Holonomy and gravity (Q1118844)

Drawing on Einstein's concept of ``space of a body'' the author attempts to derive the gravitational field in general relativity, not by prescribing the space-time manifold and its differentiable structure, metric, connection etc., but by reconstructing these objects from certain sets of paths. Following an example using Yang-Mills type theories the method is applied to general relativity and a certain amount of holonomy theory is required (and which can be found in volume 1 of \textit{S. Kobayashi} and \textit{K. Nomizu}'s book ``Foundations of differential geometry'' (New York, 1963; Zbl 0119.375). Paths of particles in space-time with a common end point are to be abstracted so as to ``represent'' that point of the spacetime manifold in the reconstruction. These paths are described in an affine space attached to that point by using the theory of developments in affine holonomy theory. Having constructed the points of the space-time manifold, the differentiable structure, the metric and the connection are introduced by a method described at the beginning of the paper for Yang- Mills type theories and which is again based on holonomy methods. (The reviewer concedes that he found this part of the argument a little obscure.) It is interesting that the topology of space-time is not fixed at the outset (i.e. after the space-time set is constructed). The author advances arguments at various points of the paper to show that his approach distinguishes gravitation from the conventional gauge theories by the physical relevance of the translational geometry of the path of the particle. He also remarks that his approach is free of the ``diffeomorphic symmetry'' (that two space-times are to be regarded as physically the same if they are globally isometric and so one has the problem of factoring out such a symmetry).