Are $Q_1(x,y)=xy$ and $Q_2(x,y)=x^2+y^2$ equivalent forms?

179 Views Asked by At

This is an old qual problem at my school. I've never dealt with quadratic forms, so I'm not sure if I've done this right.

Two quadratic forms $Q_1(x,y)$ and $Q_2(x',y')$ are said to be equivalent if they are related by a non-singular change of coordinates $(x,y)\mapsto (x',y')$. Decide whether $Q_1=xy$ and $Q_2=x^2+y^2$ are equivalent over $\mathbb{C}$ and whether they are equivalent over $\mathbb{R}$.

I'm not sure what "related by" actually means. I take it to mean there is a transformation $x=ax'+by'$ and $y=cx'+dy'$ such that $ad-bc\neq 0$. My guess is $Q_1$ and $Q_2$ are equivalent over $\mathbb{C}$, since you can take the change of coordinates $x=x'+iy'$ and $y=x'-iy'$. So $xy=(x'+iy')(x'-iy')=x'^2+y'^2$, and the change of coordinate matrix would be $$ \begin{bmatrix} 1 & 1 \\ i & -i \end{bmatrix} $$ which is nonsingular with determinant $-2i$.

However, they are not equivalent over $\mathbb{R}$. If there existed a transformation $x=ax'+by'$ and $y=cx'+dy'$ such that $xy=x'^2+y'^2$, we would have to have $$ xy=acx'^2+(ad+bc)x'y'+bdy'^2 $$ with $ac=1$, $bd=1$ and $ad+bc=0$. Writing $c=1/a$ and $d=1/b$, the last equation simplifies to $$ \frac{a}{b}+\frac{b}{a}=0\implies a^2+b^2=0 $$ for nonzero $a,b\in\mathbb{R}$, which is impossible.

Did I interpret this correctly?

1

There are 1 best solutions below

1
On BEST ANSWER

Yes, this is correct. Another way to see the inequivalence over $\mathbb R$ is to denote that (as $x$ and $y$ range over various elements of $\mathbb R$), the quadratic form $xy$ can take both negative as well as positive values, while $x^2 + y^2$ takes only non-negative values.