Hyperbolic towers and independent generic sets in the theory of free groups. (Q372628): Difference between revisions

Following the well-known works of Sela, Kharlampovich and Myasnikov, all non-Abelian free groups are elementarily equivalent and further Sela proved that, \(T_{fg}\), the elementary theory of non-Abelian free groups is stable. On the other hand, by a result of Poizat, \(T_{fg}\) is connected, i.e. it has a model which admits no proper definable subgroup of finite index. By the stablity, connectedness is equivalent to say that \(T_{fg}\) admits a unique generic type over any set of parameters, so let \(p_0\) be the generic type over the empty set. By a result of Pilly, a subset of non-Abelian free group is a maximal independent set of realizations for \(p_0\), if and only if it is a basis, so as a consequence, all such subsets have the same cardinal. Now, it is natural to ask that, is there a model of \(T_{fg}\) in which one can find two maximal independent sets of realizations for \(p_0\) with different cardinalities. The authors use hyperbolic towers to answer this and some other similar questions in the affirmative and also they prove that homogeneity is not preserved under free product, answering a question of Jaligot.