The aim of this article is threefold: to examine the communis opinio on the facts of derivation in Proto-Indo-European vis-a-vis the attested data, to analyse to what degree the prevailing approach is the result of a priori notions on the workings of internal derivation, and to propose an alternative scenario for the development of the relevant forms, compatible with both the data and the universal principles of derivation that appear to be violated in this reconstruction.1