Lefschetz properties and basic constructions on simplicial spheres (Q2269532): Difference between revisions

A homology sphere is a pure simplicial complex all of whose links have the same homology as spheres of the appropriate dimension. The \(g\)-conjecture for a homology sphere generalizes the \(g\)-theorem for a simplicial polytope, which states (in the authors' terms) that its \(g\)-vector is the \(f\)-vector of a multicomplex. (Reviewer's note: the general case was raised as a question in his original paper [\textit{P. McMullen}, Isr. J. Math. 9, 559--570 (1971; Zbl 0209.53701)] proposing the \(g\)-conjecture for polytopes, even with some doubts expressed; it was not, as the authors suggest, actually conjectured there.) The necessity of the conditions of the \(g\)-conjecture would follow from the strong Lefschetz property for the face ring (modulo the usual linear system of parameters) of a homology sphere. In this interesting paper, the authors go some way towards proving the more general \(g\)-conjecture -- at least for piecewise-linear spheres -- by showing that it is preserved by the constructions of join, connected sum and stellar subdivision. The reviewer's own proof [Invent. Math. 113, No.2, 419--444 (1993; Zbl 0803.52007)] for simple polytopes, which avoided the algebraic geometry of \textit{R. P. Stanley}'s original proof [Adv. Math. 35, 236--238 (1980; Zbl 0427.52006)], obtained the strong Lefschetz property from the quadratic Hodge-Riemann-Minkowski inequalities in the dimension below; it might be asked if there are analogous inequalities for homology spheres.