Abstract A replication of previous research, this study sets out to re-examine language choice in note-taking for consecutive conference interpreting – a topic that is widely believed to be subject to conflicting evidence. Extending the existing database considerably, the study draws on data from seven different consecutive interpreting tasks involving five different languages and both interpreting directions (B into A and A into B) performed by ten participants with seven different language combinations. Interpreters’ notes from these performances served as the main data of the study, which was complemented by questionnaire data to form a mixed-methods design. Analyses of the interpreters’ notes identified the A language as by far the strongest determinant of language choice, above and beyond other language categories; this confirmed the results of the replicated research. The questionnaire data, however, did not mirror the patterns found in the interpreting data consistently, suggesting a complex interplay of behavior and norms. The article concludes with a discussion of these and previous findings, arguing for the topic of language choice in note-taking to be reframed as one of complexity rather than contradiction.