4
$\begingroup$

If we have a category $\mathcal{C}$, then we can see it as an $\infty$-category. Furthermore, we can truncate and $\infty$-category $\mathcal{X}$ to get a category $\mathcal{X}_{\leq 1}$. My question is if these functors are adjoint, i.e. if we have $$\text{Hom}_{\mathfrak{Cat}}(\mathcal{X}_{\leq 1}, \mathcal{Y})\cong \text{Hom}_{\infty\text{-}\mathfrak{Cat}}(\mathcal{X},\mathcal{Y})$$ where $\mathcal{Y}$ is a category (which on the right is seen as an $\infty$-category).

$\endgroup$
  • $\begingroup$ I think you need to edit a bit your question. What is Y? $\endgroup$ – Mirco A. Mannucci Sep 19 at 20:28
11
$\begingroup$

There is a bit of notation to be careful about here:

$\mathcal{X}_{\leqslant 1}$ is often used to denote the full subcategory of $\mathcal{X}$ of set-truncated object. For example if $\mathcal{X}$ is an $\infty$-topos, then $\mathcal{X}_{\leqslant 1}$ is its $1$-topos reflection.

with this definition, $\mathcal{X}_{\leqslant 1}$ is a $1$-category, but it is not the one that will have the property you want (it will be a right adjoint instead of a left adjoint, and only when restricted to finite limit preserving functor).

The $1$-category you want to consider is the homotopy category $h \mathcal{X}$ of $\mathcal{X}$, sometimes also denoted $\tau \mathcal{X}$, which is the category with the same objects as $\mathcal{X}$, and with the morphism sets

$$ h\mathcal{X}(a,b) \simeq \pi_0 ( \mathcal{X}(a,b) ) $$

Which does satisfies the property you ask.

A rigorous proof of this of course depends on what model of $\infty$-category you use, but if you use quasi-categories, this follows from points 1.2 and 1.8 in Joyal notes on quasi-categories.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Just to clarify: the functor $(-)_{\leq 1}$ which restricts the fundamental $\infty$-groupoid to the fundamental groupoid of a space has no right adjoint? I'm asking because I would like to use this to give a simple proof of Seifert-van Kampen: For an open cover $U,V$ of $X$, we have a pushout $U\cap V\rightarrow U\times V\rightarrow X$. If we use the homotopy hypothesis, we get a pushout $\Pi(U\cap V)\rightarrow \Pi(U)\times \Pi(V)\rightarrow \Pi(X)$. If the truncation functor $\tau_{\leq 1}(-)$ were adjoint, then the pushout would be preserved, yielding a nice proof of S-vK. $\endgroup$ – curious math guy Sep 19 at 21:31
  • 3
    $\begingroup$ The $\pi_1$ functor taking an $\infty$-groupoid to its fundamental groupoid is a left adjoint functor and this makes the groupoid version of S-vK. somehow trivial. (or at least it hides all the difficulty under the rug) $\endgroup$ – Simon Henry Sep 19 at 21:35
  • 1
    $\begingroup$ @curious: you do not get Seifert-van Kampen this way. What you get is that the fundamental groupoid functor preserves homotopy pushouts, but Seifert-van Kampen is a genuinely point-set result telling you when a point-set pushout is "close enough" to a homotopy pushout that the fundamental groupoid is what it should be. $\endgroup$ – Qiaochu Yuan Sep 19 at 21:49
  • 2
    $\begingroup$ @curious: you're working with $\infty$-categories and adjunctions between them and in $\infty$-category land you only have $\infty$-limits and colimits and those are what right adjoints resp. left adjoints preserve. $\endgroup$ – Qiaochu Yuan Sep 19 at 21:54
  • 1
    $\begingroup$ @QiaochuYuan Ah yes, thank you! $\endgroup$ – curious math guy Sep 19 at 21:55

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.