2
$\begingroup$

I would like the find an embedding/faithful representation from the projective linear group $PGL_n\mathbb{C}\to GL_m\mathbb{C}$ for some $m$, and likewise for the other projective groups $PSL_n\mathbb{R}, PSO_n\mathbb{C}, PSO_n\mathbb{R}, Psp_{2n}\mathbb{C}$, and $Psp_{2n}\mathbb{R}$.

Is there an elementary way to deduce such an embedding, without using the machinery of Lie algebras or adjoint representations?

I know that we can make an argument using the adjoint representation of $PGL_n\mathbb{C}$ on its Lie algebra, but these haven't been introduced in my text yet.

$\endgroup$
  • $\begingroup$ Is the identity map for $GL_n\mathbb{C}$ considered acceptable? $\endgroup$ – user27126 Mar 31 '13 at 17:48
  • $\begingroup$ Is there a typo in this question? Based on the title, it seems you are looking for a representation $\mathrm{PGL}_n(\mathbb{C})\to\mathrm{GL}_m(\mathbb{C})$, not $\mathrm{GL}_n(\mathbb{C})\to\mathrm{GL}_m(\mathbb{C})$. $\endgroup$ – Jim Belk Mar 31 '13 at 18:30
  • $\begingroup$ Yes, for some reason there isn't a "P" in $GL_n\mathbb{C}$! Fixed it, thank you. $\endgroup$ – ff90 Mar 31 '13 at 18:55
3
$\begingroup$

A representation of $\text{PGL}_n(\mathbb{C})$ is, almost by definition, a representation of $\text{GL}_n(\mathbb{C})$ on which the center acts trivially. The center of a group $G$ is, almost by definition, the kernel of the homomorphism

$$G \ni g \mapsto (x \mapsto gxg^{-1}) \in \text{Aut}(G).$$

When $G = \text{GL}_n(\mathbb{C})$, this homomorphism extends to a linear representation of $\text{GL}_n(\mathbb{C})$ on $\mathcal{M}_n(\mathbb{C}) \cong \mathbb{C}^{n^2}$ on which the center acts trivially, and almost by definition, this gives a faithful representation of $\text{PGL}_n(\mathbb{C})$. (This is the adjoint representation, but you don't need to know anything about Lie algebras or even tangent spaces to recognize that $\text{GL}_n(\mathbb{C})$ embeds into $\mathcal{M}_n(\mathbb{C})$.)

$\endgroup$
  • $\begingroup$ Many thanks. As for constructing an actual representation, my understanding is that it would suffice to take any representation of $\operatorname{GL}_n(\mathbb{C})$ on $\mathcal{M}_n(\mathbb{C})$ that is invariant under multiplication by complex scalars? And the same for, say, $\operatorname{SO}_n(\mathbb{C})$ on $\mathcal{M}_n(\mathbb{C})$ for $\operatorname{PSO}_n(\mathbb{C})$ invariant under $\{ \pm I\}$? $\endgroup$ – ff90 Mar 31 '13 at 20:25
  • $\begingroup$ @ff90: yes, that would suffice. $\endgroup$ – Qiaochu Yuan Mar 31 '13 at 20:47
  • $\begingroup$ @QiaochuYuan I recently learned that a definition of a matrix Lie group is a closed subgroup of $GL_{n}(\mathbb{C})$. In this case, is the image of that faithful representation closed? $\endgroup$ – user135520 Sep 2 '15 at 3:09

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.