Amalgams of free inverse semigroups. (Q677103): Difference between revisions

It was proved by \textit{T. E. Hall} [J. Algebra 34, 375-385 (1975; Zbl 0326.20054)] that any amalgam \([S,T;U]\) of inverse semigroups is strongly embeddable in an inverse semigroup and thus in the amalgamated free product \(S*_U T\), taken within the category of inverse semigroups. For quite some time thereafter, little was known about the structure of such products. The reviewer determined that of the (unamalgamated) free products and with \textit{S. W. Margolis, J. Meakin} and \textit{J. Stephen} [Glasg. Math. J. 33, No. 3, 373-387 (1991; Zbl 0742.20057)] reinterpreted his results in the language of `Schützenberger' automata, which had been developed by the last of those authors. \textit{J. Stephen} [Amalgamated free products of inverse semigroups, (preprint)] and \textit{P. A. Bennett} [Amalgamated free products of inverse semigroups, (preprint)] both used the automata-theoretic techniques to extend this work to more general amalgamated products. The latter author developed a procedure which, for certain `lower bounded' amalgams, yields the appropriate Schützenberger automata by repeating a sequence of five steps. In the paper under review, sufficient conditions are given in order that this procedure be effective, therefore, according to Stephen's theory, solving the word problem in the amalgamated free product, relative to the initial parameters. The principal application is that the word problem is solvable in the amalgamated free product of any two free inverse semigroups, over a common finitely generated inverse subsemigroup. This then results in the solvability of the word problem in certain one-relator inverse semigroups. Various structural properties of free amalgamated products are also obtained. The authors remark that under the obvious assumptions, amalgamated free products of groups are known always to have solvable word problem; and that for semigroup amalgams, matters are decidedly more complicated than for inverse semigroups -- a situation familiar to those working in the field.