Mathematical constraints and the Tulving-Wiseman law: a rejoinder.