4
$\begingroup$

Consider the category $Sp_{T(h)}$ of $T(h)$-local spectra. Let $J, K$ be small $\infty$-categories. Recall that $J$-limits said to commute with $K$-colimits in $Sp_{T(h)}$ if, for all functors $F : J \times K \to Sp_{T(h)}$, the canonical map

$$\varinjlim_{k \in K} \varprojlim_{j \in J} F(j,k) \to \varprojlim_{j \in J} \varinjlim_{k \in K} F(j,k)$$

is an equivalence.

Question: For which small $\infty$-categories $J$ do $J$-limits commute with $K$-colimits in $Sp_{T(h)}$ for all small $\infty$-categories $K$?

There are some noteworthy examples of such $J$. In all cases, the reasoning is "$J$-limits can be re-expressed in terms colimits, which always commute with colimits by the "Fubini rule". (Indeed, such colimits are absolute, and general nonsense entails that when $J$ answers to the Question, there is always an argument of this form).

  1. Idempotent splitting is a form of limit which commutes with colimits in $Sp_{T(h)}$ (simply because $Sp_{T(h)}$ is an $\infty$-category).

  2. Finite limits commute with colimits in $Sp_{T(h)}$ (because $Sp_{T(h)}$ is stable).

  3. Limits indexed by $\pi$-finite spaces commute with colimits in $Sp_{T(h)}$ (because $Sp_{T(h)}$ is $\infty$-ambidextrous).

There is also some mixing and matching which can occur between the above examples. But what does the general picture look like?

$\endgroup$
1
  • 1
    $\begingroup$ This seems to be equivalent to J-limits preserving filtered colimits, since the functor of J-limits is exact. $\endgroup$
    – Z. M
    Commented Jun 15, 2023 at 4:46

0

You must log in to answer this question.