Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

I recently stumbled across this bilinear form: $\beta(A,B)=\operatorname{Tr}(AB)$ for $A,B \in \mathbb{R}^{n,n}$. I am searching for an orthogonal basis. With many difficulties I could find one for $\mathbb{R}^{2,2}$ namely:

$$B_{2,2}=\left\{\left( \begin{array}{cc} \frac{1}{\sqrt{2}} & 0 \\ 0 & \frac{1}{\sqrt{2}} \end{array} \right),\left( \begin{array}{cc} 0 & \frac{1}{\sqrt{2}} \\ \frac{1}{\sqrt{2}} & 0 \end{array} \right),\left( \begin{array}{cc} \frac{1}{\sqrt{2}} & 0 \\ 0 & -\frac{1}{\sqrt{2}} \end{array} \right),\left( \begin{array}{cc} 0 & -\frac{1}{\sqrt{2}} \\ \frac{1}{\sqrt{2}} & 0 \end{array} \right)\right\}$$

This form was also made that the representation matrix of $\beta$ is very elegant:

$$M_\beta=\left( \begin{array}{cccc} \beta \left(b_1,b_1\right) & \beta \left(b_1,b_2\right) & \beta \left(b_1,b_3\right) & \beta \left(b_1,b_4\right) \\ \beta \left(b_2,b_1\right) & \beta \left(b_2,b_2\right) & \beta \left(b_2,b_3\right) & \beta \left(b_2,b_4\right) \\ \beta \left(b_3,b_1\right) & \beta \left(b_3,b_2\right) & \beta \left(b_3,b_3\right) & \beta \left(b_3,b_4\right) \\ \beta \left(b_4,b_1\right) & \beta \left(b_4,b_2\right) & \beta \left(b_4,b_3\right) & \beta \left(b_4,b_4\right) \end{array} \right)=\left( \begin{array}{cccc} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & -1 \end{array} \right)$$

So I can read the positive index of inertia (basically the number of 1s on the diagonal) which is in this case $n_+=3$. I am looking for such bases for higher dimensions of $\mathbb{R}^{n,n}$ but could not succeed. Thank you in advance.

share|cite|improve this question
up vote 8 down vote accepted

You can construct an orthogonal basis with only $1$ and $-1$ on the diagonal by starting from the canonical basis $(b_{kl})_{ij}=\delta_{ik}\delta_{jl}$. The products are all zero except when the two factors are transposes of each other. Thus, each basis element $b_{kl}$ with $k=l$ is already orthogonal to all the others, and the remaining basis elements form pairs of transposes. You can form linear combinations of these pairs with coefficients $\pm 1/\sqrt{2}$ like you did for $n=2$ to from orthogonal combinations of them, one with "norm" $1$ and one with $-1$. Thus, the positive index of inertia is $n+\frac{n(n-1)}{2}=\frac{n(n+1)}{2}$.

share|cite|improve this answer
This exactly what I need, thank you!. And yes I think orthogonal is the right term I corrected this. – Listing Jun 11 '11 at 20:29
You're welcome. I removed the comment on the terminology after your correction. – joriki Jun 11 '11 at 20:31

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.