MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

How do I go about diagonalizing such a matrix.

I ask because I need to sort out the following problem:

Let $D$ be the quaternion algebra over $\mathbb{Q}$ with $i^2 = -1, j^2 = -11, ij=-ji=k$.

Let $\mathfrak{O} = \langle 1,i,\frac{1+j}{2},\frac{i+k}{2}\rangle_{\mathbb{Z}}$ be a maximal order.

Consider the Hermitian matrix:

$A = \begin{pmatrix} 11 & -j-3k \\ j+3k & 11 \end{pmatrix}$.

I need to find all matrices $u\in M_2(\mathfrak{O})$ such that $uA\bar{u}^T = A$. This is proving difficult (there should be $48$ such matrices but extensive checks have only produced $8$ and those are easy diagonal/anti-diagonal ones!)

I think I can do it easily if I diagonalize $A$.

But how do I find $P\in$ $GL_2(D)$ such that $\bar{P}^T A P$ is diagonal with rational entries? Apparently such a matrix exists. Does the non-commutativity create problems with the usual construction?

share|cite|improve this question
up vote 1 down vote accepted

Note that $U=\pmatrix{1&t\cr 0&1}$ and $X=\pmatrix{a&b\cr c&d}$ has $UX\bar U^T=\pmatrix{?&b+dt\cr c&d}\pmatrix{1&0\cr -t&1}=\pmatrix{?&b+dt\cr c-dt&?}$ (for $t$ with trace 0). So just make $t=(j+3k)/11$ in your example, killing the terms off the diagonal. Maybe I transposed this, but you can work it out.

share|cite|improve this answer
Oh I can't believe how easy it was! I was too bogged down in theory to just try simple examples...thanks. – fretty Jun 28 '13 at 11:28

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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