Confluent set-valued functions and inverse limits (Q1987229)

For a metric continuum \(Y\) let \(2^{Y}\) denote the hyperspace of all nonempty closed subsets of \(Y\). The function \(f: X \to 2^{Y}\) from a continuum \(X\) to the hyperspace \(2^{Y}\) is called \textit{set-valued function}. The set \(G(f) = \{(x,y) \in X \times Y : y \in f(x) \}\) is the graph of \(f\). In [Topology Appl. 228, 486--500 (2017; Zbl 1378.54033)] \textit{J. P. Kelly} introduced a notion of monotone and confluent set-valued function: the upper semi-continuous function \(f: X \to 2^{Y}\) is \textit{monotone} (\textit{confluent}) if the projections \(\pi _{X} : G(f) \to X\) and \(\pi _{Y} : G(f) \to Y\) are monotone (confluent). Considering inverse limits with set-valued functions, J. P. Kelly shows that if each factor space is an arc and the bonding functions are monotone set-valued functions, then the projection maps from the inverse limit and from the inverse graph are also monotone. He asks if analogous results would be true in the case of confluent set-valued functions. In the paper under review the authors give a positive answer to this question. In [Commentat. Math. Univ. Carol. 23, 183--191 (1982; Zbl 0486.54029)] \textit{W. J. Charatonik} proved that an inverse limit, whose factor spaces are continua with the property of Kelley and the bonding (single-valued) maps are confluent has the property of Kelley. In the above-mentioned paper Kelly asks if this result is still true in the case of inverse limits with set-valued functions. The authors give a negative answer to this question.