3
$\begingroup$

Suppose $X$ and $Y$ are Banach spaces, and $X \stackrel{\Phi}{\hookrightarrow} Y$ is an embedding (continuous linear injection). Let $\Phi' : Y' \rightarrow X'$ be the induced map on the dual spaces.

If $\Phi(X)$ is closed, then by a Hahn-Banach argument, $\Phi'$ is surjective, i.e. a bounded linear functional on $\Phi(X)$ can be extended to $Y$.

Question: If $\Phi(X)$ is not closed, the Hahn-Banach argument breaks down, correct? If so, counter-example where $\Phi$ is not surjective?

$\endgroup$
  • $\begingroup$ Note that "Hahn" is not spelled "Hanh". $\endgroup$ – gerw Jan 13 '17 at 7:55
3
$\begingroup$

For example, let $X$ and $Y$ be $\ell^2$, the map $\Phi$ being $(x_n)_{n=1}^\infty\mapsto (x_n/n)_{n=1}^\infty$.

The induced $\Phi'$ is defined by the same formula (essentially the same map), and it's not surjective as, for example, $(1/n)_{n=1}^\infty$ is not in its range.

$\endgroup$
  • $\begingroup$ That's a compact self adjoint operator, with spectrum containing 0...is it an embedding? $\endgroup$ – Michael Jan 13 '17 at 9:51
  • $\begingroup$ @Michael, given that you only want a continuous injection, it certainly meets these criteria. $\endgroup$ – Tomek Kania Jan 13 '17 at 10:32
  • $\begingroup$ @TomekKania Right, 0 not in the point spectrum. $\endgroup$ – Michael Jan 13 '17 at 12:51

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.