We investigate the impact of literacy on the language use of Turkish as a heritage language in two communities: Germany and the U.S. that differ from each other in the size of the community, vitality of Turkish, and access to formal education in Turkish. We concentrate on the use of non-canonical post-verbal position in heritage Turkish. In SOV-language Turkish, it is characteristic of spoken and informal registers, while in formal registers, there is normative pressure to restrict this position (Schroeder 1995). In monolingual Turkish, formal written texts show a highly limited use of this position (Schroeder et al., in review). The data come from the RUEG project (Wiese et al. 2019), with 3 groups of participants (each n=64): Turkish heritage speakers (15-18 and 23-35), raised in the U.S. and Germany, and Turkish monolinguals. The data were controlled for register (formal/informal, spoken/written). Comparison of heritage Turkish texts in the U.S. and in Germany revealed that despite the fact that the societal language (German) is SOV and English is a SVO language, heritage Turkish speakers in the U.S. did not have a higher number of structures placed post-verbally. In fact, they used post-verbal structures slightly less than monolinguals and heritage speakers in Germany (Schroeder et al. in review). On the other hand, heritage Turkish in the U.S. is levelling out the registers, although the extend of inter-speaker heterogeneity in the use of the post-verbal structure is higher in the U.S. than in Germany. We will correlate our findings with metadata, namely, information about the speakers’ formal education in Turkish, self-assessment of their productive and receptive skills, and language use in different written media. We expect that literacy will be one of the factors that is responsible for the differences between the groups.