4-dimensional compact projective planes with a 7-dimensional collineation group (Q2638523)

The study of 4-dimensional compact projective planes was initiated by H. Salzmann: the group of all continuous collineations \(\Sigma\) of such a plane is a Lie group of dimension at most 16 and the desarguesian plane over \({\mathbb{C}}\) is the only plane admitting a collineation group of dimension at least 9 (then the full collineation group has dimension 16). Further, if dim \(\Sigma\) \(=8\), the plane is isomorphic or dual to a translation plane [\textit{H. Salzmann}, Math. Z. 130, 235-247 (1973; Zbl 0239.50014)]. All 4-dimensional compact translation planes with dim \(\Sigma\) \(=8\) or 7 were completely classified by the author in a series of papers. In the paper under review the author finishes the classification of 4-dimensional compact projective planes admitting a 7- dimensional collineation group thereby contributing a major step towards the classification of all flexible 4-dimensional compact projective planes. It is proved that besides translation planes and their duals there is, up to isomorphism, only one further projective plane with a 7- dimensional collineation group; this is the shift plane of Knarr obtained by shifting the graph of the function \({\mathbb{R}}^ 2\to {\mathbb{R}}^ 2:\) \((x,y)\mapsto (xy-\frac{x^ 3}{3},\frac{y^ 2}{2}-\frac{x^ 4}{12})\)- shifts are translations of \({\mathbb{R}}^ 4\) but not necessarily translations of the new plane in the geometrical sense. Shift planes were investigated by N. Knarr and among all shift planes there is exactly one (the one generated by the function above) which has a 7-dimensional collineation group [\textit{N. Knarr}, Topologische Differenzenflächenebenen, Diplomarbeit, Kiel (1983)]. There are two partial results on the classification of 4-dimensional compact projective planes admitting a 7-dimensional collineation group obtained by others. Firstly, if the connected component of \(\Sigma\) is non-solvable, then this leads to translation planes or dual translation planes as was shown by \textit{R. Löwen} [Geom. Dedicata 36, No.2/3, 225- 234 (1990; Zbl 0712.51011)]. Secondly, the case that \(\Sigma\) contains a subgroup isomorphic to \({\mathbb{R}}^ 4\) (not necessarily translations) has been studied by \textit{N. Knarr} [J. Geom. 31, No.1/2, 114-124 (1988; Zbl 0644.51011)]; only shift planes, translation planes and dual translation planes can occur in this situation. The author deals with the remaining cases, that is a solvable connected 7-dimensional group \(\Sigma\) that contains no subgroup isomorphic to \({\mathbb{R}}^ 4\). Up to duality, \(\Sigma\) then fixes a flag (v,W) and acts transitively on \({\mathcal L}_ v\setminus \{W\}\cong {\mathbb{R}}^ 2\) (\({\mathcal L}_ v\) being the line pencil through v) or one has a translation plane. Using an explicit list of solvable connected Lie groups transitive on \({\mathbb{R}}^ 2\) and analyzing the possible actions of \(\Sigma\) it is proved that \(2\leq \dim \Sigma_{[v]}\leq 3.\) The case dim \(\Sigma\) \({}_{[v]}=2\) leads to a translation plane. An important tool to achieve this result (as well as the upper bound dim \(\Sigma\) \({}_{[v]}\leq 3\) above) is the powerful homology and elation method in which the set of centres of all non-identical homologies in a closed Lie subgroup of the group of all axial collineations with a fixed axis is compared to orbits of all elations contained in the Lie subgroup. This was thoroughly studied by \textit{H. Hähl} [Topology Appl. 12, 49-63 (1981; Zbl 0446.51010)]. The case dim \(\Sigma\) \({}_{[v]}=3\) leads to a plane whose dual plane has a 3-dimensional translation group with axis W. Such planes were studied by the author [Geom. Dedicata 16, 179-193 (1984; Zbl 0547.51012)]. With methods developed in that paper the 7-dimensional collineation group is explicitely constructed and it is proved that in this situation a subgroup isomorphic to \({\mathbb{R}}^ 4\) must result.