13
$\begingroup$

Let $n,m\in\mathbb{N}$. Is there a formula for the number of subgroups of index $n$ in $\mathbb{Z}^m$? Perhaps in terms of the divisors of $n$?

$\endgroup$
3
  • 1
    $\begingroup$ You can figure this out using Hermite normal form. See e.g. page 7 of these slides: web.northeastern.edu/suciu/slides/counting.pdf $\endgroup$ Commented Aug 3, 2022 at 16:44
  • 2
    $\begingroup$ If anyone is interested in similar questions for other finitely generated groups, then note that the area of group theory that deals with this question in general is called "subgroup growth". Here, we usually only approximate the number of subgroups of a given index, since exact formulas are hard for most f.g. groups. $\endgroup$ Commented Aug 4, 2022 at 17:10
  • $\begingroup$ Is this 1 of those questions that are easy to ask but hard to answer? $\endgroup$
    – BCLC
    Commented Aug 5, 2022 at 15:02

4 Answers 4

19
$\begingroup$

Yes. This is given by OEIS sequence A160870. The number of subgroups of index $n$ in $\mathbf{Z}^m$ is there denoted $T(n,m)$. There is a recursive formula in terms of the divisors of $n$ given at this page. The initial conditions are $$ T(n,1) = 1 \quad \textrm{ for all } n\in \mathbb{N}$$ and recursively for $m > 1$, we have either $$ \quad T(n,m) = \sum_{d \mid n} \left(\frac{n}{d}\right)^{m-1} \cdot T(d, m-1) $$ or equivalently, $$ \quad T(n,m) = \sum_{d \mid n} d \cdot T(d, m-1) $$ Note that we can solve this recurrence to get the "explicit" formula $$ T(n,m) = \sum_{\substack{(d_0,d_1,\ldots,d_m)}} d_1 \cdots d_{m-1}$$ where the sum is over all sequences of integers $(d_0,d_1,\ldots,d_m)$ with $d_0=1$, $d_m=n$, and $d_i \mid d_{i+1}$ for all $i=0,\ldots,m-1$.

For example, there are $T(4,5) = 651$ subgroups of index $4$ in $\mathbf{Z}^5$.

$\endgroup$
5
  • 2
    $\begingroup$ Well, I would not call this a closed formula - it's a recurrence. $\endgroup$ Commented Aug 3, 2022 at 17:39
  • 1
    $\begingroup$ @SamHopkins Yes, closed formula is bad phrasing! Both formulas are correct (assuming the OEIS page is right). $\endgroup$ Commented Aug 3, 2022 at 17:44
  • 1
    $\begingroup$ Hmm, I guess it is possible both recurrences are correct but that seems weird. Also, I switched your $n$ and $m$ to be more consistent with the notation of the question-asker. EDIT: okay, I see that both recursions give the same thing. $\endgroup$ Commented Aug 3, 2022 at 17:50
  • $\begingroup$ (By the way, a proof for this recurrence can be obtained, as I suggested in a comment above, using Hermite normal form; this is explained in the slides of Alex Suciu I linked.) $\endgroup$ Commented Aug 3, 2022 at 20:56
  • $\begingroup$ @SamHopkins Thanks for the excellent edits! $\endgroup$ Commented Aug 4, 2022 at 15:12
15
$\begingroup$

There is a nice Dirichlet series generating function for this number. Denoting it by $j_m(n)$, we have $$ \sum_{n\geq 1}j_m(n)n^{-s} = \zeta(s)\zeta(s-1)\cdots \zeta(s-m+1), $$ where $\zeta$ denotes the Riemann zeta function. For the history of this result, see L. Solomon, in Relations between Combinatorics and Other Parts of Mathematics, Proc. Symp. Pure Math. 34, American Mathematical Society 1979, pp. 309-330. See also Exercise 5.13(d) (and its solution) of Enumerative Combinatorics, vol. 2.

$\endgroup$
7
$\begingroup$

This is an elaboration on Stanley's reference to Exercise 5.13 in EC2. In these two blog posts you can find proofs using groupoid cardinality of the following results. For a finitely generated group $G$ let $s_n(G)$ denote the number of subgroups of index $n$ and let $c_n(G)$ denote the number of conjugacy classes of subgroups of index $n$.

Exercise 5.13a: $$\sum_{n \ge 0} \frac{|\text{Hom}(G, S_n)|}{n!} z^n = \exp \left( \sum_{n \ge 1} s_n(G) \frac{z^n}{n} \right).$$

Exercise 5.13c: $$\sum_{n \ge 0} \frac{|\text{Hom}(\mathbb{Z} \times G, S_n)|}{n!} z^n = \prod_{n \ge 1} \frac{1}{(1 - z^n)^{c_n(G)}}.$$

Taking logarithms on both sides of 5.13c and applying 5.13a gives

$$s_n(\mathbb{Z} \times G) = \sum_{d \mid n} d c_d(G)$$

which, when applied to $G = \mathbb{Z}^k$, gives the second recurrence in Carl-Fredrik's answer (which is essentially the content of Exercise 5.13d).

(Not to toot my own horn excessively and with all respect to Stanley, I believe my proof of Exercise 5.13c is better motivated than the argument that appears in EC2.)

$\endgroup$
2
  • 1
    $\begingroup$ Regarding 5.13c, Valery Liskovets once sent me an elegant proof based on Burnside's lemma (the number of orbits of a finite group action on a finite set is the average number of fixed points of the group elements). The group action here is $G$ acting on itself by conjugation. $\endgroup$ Commented Aug 5, 2022 at 1:58
  • $\begingroup$ @Richard: yes, that should be more or less equivalent to the argument I use! $\endgroup$ Commented Aug 5, 2022 at 8:21
4
$\begingroup$

The number $\sigma_d(N)$ of subgroups of index $N$ in $\mathbb Z^d$ is also given by the formula (see Gruber, B. (1997), Alternative formulae for the number of sublattices. Acta Cryst. A53, 807--808 and Y.M. Zou, Y.M. (2006), Gaussian binomials and the number of sublattices, Acta Cryst. A62, 409--410) \begin{align}\label{formexactsigma} \sigma_d(N)&=\prod_{p\vert N}\left(\begin{array}{cc}e_p+d-1\\d-1 \end{array}\right)_p \end{align} where $\prod_{p\vert N}p^{e_p}=N$ is the factorization of $N$ into prime-powers and where $$\left(\begin{array}{cc}e_p+d-1\\d-1 \end{array}\right)_p=\prod_{j=1}^{d-1}\frac{p^{e_p+j}-1}{p^j-1}$$ is the evaluation of the $q$-binomial $$\left[\begin{array}{cc}e_p+d-1\\d-1 \end{array}\right]_q=\frac{[e_p+d-1]_q!}{[e_p]_q!\ [d-1]_q!}$$ (with $[k]_q!=\prod_{j=1}^k\frac{q^j-1}{q-1}$) at the prime-divisor $p$ of $N$.

$\endgroup$
2
  • $\begingroup$ It's pretty straightforward to get from the "explicit" formula in Carl-Fredrik's answer to this one. $\endgroup$ Commented Aug 4, 2022 at 13:01
  • 3
    $\begingroup$ Yes, Zou's paper does this. I find however this formula esthetically more pleasant. It encodes the obvious fact that these numbers form a (weakly) multiplicative function of the index for a fixed dimension. $\endgroup$ Commented Aug 4, 2022 at 13:04

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .