On the width in free products with amalgamation (Q5941991)

By the width \(\text{wid}(G,\varphi)\) of a verbal subgroup \(\varphi(G)\) defined in the group \(G\) by a word \(\varphi\) the author means the least value \(m\) (finite or infinite) such that any element of the subgroup \(\varphi(G)\) can be represented as a product of \(\leq m\) values of the word \(\varphi\). In fact, such a definition is not completely correct. The last words should be changed to ``can be represented as a product of \(\leq m\) values of the words \(\varphi^{-1}\) and \(\varphi\). A. H. Rhemtulla proved that the width of any proper nontrivial verbal subgroup \(\varphi(A*B)\) is finite if and only if \(|A|,|B|\leq 2\). V. G. Bardakov proved, almost obvious in view of A. H. Rhemtulla's result, Theorem: Let \(G=A*_UB\), \(|A:U|\geq 3\), \(|B:U|\geq 2\), \(U\) a normal subgroup in \(A\), as well as in \(B\), then the width of every proper nontrivial verbal subgroup \(\varphi(A*_UB)\) is infinite. It is claimed in the paper under review that the proper nontrivial verbal subgroup \(\varphi(A*_UB)\) has infinite width in the case when \(|A:U|\geq 2\), and \(G\) contains an element \(b\) for which \(UbU\not=Ub^{-1}U\). Unfortunately the paper is closed with a non-correctly proved Proposition: Let \(G=A*_U B\); \(|A:U|,|B:U|\leq 2\), where \(U\) is a subgroup of finite width. Then the width of a proper verbal subgroup \(\varphi(G)\) is finite. The error is that \(\varphi(G)\cap U\) is not contained in general in \(\varphi(U)\).