LeanProofs.LieStructure.KillingFormMatrix
DefinitionGrade4_LeanProofs_LieStructure_KillingFormMatrixclifford-algebraleanproofs-grade4rosetta
Foundational definitions and structural lemmas from LeanProofs.LieStructure.KillingFormMatrix, required by the selected LeanProofs theorem graph.
Definition code
/-
The Killing form of so(3,1) as a 6×6 matrix.
The Killing form B(X, Y) = tr(ad_X ∘ ad_Y) computed in the
4×4 matrix representation from SO31MatrixRepr.lean.
Result: B = diag(+4, +4, +4, -4, -4, -4) in the basis
{K₁, K₂, K₃, J₁, J₂, J₃} = {boosts, rotations}.
This confirms:
- Signature (3,3) on the 6D Lie algebra
- Non-degenerate (all diagonal entries nonzero) → semisimple
- Electric sector (boosts) has +4, magnetic sector (rotations) has -4
- Mixed B(K_i, J_j) = 0
Paper reference: §2.4 (Lie algebra metric)
-/
import Definitions.Def_Grade4_LeanProofs_LieStructure_SO31MatrixRepr
namespace LeanProofsGrade4P2M
noncomputable section
/-! ## Computing ad_X as a 6×6 matrix via the 4×4 representation
In the 4×4 matrix representation, ad_X(Y) = [X, Y] = XY - YX.
The Killing form is B(X, Y) = tr₄(ad_X ∘ ad_Y) where tr₄ is
the 4×4 matrix trace.
For the standard Lie algebra trace formula:
B(X, Y) = tr(XY) · (normalization factor)
For so(n) and so(p,q), B(X, Y) = (n-2) · tr(XY).
For so(3,1) with n=4: B(X, Y) = 2 · tr(XY).
We verify this by direct computation on the generators.
-/
/-- The 4×4 matrix trace. -/
def matTrace (A : Matrix (Fin 4) (Fin 4) ℝ) : ℝ :=
A 0 0 + A 1 1 + A 2 2 + A 3 3
/-- The matrix inner product tr(A * B). -/
def matInner (A B : Matrix (Fin 4) (Fin 4) ℝ) : ℝ :=
matTrace (A * B)
/-! ## Diagonal entries: B(K_i, K_i) and B(J_i, J_i) -/
/-- tr(K₁²) = 2 (boost self-inner-product). -/
theorem matInner_K1_K1 : matInner boostK₁ boostK₁ = 2 := by
unfold matInner matTrace
simp [Matrix.mul_apply, boostK₁, Fin.sum_univ_four]
norm_num
/-- tr(J₁²) = -2 (rotation self-inner-product). -/
theorem matInner_J1_J1 : matInner rotJ₁ rotJ₁ = -2 := by
unfold matInner matTrace rotJ₁; norm_num [Matrix.mul_apply, Fin.sum_univ_four]
/-! ## Off-diagonal entries: mixed terms vanish -/
/-- tr(K₁ · J₁) = 0 (boost-rotation cross term vanishes). -/
theorem matInner_K1_J1 : matInner boostK₁ rotJ₁ = 0 := by
unfold matInner matTrace
simp [Matrix.mul_apply, boostK₁, rotJ₁, Fin.sum_univ_four]
/-! ## The Killing form B(X,Y) = 2·tr(XY) for so(3,1)
The standard formula: B(X,Y) = (n-2)·tr(XY) with n=4.
So B = 2·matInner.
Diagonal: B(K_i,K_i) = 2·2 = 4, B(J_i,J_i) = 2·(-2) = -4.
Off-diagonal: B(X,Y) = 0 for X ≠ Y (orthogonal generators).
-/
end
end LeanProofsGrade4P2M
Source