Limit groups for relatively hyperbolic groups. I. The basic tools
Limit groups for relatively hyperbolic groups. I. The basic tools
复制标题
相对双曲群的极限群。
DOI:
10.2140/agt.2009.9.1423
复制
发表时间:
2004
影响因子:
0.7
通讯作者:
D. Groves
中科院分区:
文献类型:
--
作者:
D. Groves
We begin the investigation of -limit groups, where is a torsion-free group which is hyperbolic relative to a collection of free abelian subgroups. Using the results of (16), we adapt the re- sults from (21) and (22) to this context. Specifically, given a finitely generated group G, and a sequence of pairwise non-conjugate ho- momorphisms {hn : G ! }, we extract an R-tree with a nontrivial isometric G-action. This, along with the analogue of Sela's shortening argument allows us to prove the main result of this paper, that is Hopfian. In his remarkable series of papers (38, 39, 41), Z. Sela has classified those finitely generated groups with the same elementary theory as the free group of rank 2 (see also (40) for a summary). This class includes all nonabelian free groups, most surface groups, and certain other hyperbolic groups. In particular, Sela answers in the positive some long-standing questions of Tarski (Kharlampovich and Miasnikov have another approach to these problems; see (29)). In (38), Sela begins with a study of limit groups. Sela's definition of a limit group is geometric, though it turns out that a group is a limit group if and only if it is a finitely generated fully-residually free group. Sela then produces Makanin-Razborov diagrams, which give a parametrization of Hom(G,F), where G is an arbitrary finitely gener- ated group and F is a nonabelian free group (such a parametrisation is also given in (28)). Over the course of his six papers, two of the main tools Sela uses are the theory of isometric actions on R-trees and the shortening argument. Sela's work naturally raises the question of which other classes of groups can be understood using Sela's approach. Many of Sela's meth- ods (and, more strikingly, some of the answers) come from geometric group theory. Thus it seems natural to consider, when looking for classes of groups to apply these methods to, groups of interest in geo- metric group theory. In (42), Sela considers an arbitrary torsion-free