Abstract

In communication, speakers and hearers are thought to take into account the interlocutors’ perspective and common ground – this has been called «audience design». However, it has been much debated to what extent the interlocutors do so, with which timing, whether this requires low- or high-level mindreading abilities, and whether processing is modular. We intend to show that the issue of modularity is at the heart of the debate. Then, we will argue in favor of a non-modular, constraint-based account of audience design, based on well-established knowledge about long term memory, working memory, and the dynamics of brain activation.