Symmetric Matrices and Their Entrywise Test

Symmetry Is a Match Across the Main Diagonal

A symmetric matrix has a simple visual pattern: every entry above the main diagonal matches the entry in the reflected position below it. The diagonal itself does not create a restriction, because transposing a matrix leaves its diagonal entries in place. The useful test is therefore to compare pairs of off-diagonal entries, not to inspect the diagonal or to calculate eigenvalues.

Throughout this tutorial, matrices are real and square unless stated otherwise. For an \(n\times n\) matrix \(\mathbf{A}\), the entry in row \(i\) and column \(j\) is \(a_{ij}\). Its transpose \(\mathbf{A}^{\top}\) is the matrix whose entry in row \(i\) and column \(j\) is \(a_{ji}\); transposition exchanges rows and columns.

The definition is a matrix equality, so it requires both matrices to have the same dimensions; in particular, a rectangular matrix cannot be symmetric. The entry \(a_{ij}\) is compared with \(a_{ji}\). When \(i=j\), these are the same entry, so every diagonal entry automatically passes the comparison. The substantive checks are the pairs with \(i\ne j\).

Proof. Suppose first that \(\mathbf{A}\) is symmetric. Then \(\mathbf{A}^{\top}=\mathbf{A}\). Equality of matrices means equality in each position, and the entry in position \((i,j)\) of \(\mathbf{A}^{\top}\) is \(a_{ji}\). Therefore \(a_{ji}=a_{ij}\) for every \(i,j\). Conversely, suppose \(a_{ij}=a_{ji}\) for every \(i,j\). In each position \((i,j)\), the entry of \(\mathbf{A}^{\top}\) is \(a_{ji}\), which by assumption equals the entry \(a_{ij}\) of \(\mathbf{A}\). Thus the two matrices agree in every position, so \(\mathbf{A}^{\top}=\mathbf{A}\), and \(\mathbf{A}\) is symmetric. \(\square\)

The Test in Dimensions Two and Three

For a \(2\times2\) matrix, there is just one off-diagonal pair to check. For a \(3\times3\) matrix, there are three: positions \((1,2)\) and \((2,1)\), positions \((1,3)\) and \((3,1)\), and positions \((2,3)\) and \((3,2)\). This gives the following forms.

Proof. For \(\mathbf{A}=\begin{bmatrix}a&b\\c&d\end{bmatrix}\), the entrywise test imposes \(b=c\); the diagonal entries \(a,d\) impose no conditions. Thus \(\mathbf{A}\) is symmetric exactly when it has the stated \(2\times2\) form. For a \(3\times3\) matrix \(\mathbf{A}=[a_{ij}]\), the test imposes \(a_{12}=a_{21}\), \(a_{13}=a_{31}\), and \(a_{23}=a_{32}\). These three equalities pair the off-diagonal entries, while the three diagonal entries are unrestricted. Writing the paired entries as \(b,c,e\) and the diagonal entries as \(a,d,f\) gives the stated form, and every matrix of that form satisfies all three required equalities. \(\square\)

Using the Test with Parameters

When a matrix contains unknown parameters, symmetry becomes a system of equality conditions. Each off-diagonal pair gives one equation. The diagonal parameters remain unrestricted unless the problem imposes other conditions. Solving the equations tells us which parameter choices produce symmetry, and it also reveals when no such choices exist.

Basic Operations Preserve Symmetry

The entrywise test also makes it possible to check simple operations without redoing a full transpose. The sum of two symmetric matrices of the same size is symmetric, and any real scalar multiple of a symmetric matrix is symmetric. These facts are useful when constructing or simplifying matrices while keeping track of the property.

Proof. Because \(\mathbf{A}\) and \(\mathbf{B}\) are symmetric, the entrywise test gives \(a_{ij}=a_{ji}\) and \(b_{ij}=b_{ji}\) for every \(i,j\). The entry in position \((i,j)\) of \(\mathbf{A}+\mathbf{B}\) is \(a_{ij}+b_{ij}\), while the entry in position \((j,i)\) is \(a_{ji}+b_{ji}\). These are equal because \(a_{ij}=a_{ji}\) and \(b_{ij}=b_{ji}\). Thus the sum is symmetric. The entries of \(r\mathbf{A}\) in positions \((i,j)\) and \((j,i)\) are \(ra_{ij}\) and \(ra_{ji}\). They are equal since \(a_{ij}=a_{ji}\), including when \(r=0\). Therefore \(r\mathbf{A}\) is symmetric as well. \(\square\)

Common Mistakes and Why the Test Matters

Mistake: checking only the diagonal. Consider \(\mathbf{B}\) from the earlier example. Its diagonal entries are \(1,4,6\), and transposition leaves these entries unchanged. A wrong calculation might check only that diagonal and conclude that \(\mathbf{B}^{\top}=\mathbf{B}\). But the off-diagonal entries \(b_{23}=-1\) and \(b_{32}=5\) do not match. The correction is to compare every off-diagonal pair, or to write out the full transpose.

Mistake: assuming symmetry means diagonal. A diagonal matrix has zero in every off-diagonal position, so it is symmetric; however, a symmetric matrix need not be diagonal. For instance, \(\mathbf{C}\) is symmetric and has nonzero off-diagonal entries. Symmetry requires matching reflected entries, not that those entries vanish.

Mistake: treating symmetry as an eigenvalue calculation. Symmetry is decided directly from the entries. It is unnecessary to compute a characteristic polynomial or eigenvectors to test whether \(\mathbf{A}^{\top}=\mathbf{A}\). The distinction matters because symmetry is a condition on the matrix entries, while eigenvalue computations answer different questions about the matrix.

The test is especially useful in spectral theory: it quickly identifies whether a real matrix meets the symmetry hypothesis in results about orthogonal diagonalization. It also provides a concrete verification after a matrix is assembled from data or parameters. The test itself is elementary, but applying it carefully prevents a diagonal check, a single matching pair, or a familiar matrix shape from being mistaken for a complete argument.