Let $\mathcal{C}$ be a small category and let $\mathcal{M} = \operatorname{sPre}(C)$ be the model category of simplicial presheaves on $\mathcal{C}$ with the injective model structure. Let $S$ be a set of morphisms in $\mathcal{M}$ that is stable under pullback and let $L_S\mathcal{M}$ be the left Bousfield localization of $\mathcal{M}$ with respect to $S$.

By this nlab page, $\mathcal{M}$ is a locally cartesian closed model category, and any right proper left Bousfield localization of $\mathcal{M}$ is as well. Is this true for any other left Bousfield localization $L_S\mathcal{M}$? If not, is it true when $S$ is (the saturation of) the set of covering sieves $R \hookrightarrow j(X)$ for some Grothendieck topology $t$ on $\mathcal{C}$, where $j(X)$ is the simplicial Yoneda embedding applied to $X \in \mathcal{C}$ (note that the local objects of $L_S\mathcal{M}$ are simplicial presheaves satisfying $t$-descent but not necessarily $t$-hyperdescent)? This is really the case I care about. If not, is there a counterexample?

It should be noted that $L_S\mathcal{M}$ is at least Quillen equivalent to a locally cartesian closed model category because the $\infty$-category it presents is locally cartesian closed by a result of Hoyois.

  • 1
    $\begingroup$ Any presentable $\infty$-category is obtained as a left Bousfield localization of simplicial presehaves (after Dugger's theorems). Since there are presentable $\infty$-categories that are not locally cartesian closed, the answer to the question of the title is obviously no. For the case of covering sieves, we get a proper model structure indeed. $\endgroup$ Commented Aug 21, 2023 at 14:17
  • $\begingroup$ @D.-C.Cisinski I thought the covering sieves only yield a left proper model structure, did I get that wrong? I know that for hypercovers the model structure is also right proper by an argument of Jardine. $\endgroup$ Commented Aug 22, 2023 at 13:32
  • $\begingroup$ This nLab page gives a reference to check right properness: ncatlab.org/nlab/show/Cisinski+model+structure#properness $\endgroup$ Commented Aug 23, 2023 at 6:30
  • $\begingroup$ From there, we see that any left Bousfielf localization of simplicial presheaves that induces a left exact localization at the level of $\infty$-categories is right proper. In particular, all higher topoi are obtained in this way. The left Bousfield localization by covering sieves is a particular case. $\endgroup$ Commented Aug 23, 2023 at 6:45
  • $\begingroup$ @D.-C.Cisinski Thank you so much for clarifying this, I'd been wondering about the connection to left exact localizations of $\infty$-categories. $\endgroup$ Commented Aug 23, 2023 at 10:12

1 Answer 1


Section 2 of this paper of Rezk addresses exactly the question of when the localization by S yields a Cartesian model category. For that the relevant property is that that if you take the product of a map in S and a representable object, the result must be in the saturation of S.

There are examples of S which fail this condition where the resulting localization is not a Cartesian model category. (An example is the n-fold Segal space model category constructed by Barwick - it is a simplicial model category but not Cartesian).

For local Cartesian-ness you are essentially talking about Cartesian-ness in the over categories $\mathcal{M}_{/X}$. There the same argument works where stability under products with representables is replaced with stability under fiber products of representables over X. So if your set of morphisms S is stable under all pullbacks, as mentioned in the OP, that is sufficient to give the local cartesian model category structure on $L_S\mathcal{M}$.

  • $\begingroup$ Fantastic, that's exactly what I was looking for. Thank you! $\endgroup$ Commented Aug 22, 2023 at 13:29

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.