Kernel ideals of a Stone JP-semilattice. (Q485495)

A JP-semilattice is a partial algebra \((S;\wedge,\vee)\) such that \((S;\wedge)\) is a meet-semilattice and \(\vee\) is a partial join-operation satisfying some axioms. If the relation \(\leq\) on \(S\) is defined by \(x\leq y\) if and only if \(x\wedge y=x\), then \(\leq\) is a partial order with the following properties: (a) \(\inf\{x,y\}\) exists for all \(x,y\in S\), and \(x\wedge y=\inf\{x,y\}\); (b) if \(x\vee y\) exists, then \(x\vee y\) is the supremum of \(x\) and \(y\); (c) if \(y\vee z\) exists, then \((x\wedge y)\vee(x\wedge z)\) exists for all \(x\in S\). A JP-semilattice \((S;\wedge,\vee)\) is said to be distributive if for all \(x,y,z\in S\) such that \(y\vee z\) exists we have \(x\wedge(y\vee z)=(x\wedge y)\vee(x\wedge z)\). Note that this notion does not always coincide with the customary notion of distributivity for semilattices. (In particular, every lattice can be made into a distributive JP-lattice by defining \(x\vee y\) only if \(x\) and \(y\) are comparable.) A JP-lattice with \(0\) is called pseudocomplemented (in short, PJP-semilattice) if \((S;\wedge)\) is a pseudocomplemented meet-semilattice. The pseudocomplement of \(a\) is denoted by \(a^*\). A distributive PJP-semilattice \((S;\wedge,\vee,^*,0,1)\) is said to be a Stone JP-semilattice if for any \(a\in S\), \(a^*\vee a^{**}\) exists and equals \(1\). The following characterization is proved: A distributive PJP-semilattice is a Stone JP-semilattice if and only if \(a^*\vee b^*\) exists and \((a\wedge b)^*=a^*\vee b^*\) for all \(a,b\in S\). The concepts of ideal and congruence of a PJP-semilattice \((S;\wedge,\vee,^*,0)\) are defined as usual. Clearly, every kernel of a congruence is an ideal. Conversely, it is shown that an ideal is a kernel of a congruence (called kernel ideal) if and only if for all \(x\in S\), \(x\in I\) implies \(x^{**}\in I\). Finally, the following theorem is proved: For a Stone JP-semilattice, the set of all kernel ideals forms a complete lattice, where the meet-operation is the intersection and the join-operation is given by \(\bigvee_{t\in T}I_t=\{x\in S\mid x\leq i_1\vee\cdots\vee i_n\text{ for some }i_j\in I_{t_j},\;1\leq j\leq n\}\).