8
$\begingroup$

Let $G/H\cong PSL(2,11)$, and $\theta$ be an irreducible $\mathbb{C}$-character of $H$. Suppose $\theta$ is invariant in $G$ and $\theta(1)=9$.

Question: Is $\theta$ extendible to $G$?

$\endgroup$
11
  • $\begingroup$ I take it that $\theta (1)=9$ is a typo? $\endgroup$ Commented Jun 20, 2018 at 3:29
  • $\begingroup$ @Venkataramana $\theta(1)$ is the degree of $\theta$. $\endgroup$
    – C. Simon
    Commented Jun 20, 2018 at 3:32
  • $\begingroup$ it would be very strange if this was true. look for counterexamples... $\endgroup$ Commented Jun 20, 2018 at 7:27
  • 1
    $\begingroup$ By Clifford theory, some multiple of $\theta$ extends to $G$. $\endgroup$
    – Aurel
    Commented Jun 20, 2018 at 8:50
  • 1
    $\begingroup$ Silly question- Does $PSL(2,11)$ mean the projective group of rank 1 over $\mathbb F_{11}$ or the projective group of rank $10$ over $\mathbb F_{2}$? $\endgroup$
    – kneidell
    Commented Jun 20, 2018 at 11:08

1 Answer 1

12
$\begingroup$

The answer is yes (and some of the comments were moving in the right direction): Let $T$ be a transversal to $H$ in $G,$ and let $\sigma$ afford the representation of $H.$ For each $t \in T,$ there is a matrix $M_{t} \in {\rm GL}(9,\mathbb{C})$ such that $\sigma(tht^{-1}) = M_{t}\sigma(h)M_{t}^{-1}$ for all $h \in H,$ and $M_{t}$ is unique up to scalar multiples (by Schur's Lemma) , while any non-zero scalar multiple will have the same property. Extend this to $G$ by letting $M_{th} = M_{t}\sigma(h)$ for each $t \in T, h \in H.$ (It might be convenient here for anyone interested in full detail, to assume that have multiplied $\sigma$ by a suitable power of the linear character $ \lambda = {\rm det} \sigma,$ so that ${\rm det} \sigma$ may be assumed to have multiplicative order a power of $3$).

Note also that for $x,y \in G,$ there is a scalar $\alpha(x,y) \neq 0$ such that $M_{xy} = \alpha(x,y)M_{x}M_{y} (\ast).$

Notice then $M_{x}^{|G|}$ is a scalar matrix for each $x \in G.$ Multiplying each $M_{x}$ by a suitable scalar (and we can still keep $M_{h} = \sigma(h)$ for each $h \in H$, and $M_{th} = M_{t}\sigma(h)$ for $t \in T,h \in H$), we may, and do from now on, assume that each $M_{x}$ has determinant a $3$-power root of unity.

It follows from $(\ast)$ (on taking determinants), that $\alpha(x,y)$ is a $3$-power root of unity for all $x,y \in G.$ This gives a $2$-cocycle for ${\rm PSL}(2,11)$ of $3$-power order.

Now we can finish in either of two ways: the Schur multiplier of the perfect group ${\rm PSL}(2,11)$ is well-known to have order $2.$ But a more general argument is to note that a perfect group with a cyclic Sylow $p$-subgroup always has a Schur multiplier of order prime to $p$ ( and this may be applied here with $p =3$ ).

$\endgroup$
5
  • $\begingroup$ Firstly, Thanks. But, I don't understand: (1) why $M_t$ is a scalar; (2) why could we multiply each $M_t$ by a suitable scalar in paragraph two. $\endgroup$
    – C. Simon
    Commented Jun 21, 2018 at 3:43
  • $\begingroup$ Now, I understand the proof. I think that we don’t have to prove " $M_t$ is a root of unity multiple of $σ(t^{660})$ and that each$ M_t$ has finite multiplicative order." Can you give me some reasons? @Geoff Robinson $\endgroup$
    – C. Simon
    Commented Jun 22, 2018 at 9:22
  • $\begingroup$ Maybe there is a problem. Let $t,s\in T$. However, I am not sure $st\in T$. And, if we consider a linear character $\theta\in Irr(H)$, this proof should be effective. It means that every $G$-invariant linear character is extendible to $G$. $\endgroup$
    – C. Simon
    Commented Jun 23, 2018 at 6:46
  • $\begingroup$ Now, I understand that $\alpha(x,y)$ gives a $2$-cocycle for $G$. I don't understand that $\alpha(x,y)$ gives a $2$-cocycle for $G/H\cong PSL(2,11)$. I hope you give me the reason. Thank you, again. $\endgroup$
    – C. Simon
    Commented Jun 24, 2018 at 11:29
  • $\begingroup$ We have constructed a central extension of $G$ by a finite cyclic $3$-group. This is $\{ (\omega^{i}, M_{g}): g \in G :0 \leq i \leq 3^{m}-1)$ where $omega$ is some primitive $3^{m}$-th of unity. It has a normal subgroup isomorphic to $\sigma(H),$ which is $\{(1,\sigma(h)): h \in H \}.$ When we factor out the latter normal subgroup, we get a central extension of ${\rm PSL}(2,11)$ by a cyclic group of order $3^{m}.$ $\endgroup$ Commented Jun 24, 2018 at 12:11

You must log in to answer this question.

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