Quadratic forms in 8 variables such that their Clifford algebra is of index 8 (Q1379514)

Let \(F\) be a field of characteristic \(\neq 2\) and let \(\phi\) be an anisotropic quadratic form over \(F\) of dimension \(8\) and trivial signed discriminant \(d_{\pm}\phi = 1\in F^*/F^{*2}\). The author continues his study to find necessary and sufficient conditions for \(\phi\) to become isotropic over \(F(\psi )\), the function field of another anisotropic quadratic form \(\psi\) over \(F\), which he began in [\textit{A. Laghribi}, Duke Math. J. 85, 397-410 (1996; Zbl 0865.11030)]. There, the author dealt with the case where the index of the Clifford algebra of \(\phi\) is \(1\), \(2\), or \(4\). In the present paper, the author treats the remaining case of index \(8\). The main results of both papers can be summarized as follows. Let \(\phi\) be as above and let \(\dim\psi\geq 5\). Then \(\phi\) becomes isotropic over \(F(\psi )\) if and only if there exists a form \(\psi'\) of dimension \(8\) such that \(\psi\) is similar to a subform of \(\psi'\) and either \(\psi'\) is a \(3\)-fold Pfister form and \(\phi\) contains a Pfister neighbor of \(\psi'\) (this case cannot occur if the index is \(8\)) or \(\phi\perp\psi'\) is similar to a \(4\)-fold Pfister form. The results he obtains in the cases \(\dim\psi =3, 4\) seem to be less conclusive. Some of the ingredients in the proof of the index \(8\) case are interesting in their own right. Let \(D\) be the division part of the Clifford algebra of \(\phi\). Since the index is \(8\), \(D\) is a product of three quaternion algebras. Let \(L\) be the function field of the Severi-Brauer variety associated to \(D\). Using results by Karpenko and Peyre, the author proves that the kernel of the map \(H^3F\rightarrow H^3L\) is given by \([D]H^1F\), where \(H^nF\) denotes the \(n\)th Galois cohomology group of \(F\) with coefficients in \({\mathbb Z}/2{\mathbb Z}\), and \([D]\) denotes the class of \(D\) in \(H^2F = Br_2F\), the \(2\)-torsion part of the Brauer group of \(F\). Furthermore, he uses the fact that \(\phi\) stays anisotropic over \(L\), a consequence of a deep theorem due to \textit{H. Esnault, B. Kahn, M. Levine} and \textit{E. Viehweg} [J. Am. Math. Soc. 11, No. 1, 73-118 (1998)].