Riemannian rigidity of the parallel postulate in total curvature (Q2290800): Difference between revisions

A question of great interest is, as phrased in [\textit{V. Bangert} and \textit{P. Emmerich}, J. Differ. Geom. 94, No. 3, 367--385 (2013; Zbl 1278.53038)]: ``Suppose a complete Riemannian plane \(P\) satisfies the parallel axiom, i.e., for every geodesic \(c\) on \(P\) and every point \(p \in P\) not on \(c\) there exists a unique geodesic through \(p\) that does not intersect \(c\). Does this imply that \(P\) is isometric to the Euclidean plane?'' This paper answers the question in the affirmative under the additional hypothesis that the plane admits total curvature. The key lemma is that such a plane has total curvature zero. Under the hypothesis that the plane admits total curvature, it is possible to compute total curvature by exhaustion by closed subsets. The lemma is shown by first subdividing the plane into strips, then exhausting those strips by triangles. The parallel postulate controls the excess of the large triangles and hence, by Gauss-Bonnet, their total curvature. Note that this phrasing of Euclid's fifth postulate corresponds to Playfair's formulation. The original version states that if a line segment intersects two straight lines so that the two interior angles on the same side sum to less than \(\pi\), then the two lines must intersect on that side. This paper also shows that, with this version of the postulate, the question can be answered in the affirmative under an additional hypothesis; this time that the plane has no conjugate points.