HYPERBOLICITY OF GROUPS WITH SUBQUADRATIC ISOPERIMETRIC INEQUALITY (Q4024633)

Consider an arbitrary finitely presented group (1) \(G = \langle a_ 1,\dots,a_ k \mid r_ 1,\dots,r_ l\rangle\). Define \(G = F/N\), where \(N\) is the normal closure of the set of all relators \(r_ i\) in the free group \(F = F(a_ 1,\dots,a_ k)\). Therefore for every word \(w \in N\) (i.e., \(w = 1\) in \(G\)) there is an equation (2) \(w = \prod^ m_{i = 1} s_ i r^{\pm 1}_{j_ i} s^{-1}_ i\) in \(F\), where \(s_ i \in F\). Suppose the number \(m = m(w)\) of factors in (2) is minimal for \(w\). Since \(G\) is finitely generated, there is an integer valued function \(f_ G\) (depending on the presentation (1) of \(G\)), such that (3) \(m(w) \leq f_ G(| w|)\), where \(| w|\) is the length of the word \(w \in N\). If \(G\) is hyperbolic (sometimes the terminology ``word-hyperbolic'' or ``negatively curved'' is used), then one can choose the function \(f_ G\) linear. In this case we say that (3) is a linear ``isoperimetric inequality''. For every automatic group \(G\) the function \(f_ G\) is quadratic. Therefore it is natural to ask what the structure of a group is for which the function \(f_ G\) is subquadratic. In this case Gromov has an argument that the group \(G\) is hyperbolic. That argument is scattered throughout a paper, which makes reading it difficult. In this paper the author gives a simple and coherent proof of Gromov's statement on hyperbolicity of groups with subquadratic isoperimetric inequality.