The most known examples of loops L having strong relations to geometry have classical groups as the groups generated by their left translations ( [5], [8], [7],[4], [6], Chapter 9, [10], Chapters 22 and 25, [2], [3]). These groups G may be seen as subgroups of the stabilizer of 0 in the group of affinities of suitable affine spaces An, and as the elements of the loops L may be often taken certain projective subspaces of the hyperplane at infinity of An. The semidirect products T o G, where T is the translation group of the affine space An have in many cases a geometric interpretation as motion groups of affine metric geometries. In the papers [2], [3] three dimensional connected differentiable loops are constructed which have the connected component of the motion group of the 3-dimensional hyperbolic or pseudo-euclidean geometry as the group topologically generated by the left translations and which are Bol, Bruck or left A-loops. The set of the left translations of these loops induces on the plane at infinity the set of left translations of a loop isotopic to the hyperbolic plane loop (cf. [10], Chapter 22, p. 280, [7], p. 189). This and the fact that, up to our knowledge, there are only few known examples of sharply transitive sections in affine metric motion groups, motivated us to seek a simple geometric procedure for an extension of a loop realized as the image Σ∗ of a sharply transitive section in a subgroup G∗ of the projective linear group PGL(n − 1, F ) to a loop realized as the image of sharply transitive section in a subgroup G of the group of affinities of the n-dimensional space An = F n over the commutative field F . Moreover, we desire that G contains a large normal subgroup of affine translations and that α(G) = G∗ holds for the canonical homomorphism α : GL(n, F ) → PGL(n, F ). We show that this goal can be achived if there exists for G∗ in the (n−1)-dimensional projective space Pn−1(F ) an orbitO ofm-dimensional subspaces such that Σ∗ acts sharply transitively on O, there is a subspace of dimension (n− 1−m) having empty intersection with any element of O and if the restriction of α−1 to Σ∗ defines a bijection from α−1(Σ∗) to Σ∗.
Read full abstract