2
$\begingroup$

Hi: I have a question about the tensor product of representation of Lie groups as follows:

Let $G$ be a connected, complex Lie group with Lie algebra $\mathfrak{g}$. Let $V$ and $W$ are representations with respect to $G$, and so these induce representations (with respect to $\mathfrak{g}$) on $V$ and $W$. That is, we obtain module structures: $V$ and $W$ are $\mathbb{C}[G]$-modules and $U(\mathfrak{g})$-modules as well (where $U(\mathfrak{g})$ is the universal enveloping algebra of $\mathfrak{g}$)). Is it true that $V\otimes_{\mathbb{C}[G]}W \cong V\otimes_{U(\mathfrak{g})}W$ in the category of vector spaces?

Thanks very much!

$\endgroup$
  • 1
    $\begingroup$ How did you come to ask this? Usually the tensor products one is interested in are those over the ground field. $\endgroup$ – Tobias Kildetoft Jan 4 '15 at 13:26
1
$\begingroup$

As Tobias already indicated in his comment, usually you consider the tensor product of two $G$- resp. ${\mathfrak g}$-modules over ${\mathbb C}$, but in case you mean $X\otimes_G Y$ resp. $X\otimes_{\mathfrak g} Y$ as $(X\otimes_{\mathbb C} Y)^G$ resp. $(X\otimes_{\mathbb C} Y)^{\mathfrak g}$ (that's the same as providing $X$ with the right ${\mathscr U}({\mathfrak g})$-structure given by $x.v := -v.x$ for $x\in X$ and $v\in{\mathfrak g}$ and then taking the usual tensor product $X\otimes_{{\mathscr U}({\mathfrak g})} Y$) the question splits in two parts:

  • First, check that the functor $(G\text{-mod},\otimes_{\mathbb C})\to({\mathfrak g}\text{-mod},\otimes_{\mathbb C})$ is monoidal, that is, that deriving the diagonal $G$-action $g.(v\otimes w) := (g.v)\otimes (g.w)$ on the tensor product $V\otimes_{\mathbb C}W$ of two $G$-modules results in the usual ${\mathfrak g}$-action $X.(v\otimes w) := (X.v)\otimes w + v\otimes (X.w)$ on $V\otimes_{\mathbb C} W$.

    You can do this by decomposing $G\to\text{GL}(V\otimes_{\mathbb C} W)$ as $$G\xrightarrow{\Delta} G\times G\xrightarrow{\rho_V\times\rho_W}\text{GL}(V)\times\text{GL}(W)\xrightarrow{\alpha}\text{GL}(V\otimes_{\mathbb C} W).$$ and noting that the derivative of $\Delta$ at $e$ is again $\Delta: {\mathfrak g}\to{\mathfrak g}\oplus{\mathbb g}$, while $\text{d}(\rho_V\times\rho_W)_e = \text{d}(\rho_V)_e\times\text{d}(\rho_W)_e$, and finally $\alpha$ is the restriction of the bilinear map $\text{End}(V)\times\text{End}(W)\xrightarrow{\alpha}\text{End}(V\otimes_{\mathbb C}W)$ the derivative of which is (as for any bilinear map) given by $\text{d}(\alpha)_{(f,g)}(\sigma,\tau) = \alpha(f,\tau)+\alpha(\sigma,g) = f\otimes\tau + \sigma\otimes g$.

  • Secondly, check that taking $G$-invariants $V\mapsto V^G:=\{v\in V\ |\ g.v=v\text{ for all } g\in G\}$ is the same as taking ${\mathfrak g}$-invariants $V\mapsto V^{\mathfrak g} := \{v\in V\ |\ X.v=0\text{ for all } X\in{\mathfrak g}\}$ if $G$ is connected - see $Hom_G(\pi,\sigma)$ = $Hom_{\mathfrak{g}}(d\pi,d\sigma)$?.

$\endgroup$

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.