This article explores the long-standing problem of ascribing meaning to individual acts of codeswitching. Drawing on ethnographic data from bilingual classrooms in Corsica, I situate the analysis of codeswitching within the more general question of the interpretation of speaker stance, which is defined as speakers' positioning with regard to both the content and the form of their utterances. Close analysis of recorded classroom interaction focuses on the teacher's codeswitches, and documents the way she distributes Corsican and French across different classroom discourse functions. This analysis reveals how the teacher simultaneously strives for balance and parity between the two languages, while privileging Corsican as a language of discovery and learning. In addition, the analysis reveals how this relates to the social and political context of Corsican language learning. Finally, this data is used to make an argument for the necessity of an ethnographically grounded analysis of speaker intentions and stances.