This paper will be report on a study, which is a part of the empirical project "Learning mathematics in the context of linguistic and cultural diversity in classrooms". The research is realized in seventh grade classes of six schools in Hamburg. It focuses on the question if pupils with bilingual background "learn" mathematics different from those with monolingual background. The project compares pupils with Turkish and Russian backgrounds with German speaking non-immigrant pupils. Four areas are examined: learning and teaching in mathematics classrooms, language skills in German and Turkish/Russian, problem solving processes of pupils and socio-cultural background variables. The study presented in this paper takes into account, that growing up in a multilingual family not only affects pupils` language proficiency and their coping with other languages, but the whole learning process. Hence, the focus of attention is the proficiency in both languages (German/Turkish or German/Russian) of pupils from immigrant families. The data from language testing allows to draw conclusions not only on language proficiency, but also on the knowledge and application of mathematical terminology and "decontextualised" classroom language. Moreover, the manner in which pupils deal with mathematical tasks - from the linguistic as well as from the mathematical point of view - is investigated. Based on the empirical results the conception of mathematical literacy will be discussed. It is the author's intention to point out that a conception of mathematical literacy should consider the pupils` language skills and the bilingualism as an important individual resource.