1
$\begingroup$

As algebraic groups over an algebraically closed field $K$ of characteristic not $2$, $\operatorname{GO}_{2n}$ is a closed normal subgroup of the conformal orthogonal group $\operatorname{CO}_{2n}$. Is the projective conformal orthogonal group $\operatorname{PCO}_{2n}$ isomorphic to $\operatorname{PGO}_{2n}$? Is $\operatorname{PCO}_{2n}$ a maximal subgroup of $\operatorname{PGL}_{2n}$? Same questions for the odd degree $2n+1$. I am lacking of tools to analyze this. Thank you.

Edit:

$\operatorname{GO}(V)=\{g\in \operatorname{GL}(V)\ | \ f(g(v),g(w))=f(v,w) \,\ \forall v,w \in V\}$ for a non-degenerate symmetric bilinear form on $V$. And $\operatorname{CO}_{2n}$ preserves the form up to a scalar.

$\endgroup$
4
  • 4
    $\begingroup$ I would recommend you to explain your notations. I know what ${\rm PGL}_{2n}$ is. What are the other groups? $\endgroup$ Commented Sep 28, 2023 at 7:53
  • $\begingroup$ Thanks. Doing it now. $\endgroup$
    – user488802
    Commented Sep 28, 2023 at 7:54
  • 3
    $\begingroup$ Your definition of $GO(V)$ is what I would have called the orthogonal group $O_V$. $\endgroup$
    – Ben McKay
    Commented Sep 28, 2023 at 8:13
  • $\begingroup$ Please use a high-level tag like "ag.algebraic-geometry". I added this tag now. $\endgroup$
    – GH from MO
    Commented Sep 29, 2023 at 20:52

1 Answer 1

2
$\begingroup$

$\renewcommand{\O}{{\rm O}} \newcommand{\GO}{{\rm GO}} \newcommand{\PO}{{\rm PO}} \newcommand{\PGO}{{\rm PGO}} \newcommand{\GL}{{\rm GL}} \newcommand{\PGL}{{\rm PGL}} \newcommand{\SL}{{\rm SL}} \newcommand{\SO}{{\rm SO}} \newcommand{\Id}{{\rm Id}} $Notation. $K$ is an algebraically closed field;

$V$ a finite-dimensional vector space over $K$;

$f\colon V \times V\to K$ a non-degenerate symmetric bilinear form.

We write \begin{align*} \O(V,f)&=\{g\in\GL(V)\ |\ g^*f=f\},\\ \GO(V,f)&=\{g\in\GL(V)\ |\ g^*f =\lambda f \ \ \text{for some}\ \lambda\in K^\times\}. \end{align*} Here $$(g^*f)(v_1,v_2)=f (g\cdot v_1,g\cdot v_2)\ \ \text{for}\ v_1,v_2\in V.$$ We set \begin{align*} \PO(V,f)&=\O(V,f)/\{\pm1\},\\ \PGO(V)&=\GO(V,F)/K^\times \subset \PGL(V). \end{align*}

Proposition. $\PO(V,f)\cong\PGO(V,f)$.

Proof. We have the inclusion homomorphism $$ i\colon \O(V,f)\hookrightarrow \GO(V,f)$$ which induces a homomorphism $$i_*\colon \PO(V,f)\to \PGO(V,f),\quad\ g\cdot\{\pm1\}\mapsto g\cdot K^\times.$$ We show that $i_*$ is an isomorphism.

Let $g\cdot\{\pm1\}\in\ker i_*\,$. Then $$g\in(K^\times\cdot\Id_V)\cap\O(V,f)=\{\pm1\}=1\cdot \{\pm1\}.$$ This $i_*$ is injective.

Let $g\cdot K^\times\in\PGO(V,f)$. Then $ g^*f =\lambda f$ for some $\lambda\in K^\times$. We set $g'=g/\sqrt{\lambda}$ (recall that $K$ is algebraically closed). Then $(g')^*f=f$, whence $g'\in\O(V,f)$. Thus $$g\in \O(V,f)\cdot K^\times,$$ whence $i_*$ is surjective, which completes the proof of the proposition.

Concerning maximality, it is well-known that $\SO(V,f)$ is maximal in $\SL(V)$; see, for instance YCor's answer to $\operatorname{SO}(n)$ is an (abstractly) maximal subgroup of $\operatorname{SL}(n)$ in Math.StackExchange.com in the case $K=\Bbb R$. As above, one can deduce the maximality of $\PGO(V,f)$ in $\PGL(V)$ by standard methods. If you cannot do that yourself, ask in MathStackExchange.com; your question seems to be more suitable for that forum rather than for MathOverflow.

$\endgroup$
1
  • $\begingroup$ MathJax is very poor at ignoring whitespace one would expect, so it is necessary that the $ that ends the command block be adjacent to the Notation at the beginning of the post to avoid a spurious space. I have edited accordingly. $\endgroup$
    – LSpice
    Commented Sep 29, 2023 at 20:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .