2
$\begingroup$

Let $Y_1$ and $Y_2$ be closed subspaces of a normed space $X$ (which may be infinite dimensional). How can I show that $Y_1$ and $Y_2$ must have different annihilators or else be the same subspace?

The annihilator $Y^0$ of a subspace $Y \subset X$ is defined as the set of functionals in $X'$ (the dual space of $X$) such that $f(y)=0\forall y \in Y, f \in Y^0$.

(Note: I have looked for similar questions, but all I have found are proofs that assume the space to be finite dimensional. I found a mention that this can be proved as a corollary to the Hahn-Banach theorem - how?)

$\endgroup$
4
$\begingroup$

If $Y_1 \neq Y_2$, there is some $y_1 \in Y_1$ with $y_1 \not\in Y_2$. By Hanh-Banach Theorem, there exists $f \in X^*$ such that $f(y_1) \neq 0$ and $f|_{Y_2}$ = 0. So $f \in (Y_2)^0$ but $f \not\in (Y_1)^0$.

$\endgroup$
  • $\begingroup$ Oh that is really nice. I was trying to think of a way to prove it, but I should have used hahn banach more carefully. $\endgroup$ – Andres Mejia Oct 26 '17 at 16:11
  • $\begingroup$ Hahn-Banach allows you to make an extension, but I don't understand how you automatically know that an extension exists such that $f(y_1) \neq 0$. $\endgroup$ – Johnny Hansen Oct 26 '17 at 16:43
  • $\begingroup$ @JohnnyHansen: This is a standard corollary of Hanh-Banach Theorem. In most textbooks, it is treated right after the exposition of Hanh-Banach Theorem. Check Rudin's Real and Complex Analysis or Folland's Real Analysis. $\endgroup$ – user1101010 Oct 26 '17 at 16:52

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.