Consider the usual model structure on Cat (category of small categories). Which are the fibrations of the injective model structure on the category of cosimplicial categories $Fun( \Delta ,Cat )$?

Actually, I wonder how can I compute homotopy limits of cosimplicial categories (considering them as diagrams in Cat)...

Edit: I'm most interested in understanding how to compute the homotopy limit of a cosimplicial diagram in $Cat$

Thank you in advance

  • 1
    $\begingroup$ I doubt that taking Cat as a target category may simplify the characterization of fibrations in the injective model structure on cosimplicial diagrams. I'd say you should proceed as in the general case. $\endgroup$ – Fernando Muro Feb 18 '14 at 16:53
  • $\begingroup$ @Fernando Muro, Thank you (again) for your comment. Actually, I'm more interested in understanding how to compute homotopy limits of cosimplicial diagrams in $Cat$. Do you know if there is an easy way to do so? Thank you very much $\endgroup$ – Fernando Feb 18 '14 at 19:40
  • 1
    $\begingroup$ The standard model structure on $\mathbf{Cat}$ is indeed simplicial, provided one uses as the nerve of the maximal subgroupoid of $[\mathcal{A}, \mathcal{B}]$ as the simplicial hom-space $\underline{\mathrm{Hom}}(\mathcal{A}, \mathcal{B})$: see this note of Rezk. Thus one can use the standard technique of bar and cobar constructions: see e.g. this preprint of Shulman). Note that the existence of injective/projective model structures is not needed. $\endgroup$ – Zhen Lin Feb 19 '14 at 10:25
  • 1
    $\begingroup$ Maybe section 6.5 in pages.uoregon.edu/ddugger/hocolim.pdf can help. In general, this paper is a very nice account on hocolim (just dualize for holim). $\endgroup$ – Fernando Muro Feb 19 '14 at 11:33
  • 1
    $\begingroup$ Lemma 3.4.11 in [Sketches of an elephant, Part B] shows that a certain non-full subcategory of $\mathbf{\Delta}$ is 2-coinitial. $\endgroup$ – Zhen Lin Feb 19 '14 at 21:30

I will answer the title question about computing homotopy limits. Let $A : \mathcal{C} \to \mathbf{Cat}$ be a small (strict) diagram. By expanding the definitions of the cobar construction, one eventually discovers that $\operatorname{holim} A$ can be computed as the end $$\int_{[n] : \mathbf{\Delta}} \left[ \mathbf{I} [n], \prod_{(c_0, \ldots, c_n)} \mathcal{C} (c_n, c_{n-1}) \times \cdots \times \mathcal{C} (c_1, c_0), A c_0 \right]$$ where $\mathbf{I} [n]$ is the contractible groupoid with $n + 1$ objects. In more familiar terms, this is just the hom-category of all morphisms between a certain pair of cosimplicial categories.

Here is a description of an object in $\operatorname{holim} A$:

  • For each object $c$ in $\mathcal{C}$, we have an object $a_c$ in $A c$.
  • For each morphism $f : c_1 \to c_0$ in $\mathcal{C}$, we have an isomorphism $\mu_f : f (a_{c_1}) \to a_{c_0}$ in $A c_0$; and $\mu_{\mathrm{id}_c} = \mathrm{id}_{a_c}$ for all objects $c$ in $\mathcal{C}$.
  • For each composable pair $f_1 : c_2 \to c_1, f_0 : c_1 \to c_0$ in $\mathcal{C}$, we have a commutative triangle in $A c_0$, $$\begin{array}{ccc} f_0 (f_1 (a_{c_2})) & \rightarrow & f_0 (a_{c_1}) \\ & \searrow & \downarrow \\ && a_{c_0} \end{array}$$ i.e. $\mu_{f_0 \circ f_1} = \mu_{f_0} \circ f_0 (\mu_{f_1})$.
  • For each composable triple in $\mathcal{C}$, we have a commutative tetrahedron in $A c_0$.
  • etc.

In fact, the coherence conditions above degree 2 are automatic if all the triangles commute. Thus we see that an object in $\operatorname{holim} A$ is the same thing as a pseudocone over $A$ (where we regard $A$ as a pseudofunctor with the canonical coherence data). Unsurprisingly, the morphisms are the same, so $\operatorname{holim} A$ (as constructed above) is the pseudolimit of $A$.

  • $\begingroup$ Thank you. Your answer helped me a lot. So, your conclusion was that the homotopy limit is the same concept of the pseudolimit of a strict diagram. I would like to ask, nevertheless, if there is a "end" formula like that of cobar construction for computing pseudolimits of pseudolimit of a "pseudodiagram" (non-strict diagram). $\endgroup$ – Fernando Mar 6 '14 at 3:07
  • 1
    $\begingroup$ I do not know of such a formula. But there should be no real difference other than inserting coherence isomorphisms in the appropriate places. $\endgroup$ – Zhen Lin Mar 6 '14 at 8:14
  • $\begingroup$ What is this notation inside the end, which consists of square brackets with two commas inside? $[ , , ]$? $\endgroup$ – Dean Young Feb 27 at 7:48

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.