Quotients and colimits of \(\kappa \)-quantales (Q645190)

Given a regular cardinal \(\kappa\), the paper considers the concept of \(\kappa\)-quantale, which is a partially ordered commutative monoid \((L,\cdot,1,\leqslant)\) such that, firstly, \(1\) is the top element of \((L,\leqslant)\); secondly, every subset of \(L\) of cardinality less than \(\kappa\) has a join (\(\kappa\)-join); thirdly, the monoid operation distributes over \(\kappa\)-joins. The authors additionally postulate that \(0\)-quantales are just commutative monoids (no partial order), whereas \(\infty\)-quantales are commutative strictly two-sided quantales (called in the paper quantales, for short). If the multiplication \(\cdot\) of the underlying monoid is idempotent, then \(\kappa\)-quantales are precisely the \(\kappa\)-frames of [\textit{J.~J. Madden}, ``\(\kappa\)-frames'', J. Pure Appl. Algebra 70, No. 1--2, 107--127 (1991; Zbl 0721.06006)], \(\infty\)-quantales are frames [\textit{P.~T. Johnstone}, Stone spaces. Cambridge Studies in Advanced Mathematics, 3. Cambridge etc.: Cambridge University Press (1982; Zbl 0499.54001)], \(\aleph_1\)-quantales are \(\sigma\)-frames, whereas \(\aleph_0\)-quantales are precisely the bounded distributive lattices (the bottom element comes from the join over the empty set). The main goal of the authors lies in an extension of certain results from the theory of \(\kappa\)-frames to the case of \(\kappa\)-quantales. More precisely, with the category \(\kappa\)\textbf{Qnt} of \(\kappa\)-quantales in hand, they construct a free \(\kappa\)-quantale over a given \(\lambda\)-quantale for \(\lambda\leqslant\kappa\) (thereby providing a left adjoint to the obvious forgetful functor \(\kappa\)\textbf{Qnt}\(\rightarrow\lambda\)\textbf{Qnt}), give an explicit description of colimits in the category \(\kappa\)\textbf{Qnt}, and illustrate the latter with the case of coproducts. The colimit-related machinery in turn is a particular instance of a more general construction of colimits in linear categories [\textit{B.~Banaschewski} and \textit{A.~Pultr}, ``Distributive algebras in linear categories'', Algebra Univers. 30, No. 1, 101--118 (1993; Zbl 0780.18005)] (such categories always have a tensor product of their objects in the classical sense). The cornerstone of the approach of the paper is a generalization of the technique of extending a binary relation \(R\) on a given frame to a congruence with the help of (\(R\)-)saturated elements, or, in other words, describing quotient frames (also known as sublocales, or generalized subspaces) in the language of nuclei as in, e.g., [\textit{A.~Pultr}, ``Frames'', in: Handbook of algebra. Volume 3. Amsterdam: Elsevier. 791--857 (2003; Zbl 1063.06006)]. For example, to get a free quantale over a given \(\lambda\)-quantale \(L\) (\(0<\lambda\)), one starts with a free quantale over the underlying commutative monoid of \(L\), and then constructs a quotient of the former, extending a suitable relation to a congruence, so that the natural embedding of the monoid \(L\) into its free quantale extension preserves \(\lambda\)-joins. A free \(\kappa\)-quantale (\(\lambda\leqslant\kappa\)) over \(L\) is then given by a particular restriction of the codomain of the above universal map, whose elements have a simple and convenient to handle description as \(\lambda\)-ideals.