The main concepts of gravitational lens theory are introduced on the basis of spacetime geometry without assuming approximations. The singularities of light cones, in particular their caustics, are reviewed as examples of singularities of Lagrangian resp. Legendrian maps. It is indicated how the usual approximate lens theory may be derived from the general framework. After the discovery of the double quasar QSO O957 + 561 A, B by Walsh, Carswell and Weymann in 1979, gravitational lensing rapidly developed into a major tool of astrophysics, providing infomation about cosmological parameters, masses and mass distributions on the scales of stars, galaxies, galaxy clusters and that of the universe at large. It enables astronomers to obtain information about dark matter, the structure of quasars and very distant, early generation galaxies up to redshifts of z ≈ 5. Usually, gravitational lens theory is based on plausible assumptions and various approximations designed for astrophysical applications. Physical notions and relations are expressed essentially in the framework of classical geometrical optics, with minimal input from general relativity. While such an approach is useful for the intended pur- pose, it conceals the spacetime-geometrical origin of lensing phenomena. Moreover, by using ab initio simplifications based on intuition, one foregoes the possibility to assess the accuracy of approximations, and one may not even recognize which general relativistic relations are being approximated. Besides, such presentations may render it difficult for relativists, used to think in terms of light cones, timelike world lines and the like, to understand what it's all about. Be that as it may, here I want to outline how the basic qualitative relations of gravitational lens theory may be introduced as part of Lorentzian spacetime geometry a I dedicate this paper to George Ellis, with affection and gratitude for fourty years of many stimulating encounters and sharing of ideas.
Read full abstract