Galois cohomology of fields with a dimension (Q2491833): Difference between revisions

Surgical fields were introduced in 1995 by Poizat and Pillay as fields \(K\) with a reasonably well behaved notion of dimension. For instance, it is required that, for every definable set \(X\) in \(K\) and every definable equivalence relation \(E\) on \(X\), only finitely many classes of \(E\) share the same dimension as \(X\). The paper under review shows that, for every finite extension \(L\) of a surgical field \(K\) and every finite Galois extension \(L_1\) of \(L\), the Brauer group Br\((L_1/L)\) is finite. Moreover, for every algebraic group \(G\) defined over \(L\), the cohomology group \(H^1(L_1/L, \, G)\) is also finite. The proof uses an induction argument on the degree of the extension, reducing the analysis to the cyclic case. In this restricted setting, one regards the groups under investigation as quotient of suitable equivalence relations; so, due to the surgical hypothesis, it is sufficient to prove that each class of these relations has at least the same dimension as the ambient set (which clearly implies the finiteness claims). This is done, at least in the case of the cohomogy group, by a further reduction to abelian varieties.