On Dowling geometries of infinite groups (Q1881688)

Let \(G\) be a group and let \(m\) be a natural number. Then the point set of the Dowling geometry \(Q_m(G)\) of dimension \(m\) of \(G\) consists of the points \(p_i\), \(0\leq i\leq m\) (with \(p_i\) a symbol not contained in \(G\)), and for each ordered pair \((i,j)\), \(1\leq i<j\leq m\), a copy \(G_{i,j}\) of the set of all elements of \(G\). Denoting the copy of \(\alpha\in G\) in \(G_{i,j}\) by \(\alpha_{i,j}\), the lines of \(Q_m(G)\) are the sets \(\{p_i,p_j\}\cup G_{i,j}\), for all \(0\leq i<j\leq m\), and the triples \(\{\alpha_{i,j},\beta_{j,k},(\alpha\beta)_{i,k}\}\), for \(0\leq i<j<k\leq m\). All pairs of points not contained in one of these lines constitute by definition a line of size two. In the paper under review the authors construct a finite geometric structure that embeds into the Dowling geometry of dimension 3 of an infinite group \(G\), but which cannot be embedded in any finite Dowling geometry. This is in contrast with a result of \textit{R. Rado} [Proc. Lond. Math. Soc., III. Ser. 7, 300--320 (1957; Zbl 0083.02302)] stating that every finite geometry that embeds in some projective or affine space also embeds in a finite one. This weakens the analogy between projective and Dowling geometries. Some striking similarities were noted before by \textit{J. Bonin} [Matroid theory. Contemporary Mathematics 197, 417--418 (1996; Zbl 0852.00021)], who also put forward the question negatively answered in the paper under review. The example of the paper uses the finitely generated infinite simple group discovered by \textit{G. Higman} [J. Lond. Math. Soc. 26, 61--64 (1951; Zbl 0042.02201)].