On topological actions of finite groups on \(S^{3}\) (Q1694807)

A 1941 theorem of Kérékjartò (correcting Brouwer and Eilenberg; see the 1994 proof by Constantin-Kolev) states that any compact subgroup of \(\mathrm{Homeo}(S^2)\) is conjugate to a subgroup of \(O_3\). In [Trans. Am. Math. Soc. 97, 193--212 (1960; Zbl 0144.22902)], \textit{G. M. Fisher} proved that the group \(\mathrm{Homeo}^+(S^3)\) of orientation-preserving homeomorphisms is simple as a group, is locally path-connected as a topological space, and is an open index-two subgroup of \(\mathrm{Homeo}(S^3)\). The analogous statement cannot hold for \(\mathrm{Homeo}(S^3)\), as follows. In 1954, Bing produced an example of an orientation-reversing involution of the 3-sphere with fixed-point set the Alexander horned sphere. In the same year, Montgomery-Zippin produced an orientation-preserving example with fixed-point set a wild knot. So conjugacy of any finite subgroup of \(\mathrm{Homeo}^+(S^3)\) into \(SO_4\) cannot hold either. On the other hand, in 1984 Hatcher proved the Smale conjecture: the inclusion of \(O_4\) into \(\mathrm{Diff}(S^3)\), with the \(C^\infty\) topology, is a homotopy equivalence. So there exists a stark difference between periodic \(C^0\) and \(C^\infty\) transformations, even for the 3-sphere. Given these caveats, in this article, the author shows that any finite subgroup \(G\) of \(\mathrm{Homeo}^+(S^3)\) is abstractly isomorphic to a subgroup of \(SO_3\), assuming that each element of \(G\) has a fixed point. As noted above, a conclusion of conjugacy cannot be expected. Given by extension of a homeomorphism to the one-point compactification, a corollary is that any subgroup of \(\mathrm{Homeo}^+(\mathbb{R}^3)\) is abstractly isomorphic to a subgroup of \(SO_3\). Thus, up to isomorphism, the non-smoothable finite subgroups of \(\mathrm{Homeo}^+(\mathbb{R}^3)\) occur in the list: all cyclic groups \(C_n\), all dihedral groups \(D_n = C_n \times_{-1} C_2\), the symmetric group \(S_4\), the alternating groups \(A_4\) and \(A_5\). The methods of proof are low-dimensional algebraic topology, group cohomology, and Smith theory for fixed points. Another theorem herein is that any \textit{simple} finite subgroup of \(\mathrm{Homeo}^+(S^3)\) is abstractly isomorphic to a prime-cyclic group \(C_p\) or \(A_5\). This proof is only outlined, and refers to prior work of the author [Topology Appl. 125, No. 2, 199--202 (2002; Zbl 1012.57028) and with \textit{M. Mecchia} [J. Algebra 298, No. 2, 460--467 (2006; Zbl 1098.57012)]. Reviewer's question: Is each finite subgroup of \(\mathrm{Homeo}(S^3)\) abstractly isomorphic to a subgroup of \(O_4\)?