5
$\begingroup$

Consider a homological localization $L$ of $p$-local spaces or (I think equivalently) a localization of $p$-local spectra. If a space $X$ is $L$-acyclic, then so is $K(\pi_n(X),n)$ for each $n$. To what extent does the converse hold? That is, if $X$ is a space and $K(\pi_n(X),n)$ is $L$-acyclic for each $n$, then is $X$ $L$-acyclic?

If the converse always holds, then it's not known, because this would provide an affirmative answer to the telescope conjecture. So I suppose I should ask:

Question: What is an example of a pair of homological localizations $L,L'$ of $p$-local spaces which agree on Eilenberg-MacLane spaces but not on all spaces?

If $L,L'$ are not $H\mathbb Z_{(p)}$-localization, then they kill some $K(\mathbb Z/p,n+1)$ and thus (if $B\mathbb Q$ is acyclic) all $K(A,m)$'s for $m \geq n+2$. In this case, I believe that $L,L'$ agree on $n+1$-truncated spaces if and only if they agree on Eilenberg-MacLane spaces. On the other hand, if $X$ is $n+1$-connected, then the value of $LX$ or $L'X$ depends only on the $m$-connective cover $\tau_{\geq m} X$ for any $m \geq n+2$. So any difference between $L,L'$ is kind of hard to think about in terms of homotopy groups.

$\endgroup$
5
$\begingroup$

You ask two different questions. Regarding the first, there are easy counterexamples: e.g. $\widetilde K(1)_*(S^3) \neq 0$, but $\widetilde K(1)_*(K(\pi_n(S^3),n)) = 0$ for all $n$.

$\endgroup$

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.