On regular semigroups. II: An embedding (Q1084512)

The core (a word suggested by Mario Petrich) of any regular semigroup S is the subsemigroup of S generated by E(S), where E(S) is the set of idempotents of S. For each \(e\in E(S)\) by \(\prec e\succ\) is denoted the core of the regular subsemigroup eSe. The main result is the following Theorem I. Any regular semigroup S is embeddable in a regular semigroup T such that (i) S is full in T, that is, \(E(S)=E(T)\), and (ii) for all idempotents e,f\(\in S\), we have e\({\mathcal D}f\) in T if (and only if) \(\prec e\succ\) is isomorphic to \(\prec f\succ\). Moreover, T can be taken such that further, for all e,f\(\in E(S)\) with \(\prec e\succ\) isomorphic to \(\prec f\succ\), each isomorphism from \(\prec e\succ\) onto \(\prec f\succ\) is the restriction to \(\prec e\succ\) of some inner partial automorphism of T with domain eTe and range fTf. If S is finite, then T can be taken to be finite also. This theorem answers affirmatively Question H3 of the Proceedings of the 1979 DeKalb Conference (1979; Zbl 0446.20034). For inverse semigroups the result is due to \textit{N. R. Reilly} [J. Aust. Math. Soc., Ser. A 23, 28- 41 (1977; Zbl 0365.20071)]. Theorem I enables the author to solve seven problems of the form ''characterize the idempotent-generated subsemigroups of bisimple regular semigroups'' posed in a previous paper of the author [Part I, J. Algebra 24, 1-24 (1973; Zbl 0262.20074)].