MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $G=\mathrm{SL}(2,\mathbb{R})$ and $K=\mathrm{SO}(2)$. Suppose $\pi$ is a continuous irreducible representation of $G$ in a Banach space $H$. Can one always find a nonzero $v\in H$ such that $\langle \pi(k)v : k\in K\rangle$ is finite-dimensional? Serge Lang uses this implicitly in the proof of Theorem 3 on p. 24 of his book on $\mathrm{SL}(2,\mathbb{R}).$ (I know that the answer is affirmative when $H$ is a Hilbert space and $\pi$ is unitary on $K$.)

share|cite|improve this question
Let me record here that $K$-finite vectors not only exist but are actually dense in Hilbert representations—neither irreducibility of $\pi$ nor unitarity of $\pi|_K$ is required. Indeed, one can construct a new inner product on $H$ which induces the same topology and with respect to which $\pi|_K$ is unitary. It follows then by Peter-Weyl theorem that $K$-finite vectors make up a dense subspace. – Murat Güngör Jun 7 '13 at 11:23

You must combine lemma 5 and lemma 1 in Serge Lang$(2,\mathbb{R})$: lemma 5 (which holds for Banach spaces) tells you how to construct $K$-finite vectors. Now you must rule out the case where $\pi(f)v=0$ for every $v\in H$ and $f\in \oplus_{m,n} S_{n,m}$ (notations as in Lang). But the latter direct sum is dense in $C_c(G)$, by lemma 1. And using an approximation of identity in $C_c(G)$, you see that $\pi(f)v=0$ for every $f\in C_c(G)$, forces $v=0$.

share|cite|improve this answer
Thank you for your answer, professor. The argument you give essentially reduces my question to the following one:…. – Murat Güngör Apr 21 '13 at 18:23

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.