Let $H \hookrightarrow G$ be an inclusion of semisimple, compact Lie groups. There is a measure on the homogeneous coset space $G/H$ by pulling back the Haar measure on $G$ via the projection $G \twoheadrightarrow G/H$.

The Peter-Weyl theorem states that the two-sided regular representation of $G$ on $L^2(G)$ is isomorphic to this direct sum: $$\underset{{[\rho] \in \Lambda}}{\bigoplus} \rho^* \otimes \rho$$ $\Lambda$ is the set of isomorphism classes of irreducible unitary representations.

My question now is: What does $L^2(G/H)$ look like? Is there a similar direct sum decomposition? How does it depend on the representations of $G$ and $H$?


This is a very nice, natural question, also with a pleasant answer: recalling that the expression $\bigoplus_\rho \rho^*\otimes \rho$ is the decomposition under the action of $G\times G$ by $(g,h)\cdot f(x)=f(g^{-1}xh)$, a given $\rho^*\otimes\rho$ when restricted to $G\times H$ becomes $\rho^*\otimes \rho|_H$. For such functions to descend to $G/H$ is exactly that $\rho|_H$ contains the trivial repn of $H$, giving $\rho^*\otimes \mathbb C^{m(1,\rho)}$ where $m(1,\rho)$ is the multiplicity of the trivial repn of $1$ in $\rho|_H$. The whole is the direct sum over $\rho$...

  • $\begingroup$ Great! Could we also say that the summand is $\rho^* \otimes \operatorname{Rep}_H(1, \rho|_H)$? (The second factor might be better known as $\operatorname{Hom}_{\operatorname{Rep}_H}(1, \rho|_H)$ to some people, the space of $H$-equivariant maps from the trivial representation to $\rho|_H$) $\endgroup$ – Turion Dec 10 '13 at 15:39
  • $\begingroup$ Sure, yes, that's essentially the same thing, unless potentially someone cares about a distinction between trivial repns and their duals, or something. That is, the second "factor" is the trivial $H$-repn subspace of $\rho$... A secondary issue... $\endgroup$ – paul garrett Dec 10 '13 at 15:43
  • $\begingroup$ Thanks for the useful and elegant answer. I think this implies some Schur orthogonality relations for the quotient space (given that rho is written in an appropriate basis). Would anyone please suggest a place to find this in a book or paper? Best I could find was this: etamaths.com/index.php/ijaa/article/view/792. $\endgroup$ – Victor V Albert Apr 3 '19 at 23:35

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.