Abstract: Palaeontology is for the most part a postdictive science. That is, its central task is for the most part to explain existing data. A rare instance of palaeontology being a predictive science is to be found in the history of the development of the pseudotribosphenic concept. Based on the structure of the lower molars in a single dentary of a pseudotribosphenic mammal, the structure of the then unknown upper molars was hypothesized. The subsequent discovery first of an isolated upper molar regarded as likely to be a pseudotribosphenic mammal and later, a specimen of a pseudotribosphenic mammal with its upper and lower molars in occlusion, rigorously corroborated the original hypothesis. The acceptance of the pseudotribosphenic concept was instrumental in the recognition that the tribosphenic condition, a key innovation in the history of mammals, could well have arisen more than once, despite its complexity. Living languages evolve through time. As the term “pseudotribosphenic” was utilised after being originally coined, its usage by some subsequent workers departed significantly from the original concept. The term has been used both to describe teeth of docodonts which do not occlude in the pseudotribosphenic manner as originally defined as well as teeth that were clearly tribosphenic as that term was originally defined rather than pseudotribosphenic. As a result, its definition as understood by the paleontological community became less precise, and thus the utility of the concept was significantly reduced. Consequently, a return to the original definition would be appropriate.