Course contentsShow
← Mathematics
Lesson 2272 of 3,12147. Linear Algebra — Applications and Advanced TopicsFree lesson

The Spectral Theorem for Symmetric Matrices

State and prove the spectral theorem: every symmetric matrix is orthogonally diagonalizable.

The Spectral Theorem for Symmetric Matrices

What you'll learn: You'll discover why symmetric matrices are guaranteed to be orthogonally diagonalizable—one of the most elegant results in linear algebra.

The Big Idea

The Spectral Theorem states that every symmetric matrix can be orthogonally diagonalized. In other words, if A is a real symmetric matrix (meaning A = Aᵀ), then there exists an orthogonal matrix Q and a diagonal matrix D such that:

A = QDQᵀ

This is a powerful upgrade from ordinary diagonalization: not only can we diagonalize A, but the change-of-basis matrix Q has orthonormal columns (eigenvectors). This makes the decomposition numerically stable and geometrically meaningful.

Why This Matters

Recall that orthogonal diagonalization requires:

  1. A has n linearly independent eigenvectors
  2. Those eigenvectors are mutually orthogonal

The Spectral Theorem guarantees both conditions automatically for symmetric matrices.

Proof Sketch

  1. Real eigenvalues: For any symmetric matrix A, all eigenvalues are real (you proved this earlier).

  2. Orthogonal eigenvectors: Eigenvectors corresponding to distinct eigenvalues of a symmetric matrix are orthogonal. (If λ ≠ μ and Av = λv, Aw = μw, then v · w = 0.)

  3. Complete orthonormal basis: Using the Gram-Schmidt process within each eigenspace (for repeated eigenvalues), we can construct an orthonormal basis of eigenvectors for ℝⁿ. These form the columns of Q.

  4. Orthogonal diagonalization: Since Q has orthonormal columns, Qᵀ = Q⁻¹, giving us A = QDQᵀ.

Geometric Intuition

Think of symmetric matrices as describing shapes (quadratic forms). The Spectral Theorem says every such shape has principal axes aligned with orthogonal directions—the eigenvectors. The eigenvalues tell you the "stretch" along each axis.

Key Takeaway: Every symmetric matrix is orthogonally diagonalizable; it has real eigenvalues and an orthonormal basis of eigenvectors.