5
$\begingroup$

Let $F$ be a combinatorial species. The exponential generating series of $F$ is defined to be $$ \sum_{n \geq 0} \frac{ \lvert F_n \lvert x^n}{n!} $$ It was observed by Baez and Dolan in their paper "From finite sets to Feymann diagrams" that the exponential generating series can be categorified to a 1-groupoid $$ \sum_{n \geq 0} F_n \times X^n // S_n $$ You recover the exponential generating series by taking groupoid cardinality and setting $ x = \lvert X \lvert $. The cycle index series is another important series attached to a combinatorial species. It is defined by $$ Z_F(x_1,x_2,\dots) = \sum_{n \geq 0} \frac{1}{n!} \sum_{\sigma \in S_n} \lvert {\rm Fix}(F[\sigma])\lvert x_1^{\sigma_1} x_2^{\sigma_2} \dots $$

Question: Is it possible to construct an $n$-groupoid whose groupoid cardinality is $Z_F(x_1,x_2,\dots)$?

By $n$-groupoid, I mean a homotopy $n$-type?

$\endgroup$
3
  • $\begingroup$ Why $n$-groupoids? I don't see any need to go past groupoids here. $\endgroup$ Mar 2, 2015 at 4:55
  • $\begingroup$ I have no reason to think that n > 1 except that I couldn't write down a groupoid which worked. Obviously that doesn't mean it is not possible though! $\endgroup$ Mar 2, 2015 at 6:41
  • $\begingroup$ It's straightforward to write down a groupoid that works, the question is whether it can be described in a nice way: it's the groupoid of tuples of a permutation $\sigma \in S_n$, a fixed point $f \in F_n$, and functions $c_i(\sigma) \to X_i$ where $c_i$ denotes the set of $i$-cycles and $X_i$ are some auxiliary groupoids. There is an $S_n$-action on this data by $\tau : (\sigma, f) \mapsto (\tau \sigma \tau^{-1}, \tau f)$ (and identifying $i$-cycles of $\sigma$ with $i$-cycles of its conjugate) and that's where the $\frac{1}{n!}$ comes from. $\endgroup$ Mar 2, 2015 at 18:00

0

Your Answer

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

Browse other questions tagged or ask your own question.