The behavior of the \(\nu\)-invariant of a field of characteristic 2 under finite extensions (Q919023): Difference between revisions

The \(\nu\)-invariant was introduced for fields of characteristic \(\neq 2\) by \textit{R. Elman} and \textit{T. Y. Lam} [Math. Ann. 219, 21--42 (1976; Zbl 0302.10025)]. The authors define the \(\nu\)-invariant for any field \(F\) of characteristic 2 as follows: \(\nu(F)=\min \{n\mid I^ nW_q(F)=0\}\). Here \(I^nW_q(F)\) is the submodule \(I^n(F)\cdot W_q(F)\), where \(I^n(F)\) is the \(n\)-th power of the fundamental ideal of the Witt ring \(W(F)\) of symmetric bilinear forms over \(F\) and \(W_q(F)\) is the \(W(F)\)-module of quadratic forms over \(F\). The authors prove that for any finite field extension \(L/F\) one has \(\nu(F)\leq \nu(L)\leq \nu(F)+1.\) When \(\operatorname{char} F\neq 2\), it is still an open question whether \(\nu(L)\leq \nu(F)+1\). If \(L/F\) is separable, the authors establish that \(s_*(I^nW_q(L))=I^nW_q(F)\), where \(s_*\) is the transfer homomorphism, and also have a corresponding exact result for the extension of the scalar homomorphism \(i^*\). If \(L/F\) is a purely inseparable extension, they show \(\nu(F)=\nu(L)\).