On the descriptive complexity of homogeneous continua (Q2665214)

For a~Polish space~\(X\) the space \(2^X\) of all nonempty compact subsets of~\(X\) is a~Polish space. Let \(C(X)\) and \(\mathcal{H}(X)\subset2^X\) denote the set of all subcontinua of~\(X\) and the set of all homogeneous subcontinua of~\(X\), respectively. The set \(C(X)\) is a~closed subset of~\(2^X\) but to determine the complexity of \(\mathcal{H}(X)\) is not trivial. In the paper the author tries to determine the complexity of \(\mathcal{H}(I^n)\) for \(2\le n\le\omega\) where \(I=[0,1]\). He concludes that \(\mathcal{H}(I^2)\) is Borel in \(C(I^2)\). A~naive direct analysis of the homogeneity property shows that \(\mathcal{H}(I^n)\) is~\(\Pi^1_2\) for \(n\ge3\). In the paper the author proves that \(\mathcal{H}(I^n)\) is a~\(\Sigma^1_1\)-subset of \(C(I^n)\) and \(\mathcal{H}(I^n)\) has a~closed subset homeomorphic to~\(c_0\) which is an absolute \(F_{\sigma\delta}\)~set. Consequently \(\mathcal{H}(I^n)\) is not \(G_{\delta\sigma}\) for all \(n\ge2\).