From Encyclopedia of Mathematics
Jump to: navigation, search
The printable version is no longer supported and may have rendering errors. Please update your browser bookmarks and please use the default browser print function instead.

In 1949, G. Higman, B.H. Neumann and H. Neumann [a4] proved several famous embedding theorems for groups using a construction later called the HNN-extension. The theory of HNN-groups is central to geometric and combinatorial group theory and should be viewed in parallel with amalgamated products (cf. also Amalgam of groups).

The easiest way to define an HNN-group is in terms of presentations of groups.

Presentation of groups.

A presentation of a group $G$ is a pair $\langle X | R \rangle$ where $R$ is a subset of $F(X)$, the free group on the set $X$, and $G$ is isomorphic (cf. also Isomorphism) to the quotient group $F(X)/N(R)$, where $N(R)$ is the intersection of all normal subgroups of $F(X)$ containing $R$. The subgroup $N(R)$ is called the normal closure of $R$ in $F(X)$.

Given an arbitrary group $G$, there is an obvious homomorphism $\tau_G : F(G) \rightarrow G$ such that $\tau_G(g) = g$ for all $g \in G$. Clearly, $\langle G | \ker \tau_G \rangle$ is a presentation for $G$.


Suppose $\mu : A \rightarrow B$ is an isomorphism of subgroups of a group $G$ and $t$ is not in $G$. The HNN-extension of $G$ with respect to $\mu$ has presentation

\begin{equation*} \left\langle G \bigcup \{ t \} : ( \operatorname { ker } ( \tau _ { G } ) ) \bigcup \left\{ t ^ { - 1 } a ^ { - 1 } t \mu ( a ) : \forall a \in A \right\} \right\rangle. \end{equation*}

The generator $t$ is called the stable letter, $G$ the base group and $A$ and $B$ the associated subgroups of this HNN-extension. When $A = G$, the HNN-extension is called ascending.

Shorthand notation for the above group is $\langle G , t : t ^ { - 1 } A t = B , \mu \rangle$ or $G ^ { * } \mu$.

In [a4] it was shown that the mapping $G \rightarrow G ^ { * } \mu$ taking $g \rightarrow g$ for all $g \in G$ is a monomorphism. The rest of the normal form theorem for HNN-extensions was proved by J.L. Britton in 1963 [a1] (Britton's lemma): Let $g _ { 0 } , \ldots , g _ { n }$ be a sequence of elements of $G$ and let the letter $\epsilon$, with or without subscripts, denote $\pm 1$. A sequence $g_0, t^{\epsilon_1}, g_1, \cdots, t^{\epsilon_n}, g_n$ will be called reduced if there is no consecutive subsequence $t^{-1} , g_{i} , t$ with $g _ {i} \in A$ or $t, g_{i} , t^{-1}$ with $g_i \in B$. For a reduced sequence and $n \geq 1$, the element

$$g_0 t^{\epsilon_1} g_1 \cdots t^{\epsilon_n} g_n$$

of $G _ { \mu } ^ { * }$ is different from the unit element.

In the original reference [a4], the following theorem is proved: Every group $G$ can be embedded in a group $G ^ { * }$ in which all elements of the same order are conjugate (cf. also Conjugate elements). In particular, every torsion-free group can be embedded in a group $G ^ { * * }$ with only two conjugacy classes. If $G$ is countable, so is $G ^ { * * }$. Also, every countable group $C$ can be embedded in a group $G$ generated by two elements of infinite order. The group $G$ has an element of finite order $n$ if and only if $C$ does. If $C$ is finitely presentable, then so is $G$.

For an excellent account of the history of HNN-extensions, see [a2]. See [a5], Chap. IV, for basic results and landmark uses of HNN-extensions, such as: the torsion theorem for HNN-extensions; the Collins conjugacy theorem for HNN-extensions; the construction of finitely-presented non-Hopfian groups (in particular, the Baumslag–Solitar group $\langle b , t : t ^ { - 1 } b ^ { 2 } t = b ^ { 3 } \rangle$ is non-Hopfian; cf. also Non-Hopf group); decompositions of $1$-relator groups; Stallings' classification of finitely-generated groups with more than one end in terms of amalgamated products and HNN-extensions; and Stallings' characterization of bipolar structures on groups.

HNN-extensions are of central importance in, e.g., the modern version of the Van Kampen theorem (based on topological results in [a8], [a9]); the Bass–Serre theory of groups acting on trees and the theory of graphs of groups (see [a7]); Dunwoody's accessibility theorem [a3]; and JSJ decompositions of groups [a6].


[a1] J.L. Britton, "The word problem" Ann. of Math. , 77 (1963) pp. 16–32
[a2] B. Chandler, W. Magnus, "The history of combinatorial group theory: A case study in the history of ideas" , Studies History Math. and Phys. Sci. , 9 , Springer (1982)
[a3] M.J. Dunwoody, "The accessibility of finitely presented groups" Invent. Math. , 81 (1985) pp. 449–457
[a4] G. Higman, B.H. Neumann, H. Neumann, "Embedding theorems for groups" J. London Math. Soc. , 24 (1949) pp. 247–254; II.4, 13
[a5] R. Lyndon, P. Schupp, "Combinatorial group theory" , Springer (1977)
[a6] E. Rips, Z. Sela, "Cyclic splittings of finitely presented groups and the canonical JSJ decomposition" Ann. of Math. (2) , 146 : 1 (1997) pp. 53–109
[a7] J.P. Serre, "Arbres, amalgams, $S L _ { 2 }$" Astéerisque , 46 (1977)
[a8] E.R. Van Kampen, "On the connection between the fundamental groups of some related spaces" Amer. J. Math. , 55 (1933) pp. 261–267
[a9] E.R. Van Kampen, "On some lemmas in the theory of groups" Amer. J. Math. , 55 (1933) pp. 268–273
How to Cite This Entry:
HNN-extension. Encyclopedia of Mathematics. URL:
This article was adapted from an original article by Mike Mihalik (originator), which appeared in Encyclopedia of Mathematics - ISBN 1402006098. See original article