Hereditarily irreducible maps (Q2216598)

A mapping between continua \(f:X\rightarrow Y\) is hereditarily irreducible, if \(f(A)\subsetneq f(B)\) for any subcontinua \(A\) and \(B\) such that \(A\subsetneq B\). For the case that \(X\) is an arc, \textit{B. Espinoza} and \textit{E. Matsuhashi} [Topology Appl. 190, 74--92 (2015; Zbl 1321.54058)] studied these mappings under the name of arcwise increasing mappings. In the paper under review, the authors introduce a new way to define the order of a point in a continuum, they use this notion to make a detailed study of hereditarily irreducible mappings between graphs. They extend the results by Espinoza and Matsuhashi. They also investigate the existence of these mappings from a graph onto dendrites (locally connected continua without simple closed curves). A very interesting result in this direction is the following (Theorem 6.5). THEOREM. For a dendrite \(D\), the following conditions are equivalent. (a) The set of ramification points of \(D\) is dense in \(D\). (b) \(D\) contains no free arc. (c) The set of end points of \(D\) is dense in \(D\). (d) For any graph \(G\) there is a hereditarily irreducible mapping from \(G\) onto \(D\). (e) There is a hereditarily irreducible mapping from a simple closed curve onto \(D\). (f) There is a graph \(G\) with no cut points and a hereditarily irreducible mapping from \(G\) onto \(D\). The authors also pose the following problem (Problem 3.23). PROBLEM. Suppose \(X\) is a locally connected continuum in which the union of free arcs is dense, all points of \(X\) are of finite order, and \(X\) has at most two points of odd order. Does there exist a hereditarily irreducible mapping from \([0,1]\) onto \(X\)?