Why Symmetry and Complex Eigenvalues Matter
Introduction
When working with linear algebra, eigenvectors often raise the question of uniqueness. In many cases, students notice that eigenvectors are not always unique. This happens due to the properties of the matrix itself and the nature of its eigenvalues. Understanding the reason can help in handling such cases correctly.
One reason eigenvectors are non-unique is when the matrix is symmetric. Symmetric matrices can produce multiple valid eigenvectors corresponding to the same eigenvalue. In such cases, different eigenvectors can still span the same eigenspace, meaning uniqueness is not guaranteed.
Another case is when the eigenvalues are complex. For matrices with complex eigenvalues, the eigenvectors are determined up to a scalar multiple, and they may not appear unique in their real form. However, this is normal, since multiplying an eigenvector by any scalar (including complex numbers) still results in a valid eigenvector.
Thus, eigenvectors are unique only up to scalar multiples, and in cases of repeated eigenvalues, entire eigenspaces exist, leading to multiple valid eigenvectors.
Symmetric Matrices
A symmetric matrix can have repeated eigenvalues. When this happens, more than one independent eigenvector exists for the same eigenvalue. These vectors are not unique, but they still span the same eigenspace. Example:
Consider the symmetric matrix: A=[2002]A = \begin{bmatrix} 2 & 0 \\ 0 & 2 \end{bmatrix}A=[2002] Its eigenvalue is 2 (repeated). Any vector of the form [xy]\begin{bmatrix} x \\ y \end{bmatrix}[xy] is an eigenvector. This shows infinitely many valid eigenvectors exist.
Complex Eigenvalues
For matrices with complex eigenvalues, eigenvectors are only unique up to a scalar multiple. Multiplying an eigenvector by any real or complex number still produces a valid eigenvector. Hence, uniqueness is not possible. Example:
Consider B=[0−110]B = \begin{bmatrix} 0 & -1 \\ 1 & 0 \end{bmatrix}B=[01−10] Its eigenvalues are iii and −i-i−i. The eigenvectors are complex, and multiplying them by any scalar (say 2i2i2i) still gives a valid eigenvector.
Conclusion
Eigenvectors are not meant to be unique. They represent directions, not fixed points. In symmetric matrices, repeated eigenvalues lead to multiple eigenvectors. With complex eigenvalues, eigenvectors differ only by scalar multiples. The real takeaway: eigenvectors describe a whole family of directions, and that’s what makes them powerful.