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 R be the set of real numbers, then R acts on L²(R) by translation. This is a unitary representation, which is far away from being irreducible. So what are the closed irreducible subspaces?

This might be a too naive question and not appropriate here.

share|cite|improve this question
Isn't this Fourier analysis? – Steve Huntsman Mar 13 '10 at 0:51
And a nice application too (IMHO). Motivates group C*-algebras and all that jazz. – Yemon Choi Mar 13 '10 at 0:55
Not only that but it's the fundamental example that shows that if you want to decompose a representation of a Lie group on a Hilbert space then you need the notion of a direct integral as well as a direct sum. @unknown: if you're happy to accept that an irreducible representation of an abelian group is 1-dimensional, then you're asking for a character chi:R-->C^* and a non-zero L^2 function f such that f(x+z)=chi(z)f(x) for all x. But this latter property forces f to be non-L^2. This at least proves there are no 1-d subspaces. And then you go on to invent the Fourier transform to fix this! – Kevin Buzzard Mar 13 '10 at 9:07

Taking Fourier transforms, you get the action by multiplication. (Really, $x$ acts by multiplying with $e^{ix}$.) The closed invariant subspaces will thus be of the form $\{f\in L^2(\mathbb{R})\colon\hat f|_E=0\}$ for measurable sets $E$.

But you were asking for closed irreducible subspaces: There aren't any, because the Lebesgue measure has no atoms.

share|cite|improve this answer
I was typing a more cumbersome version, but HHO beat me to it :) I'd also suggest considering the discrete version, with $Z$ acting on $l^2(Z)$. The resulting characterisation is exactly analogous, but the Fourier analysis you'd want to use is a bit easier since we only need to think about Fourier series rather than Fourier transforms, etc. – Yemon Choi Mar 13 '10 at 0:55
@Yemon: I added the second paragraph while you were commenting. (I had forgotten about the irreducible part of the question.) To get irreducible subspaces, go for the circle group T acting on L²(T), since the dual group Z certainly has atoms. – Harald Hanche-Olsen Mar 13 '10 at 1:04
@Harald: What about $L^2(\mathbb{R},\mathcal{B},\nu)$ for $\nu$ different than Lebesgue measure? – Steve Huntsman Mar 13 '10 at 2:18
@Steve: but you still want R to act on L^2(R) in a norm-preserving way, right? and this limits the choices... – Yemon Choi Mar 13 '10 at 4:27
@Yemon--Good point. – Steve Huntsman Mar 13 '10 at 14:55

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.