2
$\begingroup$

Consider a compact Lie algebra $\mathfrak{g} \subset \mathfrak{u}(n)$ and its associated connected, compact Lie group $G$. Let $\mathfrak{g}^{\perp}$ denote $\mathfrak{g}$'s orthogonal complement (as a vector space) in $\mathfrak{u}(n)$. Due to invertibility, conjugation by $G$ preserves both $\mathfrak{g}$ and $\mathfrak{g}^{\perp}$, and $\mathfrak{g}^{\perp}$ decomposes into a direct sum of irreducible representations of $G$.

Are there any known results as to whether any of the irreducible $G$-invariant subspaces of $\mathfrak{g}^{\perp}$ can be isomorphic (as $G$-modules) to a simple ideal of $\mathfrak{g}$, in general? Or can such statements only be made when branching rules are available?

$\endgroup$

1 Answer 1

1
$\begingroup$

I don't know what you can say in general but there is a nice family of examples given by decomposing into blocks.

I'll work over $\mathfrak{su}(n)$ as you can just add the centre back in to $\mathfrak{g}^\perp$. Consider the (orthogonal) decomposition of $\mathfrak{su}(2n) = \mathfrak{s}\oplus \mathfrak{t} \oplus \mathfrak{m}$ where $$ \mathfrak{s} = \left\{\left.\begin{pmatrix}aI & bI \\ cI & dI\end{pmatrix}\right|\begin{pmatrix}a & b \\ c & d\end{pmatrix}\in \mathfrak{su}(2)\right\} \cong \mathfrak{su}(2)$$ $$ \mathfrak{t} = \left\{\left.\begin{pmatrix}A & 0 \\ 0 & A\end{pmatrix}\right|A\in \mathfrak{su}(n)\right\}\cong \mathfrak{su}(n)$$ $$ \mathfrak{m} = \mathfrak{su}(2n) \cap \left\{\left.\begin{pmatrix}A & B \\ C & -A\end{pmatrix}\right|A,B,C\in \mathfrak{sl}(n, \mathbb{C})\right\}$$

Then $\mathfrak{m}$ is isomorphic to $\mathfrak{s}\otimes \mathfrak{t}$ as a $\mathfrak{s}\oplus \mathfrak{t}$-module. Thus you can take $\mathfrak{g} = \mathfrak{s}$ to get a decomposition of $\mathfrak{g}^\perp \cong \mathfrak{s}^{n^2-1} \oplus \mathfrak{t}$ or take $\mathfrak{g} = \mathfrak{t}$ to get $\mathfrak{g}^\perp \cong \mathfrak{t}^{3} \oplus \mathfrak{s}$.

Note for context: I know this decomposition from the non-compact and complex setting but I think it still works in this compact setting. More generally, this comes from a family of decompositions of the form $\mathfrak{sl}_2 \oplus \mathfrak{t} \oplus (S \otimes T)$ where $\mathfrak{sl}_2$, $\mathfrak{t}$ commute, $S,T$ are $\mathfrak{sl}_2$, $\mathfrak{t}$ modules respectively and $S$ is isomorphic to $\mathfrak{sl}_2$ ($T$ is not always isomorphic to $\mathfrak{t}$). These exist in $\mathfrak{sl}_{2n}, \mathfrak{sp}_{2n}, \mathfrak{so}_{n}, \mathfrak{so}_{4n}$ (different form in the latter) and $\mathfrak{e}_7$. Thus you should be able to construct more examples by taking the compact real form of each of these embedded inside some $\mathfrak{su}(n)$. Even more generally, I think you can construct examples similar to this by choosing a copy of $\mathfrak{sl}_2$ sitting across a set of "root $\mathfrak{sl}_2$'s" i.e. $\mathfrak{s}^\alpha := \mathfrak{g}^\alpha \oplus \mathfrak{g}^{-\alpha} \oplus [\mathfrak{g}^\alpha, \mathfrak{g}^{-\alpha}]$ and $\mathfrak{sl}_2 \cong \mathfrak{g} \leq \bigoplus \mathfrak{s}^\alpha$ (you may need the set of roots to be strongly orthogonal).

$\endgroup$

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.