Isometries in hyperbolic spaces (Q2379246)

The characterization of geometric transformations under mild hypotheses is a well-established area of research, going back to Liouville and Darboux, on which \textit{Walter Benz} has published three comprehensive monographs [Geometrische Transformationen. Unter besonderer Be\-rück\-sich\-ti\-gung der Lorentztransformationen. Mannheim: B. I. Wissenschaftsverlag (1992; Zbl 0754.51005), Real geometries, B. I. Wissenschaftsverlag, Mannheim (1994; Zbl 0819.51002), and Classical geometries in modern contexts. Geometry of real inner product spaces. Basel: Birkhäuser (2005; Zbl 1084.51001), (2008; Zbl 1135.51002)]. Based on deep results obtained by \textit{A. Brezuleanu} and \textit{D. C. Rădulescu} [Abh. Math. Semin. Univ. Hamb. 55, 171--181 (1985; Zbl 0595.51006)], \textit{R. Höfer} [J. Geom. 61, No.~1--2, 56--61 (1998; Zbl 0899.51014)] proved that a surjective self-map \(f\) of the \(n\)-dimensional real hyperbolic space which maps collinear points into collinear points must be a hyperbolic motion (he also showed that the surjectivity condition can be removed if the image of \(f\) is not a line, and \(f\) maps lines onto lines). For \(n=2\), the reviewer has shown in [Acta Math. Hung. 100, No.~1--2, 63--67 (2003; Zbl 1027.51021)], by providing positive definitions of point-distinctness \(\neq\) and of non-collinearity in terms of collinearity, that the result holds in hyperbolic planes over Archimedean ordered fields. The logical aspect of characterizations of geometric transformations under mild hypotheses has been emphasized by the reviewer in [Indag. Math., New Ser. 11, No.~3, 453--462 (2000; Zbl 0987.51010)] and elsewhere. Ignoring all references cited above, the present paper is devoted to proving Höfer's result and a slight generalization thereof (surjective self-maps \(f\) of the \(n\)-dimensional real hyperbolic space which map \(r\)-dimensional linear subspaces into \(r\)-dimensional subspaces must be hyperbolic motions --- which follows easily from Höfer's result by using Proposition 3.1 of this paper, which is deemed ``obvious'' by the authors, that if \(f\) maps \(r\)-dimensional linear subspaces into \(r\)-dimensional subspaces, then it must map lines into lines), which was called ``Conjecture 2'' in [\textit{B. Li} and \textit{G. Yao}, Math. Proc. Camb. Philos. Soc. 147, No.~2, 439--446 (2009; Zbl 1178.51014)], and where a remark lets the reader know that it was ``recently'' proved by \textit{G. Yao} [in a paper to appear in Math. Nachr.]. Given this history, the interest the paper might have can be seen in providing an alternate proof to Höfer's theorem, by proving the result (for \(n>2\), as in case \(n=2\), the reviewer's paper already provided a very simple self-contained proof) without using the result of the A. Brezuleanu and D. C. Rădulescu paper referred to in Höfer's paper.