11
$\begingroup$

I am looking for interesting examples of categories admitting multiple monoidally inequivalent closed (or compact closed) symmetric monoidal structures. We know how to construct disconnected toy models [1], but none of direct practical interest. I also seem to recall from [2] that the category of $G$-sets admits two inequivalent such structures, but I am curious to know whether this is a sporadic occurrence or whether there is a large family of interesting examples lurking somewhere out there.

[1]: see e.g. https://arxiv.org/abs/1803.00708

[2]: Borceux, Francis. Handbook of Categorical Algebra

$\endgroup$

3 Answers 3

15
$\begingroup$

A pretty interesting class of examples comes about by classifying compact monoidal groupoids. Given a group $G$, a $G$-module $M$, and a (normalized) 3-cocycle $a: G \times G \times G \to M$, one can manufacture a compact monoidal groupoid whose category of objects is $G$, whose morphisms are ordered pairs $(g, m) \in G \times M$ where we define $\text{dom}(g, m) = \text{cod}(g, m) = g$ and where endomorphism composition is defined by addition in $M$, and where we define the tensor product by $(g, m) \otimes (h, n) = (g h, m + g n)$. It's the 3-cocycle that furnishes the associativity data, and we get monoidally inequivalent groupoids whenever the 3-cocycles are not cohomologous. This was observed by Joyal and Street in their paper Braided Monoidal Categories.

Now you were asking about the symmetric monoidal case (where we now assume $G$ is abelian). These are also known as Picard groupoids. In the simplified scenario where we demand strict associativies and consider only the case where $G$ acts trivially on $M$, in which case the underlying category becomes the product $K G \times B M$ of the evident discrete monoidal category $K G$ with the evident one-object category $B M$, any symmetric bilinear pairing $G \otimes G \to M$ can be used to manufacture a symmetry isomorphism for a symmetric monoidal structure on $K G \times B M$, and these examples are generally symmetric-monoidally inequivalent. I can't tell how far away such examples are from the "toy"examples" you have in mind, but Picard groupoids are surely of interest -- see for example applications to 2-stage Postnikov systems of spectra here.

$\endgroup$
7
$\begingroup$

The category $Cat$ of small categories admits two inequivalent tensor products: the Cartesian product and the funny product. This generalizes up to 2-cat, 3-cat, etc. By the way, a word of advise to those like me with a murky memory: googling "cat funny product" is going to just give you pictures of cats.

$\endgroup$
3
  • 1
    $\begingroup$ But he asked for compactness. $\endgroup$
    – Todd Trimble
    Sep 1, 2018 at 18:01
  • 2
    $\begingroup$ Omega-cat admits the Crans-Gray-Steiner lax tensor product. I dunno if this meets the compactness requirement, but it is in the same vein. $\endgroup$ Sep 1, 2018 at 18:10
  • 2
    $\begingroup$ At present, the question asks for compact or compact closed, but anyway, I think all the answers are pointing the OP towards examples from category theory, and that's the important thing. $\endgroup$ Sep 1, 2018 at 18:39
4
$\begingroup$

The large class of examples I have in mind, though I am not sure if it meets your compactness requirement (definition?) are the Lax tensor products on $n$-Cat. It has been constructed in the cases $n=2$ and $n=\omega$ by Gray in the case $n=2$ and Crans, Steiner, and Verity independently in the case $n=\omega$.

Edit: To clarify, these are all biclosed, and their right adjoints are the $n$-categories whose objects are strict $n$-functors and whose higher cells are (op)lax natural transformations and (op)lax modifications between them.

$\endgroup$
1
  • $\begingroup$ I have edited my question to clarify what I mean by closed and by compact closed. $\endgroup$ Sep 1, 2018 at 18:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.