Recurrent and weakly recurrent points in \(\beta\) G (Q1079770)

Let G be a (discrete) group and \(\beta\) G its Čech-Stone compactification. For a set \(A\subset G\) \(\bar A\) means the closure of A in \(\beta\) G and \(\hat A=\bar A\setminus G\). A point \(\omega\in \hat G\) is said to be discrete iff there exists a neighborhood U of \(\omega\) such that \(g\omega\) \(\not\in U\) for every \(g\in G\), \(g\neq e\). \(\omega\) is called strongly discrete iff, for some neighborhood U of \(\omega\), for any \(g_ 1,g_ 2\in G\), \(g_ 1\neq g_ 2\), one has \(g_ 1U\cap g_ 2U=\emptyset\); \(\omega\) is almost periodic iff for any neighborhood U of \(\omega\) there are sets A and K in G such that K is finite, \(G=KA\), \(A\omega\) \(\subset U\). A not (strongly) discrete point of \(\hat G\) is called (weakly) recurrent. A set \(A\subset G\) is thin iff \(g_ 1A\cap g_ 2A\) is finite whenever \(g_ 1\neq g_ 2.\) The author proves that for the set SD of strongly discrete points one has (*) \(SD=\cup \{\hat A:\) A is thin\(\}\). He makes use of this theorem for proving the existence of a weakly recurrent not almost periodic point in \(\hat G\) for G belonging to a class of groups introduced by Fairchild which contains all infinite abelian or solvable groups. That proof is not convincing because one does not see how (*) implies that for a not thin set \(A\subset G\) one necessarily has \(\hat A\not\subset SD\) as the author claims. However, assuming additionally that G has an element of infinite order he proves independently the existence of a recurrent not almost periodic point in \(\hat G.\) - Some inconvenience for the reader is caused by bad typography of symbols, misprints and a rather neglected grammar.