4
$\begingroup$

The classical setting.

Given a monoid $A$, there's a category $\mathbf{B}A$, called the delooping of $A$, having a single object $\star$ and satisfying $\mathrm{Hom}_{\mathbf{B}A}(\star,\star)\overset{\mathrm{def}}{=}A$, with composition given by multiplication and the sole identity $\mathrm{id}_{\star}$ given by $1_A$.

This construction is characterised by the following property: for any other category $\mathcal{C}$, we have a bijection of sets $$ \left\{ \begin{gathered} \text{functors}\\ \mathbf{B}A\to\mathcal{C} \end{gathered} \right\} \cong \left\{ \begin{aligned} &\text{pairs $(X,\phi)$ with}\\ &\,\,\,\,\,\,\,\text{- $X$ an object of $\mathcal{C}$;}\\ &\,\,\,\,\,\,\,\text{- $\phi$ a morphism of monoids}\\ &\text{from $A$ to $\left(\mathrm{Hom}_{\mathcal{C}}(X,X),\circ,\mathrm{id}_{X}\right)$.} \end{aligned} \right\}. $$ For example:

  • A functor $\mathbf{B}\mathbb{N}\to\mathcal{C}$ is the same as an endomorphism $A\to A$ of $\mathcal{C}$;
  • A functor $\mathbf{B}\mathbb{Z}\to\mathcal{C}$ is the same as an automorphism $A\to A$ of $\mathcal{C}$;
  • A functor $\mathbf{B}\mathbb{Z}/2\to\mathcal{C}$ is the same as an involution $A\to A$ of $\mathcal{C}$;
  • A functor $\mathbf{B}\mathbb{B}\to\mathcal{C}$ is the same as an idempotent $A\to A$ of $\mathcal{C}$, where $\mathbb{B}=(\{0,1\},\text{OR},1)$.

The $\infty$-categorical setting.

Preliminary Question. Given an $\infty$-category $\mathcal{C}$, is there a natural monoidal $\infty$-groupoid structure on $\mathrm{Hom}_{\mathcal{C}}(X,X)$?

Question. Is there an analogue of deloopings for $(\infty,1)$-categories, where we start with a monoidal $\infty$-groupoid $\mathcal{C}$ and construct an $(\infty,1)$-category $\mathbf{B}\mathcal{C}$ such that

  • A functor $\mathbf{B}\mathcal{C}\to\mathcal{D}$ from $\mathbf{B}\mathcal{C}$ to another $(\infty,1)$-category $\mathcal{D}$;

is the same thing as

  • An object of $\mathcal{D}$ together with a functor of monoidal (?) $\infty$-groupoids $\mathcal{C}\to\mathrm{Hom}_{\mathcal{D}}(X,X)$,

with $\mathrm{Hom}_{\mathcal{D}}(X,X)$ the morphism space of $\mathcal{D}$ from $X$ to itself, and where this bijection can be made into a full-fledged isomorphism of (appropriate) $\infty$-groupoids?

$\endgroup$
3
  • $\begingroup$ See also Delooping monoidal $(\infty,1)$-categories into $(\infty,2)$-categories. $\endgroup$
    – Emily
    Commented Sep 18, 2021 at 2:26
  • $\begingroup$ If you think of Infinity categories as simplicial categories and Infinity groupoids as simplicial sets you'll realize that the answer is clearly 'yes' $\endgroup$ Commented Sep 18, 2021 at 9:30
  • 2
    $\begingroup$ @FernandoMuro To be fair this argument passes through a non-trivial strictification theorem (that every $E_1$-space can be represented by a simplicial monoid) $\endgroup$ Commented Sep 18, 2021 at 9:31

1 Answer 1

3
$\begingroup$

I assume that with ``monoidal ∞-groupoid'' you mean an $E_1$-space. In this case the answer is yes. It is well known that $E_1$-spaces can be modeled by functors

$$X:\Delta^{\mathrm{op}}\to \operatorname{Space}$$ satisfying the Segal conditions. Now if you are given an $\infty$-category $\mathcal{C}$ you can define a simplicial space $$s(\mathcal{C}): [n]\mapsto\operatorname{Map}_{\operatorname{Cat}_∞}(\Delta^n,\mathcal{C})\,.$$ In fact this functor is fully faithful and identifies $\operatorname{Cat}_∞$ with the category of complete Segal spaces. For the following we won't need all this though - we will use only that it takes values in Segal spaces (which follows immediately from $\Delta^n\amalg_{\Delta^0} \Delta^m\simeq\Delta^{n+m-1}$ in $\operatorname{Cat}_∞$).

Now let $x\in\mathcal{C}$ be an object of $\mathcal{C}$. Then we can define the simplicial space $$ \operatorname{End}_{\mathcal{C}}(x):\Delta^{\mathrm{op}}\to \operatorname{Space}\qquad [n]\mapsto \{x\}\times_{\operatorname{Map}(\{0,\dots,n\},\mathcal{C})} \operatorname{Map}_{\operatorname{Cat}_∞}(\Delta^n,\mathcal{C})\,.$$ That is it sends $[n]$ to the (∞-)groupoid of functors $F:\Delta^n\to \mathcal{C}$ that sends all objects to $x$. It is easy now to see that $\operatorname{End}_{\mathcal{C}}(x)$ satisfies the Segal conditions and so it is an $E_1$-space.


This takes care of your preliminary question. To go back to your main question, the functor $(\mathcal{C},x)\mapsto \operatorname{End}_{\mathcal{C}}(x)$ obviously preserves all limits and filtered colimits, and so it has a left adjoint $B$ exactly as you wanted. To get a more ``concrete'' description $B$ sends an $E_1$-space $X$ to the ∞-category corresponding to the completion of $X$ seen as a Segal space. That is $$BX:=\int^{[n]\in\Delta^{\mathrm{op}}} X([n])\times \Delta^n$$ where the coend is computed in $\operatorname{Cat}_∞$.

With more care one can show that $B:E_1-\operatorname{Space}\to(\operatorname{Cat}_∞)_{\Delta^0/}$ is fully faithful with essential images those arrows $\Delta^0\to\mathcal{C}$ that are essentially surjective (that is such that $\mathcal{C}$ has only one equivalence class of objects). Indeed this is a special case of the equivalence between the ∞-category of Segal spaces and the ∞-category of flagged ∞-categories.

$\endgroup$
1
  • $\begingroup$ This is perfect! Thank you so much, Denis! :) $\endgroup$
    – Emily
    Commented Sep 18, 2021 at 18:32

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.