A characterization of the \(\hat A\)-genus as a linear combination of Pontrjagin numbers (Q2634774): Difference between revisions

Let \(\mathcal A\) be the class of all compact \(4k\)-dimensional spin manifolds admitting a non-trivial circle action. \textit{M. F. Atiyah} and \textit{F. Hirzebruch} [in: Essays Topol. Relat. Top., Mem. dédiés à Georges de Rham 18--28 (1970; Zbl 0193.52401)] proved that the rational linear combinations of Pontrjagin numbers which vanish on \(\mathcal A\) are precisely the multiples of the \(\hat A\)-genus. Later, \textit{M. Gromov} and \textit{H. B. Lawson jun.} [Ann. Math. (2) 111, 423--434 (1980; Zbl 0463.53025)] obtained a similar result with \(\mathcal A\) replaced by the class \(\mathcal B\) of compact simply connected spin manifolds admitting a Riemannian metric with positive scalar curvature. They obtained this as a corollary to their result that \(\mathcal B\) is closed with respect to spin cobordism. The present author obtains the corollary by a different method. He uses the original ideas of Atiyah and Hirzebruch together with the Calabi-Yau theorem to prove that a rational combination of Pontrjagin numbers must be a multiple of \(\hat A\) if it vanishes on all \(4k\)-dimensional simply connected spin manifolds admitting a metric whose Ricci curvature is nonnegative and not identically zero.