3
$\begingroup$

Fix a category or $\infty$-category $C$ with all small limits.

We call a natural transformation $\alpha\colon f\to g$ between two functors $f,g\colon K\to C$ Cartesian, if for every arrow $k\colon a\to b$ in $K$, the naturality square of $\alpha$ over $k$ identifies $f(a)$ with the fiber product $g(a)\times_{g(b)}f(b)$, i.e. every naturality square of $\alpha$ is a pullback square in $C$.

Denote by $\overline K$ the cone on $K$ (i.e. $K$ with an additional initial object $\star$). For every diagram (a.k.a. functor) $f\colon K\to C$ we denote by $\overline f\colon\overline K\to C$ the right Kan extensions of $f$ along the inclusion $K\hookrightarrow \overline K$; in other words, $\overline f$ is a cone that exhibits $f(\star)\in C$ as the limit of the diagram $f$.

The limit cone $\overline f\colon\overline K\to C$ is called absolute, if for every functor $F\colon C\to D$, the composition $F\circ \overline f\colon \overline K \to D$ is still a limit cone.

Given a natural transformation $\alpha\colon f\to g$ between two diagrams $f,g\colon K\to C$, there is a unique transformation $\overline \alpha\colon \overline f\to\overline g$ which restricts to $\alpha$ and is induced by the universal property of the limit cone.

Question: Let $\overline f,\overline g\colon \overline K\to C$ be absolute limit cones extending diagrams $f,g\colon K\to C$. If $\alpha\colon f\to g$ is a Cartesian transformation, does it follow that $\overline \alpha\colon\overline f\to \overline g$ is also Cartesian?

While I am interested in the general statement, I am mostly concerned about the special case where $K$ is the simplex category $\Delta$, hence $\overline K$ is the augmented simplex category $\Delta_+$.

$\endgroup$
2
  • 3
    $\begingroup$ I don't think this is true. Take $K=\{a,b\}$ with no non-identity morphisms. Then a natural transformation being cartesian is not a condition any more. The right kan extensions will record the product in both cases, and then you're asking that the induced map on products is pulled back along either of the projections- which certainly needn't be true. $\endgroup$ Commented Feb 7, 2019 at 19:18
  • $\begingroup$ You are right. I'll edit the question and only ask about the absolute limit cones. $\endgroup$ Commented Feb 7, 2019 at 19:48

0

You must log in to answer this question.