Symmetric Matrices — Mathematics Knowledge System

Overview

A symmetric matrix is a square matrix equal to its own transpose (A = Aᵀ), meaning the entry in row i, column j always matches the entry in row j, column i. Symmetric matrices are exceptionally well-behaved: the spectral theorem guarantees that every real symmetric matrix has only real eigenvalues and possesses a full set of orthogonal (or orthonormal, after scaling) eigenvectors, so it is always diagonalizable — no defective cases exist. This makes symmetric matrices the workhorse of applications from covariance matrices in statistics to stiffness matrices in engineering, all of which are symmetric by construction.

Intuition

Symmetric matrices behave like a 'balanced' stretching machine: they never introduce rotation without a matching mirror-image effect, so their fundamental stretching directions (eigenvectors) can always be chosen mutually perpendicular. This is why a symmetric matrix's spectral decomposition A = QΛQᵀ looks so clean — Q is an orthogonal 'change of perspective' to the natural axes, Λ does pure independent stretching along those axes, and Qᵀ changes back. Contrast this with a general (non-symmetric) matrix, whose eigenvectors can be skewed relative to each other or even fail to exist in full.

Formal Definition

Definition

A real n×n matrix A is symmetric if it equals its transpose. The spectral theorem states:

A = A^\top \iff a_{ij} = a_{ji} \ \text{for all } i,j

Symmetry condition

A = A^\top \implies \text{all eigenvalues of } A \text{ are real}

Real eigenvalues

A = Q\Lambda Q^\top, \quad Q^\top Q = I, \ \Lambda = \operatorname{diag}(\lambda_1,\ldots,\lambda_n)

Spectral decomposition (Q orthogonal)

Worked Examples

Check A = Aᵀ: the (1,2) and (2,1) entries both equal 1, so A is symmetric.
$A^\top = \begin{pmatrix}2&1\\1&2\end{pmatrix} = A$
Eigenvalues from det(A-λI) = (2-λ)²-1 = 0 give λ=1 and λ=3 (both real, as guaranteed).
$\lambda_1 = 1, \ \lambda_2 = 3$
Eigenvector for λ=1: (A-I)v=0 → [[1,1],[1,1]]v=0 → v=(1,-1). For λ=3: (A-3I)v=0 → [[-1,1],[1,-1]]v=0 → v=(1,1). These are orthogonal: (1,-1)·(1,1)=0.
$v_1=(1,-1),\ v_2=(1,1),\quad v_1 \cdot v_2 = 0$

Answer: Eigenvalues 1 and 3 with orthogonal eigenvectors (1,-1) and (1,1) — confirming the spectral theorem's guarantee.

Practice Problems

Difficulty 4/10

Is A = [[0,5],[5,0]] symmetric? Find its eigenvalues.

Difficulty 6/10

Find the eigenvectors of A = [[0,5],[5,0]] for λ=5 and λ=-5, and verify they are orthogonal.

Difficulty 6/10

In statistics, a covariance matrix Σ is always symmetric. Why does this guarantee that Principal Component Analysis (PCA) can always find orthogonal directions of maximum variance?

Quiz

Summary

A symmetric matrix satisfies A = Aᵀ, meaning entries mirror across the main diagonal.
The spectral theorem guarantees real eigenvalues and orthogonal eigenvectors for every real symmetric matrix.
Every symmetric matrix diagonalizes as A = QΛQᵀ with Q orthogonal — the basis for PCA and stress/vibration analysis.

References

WebsiteWikipedia — Symmetric matrix