I am working in the vector space $\mathbb F^{2\times 2}$, which denotes the space of $2\times 2$ matrices over the field $\mathbb F$ (the characteristic of $\mathbb F$ is not equal to $2$) and I would like to proof that: $\det:A\mapsto\det A$ is a quadratic form. Furthermore, I would like to show that $$\det(A+B) - \det(A) - \det(B) = tr(AB^\#)$$ whereas $tr$ denotes the trace of the matrix and $B^\#$ denotes the cofactor matrix of $B$.

My definition of a quadratic form $q$ is that $q(cx) = c^2q(x)$ and that the map $$(x,y) \mapsto q(x+y)-q(x)-q(y)$$ is a bilinear form. I could easily prove the first criterion, but I am really stuck with the second. I also tried to first show the equality above (the trace equality), but for some reason, for matrices $$A= \begin{pmatrix} a & b\\c & d \end{pmatrix} \qquad B = \begin{pmatrix} w & x\\y & z \end{pmatrix}$$

the left side gives me $az+dw-by-cx$ and the right side leaves me with $az+dw-bx-cy$

I am pretty sure that I computed the determinants and the cofactor matrix correctly, so I am genuinely confused where my mistake lies and I also do not know how to prove the bilinear form criterion. Any help is greatly appreciated!

  • $\begingroup$ "...in a field $\;\Bbb F^{2\times2}\;$ "....and its characteristic isn't $\;2\;$ ...then what in the world is that field, anyway!? $\endgroup$ – DonAntonio Apr 1 '17 at 9:50
  • 1
    $\begingroup$ @DonAntonio I suppose M3xr means the algebra of $2\times 2$ matrices over the field $\Bbb F$, where $\operatorname{char}F\ne 2$ $\endgroup$ – Hagen von Eitzen Apr 1 '17 at 9:52
  • $\begingroup$ updated it now, sorry. $\endgroup$ – AxiomaticApproach Apr 1 '17 at 9:53
  • $\begingroup$ @HagenvonEitzen Yup, that is now edited that way. Thanks. $\endgroup$ – DonAntonio Apr 1 '17 at 9:53
  • $\begingroup$ The determinant of a $\;2\times2\;$ matrix is a homogeneous polynomial of degree two in the matrix's entries and thus a quadratic form. $\endgroup$ – DonAntonio Apr 1 '17 at 9:55

Your computation of the left hand side, $az+dw-by-cx$ is correct. Note that this sum of products can be interpreted as trace of $$\tag1\begin{pmatrix}az-by&?\\?&dw-cx\end{pmatrix}=\begin{pmatrix}a&b\\c&d\end{pmatrix} \begin{pmatrix}z&-x\\-y&w\end{pmatrix}$$ So maybe you check again the definition of $B^\#$. At any rate, the second factor in $(1)$ is clearly linear in $B$, hence $(1)$ is linear in $A$ for fixed $B$ and linear in $B$ for fixed $A$ - in other words, it is bilinear, as desired.

  • $\begingroup$ First of all, thank you very much! Maybe it's just me, but I am sure that for $B = \begin{pmatrix} w\ x\\y\ z \end{pmatrix}$ $B^\# = \begin{pmatrix} z\ -y\\-x\ w \end{pmatrix}$. I double checked definitions from various sources but I dont really get it. $\endgroup$ – AxiomaticApproach Apr 1 '17 at 10:08
  • $\begingroup$ @M3xr Yes, I meant that maybe your textbook might have a different definition than most of the rest. Alternatively, they may have accidentally left out a ${}^T$ somewhere in the problem statement. $\endgroup$ – Hagen von Eitzen Apr 1 '17 at 10:15
  • $\begingroup$ okay, I think then they just might have forgotten a transpose sign. Thanks! $\endgroup$ – AxiomaticApproach Apr 1 '17 at 10:19

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.