Take the 2-minute tour ×
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.

It follows from Krivine's theorem for a given Banach space $X$, either some $\ell_p$ or $c_0$ is finitely represented in $X$. Since finite dimensional $\ell_1$ and $\ell_\infty$-spaces are dual to each other, can we expect that

$\ell_1$ is finitely representable in $X$ if and only if $c_0$ is finitely representable in $X^*$?

share|improve this question

1 Answer 1

up vote 2 down vote accepted

No, $X=c_0$ with $X^*=\ell_1$ is a counterexample. The space $\ell_1$ isometrically embeds into $\ell_\infty$. For any Banach space $Y$ the bidual $Y^{**}$ is finitely representable in $Y$ (the principle of local reflexivity). Since $\ell_\infty=c_0^{**}$ we conclude that $\ell_\infty$, and a fortiori $\ell_1 $ is finitely representable in $X$.

On the other hand, $c_0$ is not finitely representable in $X^*$. Indeed, $\ell_1$ has cotype $2$, while $c_0$ does not; in fact both type and cotype of $c_0$ are trivial because $\ell_\infty$ is finitely representable in it.

But the other direction holds: if $\ell_1$ is not finitely representable in $X$, then the type of $X$ is strictly greater than $1$, by the Maurey-Pisier theorem. It follows that the cotype of $X^*$ is finite, and therefore $c_0$ is not finitely representable in $X^*$.

Reference on type and cotype.


Added a reflexive example. Let $X=\bigoplus_{n=2}^\infty \ell_n^n$; so that $X^*=\bigoplus_{n=2}^\infty \ell_{n/(n-1)}^n$. (The direct sums are $\ell_2$-normed.) Again, $c_0$ is not finitely representable in $X^*$ because $X^*$ has cotype $2$. As for $X$, it has arbitrarily large subspaces that are arbitrarily close to being in $\ell_\infty$, which implies that $\ell_1$ is finitely representable in it.

share|improve this answer
    
Thank you... Actually I am mostly interested in reflexive spaces. Do we know anything in this realm? –  user60253 Feb 21 '13 at 21:09
    
@user60253 Reflexivity does not help with local properties of Banach spaces, unless it's super-reflexivity. // Also, I see you asked 10 questions on this site; why not create a registered account here? It can be just as anonymous as an unregistered account, but more stable: unregistered accounts tend to get lost when cookies are wiped out. –  user53153 Feb 21 '13 at 21:18

Your Answer

 
discard

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.