On countable completions of quotient ordered semigroups (Q1006337)

A poset is called \(\omega\)-chain complete if each of its chains that is isomorphic to \(\omega\) (equivalently, each countable chain or even each countable directed subset) has a supremum. As shown by \textit{G. Markowsky} [``Chain-complete posets and directed sets with applications'', Algebra Univers. 6, 53--68 (1976; Zbl 0332.06001)], any poset has a universal \(\omega\)-completion. The author calls a poset \(P\) strict if each element of the completion is a directed join of an \(\omega\)-chain in \(P\), and all elements of \(P\) are \(\omega\)-compact in the completion. By an easy example it is demonstrated that in this setting, the formation of quotients does not always commute with the \(\omega\)-completion. In contrast to that negative observation, the author proves a positive result for what he calls naturally ordered semigroups. These are monoids for which the preorder given by \(x\sqsubseteq y \Leftrightarrow \exists\,z \, (y = x\cdot z)\) is a partial order. More generally, he means by a preordered (resp. partially ordered) semigroup one with an extra preorder (resp. partial order) such that the neutral element is the least element and the left translations preserve the order relation. [Reviewer's remark. This one-sided approach causes some difficulties; for example, the statement that ``there is an obvious adjunction between \textsf{POSG}, the category of partially ordered semigroups, and \textsf{PreOSG}, the category of preordered semigroups, obtained by restriction'' of the adjunction between the categories of posets and of preordered sets, is obscure: certainly one cannot speak of a ``restriction'', and why is the equivalence relation obtained by symmetrization of the preorder a (two-sided!) semigroup congruence? This would be crucial for the intended adjunction between \textsf{POSG} and \textsf{SG}, the category of semigroups with neutral element (monoids).] The study of the interplay between order and semigroup structures culminates in the statement that the \(\omega\)-completion of any quotient of a strict naturally ordered semigroup \(S\) coincides with the so-called pseudo-completion. Roughly speaking, the latter is obtained by canonically extending the given congruence on \(S\) to its \(\omega\)-completion and then forming the \textsf{POSG}-quotient with respect to the extended congruence. Hence, in this case one might say that the completion commutes with the formation of quotients. Moreover, there is a natural action of the ground semigroup on its \(\omega\)-completion. In particular, these results apply to quotients of free semigroups. More transparent and technically simpler is the last part, dealing with (two-sided) partially ordered monoids, for which not only the left but also the right translations are order-preserving. The author speaks here of strongly ordered semigroups. After some instructive examples he shows that the \(\omega\)-completion of a strict and strongly ordered semigroup carries a unique semigroup structure extending the given one and making the completion a topological semigroup with the \(\omega\)-Scott topology, an obvious variant of the usual Scott topology: the \(\omega\)-Scott open subsets are those upper sets which contain with the supremum of any \(\omega\)-chain \(C\) already some element of \(C\).