Let $G$ be a countable (that is edit) residually-$p$ group and let $\hat{G}_p$ be its pro-$p$ completion.

  1. If $\hat{G}_p$ is finitely generated does it mean that $G$ is finitely generated?

  2. If $\hat{G}_p$ is finitely presented does it mean that $G$ is finitely presented? (I think the Grigorchuk group is not finitely presented, what about its pro-$p$ completion?)

Edit: Following @Ycor comment. Let me add that in Question 2, I would like $G$ to be finitely generatd. Also, let me be explicit:

  1. Is the pro-$p$ completion of the Grigorchuk group finitely presented?
  • 1
    $\begingroup$ 1,2 have a trivial negative answer. Consider $G=\mathbf{Z}[1/q]$ for some prime $q\neq p$. Then the embedding of $\mathbf{Z}$ in $G$ induces an isomorphism of pro-$p$ completions. $\endgroup$
    – YCor
    Jan 20, 2018 at 13:27

2 Answers 2


As I said in a comment, the answer to 1 is trivially no.

The answer to 2 (as edited) is also no. Platonov and Tavgen produced a finitely generated, infinitely presented subgroup $H$ in the square $F\times F$ of a free groups such that the inclusion induces an isomorphism of profinite completions (and hence of pro-$p$-completions for all $p$).

Here, $F$ can be constructed as follows. Consider any epimorphism $p:F\to P$ from $F$ to an infinite finitely presented group $P$ with no nontrivial finite quotient; then $H$ is defined as the fibre product $\{(g,h)\in F\times F:p(g)=p(h)\}$. That $P$ is finitely presented is used to ensure that $H$ is finitely generated. That $P$ has no nontrivial proper quotient is used by Platonov and Tavgen to prove that the inclusion induces an isomorphism of profinite completions. That the fibre product $H$ is infinitely presented is a classical fact on fiber products.

Reference: V. Platonov O.Tavgen. On the Grothendieck problem of profinite completions of groups. (Russian) Dokl. Akad. Nauk SSSR 288 (1986), no. 5, 1054-1058. English translation: Soviet Math. Dokl. 33 (1986), no. 3, 822-825. (MR link)


The answer to question 3 is "no" (for $p=2$) : https://arxiv.org/abs/1011.3880 and "yes" for every other $p$.

  • $\begingroup$ What do you mean "yes" for every other $p$? Did you mean I should have said pro-$2$ completion? $\endgroup$ Jan 20, 2018 at 19:53
  • $\begingroup$ I think so: Mark mentioned the trivial case of odd $p$. $\endgroup$
    – YCor
    Jan 20, 2018 at 21:44
  • $\begingroup$ I can only accept one answer, so I choose Yves' one as he answer two of my questions. However, Mark answered my third one. $\endgroup$ Jan 20, 2018 at 23:06

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.