How highly connected can an orbifold be? (Q6118203)

Orbifolds are a generalization of smooth manifolds, they are allowed to have singularities of the form of fixed points of finite group actions. Every reduced orbifold can be realized as a quotient space of an almost free action of a compact Lie group on a smooth manifold. An orbifold that is covered by a manifold is called good. An orbifold that is not good is called bad. It is natural to ask if there are ways to tell good and bad orbifolds apart and how to identify manifolds among orbifolds. \textit{M. W. Davis} [Adv. Lect. Math. (ALM) 16, 63--93 (2011; Zbl 1273.57018)], asked whether a contractible orbifold is a manifold. In [Proc. Am. Math. Soc. 141, No. 9, 3303--3304 (2013; Zbl 1272.57020)], \textit{A. Lytchak} proved that a \((2n-2)\)-connected \(n\)-dimensional orbifold is a manifold, and hence answered Davis's question in the positive. In the same paper, Lytchak asked whether there are arbitrarily highly connected bad orbifolds. In this paper, the authors provide a positive answer to Lytchak's question by showing that for any \(n\geq 4\), there exist compact \(\lfloor n/2-1\rfloor\)-connected bad orbifolds in dimension \(n\). They also improve the bound obtained by Lytchak. Namely, they prove that an \(n\)-connected \(n\)-orbifold is a manifold for any \(n\geq 1\). In the odd-dimensional case, they show that a \(2n\)-connected \((2n+1)\)-orbifold is a manifold for any \(n\geq 1\). For compact orbifolds the authors obtain better bounds. They show that a compact \((2n-2)\)-connected \(2n\)-orbifold is a manifold for any \(n\geq 3\). If \(n\geq 3\) and \(n\) is not a power of \(2\), then a compact \((2n-2)\)-connected \((2n+1)\)-orbifold is a manifold. Finally, let \(k(n)\) (resp. \(k_c(n)\)) denote the maximum \(k\) such that there exists a \(k\)-connected, (resp. compact, \(k\)-connected) bad \(n\)-orbifold. The authors give bounds for \(k(n)\) and \(k_c(n)\).