Let $T\in \ell^1$, $Tx = (\lambda_1x_1,\dots,\lambda_nx_n,\dots)$. Want to show that if $T$ is compact, then $\lambda_n\to0$.

I know for $p\in(1,\infty]$, canonical basis $e_n \rightharpoonup 0$ (so $T(e_n)\rightharpoonup 0$), and $T(e_n)\in \overline{T(B(0,1))}$, which is compact, so we can get $T(e_n)\to 0$, and $\lambda_n\to 0$.

But what should I do with this problem when $p=1$, there is no weak convergence to $0$. Can someone help me with this? Thanks


Suppose that $T\colon\ell^1 \to \ell^1$ is compact. Then the sequence $\left(Te_n\right)_{n\geqslant 1}$ admits a subsequence $\left(Te_{n_k}\right)_{k\geqslant 1}$ which converges to some $v$ (strongly) in $\ell^1$. Look at $\left\lVert Te_{n_{k+1}}-Te_{n_k}\right\rVert_1$ to conclude that $\lambda_{n_k}\to 0$.

Now apply the previous result to $\left(Te_{N_j}\right)_{j\geqslant 1}$ for a fixed sequence $N_j\uparrow \infty$ instead of $\left(Te_n\right)_{n\geqslant 1}$ to see that each subsequence of $\left(\lambda_n\right)_{n\geqslant 1}$ admit a further subsequence with converges to $0$.


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.