Compactness of the automorphism group of a topological parallelism on real projective 3-space (Q2406498)

The first author and \textit{R.~Riesinger} [Adv. Geom. 14, 175--189 (2014; Zbl 1291.51011)] showed that the automorphism group \(\Phi=\Aut\Pi\) of a topological parallelism \(\Pi\) of the real projective 3-space \(\mathrm{PG}(3,\mathbb{R})\) is a closed subgroup of the collineation group \(\mathrm{PGL}(4,\mathbb{R})\) of \(\mathrm{PG}(3,\mathbb{R})\), and thus is a Lie group. Moreover, the dimension of \(\Phi\) is at most 6, and one has a Clifford parallelism if \(\Phi\) is at least 5-dimensional. In the paper under review, the authors prove that the identity component \(\Phi^1\) of \(\Phi\) is compact and hence is conjugate to a (closed, connected) subgroup of \(\mathrm{PSO}(4,\mathbb{R})\). They furthermore conjecture that the automorphism group of every topological parallelism on \(\mathrm{PG}(3,\mathbb{R})\) is compact. To prove their result, the authors examine all non-compact one-parameter subgroups of \(\mathrm{PGL}(4,\mathbb{R})\) up to conjugation and one-by-one determine whether such a group can leave a topological parallelism invariant. After some further reduction, there are nine cases depending on the eigenvalues of an infinitesimal generator of a one-parameter subgroup. Many cases are excluded by eliminating certain unbounded cyclic subgroups acting on 4-dimensional translation planes obtained from suitable spreads in the parallelism, further supporting the authors' conjecture. The second author [``A characterization of Clifford parallelism by automorphisms'', Innov. Incidence Geom. (to appear)] uses the present result to show that \(\dim \Phi \geq 4\) characterizes the Clifford parallelism.