Determining how and why adult outcomes of heritage speaker (HS) bilingualism differ from monolinguals is difficult because it requires the reconstruction of developmental paths from end-state data. In an effort to address this issue, we examine HSs of Turkish in Germany at an early age of development (10-15 years old, n=22), as well as age-matched monolingual controls in Turkey (n=20) and Germany (n=20), using a structured elicitation task for production of passives. The goal is to see whether HSs have the representation of passives in their mental grammars and to better understand the relative weight of factors (age at time of testing, immigration status of the Turkish parents (first or second generation), and literacy in the L1) that potentially contribute to the formation of HSs’ grammatical competence. The results show that all HSs have the underlying representation for passives in both Turkish and German. There was a significant effect of only literacy; high level of L1 literacy has a positive effect on monolingual-like production as compared to those with no literacy. We discuss these results pertaining to explicating ultimate attainment outcomes in heritage language acquisition in relation to larger debates in the field.