅x, y⁆).1 : L) =
⁅((e x).1 : L), ((e y).1 : L)⁆ +
⁅s ((e x).2), ((e y).1 : L)⁆ -
⁅s ((e y).2), ((e x).1 : L)⁆)) ∧
(∀ (n : ℕ) (J : Fin n → LieIdeal ℝ L),
(∀ i, LieAlgebra.IsSimple ℝ (J i)) →
∀ e : L ≃ₗ⁅ℝ⁆ (LieAlgebra.center ℝ L × ((i : Fin n) → J i)),
(∀ x : L, x = ((e x).1 : L) + ∑ i : Fin n, ((e x).2 i : L)) →
∃ eV : UniversalFormSpace L ≃ₗ[ℝ]
(UniversalFormSpace (LieAlgebra.center ℝ L) ×
((i : Fin n) → UniversalFormSpace (J i))),
∀ x y : L, eV (universalForm L x y) =
(universalForm (LieAlgebra.center ℝ L) (e x).1 (e y).1,
fun i => universalForm (J i) ((e x).2 i) ((e y).2 i))) ∧
(∀ s : LieSubalgebra ℝ L,
LieAlgebra.IsSemisimple ℝ s →
IsCompl (LieAlgebra.radical ℝ L).toSubmodule s.toSubmodule →
Function.Injective (universalMap s.incl)) := by sorry
end Functor
end QAlgorithms.Papers.Independent.Intakenext100.P67
Read-back
What the Lean code literally says, in plain math · gpt-6.1-sol
For any three types , each equipped with a real Lie algebra structure and finite-dimensional over , use the following notation for any real Lie algebra : is the real linear span in of all and , , , and for any real Lie algebra homomorphism . The theorem asserts all six of the following: (1) . (2) For every pair of real Lie algebra homomorphisms and , . (3) If the real linear span of all brackets , with , is all of , then there exists a real linear isomorphism such that, for every and in ,
(4) For every pair of real Lie algebra homomorphisms and satisfying , the map is injective, , and there exists a real linear isomorphism , where , with the following properties: writing and viewing as an element of , one has and for every , and
(5) For every natural number , every family of Lie ideals indexed by , each simple as a real Lie algebra, and every real Lie algebra isomorphism , where is the center, if
after including the center and ideals into , then there exists a real linear isomorphism satisfying
for all ; is included, with an empty sum, a one-element empty product, and a vacuous simplicity hypothesis. (6) For every Lie subalgebra , if is semisimple and its underlying vector subspace is complementary to the underlying vector subspace of the radical of , meaning their intersection is zero and their sum is , then the map induced by inclusion is injective. The zero-dimensional cases are included throughout.
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.