Zoom link: https://sasupenn.zoom.us/my/kodnerglow Presentation times: - April 10 16:00 CEST / 10:00 EST - April 15 15:00 CEST / 9:00 EST The forms of the Latin past participle are notoriously varied and difficult to predict. As such, they have inspired decades-worth of treatments in theoretical morphology and have even served as motivating case studies for some approaches (e.g., Matthews 1972, Aronoff 1994, Embick 2000, Steriade 2016). One reoccurring issue is the role of regularity in the system. It is universally agreed that some forms are not predictable and have to be listed in some fashion. It is also clear that some generalizations exist in the system regarding theme vowels, root length in syllables, perfect stem formations, or others, but it is far less obvious which of these patterns deserve status in the grammar. The task that a theoretician faces in deciding what to include as regular is analogous to the one that children encounter when acquiring their native grammars, which suggests a learners' perspective approach to the problem. By grounding the split between regulars and irregulars in child language acquisition rather than by researchers' intuition, we can better distinguish between contrasting theoretical accounts and gain new insights into the diachrony of the system.