Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

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

My question seems to be easy but I cannot spot the answer. I am interested in ranges of operators defined on $c_0$. The celebrated "operator version" of Sobczyk's theorem says that if we are given a separable Banach space $X$ and its subspace $Y$, then every bounded operator $T\colon Y\to c_0$ can be extended to a bounded operator $\overline{T}$ with $\|\overline{T}\|\leq 2\|T\|$ (categorically speaking, $c_0$ is "separably injective"). I am wondering if I could use this theorem (or anything else) to (dis)prove the following conjecture:

If $X$ is a $c_0$-saturated separable Banach, then the range of every operator $T\colon c_0\to X$ embeds into $c_0$. We know that (consult Lindenstrauss and Tzafriri's book) every quotient of $c_0$ embeds into $c_0$ but how about ranges of this sort of operators?

share|cite|improve this question
There is a quite recent survey article on Sobczyk's theorem by Cabello Sánchez, Castillo and Yost. While I haven't read it in detail, it might contain some pointers to the literature. – t.b. Jun 16 '11 at 0:21
I'm really not sure if I understand your question correctly, but as it is stated I don't really see how the information that $X$ contains a complemented copy of $c_0$ should help. If $T: c_0 \to Y$ is any operator to a separable Banach space, simply consider $X = Y \oplus c_0$ and compose $T$ with the inclusion $Y \to X$. The range of $T$ will remain the same and certainly $X$ is separable. Also, it would be nice if you could make your question a bit more precise (e.g. what exactly does it mean for the range of $T$ to embed into $c_0$)? – t.b. Jun 16 '11 at 0:40
$T(X)$ isomorphic to a subspace of $c_0$. Right, I was thinking about $c_0$-saturation, since if $X$ contains a copy $c_0$ then in the separable setting it is automatically complemented. Btw, I know this paper quite well. – dziobak Jun 16 '11 at 6:45

Like Theo, I don't really understand what exactly it is that is being asked. I am quite sure that the OP knows that regardless of whether or not $X$ is $c_0$-saturated, an operator $T:c_0 \longrightarrow X$ with closed range has the property that $T(c_0)$ embeds into $c_0$ by the Johnson-Zippin result cited above in the comments to the question.

I am somewhat guessing that the question is: for a $c_0$-saturated $X$ and arbitrary $T:c_0 \longrightarrow X$, is the Banach space $\overline{T(c_0)}$ isomorphic to a subspace of $c_0$?

The answer to this question is no. For a counterexample, take $X$ to be $C(\omega^\omega)$ (or replace $\omega^\omega$ by any larger countable ordinal) and let $(\alpha_n)_{n=1}^\infty$ be an enumeration of $\omega^\omega+1$. For each $n\in \mathbb{N}$ let $\alpha_n^- = \min \{\beta \mid \exists \nu \mbox{ such that } \beta + \omega^\nu = \alpha_n\}$, so that the (continuous) characteristic functions $\chi_{(\alpha_n^-, \alpha_n]}$, $n\in\mathbb{N}$, span a dense linear subspace of $C(\omega^\omega)$. The map $e_n \mapsto 2^{-n}\chi_{(\alpha_n^-, \alpha_n]}$ extends to a (compact) continuous linear operator $T: c_0 \longrightarrow C(\omega^\omega)$ - with dense range. Since $C(\omega^\omega)$ does not embed in $c_0$, the claimed counterexample is achieved.

P.S. If you write $\alpha_n$ as a sum of powers of $\omega$ - i.e., in Cantor normal form - then $\alpha_n^-$ is just the ordinal attained by leaving off the last summand.

P.P.S. Apologies to anyone whose time has been wasted by this answer if I have guessed wrong with respect to the intent of the OP.

share|cite|improve this answer

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.