In 1959, \textit{J. Stallings} [Fundam. Math. 47, 249--263 (1959; Zbl 0114.39102)] asked the following question: Suppose that \(\tau\) is a connected topology on \([0,1]\) finer than the usual Euclidean topology and \(\tau_L\) (respectively, \(\tau_R\)) the topology generated by \(\tau\) together with all half-closed intervals of the form \([a,b)\) (respectively \((a,b])\); if \(L,R\) are subsets of \([0,1]\) such that \(0\in L\), \(1\in R\), \(L\cup R=[0,1]\) and \(L\in\tau_L\), \(R\in\tau_R\), then is it true that \(L\) and \(R\) must intersect? A negative answer was given to this question by \textit{S. K. Hildebrand} [Fundam. Math. 61, 133--140 (1967; Zbl 0185.26101)], but in the paper under review, the authors prove that if the topology \(\tau\) is maximal connected (that is, no strictly finer topology than \(\tau\) is connected), then the question has an affirmative answer.
This page was built for publication: On a problem of J. Stallings
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2850640)
