It is commonly assumed in language contact studies that languages spoken by multilinguals become more like one another, which leads to long-term convergence between the languages. This does not lead to a sweeping linguistic uniformity, however: there is still an incredible diversity across the world’s languages. In this talk, I present my ongoing study of language contact between the Mano and the Kpelle languages of Guinea and show how, despite the potential of convergence, languages remain distinct. Language dominance seems to play a decisive role in the extent of cross-linguistic influence, which is the strongest in individuals born in monolingual Kpelle families and residing in Kpelle-dominant settings. One unexpected result, however, is that speakers residing in a balanced bilingual setting pattern together with speakers residing in a Mano setting, regardless of their family language, contradicting frequent claims of language convergence in communities in contact. Another surprising finding is that speaker’s comprehension and expectations of form-to-meaning correspondence are strongly affected by the way language is spoken around them, which is an indication of (perhaps, implicit) awareness of sociolinguistic variation. I further discuss the implications of both these findings for the studies of social meaning of grammatical variation.