%! TEX root = RT.tex % vim: tw=50 % 24/11/2023 11AM \begin{proposition*}[Cletsch-Gordon rule] For $n, m \in \NN$, \[ V_n \tprod V_m \simeq V_{n + m} \oplus V_{n + m - 2} \oplus \cdots \oplus V_{|n - m|} \] \end{proposition*} \begin{proof} Without loss of generality $n \ge m$. Then \begin{align*} (\chichar_n \chichar_m)(z) &= \left( \frac{z^{n + 1} - z^{-(n + 1)}}{z - z^{-1}} \right) (z^m + z^{m - 1} + \cdots + z^{-m}) \\ &= \sum_{j = 0}^m \left( \frac{z^{n + m + 1 - 2j} - z^{-n + m + 1 - 2j}}{z - z^{-1}} \right) \\ &= \sum_{j = 0}^m \chichar_{n + m - 2j}(z) \end{align*} \end{proof} \subsection{Representations of $\SO(3)$} \begin{proposition*} The action of $\SU(2)$ on the 3D $\RR$-normed vector space \[ \left\{ \begin{pmatrix} a & b \\ -\ol{b} & \ol{a} \end{pmatrix} ~\bigg|~ a + \ol{a} = 0 \right\} \subseteq M_2(\CC) \] with norm $\|A\|^2 = \det A$ by conjugation induces an isomorphism of \glspl{top_group} \[ \frac{\SU(2)}{\{\pm I\}} \to \SO(3) .\] \end{proposition*} \begin{proof} See \es[4]{4} (for more hints see lecturers notes from 2012). \end{proof} \begin{corollary*} Every \gls{irred} \gls{top_rep} of $\SO(3)$ is of the form $V_{2n}$ for some $n \ge 0$. \end{corollary*} \begin{proof} It follows from the previous proposition that the \gls{irred} \gls{rep} of $\SO(3)$ correspond to \gls{irred} \glspl{rep} of $\SU(2)$ whose kernel contains $\pm I$. But it is easy to see that $-I$ acts on $V_n$ by $(-1)^n$. \end{proof} \newpage \section{Character table of $\GL_2(\FF_q)$} \subsection{$\FF_q$} Let $p > 2$ be a prime, $q = p^a$ a power of $p$ for some $a > 0$ and let $\FF_q$ be the field with $q$ elements. We know that $\FF_q^\times \simeq C_{q - 1}$ and $\FF_q^\times \to \FF_q^\times$, $x \mapsto x^2$ is a homomorphism with kernel $\{\pm 1\}$. Thus half the elements are squares and half are not. Moreover, $x \mapsto x^{\frac{q - 1}{2}}$ sends squares to $1$ and non-squares to $-1$. Let $\eps \in \FF_q^\times$ be a fixed non-square. So $\eps^{\frac{q - 1}{2}} = -1$ and let \[ \FF_{q^2} = \{a + b\sqrt{\eps} \mid a, b \in \FF_q\} \] the field extension of $\FF_q$ (with respect to the obvious operations) of order $q^2$. Every element of $\FF_q$ has a square root in $\FF_q^2$, since if $\lambda \in \FF_q$ is a non-square then $\frac{\lambda}{\eps}$ is a square, $\mu^2$ say, and $(\sqrt{\eps} \mu)^2 = \eps \mu^2 = \lambda$. Thus every quadratic polynomial with coefficients in $\FF_q$ factorises over $\FF_q^2$. Notice $(a + b\sqrt{\eps})^q = a^q + b^q \eps^{\frac{q - 1}{2}} = a - b\sqrt{\eps}$ since $p \mid {q \choose i}$ for $0 < i < q$. Thus the roots of an irreducible quadratic over $\FF_q$ are of the form $\lambda, \lambda^q$ ($\lambda \mapsto \lambda^q$ is like complex conjugation). \subsection{$\GL_2(\FF_q)$ and its conjugacy classes} We want to compute the \gls{char_table} of the group \[ \GL_2(\FF_q) = \left\{ \begin{pmatrix} a & b \\ c & d \end{pmatrix} ~\bigg|~ a, b, c, d \in \FF_q, ad - bc \neq 0 \right\} .\] The order of $\GL_2(\FF_q)$ is equal to the number of bases for $\FF_q^2$ over $\FF_q$. This is $(q^2 - 1)(q^2 - q) = q(q - 1)^2(q + 1)$. First we compute the conjugacy classes of $\GL_2(\FF_q) \eqdef G$. We know from linear algebra (rational canonical form) that for $A \in G$, $[A]_G$ is determined by $m_A(x)$, the minimal polynomial and $\deg m_A(x) \le 2$ (Cayley-Hamilton). Moreover $m_A(0) \neq 0$. There are $4$ cases: \begin{enumerate}[\bfseries Case 1:] \item $m_A(x) = (x - \lambda)$ for some $\lambda \in \FF_q^\times$. Then $A = \lambda I$ so $C_G(A) = G$ and $|[A]_G| = |\{\lambda I\}| = 1$. There are $q - 1$ such classes -- one for each $\lambda$. \item $m_A(x) = (x - \lambda)^2$ for some $\lambda \in \FF_q^\times$. Then \[ [A]_G = \left[ \begin{pmatrix} \lambda & 1 \\ 0 & \lambda \end{pmatrix} \right]_G \] Now \[ C_G \left( \begin{pmatrix} \lambda & 1 \\ 0 & \lambda \end{pmatrix} \right) = \left\{ \begin{pmatrix} a & b \\ 0 & a \end{pmatrix} ~\bigg|~ a, b \in \FF_q, a \neq 0 \right\} \] (compute!). So \[ |[A]_G| = \frac{q(q - 1)^2 (q + 1)}{(q - 1)q} = (q - 1)(q + 1) .\] There are $q - 1$ such classes -- one for each $\lambda$. \item $m_A(x) = (x - \lambda)(x - \mu)$ for $\lambda, \mu \in \FF_q^\times$ distinct. So \[ [A]_G = \left[ \begin{pmatrix} \lambda & 0 \\ 0 & \mu \end{pmatrix} \right] = \left[ \begin{pmatrix} \mu & 0 \\ 0 & \lambda \end{pmatrix} \right]_G \] Moreover \[ C_G \left( \begin{pmatrix} \lambda & 0 \\ 0 & \mu \end{pmatrix} \right) = \left\{ \begin{pmatrix} a & 0 \\ 0 & d \end{pmatrix} ~\bigg|~ a, d \in \FF_q^\times \right\} \eqdef T \] so \[ |[A]_G| = \frac{q(q - 1)^2(q + 1)}{(q - 1)^2} = q(q + 1) \] There are ${q - 1 \choose 2}$ such classes -- one for each pair $\lambda, \mu$. \item $m_A(x)$ is irreducible over $\FF_q$ of degree $2$. So for some $\alpha \in \FF_{q^2} \setminus \FF_q$, $\alpha = \lambda + \mu \sqrt{\eps}$ for some $\lambda, \mu \in \FF_q$, $\mu \neq 0$, \begin{align*} m_A(x) &= (x - \alpha)(x - \alpha^q) \\ &= (x^2 - (\alpha + \alpha^q) x + \alpha \alpha^q) \\ &= (x^2 - (\Trace A) x + \det A) \end{align*} Then \[ [A]_G = \left[ \begin{pmatrix} \lambda & \eps \mu \\ \mu & \lambda \end{pmatrix} \right]_G = \left[ \begin{pmatrix} \lambda & -\eps\mu \\ -\mu & \lambda \end{pmatrix} \right]_G \] Since these matrices have trace $2\lambda = \alpha + \alpha^q$ and $\det \lambda^2 - \eps \mu^2 = \alpha \alpha^q$. Now \[ C_g \left( \begin{pmatrix} \lambda & \eps \mu \\ \mu & \lambda \end{pmatrix} \right) = \left\{ \begin{pmatrix} a & \eps b \\ b & a \end{pmatrix} ~\bigg|~ a, b \in \FF_q, a^2 - \eps b^2 \neq 0 \right\} \eqdef K \] If $a^2 - \eps^2 b = 0$, then $a^2 = \eps b^2$ so $\eps$ is a square or $a = b = 0$. So $|K| = q^2 - 1$ and \[ |[A]_G| = \frac{q(q - 1)^2(q + 1)}{(q - 1)(q + 1)} = q(q - 1) .\] There are $\frac{q(q - 1)}{2} = {q \choose 2}$ such classes -- one for each pair $\{\alpha, \alpha^q\} \subset \FF_{q^2} \setminus \FF_q$. \end{enumerate} In summary: \begin{center} \begin{tabular}{c|c|c|c} Rep $A$ & $C_G$ & $|[A]_G|$ & \# classes \\ \hline $\lambda I$ & $G$ & $1$ & $q - 1$ \\ $\begin{pmatrix} \lambda & 1 \\ 0 & \lambda \end{pmatrix}$ & $\left\{ \begin{pmatrix} a & b \\ 0 & a \end{pmatrix}\right\}$ & $(q - 1)(q + 1)$ & $q - 1$ \\ $\begin{pmatrix} \lambda & 0 \\ 0 & \mu \end{pmatrix}$ & $T$ & $q(q + 1)$ & ${q - 1 \choose 2}$ \\ $\begin{pmatrix} \lambda & \eps\mu \\ \mu & \lambda \end{pmatrix}$ & $K$ & $q(q - 1)$ & ${q \choose 2}$ \end{tabular} \end{center} The groups $T$ and $K$ are both called \emph{maximal tori}, i.e. they are maximal subgroups such that they are conjugate to a diagonal subgroup in $\GL_2(\FF)$ for some $\FF / \FF_q$. $T$ is called \emph{split} and $K$ is called \emph{non-split}. Some other important subgroups of $G$ are \begin{itemize} \item The subgroup of scalar matrices (the centre of $G$): \[ Z = \{\lambda I \st \lambda \in \FF_q^\times\} .\] \item A Sylow-$p$-subgroup of $G$ \[ N = \left\{ \begin{pmatrix} 1 & b \\ 0 & 1 \end{pmatrix} ~\bigg|~ b \in \FF_q \right\} \] so \[ ZN = \left\{ \begin{pmatrix} a & b \\ 0 & a \end{pmatrix} ~\bigg|~ a, b \in \FF_q, a \neq 0 \right\} \] \item A \emph{Borel subgroup} of $G$ \[ B = \left\{ \begin{pmatrix} a & b \\ 0 & d \end{pmatrix} ~\bigg|~ a, d \in \FF_q^\times, b \in \FF_q \right\} \] \end{itemize} Then $N \normalsub B$ and $B / N \simeq T \simeq \FF_q^\times \times \FF_q^\times \simeq C_{q - 1} \times C_{q - 1}$.