$T:V\rightarrow V$ is over $\mathbb{R}$ and $V$ of finite dimension $n$, and I know that it is orthogonally diagonalizable.

The Matrix that represents it - call it $A$ ,in orthonormal basis is orthogonally diagonalizable as well. So $A=PDP^*$ where $P^*$ is a unitary matrix, and $D$ a diagonal matrix.

Is it true that $A$, $D$, and $P$ are in $M_{n \times n}(\mathbb{R})$ (real matrices) because the transformation is over $\mathbb{R}$? I want to conclude that $P^*$ = $P^t$.

Thanks a lot!!


No it is not in general true.

Let $A=\left(\begin{matrix}0&1\\ -1&0\end{matrix}\right)$.

The $A$ is anti-symmetric, and therefore it is orthogonally diagonalizable, and its eigenvalues are $\pm i$. Therefore, $D$ contains purely imaginary elements in the diagonal. Clearly, $P$ is not a real matrix either. In fact $$ P=\left(\begin{matrix} 1& i\\ i&1\end{matrix}\right) $$

Nevertheless it is true is $A$ is symmetric (real symmetric). In such case, the eigenvalues are real (and so is $D$) and a diagonalization is achieved via a real unitary matrix $P$.

  • $\begingroup$ Is it true for symmetric matrices (if A is symmetric)? $\endgroup$ – CnR Feb 10 '14 at 7:15
  • $\begingroup$ See updated version. $\endgroup$ – Yiorgos S. Smyrlis Feb 10 '14 at 7:19
  • $\begingroup$ You really help me to pass my final exam. thanks :) $\endgroup$ – CnR Feb 10 '14 at 7:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.