No label defined (Q3216728)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 3882519
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | No label defined |
scientific article; zbMATH DE number 3882519 |
Statements
1984
0 references
Merkur'ev theorem
0 references
quadratic forms
0 references
isometry
0 references
invariants
0 references
graded Witt ring
0 references
Galois cohomology
0 references
product of binary forms
0 references
0.8913713097572327
0 references
0.8701564073562622
0 references
0.8439022898674011
0 references
0.8352490663528442
0 references
[For the entire collection see Zbl 0542.00004.] In the algebraic theory of quadratic forms over a field \(F\), a major problem is to find invariants which will classify quadratic forms up to isometry. The present paper continues the search for new invariants and for those fields where the invariants do classify forms. Let \(F\) be a field of characteristic not 2, and \(WF\) the Witt ring of anisotropic forms over \(F\). The graded Witt ring is \(GWF=\coprod_{n\geq 0}\bar I^ nF\), where \(\bar I^ nF=I^ nF/I^{n+1}F\) is the quotient of successive powers of the fundamental ideal IF. The Galois cohomology ring is \(H^*F=\coprod_{n\geq 0}H^ nF\), where \(H^ nF=H^ n(G_ F,{\mathbb{Z}}/2{\mathbb{Z}})\) is the cohomology of the absolute Galois group \(G_ F\). The desired invariants should be maps \(\bar e^ n_ F: \bar I^ nF\to H^ nF\). The map \(\bar e^ 0_ F\) is the trivial isomorphism induced by dimension. If \(a\in \dot F\) let \((a)\in H^ 1F\) be the 1- cocyle corresponding to \(a\dot F^ 2\), using the isomorphism \(H^ 1F\cong \dot F/\dot F^ 2\). Pfister showed that the signed discriminant induces an isomorphism \(\bar I^ 1F\to \dot F/\dot F^ 2\). The isomorphism \(\bar e^ 1_ F: \bar I^ 1F\to H^ 1F\) is then defined by setting \(\bar e^ 1_ F(<\overline{1,-a}>)=(a)\), for \(a\in \dot F.\) Generally, if \(\phi =<1,-a_ 1>... <1,-a_ n>\) is a product of binary forms in \(WF\) then \(\phi \in I^ nF\) and we hope to find maps \(\bar e^ n_ F\) such that \(\bar e^ n_ F({\bar \phi})=(a_ i)\cup... \cup (a_ n)\) in \(H^ nF\). When \(n=2\) this map is a well-defined homomorphism given by the classical Clifford invariant. The major problem here is to determine whether \(\bar e^ n_ F\) is well-defined, and if it is, to determine its kernel and cokernel. If \(\bar e^ n_ F\) is well-defined for all n, then they provide a ring homomorphism \(\bar e^*_ F: GWF\to H^*F.\) The first author [J. Algebra 36, 448-491 (1975; Zbl 0314.12104), Satz 5.7] proved that \(\bar e^ 3_ F\) is a well-defined homomorphism. \textit{A. S. Merkur'ev} [Dokl. Akad. Nauk. SSSR 261, 542-547 (1981; Zbl 0496.16020)] proved that \(\bar e^ 2_ F\) is an isomorphism. The main concern of the paper under review is to discover classes of fields \(F\) where \(\bar e^ n_ F\) is a well-defined isomorphism for all \(n\), that is, where \(\bar e^*_ F: GWF\to H^*F\) is an isomorphism. Theorem. Suppose either (i) \(I^ 4L=0\) for every odd degree extension \(L\) of \(F(\sqrt{-1})\), or (ii) \(I^ 4L\) is linked for every finite extension \(L\) of \(F\). Then \(\bar e^*_ F\) is a well-defined isomorphism. - For example, the hypothesis is satisfied by any field \(F\) of transcendence degree at most 4 over an algebraically closed field, at most 3 over a real closed field or at most 3 over a finite field.
0 references