0
$\begingroup$

If a Banach space is

isomorphic in the vector and norm sense

to another Banach space, does this imply the two are isometrically isomorphic?

I think so, because: from Wikipedia

  • "an isometry [...] is a linear map f which preserves the norm"
  • "a surjective isometry [...] is called an isometric isomorphism"

and Definition 1.3 from here

  • an isomorphism between two vector spaces is a map that is one-to-one and onto and preserves structure.

So if two Banach spaces are isomorphic in both the vector sense and the norm sense, they must be isometrically isomorphic?

$\endgroup$
  • $\begingroup$ Isomorphic in the norm sense? Does it mean homeomorphic? Or that there is a norm preserving bijection (not necessarly linear)? $\endgroup$ – freakish Sep 13 '17 at 8:29
  • $\begingroup$ @freakish, I don't know. This is precisely my question, how to interpret the clause "isomorphic in the vector and norm sense". $\endgroup$ – Carucel Sep 13 '17 at 8:31
2
$\begingroup$

Let $X$ and $Y$ be the two Banach spaces in question.

Isomorphic in the vector sense means that there exists a linear bijection $\Phi : X \to Y$. Notice that $\Phi^{-1}$ is also a linear bijection.

Isomorphic in the norm sense could mean several things:

  1. There exists a bijective linear isometry $T : X \to Y$, meaning that $T$ is norm preserving: $$\|Tx\|_{Y} = \|x\|_X, \quad\forall x\in X$$ Notice that $T^{-1}$ is also a norm preserving linear bijection.
    In this case we would say that $X$ and $Y$ are isometrically isomorphic.
  2. There exists a bounded (continuous) linear bijection $T : X \to Y$. As a consequence of the Open Mapping Theorem, $T^{-1}$ is also a bounded liner bijection.
  3. There exists a bijective isometry $f : X \to Y$, i.e. norm preserving but not necessarily linear. $f^{-1}$ is also isometric. As a consequence of the Mazur-Ulam Theorem, the map $f$ is necessarily affine. Thus, if we also have $f(0) = 0$, then $f$ is a linear isometry so we are it the situation $(1)$.

In fact, if $X$ and $Y$ are real spaces then $(3) \implies (2)$.

$\endgroup$
  • $\begingroup$ Thank you! If I understand the clause isomorphic in the vector and norm sense is ambigiuous. $\endgroup$ – Carucel Sep 13 '17 at 10:16
  • 1
    $\begingroup$ @Carucel Yes, although the most probable intention is that they are isometrically isomorphic, IMO. $\endgroup$ – mechanodroid Sep 13 '17 at 10:18

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.