The Higman Embedding theorem says that any finitely generated and recursively presented group can be embedded in a finitely presented group.

My question is if one can embed such a group as a normal subgroup into a finitely presented group?


No. Take a f.g. non-finitely presented group $G$ with trivial $Out(G)$ and trivial center (such groups clearly exist; in fact one can even assume that $Out(G)$ is locally finite, say $G$ is the Grigorchuk group of intermediate growth, by the result of Grigorchuk and Sidki). Suppose that $G$ is a normal subgroup of $H$. Then every $h\in H$ acts on $G$ by conjugation $x\to x^h$. Since $Aut(G)=Inn(G)$, there exists $g\in G$ such that $x^g=x^h$ for every $x\in G$. Hence $g^{-1}h$ centralizes $G$. Therefore $H=G Z_H(G)$ (if $Out(G)$ is locally finite, then instead of $G$ here you will get a finite extension of $G$, the image of $H$ in $Aut(G)$ under the natural homomorphism) where $Z_H(G)$, the centralizer of $G$ in $H$, intersects $G$ trivially (since the center of $G$ is trivial) and is a normal subgroup. Hence $H$ is a direct product of $G$ and $Z=Z_H(G)$. Since $H$ is finitely generated, $Z$ is finitely generated (being a homomorphic image of $H$). Hence if $H$ is finitely presented, then $G$ is presented by the presentation of $H$ plus finitely many relations killing the generators of $Z$, so $G$ must be finitely presented, a contradiction.

  • $\begingroup$ Thank you. Is there a characterization of finitely generated normal subgroups of f.p. groups? $\endgroup$ – Mustafa Gokhan Benli Aug 17 '11 at 17:33
  • $\begingroup$ @Mustafa: I do not think a characterization which is better than the definition exists. $\endgroup$ – Mark Sapir Aug 17 '11 at 18:12

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

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