Spaces whose only finite-sheeted covers are themselves. I (Q1854130)

This paper focuses on those connected topological spaces that have the property that all of their finite-sheeted connected covering spaces have a total space homeomorphic to the base space. There are two ways that a space \(M\) can satisfy this property: either (1) \(M\) has no trivial finite-sheeted covers or (2) \(M\) has a \(k\)-fold connected cover \(p: X\to M\) for some \(k\geq 2\), and the total space of every connected finite-sheeted cover \(p: X\to M\) is such that \(X\) is homeomorphic to \(M\). In this part I of the paper the author considers those spaces that satisfy the first condition. Usually, the considered space \(M\) is a metric continuum, a compact connected metric space. Then in analogy to profinitely trivial groups, a connected metric space \(M\) is called cofinitely self-similar if whenever \(p:X\to M\) is a finite-to-one connected covering space, it follows that \(X\) is homeomorphic to \(M\); \(M\) is cofinitely trivial if and only if it has no trivial finite-sheeted covers. As a trivial example, any simply connected space is cofinitely trivial. Also, other interesting examples are mentioned: the topologist's sine curve, any continuum that is either tree-like or a nonseparating planar continuum, cf. \textit{A. Y. W. Lau} [Bull. Acad. Pol. Sci., Sér. Sci. Math. Astron. Phys. 26, No. 4, 315-317 (1978; Zbl 0394.54017)], the spaces constructed by \textit{H. B. Griffiths} [Q. J. Math., Oxf. II. Ser. 5, 175-190 (1954; Zbl 0056.16301)], or the join of two copies of the cone on the Hawaiian earring, cf. \textit{E. H. Spanier} [Algebraic topology (1966; Zbl 0145.43303), p. 84]. Then the author presents what is known about cofinitely trivial spaces: ``A calculus of noncovering'', a Lau's theorem [loc. cit.] and a theorem of \textit{G. Jungck} and the author [Topol. Proc. 24, 173-187 (1999; Zbl 0970.54016)] about a covering projection \(p: X\to M\) with \(M\) the inverse limit of an inverse sequence \((M_n, f_n)\) in some special conditions. As a corollary it follows that a countable product of cofinitely trivial metric compacta is also cofinitely trivial. Another mentioned ``test'' to determine that a metric continuum is cofinite is a theorem of \textit{A. Tominaga} [Fundam. Math. 119, No. 1, 1-6 (1983; Zbl 0537.54024)] about a local homeomorphism \(f: X\to M\) with \(X\) an inverse limit. These results are applied to obtain new examples of cofinitely trivial spaces. One such example, suggested by a paper of \textit{S. Mardešić} and \textit{V. Matijević} [Topology Appl. 113, No. 1-3, 167-209 (2001; Zbl 0989.57002)] about overlay structures, is the following: Suppose that \(X_\infty= \underleftarrow{\lim} {\mathcal H}\) is a solenoid from \({\mathcal H}= (S^1,f_n)\), where \(f_n: S^1\to S^1\) is a prime order covering projection of order \(p_n\) and such that for each prime \(p\) there are infinitely many \(n\in\mathbb{N}\) for which \(p_n= p\). Then \(X_\infty\) is cofinitely trivial. The last section presents what is known in the context of spaces with more structure such as low-dimensional manifolds and cell complexes. These results refer to some finitely generated 3-manifold groups and have as conclusion the following conjecture of the author: (a) There do not exist any nonsimply connected cofinitely trivial compact 3-manifolds. (b) The only compact cofinitely trivial 3-manifolds are the closed 3-ball, \(S^2\times I\), and \(S^3\). The paper also contains statements of other interesting related questions and problems which the author was concerned with.