On non-metric continua that support Whitney maps (Q2451998)

In the present discussion, a \textit{continuum} is a connected compact Hausdorff space. For any compact Hausdorff \(X\), \(2^X\) is the hyperspace of closed nonempty subsets of \(X\), endowed with the Vietoris topology. \(2^X\) is also well known to be compact Hausdorff in this topology; important closed subspaces include \(F_n(X)\) (the subsets with at most \(n\) points), and \(C(X)\) (the subcontinua). If \(\mathcal{H}\) is a subspace of \(2^X\), a \textit{Whitney map on} \(\mathcal{H}\) is a continuous function \(\mu:\mathcal{H}\to [0,\infty)\) such that: (i) \(\mu(\{x\})=0\) for \(x\in X\); and (ii) if \(A\) and \(B\) are distinct members of \(\mathcal{H}\) and \(A\subseteq B\), then \(\mu(A)<\mu(B)\). In [\textit{J. J. Charatonik} and \textit{W. J. Charatonik}, Colloq. Math. 83, No. 2, 305--307 (2000; Zbl 0953.54013)], it is proved that for a compact Hausdorff \(X\), \(X\) is metrizable iff there is a Whitney map on \(2^X\), iff there is a Whitney map on \(F_2(X)\); and the problem is posed to characterize those nonmetrizable continua \(X\) for which there is a Whitney map on \(C(X)\). They show that such a continuum \(X\) cannot be a dendron, as in that case there is an embedding of \(F_2(X)\) into \(C(X)\). They also show that the continuum \(K\) featured in [\textit{A. Gutek} and \textit{C. L. Hagopian}, Proc. Am. Math. Soc. 86, 169--172 (1982; Zbl 0489.54031)] is indeed such a continuum. In the paper under review, this example \(K\) is developed in several clever ways to produce a host of nonmetrizable continua that have Whitney maps on their subcontinuum hyperspaces. The basic idea behind \(K\) is: (i) start with a totally disconnected compact Hausdorff \(Z\) (a double-arrow space, say); (ii) carefully choose a homeomorphism \(h:Z\to Z\) with dense orbits; and (iii) form the quotient space \((Z\times [0,1])/\sim\), where points \(\langle z,0\rangle\) and \(\langle h(z),1\rangle\) are identified by \(\sim\). (Such spaces may be justifiably thought of as ``generalized solenoids''.) Among the topological properties possible for examples produced along these -- and similar -- lines include: (1) being perfectly normal, and having all proper subcontinua be metrizable; (2) failing to be perfectly normal, but still having all proper subcontinua be metrizable; and (3) having all nondegenerate subcontinua be nonmetrizable. The authors observe that if \(X\) is a continuum for which \(C(X)\) has a Whitney map, then each order arc in \(C(X)\) is metrizable; they then ask whether this necessary condition is also sufficient.