In Brian Day's thesis, he gives a definition of a pro-monoidal structure based on a particular motivating example:

Suppose $M:\cal{A}\hookrightarrow \cal{B}$ is the inclusion of a small full subcategory, and suppose $\cal{B}$ is equipped with a monoidal structure $1\in \cal{B}$ and $\otimes:\cal{B} \times \cal{B} \to \cal{B}$. Summarizing the discussion, he essentially gives a condition for the monoidal structure on $\cal{B}$ to restrict to a pro-monoidal structure on $\cal{A}$, which amounts to showing that the four induced maps on coends:

$$\int^X B(MX,1) \times B(M(-), MX\otimes MA) \\ \downarrow \\ \int^Y B(Y,1) \times B(M(-), Y\otimes MA) $$

$$\int^X B(MX,1) \times B(M(-), MA\otimes M(X)) \\ \downarrow \\ \int^Y B(Y,1) \times B(M(-), MA\otimes Y) $$

$$\int^X B(MX,MA\otimes MA') \times B(M(-), MX\otimes MA'') \\ \downarrow \\ \int^Y B(Y,MA\otimes MA') \times B(M(-), Y\otimes MA'') $$

$$\int^X B(MX,MA'\otimes MA'') \times B(M(-), MA\otimes MX) \\ \downarrow \\ \int^Y B(Y,MA'\otimes MA'') \times B(M(-), MA\otimes Y) $$

are isomorphisms for all $A,A',A''\in \cal{A}$. (I've used the opposite variance for the Yoneda reduction to the one in Day's paper). Morally this seems to say that when we restrict to tensor products of objects of $\cal{A}$, it is enough to compute the coend over $\cal{A}$ rather than all of $\cal{B}$.

In the situation that I care about, $\cal{A}$ is actually also dense in $\cal{B}$, the monoidal structure on $\cal{B}$ is biclosed, and the unit of the monoidal structure lies in $\cal{A}$ and is the terminal object of both $\cal{A}$ and $\cal{B}$, so the first two maps seem to end up working out to be isomorphisms pretty easily (I haven't proven it, but I've convinced myself that it works).

I have two questions then:

1.) If $\cal{A}$ is also dense in $\cal{B}$, can these conditions be simplified in any way? What about if any of the other conditions relevant to my case of interest hold ($\cal{B}$ biclosed or $1\in \cal{A}$ and terminal in both categories)?

2.) Are there any examples in the literature where the existence of the induced pro-monoidal structure on $\cal{A}$ has been proven by using these coend maps or a related technique (specifically in the nontrivial case where $\cal{B}$ is not just the category of presheaves on $\cal{A}$ and $\cal{A}$'s induced pro-monoidal structure is not representable).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.