In a sense, this is a follow up on this question, but one PhD programme later.

Let $\mathcal{C}$ be ribbon fusion. By $\mathcal{C}'$, we denote the symmetric centre, i.e. the full subcategory of objects that trivially braid with every other object in $\mathcal{C}$. Let us require that $\mathcal{C}$ is modularisable, i.e. $\mathcal{C}' \simeq \operatorname{Rep}G$ for a finite group $G$. We then have the following "exact sequence of ribbon fusion categories": $$\mathcal{C}' \hookrightarrow \mathcal{C} \twoheadrightarrow \widetilde{\mathcal{C}}$$ Here, $\widetilde{\mathcal{C}}$ is the modularisation of $\mathcal{C}$. I'm calling this an exact sequence since $\mathcal{C}'$ is being sent to the copy of vector spaces sitting inside the modularisation as the category spanned by the tensor unit, and in a certain sense, $\mathcal{C}'$ is the biggest such subcategory of $\mathcal{C}$.

This looks a lot like a central extension. So another way of describing a (modularisable) ribbon fusion category would then be as a central extension of a modular category by the representations of a finite group (i.e. a Tannakian symmetric fusion category).

It's known that central extensions of groups are parametrised by second group cohomology. Is there a similar characterisation of ribbon fusion categories in terms of "central extensions"? I.e. can every modularisable ribbon fusion category be described by a modular category, a symmetric fusion category and some extra data?


Deequivariantization (of which modularization is a special case) is inverse to equivariantization. That is, you can recover $\mathcal{C}$ as the category of G-equivariant objects in $\tilde{\mathcal{C}}$ for some action of G on $\tilde{\mathcal{C}}$. So to classify such extensions you want to classify G-actions. (Note that this can be a bit tricky because G can act nontrivially on the category even if it acts trivially on iso classes of objects.)

  • $\begingroup$ Awesome! Is there work on classifying $G$-actions on a given modular category? (A quick search didn't bring up anything, but I might searching for the wrong keywords.) $\endgroup$ – Manuel Bärenz May 5 '16 at 17:09
  • 2
    $\begingroup$ I don't think there's a ton of work in that direction. One place to start is arxiv.org/abs/0712.0585 $\endgroup$ – Noah Snyder May 5 '16 at 19:34

Your Answer

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

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