Individuals unconsciously simulate others’ actions, a crucial element of prediction in joint action (Sebanz, Knoblich & Prinz, 2005). Studies reflect this tendency in tasks in which an individual’s performance is affected by the demands of a coactor’s task, even when the coactor’s task is not relevant to achieving one’s own goal (Sebanz, Knoblich & Prinz, 2003). This is taken to suggest that individuals represent a coactor’s actions (Knoblich, Butterfill & Sebanz, 2011). It is proposed that in the joint action of language, conversation is facilitated by a similar use of prediction (Garrod & Pickering, 2004). As in joint action tasks, the performance of the individual is likely to be affected by a partner’s task. The current study used a consecutive production design whereby participants produced a description of a picture, with the knowledge that their partner would either produce the same or a different description of the picture, or remain silent. It was hypothesised that if their partner was to produce a different utterance, then the participant’s performance would be negatively affected, as opposed to if their partner was to produce the same utterance or to remain silent. The results were close to statistical significance in support of the hypothesis .The other aspect of the study aimed to explore specific social affects on corepresentation. The interpersonal relationship between the participants was manipulated, with participant pairs either being friends or strangers. It was hypothesised that friends would be more likely to represent what their partner was to say and that their performance in this task would be more affected by the nature of their partner’s task. The results partially supported this hypothesis.