Insertion of lattice-valued and hedgehog-valued functions (Q820101)

The present paper continues the development of insertion-type theorems for lattice-valued functions. For this study, one has to begin with an appropriate definition of semicontinuity for functions taking values in a complete lattice. The most flexible versions of semicontinuity for functions defined on a topological space seem to be the continuities with respect to the upper topology \(\nu(L)\) and the lower topology \(\widetilde v(L)\), where the range space \(L\) is a completely distributive lattice with a countable join-dense subset consisting of non-supercompact elements (such a lattice from now on will be called a \(\triangleleft\)-separable lattice). The authors prove an analogue of the classical insertion theorem of Lane for \(L\)-valued functions, where \(L\) is as above. As a consequence they establish several nice results, namely the \(L\)-version (with a considerably simpler proof) of the Katětov-Tong-type theorem of Liu and Luo and the \(L\)-valued version of the insertion theorem of Stone. Also, the authors, in their more detailed study of completely distributive lattices, prove that these lattices are closed under the formation of countable products. More precisely they show that the Hilbert cube is a \(\triangleleft\)-separable lattice and contains a join dense subset, which is both order and topologically isomorphic to the hedgehog \(J(\omega)\) with appropriately defined topologies. Note that their insertion theorem for hedgehog-valued functions is independent of that of Blair and Swardson. At the end, they prove a necessary and sufficient condition for a double insertion theorem and obtain, as corollaries, among others, characterizations of hereditarily normal spaces and extremally disconnected ones.