The book emphasizes that in collaborative learning environments particular forms of interaction are needed to trigger the desired learning mechanisms. There is, however, no guarantee that those interactions will actually occur. Hence, a general concern across the book is to develop ways of increasing the probability that they will happen. The chapters illustrate these points relatively well. Particularly, the book focuses on collaboration between two or a few human or artificial agents for a well-defined learning or problem-solving task.
In chapter 2, Littleton and Hakkinen outline some of the theoretical approaches currently informing the analysis of learning interactions and demonstrate how the study of collaborative activity in computer-based settings can enhance the understanding of the processes of productive interaction.
Chapter 3 discusses relations between grounding, collaboration and learning. According to Baker, Hansen, Joiner and Traum, grounding is the process by which the participants of a collaborative learning task maintain some degree of mutual understanding. They illustrate the perspective with reference to a particular computer-mediated collaborative learning situation in the domain of physics.
Chapters 4 and 5 focus on multi-agent learning. First, Weiss and Dillenbourg in chapter 4 provide a general characterization of multi-agent learning from two different perspectives: the perspective of single-agent learning (the ``machine learning perspective'') and the perspective of human-human collaborative learning (the ``psychological perspective''). They also offer a brief guide to agents and multi-agents systems as studied in artificial intelligence, and suggest directions for future research on multi-agent learning. In chapter 5, Joiner, Issroff and Demiris describe some differences between multi-agent systems and cognitive psychology. The chapter reviews recent research concerning the analysis of multi-robot systems and research on human-human computer-supported collaborative learning and illustrates how researchers in both areas can benefit from each other's fields.
The relation between dialogue with oneself and with a peer is addressed by Ploetzer, Dillenbourg, Preier and Traum in chapter 6 and by Nguifo, Baker and Dillenbourg in chapter 7. They compare the basic operators used to model the co-construction of knowledge through dialogue and those used in machine-learning research to model individual learning. These chapters are rather modest in their scope and in their results. Their modesty has, however, to be compared with the width of the gap that separates cognitive psychology and DAI (Distributed Artificial Intelligence).
One of the weakness of the book is that it focuses only on cognitive psychology and neglects the socio-affective aspects of collaborative learning. Social and affective aspects are crucial to understanding human-human collaboration. Yet, the editor says that this was intentional,
