6
$\begingroup$

$\DeclareMathOperator\Hom{Hom}$Let $X$ be a condensed set in the sense of Clausen-Scholze. If there is a universal anima $Y$ (or $\infty$ groupoid, or homotopy type) together with a map of condensed anima $X \to \pi^*Y$ that induces an equivalence $$\Hom(Y,Z) \to^\sim \Hom(X,\pi^* Z)$$ for every anima $Z$, we say that $Y = \pi_\#X$ is anima (or homotopy type) associated to $X$ and that the left adjoint $\pi_\#$ is defined on $X$

In Section 11 of "Lectures on Analytic Geometry" it is proved that $\pi_\#$ is defined on the subcategory of CW-complexes. Is it true more generally that $\pi_\#$ is defined (and equals the usual weak homotopy type) on the subcategory of locally contractible compactly generated Hausdorff spaces?

If not, what is the largest class of topological spaces on which this partial adjoint is defined?

$\endgroup$
5
  • 1
    $\begingroup$ Closely related (in particular, the comments) mathoverflow.net/q/388767 $\endgroup$
    – Z. M
    Commented Nov 10, 2022 at 17:39
  • $\begingroup$ @Z.M. Which comments? On a first look, I don't see anything going beyond CW-complexes $\endgroup$ Commented Nov 10, 2022 at 17:53
  • 1
    $\begingroup$ In 7.3 of arxiv.org/pdf/2105.07888.pdf, the authors prove that the inclusion Ani $\hookrightarrow$ Cond(Ani) does not preserve limits and hence the left adjoint can’t be defined everywhere. On the level of locally contractible CH Hausdorff spaces, you should be able to assume the space is compact by a filtered colimit argument. I’m not entirely sure if it’s true or not for those spaces, but hopefully this is at least a little helpful. $\endgroup$ Commented Nov 10, 2022 at 18:28
  • 1
    $\begingroup$ Dustin Clausen's comment (you need to click that link). Note that this property is false for Cantor sets. $\endgroup$
    – Z. M
    Commented Nov 10, 2022 at 18:59
  • $\begingroup$ Ok, for reference I have looked at the discussion at n-category cafe and the linked paper (in part based on that discussion). So I am aware that the left adjoint is not defined in general, but I would like to know about locally contractible spaces. $\endgroup$ Commented Nov 10, 2022 at 19:22

1 Answer 1

5
$\begingroup$

Great question!

The answer is Yes. Let me elaborate a little. The question is more generally about the left adjoint to the inclusion $\mathrm{An}\to \mathrm{CondAn}$ from anima to condensed anima. This left adjoint is only partially defined; what exists in general is the functor $F: \mathrm{CondAn}\to \mathrm{Pro}(\mathrm{An})$ from condensed anima to pro-anima. Then $F(X)$ is an anima precisely when the left adjoint to $\mathrm{An}\to \mathrm{CondAn}$ exists on $X$, in which case it equals $F(X)$.

Lemma. The functor $F$ inverts the map $X\times [0,1]\to X$ for any condensed anima $X$.

In other words, for any condensed anima $X$ and any anima $Y$, the map $$ \mathrm{Hom}(X,Y)\to \mathrm{Hom}(X\times [0,1],Y) $$ is an isomorphism. Writing $X$ as a colimit of profinite sets $S$, this reduces to the case $X=S$. But then the condensed anima $\mathrm{Hom}(S,Y)$ is actually itself an anima (if $S=\mathrm{lim}_i S_i$, it is the colimit $\mathrm{colim}_i \mathrm{Hom}(S_i,Y)$), and by adjunction one reduces to the case that $X=\ast$. In this case, this is part of Lemma 11.9 in Analytic Geometry.

The lemma shows that $F$ factors over a functor $$\mathrm{CondAn}[W^{-1}]\to \mathrm{Pro}(\mathrm{An})$$ from the $\infty$-category obtained from $\mathrm{CondAn}$ by inverting homotopy equivalences.

In particular, any condensed anima $X$ which is homotopy equivalent to an anima $Y$ has the property that $F(X)=Y$. In particular, if $X$ comes from a contractible topological space, then $F(X)=\ast$.

Now if $X$ is just locally contractible, then one can find a cover $X=\bigcup_i U_i$ by contractible $U_i$, and then covers $U_i\cap U_j=\bigcup_k U_{ijk}$, etc., leading to a hypercover of $X$ by disjoint unions of contractible $U\subset X$. As $F$ commutes with colimits, this writes $F(X)$ as a colimit of disjoint unions of points, and hence $F(X)$ is an anima (which is the usual (weak) homotopy type).

$\endgroup$
3
  • $\begingroup$ When does this pro-anima coincide with the étale topos of $X$? I might have seen something like this. $\endgroup$
    – Z. M
    Commented Nov 11, 2022 at 21:44
  • $\begingroup$ I'm not sure exactly what you're asking, but for any condensed anima $X$, one can define an $\infty$-topos of relatively discrete condensed anima over $X$ (this is some kind of etale topos of $X$). Then this pro-anima is the shape of this $\infty$-topos. It's also the shape of the slice (condensed anima over $X$). $\endgroup$ Commented Nov 12, 2022 at 20:15
  • $\begingroup$ Thanks, that was precisely what I wanted to ask (weird that I removed the word "shape" for some reasons), and Lurie's Higher Algebra, Appendix A is devoted to this. In particular, there are very general conditions (incorporating locally contractible spaces) under which this shape is precisely given by the singular complex. $\endgroup$
    – Z. M
    Commented Nov 12, 2022 at 20:58

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .