The Spectral Theorem for Normal Matrices
You'll learn how the spectral theorem extends from symmetric matrices to all normal matrices over the complex numbers, enabling unitary diagonalization.
What Makes a Matrix Normal?
A matrix A is called normal if it commutes with its conjugate transpose: A*A = AA*. This elegant condition captures a broad family of matrices:
- Symmetric matrices (A = Aᵀ) are normal over the reals
- Hermitian matrices (A = A*) are normal over the complexes
- Unitary matrices (A* = A⁻¹) are normal
- Skew-symmetric and skew-Hermitian matrices are normal
The Spectral Theorem: Complex Version
The spectral theorem for symmetric real matrices guaranteed orthogonal diagonalization. For normal matrices over ℂ, we get something equally powerful:
Every normal matrix A is unitarily diagonalizable.
This means there exists a unitary matrix U (where U*U = I) and a diagonal matrix Λ such that:
A = UΛU*
The columns of U form an orthonormal basis of eigenvectors, and Λ contains the (possibly complex) eigenvalues.
Why This Matters
Normality is precisely the condition needed for a matrix to have a complete orthonormal set of eigenvectors. Without normality, matrices may lack enough eigenvectors to diagonalize, or their eigenvectors may not be orthogonal.
Think of unitary diagonalization as a rotation of your coordinate system (via U) that perfectly aligns with the matrix's natural directions, revealing its action as simple scaling along orthogonal axes.
The Bridge
Over the reals with symmetric matrices, you got orthogonal diagonalization. Over the complexes with normal matrices, you get unitary diagonalization—the natural generalization where "orthogonal" becomes "unitary."
Key Takeaway: Normal matrices (those satisfying AA = AA) are exactly those that can be unitarily diagonalized, extending the spectral theorem to the complex field with orthonormal eigenvector bases.