This paper investigates a range of instances of collaborative talk in English and French in which both participants use similar talk at roughly the same time. It is demonstrated that prediction is being employed in these utterances to determine the form of the next lexical item and to identify larger syntactic structures.