1
$\begingroup$

Let $\textbf{CART}$ be a category where the objects are all Cartesian closed categories (henceforth shortened as CCC). Is there any way to define the arrows so that $\textbf{CART}$ itself becomes Cartesian closed? I am particularly interested in arrows that preserve the dinatural transformations of the simply typed lambda calculi in the different CCCs.

Possible start for solution [EDIT: ruled out by მამუკა ჯიბლაძე , see comment]: There is one rather obscure family of functors called Cartesian closed functors, that preserves the products and exponentials of the domain CCC. Given two CCCs $\textbf{C}$ and $\textbf{D}$, the collection of all Cartesian closed functors from $\textbf{C}$ to $\textbf{D}$ seems like a good candidate for the exponential $\textbf{D}^\textbf{C}$, but will the natural transformations between those functors make arrows in the category $\textbf{D}^\textbf{C}$ so that it becomes Cartesian closed? That is, $\textbf{D}^\textbf{C}$ is only a good exponential in $\textbf{CART}$ if it is an object in $\textbf{CART}$, but by definition of $\textbf{CART}$, it then must be a CCC.

$\endgroup$
3
  • 2
    $\begingroup$ Regarding possible start - no way, I believe. Boolean algebras are particular CCCs, and for such an algebra B the category CART(B,2) can be identified with the poset of all ultrafilters of B together with the improper filter (the whole of B). This category does not even have any nontrivial binary products. $\endgroup$ Commented Aug 3, 2023 at 8:39
  • $\begingroup$ @მამუკაჯიბლაძე ჯიბლაძე Does your CART(B,2) contain all functors from B to 2, or just the functors that are Cartesian closed, i. e. the ones that preserves products and exponentials in B? $\endgroup$ Commented Aug 3, 2023 at 9:16
  • 1
    $\begingroup$ Only Cartesian closed functors. All functors would give the poset of all upper sets of B, which is indeed Cartesian closed. Also the functors which only preserve products would give the poset of all filters of B, which is Cartesian closed too. $\endgroup$ Commented Aug 3, 2023 at 9:31

0

You must log in to answer this question.