Pseudo-Riemannian manifolds with recurrent spinor fields (Q376197)

The author deals with Riemannian and pseudo-Riemannian spin-manifolds. In chapter 2 ``Preliminaries'' he considers the complexification \(CCl(r,s)\) of the Clifford algebra being assigned to a real pseudo-Euclidian vector space with signature \((r,s)\) and indicates a confusing definition of a true representation of \(CCl(r,s)\) depending on \(r\) and \(s\). Why? In order to get ``\(Sp(r,s)\)''? It cannot change the following fact: The algebra \(CCl(r,s)\) is also the Clifford algebra assigned to the complexifications of the underlying vector space and the quadratic form. But here the signature gets lost: a simple transformation of the basis extends to an automorphism of \(CCl(r,s)\) and leads to \(Sp(n)\)!