On mappings preserving orthogonality of non-singular vectors (Q1176176): Difference between revisions

Let \(V\) be a vector space over a commutative field \(F\) carrying a quadratic form \(q\). Let \(V^*:=\{v\in V\mid q(v)\neq0\}\) and define \(\Omega(v,w):=f(v,w)^ 2 q(v)^{-1}q(w)^{-1}\) where \(v,w\in V^*\) and \(f\) denotes the symmetric bilinear form associated with \(q\). For a mapping \(\varphi: V^*\to V^*\) the authors study the following properties (I), (II) and (III). (I) \(f(v,w)=0\Leftrightarrow f(v\varphi,w\varphi)=0\) for all \(v,w\in V^*\); (II) \(\Omega(v,w)=\Omega(v\varphi,w\varphi)\) for all \(v,w\in V^*\) (III) \(f(v,w)^ 2=f(v\varphi,w\varphi)^ 2\) for all \(v,w\in V^*\). The theorems include infinite-dimensional vector spaces and yield e.g. the following result. If \(V\) has finite dimension \(\geq3\) and \(| F|>3\) then each of the assumptions (I), (II) and (III) implies that there is a \(\rho\)-linear injective mapping \(\psi: V\to V\) and some \(\lambda\in F\backslash\{0\}\) such that \(Fv\varphi=Fv\psi\) and \(f(v\psi,w\psi)=\lambda(f(v,w)\rho)\) for all \(v,w\in V^*\). Furthermore, if (II) and (III) holds then \(\rho=id_ F\).