Topological projective geometries (Q1079810): Difference between revisions

There had been various definitions of a topological finite-dimensional projective geometry [\textit{H. Lenz}, Vorlesungen über projektive Geometrie (1965; Zbl 0134.162), \textit{J. Misfeld}, Abh. Math. Sem. Univ. Hamburg 32, 232-263 (1968; Zbl 0164.208), \textit{K. Sörensen}, Math. Z. 109, 239-245 (1969; Zbl 0181.233)]. The crucial point was to have a topology on the point set P such that projections from points onto hyperplanes becomes continuous, or to have topologies on the point set, the set of all lines,..., the set of all hyperplanes such that the geometric operations of joining and intersecting become continuous with respect to these topologies. The author does not longer stick to the geometric setting of a projective geometry with points, lines, planes etc. for his definition of a topological projective geometry (TPG). He starts from the well known interpretation of a projective geometry as an irreducible modular lattice \({\mathcal L}=(L,\vee,\wedge,0,1)\) of length at least 3 with \(\vee,\wedge\) being the lattice operations of joining and intersecting, and 0,1 being the lowest resp. greatest element of \({\mathcal L}\). Let \((L\times L)_ 0=\{(s,t)\in L^ 2: s\wedge t=0\},\) \((L\times L)_ 1=\{(s,t)\in L^ 2: s\vee t=1\}\), and let \(\tau\) be a topology on L. Then the author defines (\({\mathcal L},\tau)\) to be a TPG if and only if \[ {\dot \vee}: (L\times L)_ 0\to L: (s,t)\mapsto s\vee t\text{ and }{^{q}\wedge}: (L\times L)_ 1\to L: (s,t)\mapsto s\wedge t \] are continuous mappings with open domains. Note that this definition does not make any finiteness restriction to the dimension of the geometry, and that only one simultaneous topology is needed, which induces topologies on the sets \({\mathcal L}_ k\) of all k- dimensional subspaces. As it is shown this definition of a TPG comprises the standard definition of a topological projective plane as well as in the finite-dimensional case the definitions of a TPG mentioned above. The author shows that the usual construction of a (d-1)-dimensional projective geometry over an infinite topological division ring D gives a TPG in his sense if one defines the topology \(\tau\) in the following way: Let \(\tau_ D\) be the topology of D and let \(\tau^ d_ D\) be the product topology of the d-dimensional right vector space \(V=D^ d\) over D; for \(M\in \tau^ d_ D\) let \(\tilde M\) be the set of all subspaces of V which intersect M not trivially. Then \(\{\) \(\tilde M:\) \(M\in \tau^ d_ D\}\) together with \(\{\) \({\mathcal L}_ k: k=0,...,d-1\}\) forms a subbase for \(\tau\). Finally it is shown that these are the only arguesian TPG's of finite dimension, i.e. an arguesian TPG always possesses a coordinatization over a topological division ring if the dimension is finite.