Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedprecheck pass
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

A real normal endomorphism is orthogonally block-diagonalisable with 1x1 real blocks and 2x2 rotation-scaling blocks

Statement

Let V be a finite-dimensional real inner product space and let T:V→V be normal. Then V has an orthonormal basis in which the matrix of T is block diagonal with blocks of the two forms

[a]and(ab−ba) with b≠0,

where a,b∈R.

Facts & Assumptions

Given: A finite-dimensional real inner product space V and a normal endomorphism T:V→V.

[L1]

In an orthonormal basis, a real operator is normal exactly when its matrix commutes with its transpose (In an orthonormal basis, self-adjoint means conjugate-transpose symmetry and normal means commuting with the conjugate transpose).

[L5]

Gram–Schmidt turns a finite real or complex independent list into an orthonormal list with the same span, including the empty list (Gram–Schmidt turns every finite independent list into an orthonormal list with the same successive spans).

Proof

technique · direct
1.1L1L2L3L5choose

Apply [L5] to a basis of V to choose an orthonormal basis, and let A∈Mn(R) be the matrix of T. By [L1], AAT=ATA. Since A has real entries, its conjugate transpose over C is AT, so A is normal on Cn with the standard Hermitian inner product from [L3]. Thus [L2] gives an orthonormal eigenbasis (z1,…,zn), say Azj=μjzj. If n=0, all subsequent lists are empty and give the asserted empty block decomposition.

2.1step 1.1L4algebra

For each eigenvalue λ, write Eλ=ker⁡C(A−λI). Expanding x=∑jcjzj gives (A−λI)x=∑j(μj−λ)cjzj, so Eλ is exactly the span of the zj with μj=λ. Consequently these vectors form an orthonormal basis of the entire eigenspace, the distinct eigenspaces are mutually orthogonal, and their sum is Cn. Also Ax‾=Ax‾ by reality of A, so conjugation maps Eλ onto Eλ‾, with itself as inverse.

3.1step 2.1L4L5choosealgebra

For real λ, put Rλ=ker⁡R(A−λI). If x=u+iw∈Eλ with u,w∈Rn, comparison of real and imaginary parts gives u,w∈Rλ; conversely every such u+iw lies in Eλ. Choose a real basis of the finite-dimensional subspace Rλ and apply [L5] to obtain a real orthonormal basis. It also spans Eλ over C by the decomposition just proved, and is complex-linearly independent: the real and imaginary parts of a complex linear relation are real relations among that real basis. Thus it is a full orthonormal complex basis of Eλ, consisting of real eigenvectors.

3.2step 2.1L3L4choosealgebra

For each conjugate pair of nonreal eigenvalues, select the representative λ=a+ib with b>0. Choose an orthonormal basis v1,…,vm of Eλ from step 2.1 and use v‾1,…,v‾m for Eλ‾. Conjugation is bijective between these eigenspaces and satisfies ⟨v‾j,v‾k⟩=⟨vj,vk⟩‾=δjk, so this is a full orthonormal basis there. Step 2.1 also gives ⟨vj,v‾k⟩=0 for every j,k.

4.1step 3.2L3L4algebraconstruct

Write vj=uj+iwj with real uj,wj. Using the coordinate pairing, the identities ⟨vj,vk⟩=δjk and ⟨vj,v‾k⟩=0 give respectively uj⋅uk+wj⋅wk=δjk, wj⋅uk−uj⋅wk=0, uj⋅uk−wj⋅wk=0, and wj⋅uk+uj⋅wk=0. Hence uj⋅uk=wj⋅wk=δjk/2 and uj⋅wk=wj⋅uk=0. Therefore ej=2uj and fj=2wj form an orthonormal real list. Comparing parts in A(uj+iwj)=(a+ib)(uj+iwj) gives Aej=aej−bfj and Afj=bej+afj. Thus each ordered pair (ej,fj) spans an invariant plane with matrix (ab−ba). Moreover uj=(vj+v‾j)/2 and wj=(vj−v‾j)/(2i), so these real vectors have complex span exactly Eλ⊕Eλ‾.

5.1step 1.1step 2.1step 3.1step 3.2step 4.1algebra∎

Collect the real bases from step 3.1 and the real pairs from step 4.1 for every selected nonreal representative. Vectors from different groups are orthogonal: each group lies in the complex span of its own eigenspace or conjugate eigenspace pair, and these disjoint groups of eigenspaces are mutually orthogonal by step 2.1; for real vectors the Hermitian pairing is the real dot product. Within each group orthonormality was proved in steps 3.1 and 4.1. Their complex span is all of Cn by step 2.1. Since the collected vectors are real, taking real parts of a complex expansion of any x∈Rn shows that they span Rn over R. They therefore give a real orthonormal basis, including the empty basis when n=0. Each basis vector from step 3.1 contributes [λ], and each pair from step 4.1 contributes the displayed block with b>0, hence b≠0. Transporting this basis through the coordinates of step 1.1 gives the required orthonormal basis of V.

Depends on

Dependency tree · 2 levels

Sources