Proto-Indo-European verbal morphology is generally described as consisting of at least a “thematic” and an “athematic” conjugations, which differ in whether or not a fixed vocalic piece adjacent to the root (traditionally known as “thematic vowel”) is present. This paper investigates the behavior of the outcomes of the thematic vowels in three ancient Indo-European languages: Sanskrit, Ancient Greek, and Latin. We show that, on the one hand, Latin thematic vowels are typically “ornamental”, in that they lack any morpho-syntactico-semantic information, and are used exclusively for classification purposes; on the other hand, Sanskrit and Ancient Greek thematic vowels are fully functional, in that they expone (at least) aspectual information. We argue for a diachronic account of such differential treatment of thematicity across the three languages, whereby the original functionality of such pieces, albeit fully preserved in Sanskrit, was gradually lost over time, and finally gave rise to the Latin verbal ornamental system.