Many languages exhibit Differential Argument Marking (DAM), where nouns are marked for their grammatical role only in certain contexts. DAM is subdivided into Differential Object Marking (DOM) and Differential Subject Marking (DSM), both of which are conditioned by factors such as animacy across languages and hypothesized to arise via ambiguity avoidance and tendencies to mark atypical situations. While previous studies have primarily focused on production, this study employs artificial language learning to investigate comprehension in both DOM and DSM, examining whether learners rely on typicality biases for grammatical role assignment in unmarked sentences, and whether the type of marking learned affects their interpretation. Results indicate a strong givenness effect and a smaller animacy effect, with no significant differences between DOM and DSM conditions. These findings suggest that typicality biases play a key role in shaping DAM systems and that their emergence is best understood as a communicative phenomenon.