Totally decomposable involutions and quadratic pairs
Determining whether a central simple algebra is isomorphic to the tensor product of quaternion algebras is a classical question. One can also ask similar decomposability questions when there is additional structure defined on the central simple algebra, for example an involution. We may ask whether an involution on a central simple algebra is isomorphic to the tensor product of involutions defined on quaternion algebras, i.e. whether the involution is totally decomposable.
Algebras with involution can be viewed as twisted symmetric bilinear forms up to similarity, and hence also as twisted quadratic forms up to similarity if the characteristic of the underlying field is different from 2. In a paper of Bayer, Parimala and Quéguiner it was suggested that totally decomposable involutions could be a natural generalisation of Pfister forms, a type of quadratic form of central importance to the modern theory of quadratic forms. In this talk we will discuss recent progress on the connection between totally decomposable involutions and Pfister forms.
We will also discuss fields of characteristic 2, where, since symmetric bilinear forms and quadratic forms are no longer equivalent, involutions are not twisted quadratic forms. Instead, if one wants a notion of a twisted quadratic form with analogous properties to involutions, one works with objects introduced in the Book of Involutions, known as a quadratic pairs. One can define an analogous notion of total decomposability for quadratic pairs, and there is a connection to Pfister forms very similar to that found between involutions and Pfister forms in characteristic different from 2.