The chapter concludes the book bu summarising the main points and indicating future directions and possible applications. We suggest applications such asdisordered dialogue (e.g., involving young children or people with autism) and remote conferencing facilities.

The chapter shows how interlocutors achieve alignment of dialogue models -- that is, both situation models and dialogue game models. Such alignment is the basis of successful dialogue. We discuss the importance of co-reference for alignment of situation models. We then consider the role of meta-representation of aiignment in dialogue and how this controls what people choose to say next. We consider the relationship between focal alignment of dialogue models and what is in the shared workspace. Finally, we discuss the relationship between alignment and common ground.

The chapter discusses efficient use of the limited shared workspace and how interlocutors work together to achieve it. We introduce the notion of commentary (both positive and negative) that reflects interlocutors' confidence in focal alignment. We show how positive commentary leads directly to succinct expressions and negative commentary leads to fuller expressions. As a result, positive and negative commentary promote alignment while minimizing collaborative effort.