Congruence subgroups of $SL_2(\mathbb{Z})$ usually seem to be defined as a subgroup that contains $\Gamma(N) = \left\{\begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix} \mod N \right\}$ for some positive integer $N$. But, I have also seen it defined as a subgroup defined by congruence conditions on the entries. The problem is, this isn't very specific. For example, what if I look at the set $\left\{\begin{pmatrix} a & b \\ c & d \end{pmatrix} : a^b \equiv c^2 d^{15} \mod 17 \right\}$? I just made that up off the top of my head and odds are it's not closed under multiplication and thus not a subgroup. Maybe the solution set is empty. But, assuming it is a nonempty subgroup, would that be considered a congruence subgroup? Or, more generally, are there more specific conditions on what types of congruences are allowed, i.e., is there a rigorous definition somewhere?

  • 1
    $\begingroup$ I don't think the notion of "congruence relation on the entries" is supposed to include exponential functions. I think it's meant to be linear congruence relation on the entries, or perhaps some larger subset. $\endgroup$ – Arturo Magidin Apr 5 '11 at 19:01
  • $\begingroup$ I think the only examples I have ever seen have been of the form $\left\{\begin{pmatrix} a & b \\ c & d \end{pmatrix} \equiv x : x \in X\right\}$, where $X$ is some set of matrices. $\endgroup$ – Graphth Apr 5 '11 at 19:16

If the subgroup $\Gamma$ contains $\Gamma(N)$, then we can find a coset decomposition $\Gamma = \displaystyle \cup_{\alpha \in A} \alpha \Gamma(N)$ where $A$ is a set of coset representatives.

Then we have $\Gamma = \left\{ \begin{pmatrix} a & b \\ c & d \end{pmatrix} \equiv \alpha \pmod{N}: \alpha \in A \right\}$.

If you have any set of linear congruence relations on $a, b, c, d$ that define a subgroup of $SL_2(\mathbb{Z})$, you can just take the lcm of the moduli of congruence and construct a set $A$ of coset representatives.

I believe that this is the content of the comment that "congruence subgroups are those defined by congruence relations." I don't know of anything to say about other forms of congruences, such as the one you suggest.


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.