The Dutch and German morphological feminisation systems, i.e., the morphology serving the marking of female sex on personal nouns, are analysed in this manuscript. For this, diachronic data were collected, which substantiate the observation that feminisation is being less actively made use of in Dutch than in German. The reasons for these diverging phenomena were explored empirically in three different corpus studies: 1) a study that aims to describe the Dutch and German feminisation systems in various text genres/registers (newspapers, chats, tweets) synchronically, and which also focuses on the variable productivity degrees of different morphemes (e.g., -in and -ster) diachronically from Middle Dutch / Middle High German until Modern Dutch and New High German; 2) a diachronic study, comprising nearly 200 years of data, concerning the use of feminising morphology in non-referential predicative constructions in newspapers in four language areas, namely Netherlandic Dutch, Flemish, and East and West Germany between 1946 and 1990; 3) a corpus study concerned with feminised personal nouns that have an inanimate referent (e.g., German die Partei als Verliererin der Wahlen). These studies demonstrate that the main factors contributing to loss or stability of feminisation systems are grammatical gender, language policy, referentiality, semantics of the personal noun, and social and register-bound connotations (as is the case for the Dutch suffix -e). These findings have some implications for current debates concerning gender-fair language use in that they show that the binarily coded German system of expressing gender on nouns (e.g., Autor–Autorin) is diachronically very stable, which in turn helps explain why some speakers see the necessity use this system as the basis for further non-binary differentiation (e.g., Autor–Autorin–Autor*in). The reverse is (increasingly) true of Dutch, where non-feminised personal nouns are increasingly being used in female contexts.
