Gender Agreement in Heritage Serbian: A First Study

Abstract

This study investigates nine child heritage speakers’ gender agreement in Serbian, with German being the dominant language. We hypothesized that our participants will display different stages of the gender system development found with (Slavic) monolinguals and bilinguals, in which low-frequent non-canonical grammatical suffixes get to be interpreted as regular, canonical endings, resulting in attributive agreement errors among speakers. The results from an elicited production task confirm that speakers rely on morphophonological cues to determine noun gender, the lower their proficiency is. On the other hand, the advanced speakers exposed agreement patterns similar to our monolingual control group. Expectedly, the overall age was found to have a positive effect (when the proficiency is not disparate), as both older child bilinguals and monolinguals (7‑10) demonstrated a more target-like gender agreement system. Finally, our findings show that the advanced participants utilized a three-gender system, slightly simplified than the elaborate one found with monolinguals, while the lowest-ranked subjects exposed a two-gender system (masculine vs. feminine).