11
$\begingroup$

The Thomason model structure on the category of small categories is transferred from the Quillen model structure on simplicial sets along the right adjoint $Ex^2 \circ N$ (where $N$ is the nerve), i.e. weak equivalences and fibrations are created by this functor. It is known that this model structure is cofibrantly generated (by the left adjoint $c \circ Sd^2$ applied to the usual generators in $sSet$) and is Quillen equivalent to $sSet$ along the adjoint pair $(c \circ Sd^2, Ex^2 \circ N)$. Furthermore, $Cat$ has the structure of a category enriched, tensored, and cotensored over $sSet$ (one such structure is described in Rezk's note).

Is there such a structure on $Cat$ that satisfies Quillen's SM7 axiom to make the Thomason model structure into a simplicial model category?

Some sufficient conditions to check can be found in Lurie's Higher Algebra, Prop 4.1.8.3, but don't seem likely to be satisfied in this case.

Similarly, $Cat$ has two monoidal structures: the Cartesian one and the "funny" one.

Do either of the monoidal structures satisfy the pushout product axiom to make the Thomason model structure into a monoidal model category?

On the one hand, its homotopy category certainly is monoidal. On the other hand, the cofibrations don't seem to be super well understood (though, it is known that every cofibrant object is a poset), and I don't know much about the monoidal properties of the functors $Ex^2$ and $N$.

UPDATE: Tim's answer below has an attractive argument, but the comments lay out the ways it's wrong. First, Tim's argument fails between the two displayed diagrams. Secondly, even if the Cartesian structure did satisfy the pushout product axiom, the counterexample shows that one simplicially enriched structure on Cat (the one you get from applying the nerve functor to the Cartesian internal hom) fails the SM7 axiom. However, there does not seem to be a counterexample yet to the question of whether or not the Cartesian structure satisfies the pushout product axiom, and we still don't know if there exists a simplicial structure that satisfies the SM7 axiom. As far as I can tell, both questions in the current thread are still open.

Improve this question
$\endgroup$
5
  • 4
    $\begingroup$ $\operatorname{Ex^2}$ preserves products, as does the nerve. Both are right adjoints. Does this help? $\endgroup$
    – Harry Gindi
    Commented Nov 25, 2018 at 23:19
  • 3
    $\begingroup$ The natural way to do this would be to have the given Quillen adjunction be an equivalence of $\mathrm{SSet}$-enriched model categories, which would amount to it preserving the simplicial enrichment. The usual way this would happen would be if left adjoint preserved finite products, which it does not-for instance, look at a square. So I'm skeptical whether this can be done. $\endgroup$
    – Kevin Carlson
    Commented Nov 26, 2018 at 4:36
  • $\begingroup$ What is the "funny" monoidal structure? $\endgroup$
    – Cihan
    Commented Nov 29, 2018 at 15:08
  • 2
    $\begingroup$ @Cihan ncatlab.org/nlab/show/funny+tensor+product $\endgroup$
    – Tim Campion
    Commented Dec 5, 2018 at 20:46
  • 2
    $\begingroup$ I just want to record here a lesson I learned from Kevin Carlson in my now-deleted answer: $sd\dashv Ex : sSet \to sSet$ is an adjunction, but it's not a simplicial adjunction! This sinks the attempted proof I had. $\endgroup$
    – Tim Campion
    Commented Jan 12, 2019 at 2:49

1 Answer 1

Reset to default
4
$\begingroup$

The answer to this question is NO. In particular, the discussion after Corollary 3.7 of George Raptis's paper Homotopy theory of posets provides a concrete example of a cofibration (namely, $f: [0] \to [1]$) such that the pushout product of $f$ with itself with respect to the Cartesian monoidal structure, is not a cofibration. I think this example also shows that the Thomason model structure is not a simplicial model category (fails the SM7 axiom). Furthermore, the same example shows the funny monoidal structure fails the pushout product axiom, because taking the funny pushout product of this $f$ with the identity on $[1]$ (which is a cofibrant object) yields a map from $[1]$ to a square, which is not a weak equivalence.

Improve this answer
$\endgroup$
2
  • $\begingroup$ In Section 5 of the preprint arxiv.org/abs/1605.07983v1, Yuzhou Gu relies on the "Ex^2 N" simplicial enrichment, and seems to fall into the trap from Tim's deleted answer. Gu uses a 1987 paper of Farjoun that extends a 1984 paper of Dwyer and Kan, to produce a simplicial model structure (with SM7) and in Remark 5.11 claims this recovers the Thomason model structure (with SM7). Rather than a conversation with myself on MO, I will investigate further and include a remark in a future paper, explaining whether or not there's an error in Gu's paper. $\endgroup$
    – David White
    Commented May 25 at 15:45
  • $\begingroup$ A related paper (that also builds on Raptis's example) is arxiv.org/pdf/2208.02954. It also shows that the Thomason model structure does satisfy the acyclic cofibration part of the pushout product axiom, despite failing the nonacyclic part. $\endgroup$
    – David White
    Commented Aug 19 at 16:55

You must log in to answer this question.

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