Let $H$ be a Hilbert Space, and $M$ a closed subspace. Is it true that

$H = M \bigoplus M^{\perp}$

Does this hold if $M$ is not closed? Or only if $H$ is finite/infinite dimensional?

  • $\begingroup$ How exactly are you defining $\oplus$? $\endgroup$ – Nate Eldredge May 5 '12 at 23:57
  • $\begingroup$ I hadn't considered this, just the usual Cartesian product of the underlying sets I assumed? My lecturer didn't define it. What do you think? $\endgroup$ – rk101 May 6 '12 at 0:44
  • $\begingroup$ Well, some definitions say you take the completion afterwards. If you don't then it certainly won't be true for non-closed $M$. (For example, if $M$ is dense in $H$ then $M^\perp = 0$.) But if you do, then it will be true. $\endgroup$ – Nate Eldredge May 6 '12 at 1:54

Yes, this is true. Here's how you can go about proving this:

  1. Prove that if $N\leqslant M^\perp$ then $M\oplus N$ is closed.
  2. Prove that if $N\leqslant H$ is closed and $N^\perp=0$ then $N=H$.
  3. Conclude that $M\oplus M^\perp$ is closed and $(M\oplus M^\perp)^\perp=0$.
  • $\begingroup$ SO this is true in both infinite and finite dimensional Hilbert Spaces? What about just Inner product spaces? $\endgroup$ – rk101 May 5 '12 at 23:39

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.