# Eigenvalues of $A^*A$

How can we show that eigenvalues for $A^*A$ are real and positive without using the Singular Value Decomposition theorem (where $A$ is a complex square matrix and $A^*$ its Hermitian conjugate)?

-
Do you know how to show that the eigenvalues of a self-adjoint matrix are real? – Qiaochu Yuan May 2 '11 at 23:08
Hi, have you tried to prove it by proving $A^* A$ is positive-semidefinite? a hint would be $(A^* Ax, y) = (Ax, Ay)$, where $(\cdot, \cdot)$ is the standard inner product – Shuhao Cao May 2 '11 at 23:11

The eigenvalues of $A^*A$ can be zero, so what you want to show is that they are non-negative: Take any eigenvalue $\lambda$ of $A^*A$, Let $v$ be an associated non-zero eigenvector. $0\leq\langle Av,Av\rangle=\langle A^*Av,v \rangle=\lambda ||v||^2$, which implies that $\lambda\geq 0$. In particular, $\lambda$ is real and non-negative.
Another way is to say that for every matrix $B$, $\lambda$ is an eigenvalue of $B$ if and only if $\bar{\lambda}$ is an eigenvalue of $B^*$, which implies (how?) that the eigenvalues of $B$, if $B$ is self-adjoint, are real ($B=B^*$). Since $A^*A$ is self-adjoint (check!) the result follows. – Dennis Gulko May 2 '11 at 23:14