Squares of positive \((p,p)\)-forms (Q1942072)

A \((p,p)\)-form \(\alpha\) in \(\mathbb C^n\) is called positive (\(\alpha\geq0\)) if for \((1,0)\)-forms \(\gamma_1,\dots,\gamma_{n-p}\) one has \[ \alpha\wedge i\gamma_1\wedge\bar\gamma_1\wedge\dots\wedge i\gamma_{n-p}\wedge\bar\gamma_{n-p}\geq0. \] It is well known that if \(\alpha\) is positive and \(\beta\) is a positive \((1,1)\)-form, then \(\alpha\wedge\beta\) is positive, too. However, this is no longer true for arbitrary positive \((q,q)\)-forms \(\beta\) (cf.~[\textit{R.~Harvey} and \textit{A.~W.~Knapp}, Value-Distrib. Theory, Proc. Tulane Univ. Progr. Value-Distrib. Theory complex Analysis related Topics differ. Geom., Part A, 43--62 (1974; Zbl 0287.53046)]). In 2006 \textit{S. Dinew} [``On positive \(\mathbb C_{(2,2)}(\mathbb C^4)\) forms'', Preprint, 2006] constructed positive \((2,2)\)-forms \(\alpha,\beta\) in \(\mathbb C^4\) such that \(\alpha\wedge\beta<0\). The aim of the paper under review is to show that \(\alpha^2\) is positive whenever \(\alpha\) is a positive \((2,2)\)-form. Moreover, this phenomenon holds only for \((2,2)\)-forms. Indeed, the authors show that for every \(p\geq3\) there is a positive \((p,p)\)-form \(\alpha\) in \(\mathbb C^{2p}\) such that \(\alpha^2<0\).