Supporting the generation of argument structure within video sequences