3
$\begingroup$

Let $\mathsf{Top}_*$ be the category of well-based spaces and $\mathsf{TopMon}$ the category of topological monoids. Recall the James construction $\mathcal{J}:\mathsf{Top}_*\to \mathsf{TopMon}$ which is left adjoint to the forgetful functor.

Now let $A\hookrightarrow X$ be a cofibration of based spaces and let $M$ be a topological monoid as well as $f:A\to M$ a based map. I want to “attach” the free topological monoid $\mathcal{J}X$ to $M$ along $\mathcal{J}A$. More precisely, I want to consider the adjoint map $\mathcal{J}A\to M$ and form the pushout of $(\mathcal{J}X\leftarrow \mathcal{J}A\to M)$ in the category $\mathsf{TopMon}$.

Call this pushout $M'$. I am interested in the classifying space $BM'$. The naïve hope would be of course something like $$BM' = \mathrm{hocolim}_{\mathsf{Top}_*}\left(B\mathcal{J} X\leftarrow B\mathcal{J} A\to BM\right),$$ and then one would identify up to homotopy $B\mathcal{J}\simeq \Sigma$, where the occuring map $\Sigma A\to BM$ is just the composition $\Sigma A\to \Sigma M\to BM$ with $\Sigma M\to BM$ being induced by the inclusion $\Delta^1\times M\to |NM|=BM$ of the $1$-simplices (or in other words, $\Sigma M\to BM$ is the adjoint of the group completion $M\to \Omega B M$).

However, I guess that this is too nice to wish for. I found the following paper by Fiedorowicz in which he gives explicit conditions under which the classifying space of a discrete amalgamated monoid is a (homotopy) pushout of classifying spaces.

Is there anything known for the non-discrete case? What would we additionally have to assume for the inclusion $A\hookrightarrow X$ or the attaching map $f:A\to M$ to hold?

$\endgroup$
4
  • 2
    $\begingroup$ The paper of Fiedorowicz has some issues. See mathoverflow.net/questions/297583/… $\endgroup$ Sep 2, 2020 at 18:58
  • 1
    $\begingroup$ Possibly of interest: arxiv.org/abs/1203.4978 $\endgroup$ Sep 4, 2020 at 6:48
  • $\begingroup$ Thank you, @GustavoGranja! I think that this should help: Vogt proves that B preserves homotopy colimits up to homotopy equivalence. Now there is a left proper model structure on $\mathsf{TopMon}$ such that $\mathcal{J}$ is left Quillen, so $\mathcal{J}(A\hookrightarrow X)$ is a cofibration, whence $M'$ is indeed a homotopy pushout. Or am I missing something? $\endgroup$
    – FKranhold
    Sep 6, 2020 at 18:25
  • $\begingroup$ Well … one would have to argue why $M'$ is not only a homotopy pushout with respect to the model structure on $\mathsf{TopMon}$, but also a homotopy pushout in Vogt’s sense. Or is this clear? $\endgroup$
    – FKranhold
    Sep 6, 2020 at 22:12

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.