Reviewed by Remi van Trijp, Sony Computer Science Laboratory Paris
One of the most unfortunate facts about the field of linguistics is that the importance of artificial intelligence (AI) and robotics has been erased from the field’s collective memory, even though many ideas in contemporary linguistics were directly inspired by the work of AI researchers such as Ron Kaplan, Roger Schank, Eric Wanner, Terry Winograd, and William Woods.
This book is a good place to start for making amends. In the introductory chapter, the editors explain the major challenges of human-robot interaction (HRI). Not only does HRI require an ‘embodied agent’ that can perceive and act intelligently in a real-world environment, the field also needs to understand how communication with and learning from an often unpredictable human interlocutor is possible. Linguists will thus find operational solutions to hard problems, such as how to achieve joint attention and how to make use of contextual information for communication.
The book is divided into three parts. The first part, ‘The human in the loop’, focuses on how HRI is possible in everyday situations with non-expert human ‘users’. Apart from their technical significance, these chapters also provide important insights to linguists interested in dialogue. For instance, Ch. 1 ‘Helping robots imitate’, by Aris Alissandrakis, Dag S. Syrdal, and Yoshihiro Miyake, shows the importance of appropriate feedback to achieve more natural communication. Manja Lohse’s chapter, ‘The role of expectations and situations in human-robot interaction’, demonstrates how successful communication requires the capacity to predict the behavior of the other interlocutor. The remaining chapters touch upon subjects such as the human interlocutor’s attitudes and sociality.
The second part of the book, ‘Joint action, collaboration and communication’, handles social learning, collaborative activities, and language acquisition, and is most relevant for linguists working from a constructivist perspective. The most notable chapter is Ch. 10 ‘The acquisition of word semantics by a humanoid robot via interaction with a human tutor’, in which Joe Saunders, Chrystopher L. Nehaniv, and Caroline Lyon show that the linguistic behavior of human tutors resembles child-directed speech, and that a robot can use the tutor’s input to acquire words for referring to objects in its environment.
The final part, ‘Robots in therapy, safety and communication’, discusses how robots open up exciting new possibilities in a.o. healthcare and is, therefore, of interest to linguists who are seeking new ways to investigate and treat speech and language impairments. In Ch. 14 ‘Rehabilitation robots’, Farshid Amirabdollahian surveys how robots can be employed for the rehabilitation of humans who suffer from brain injury or spinal cord injuries. Similar work has been performed by one of the book’s editors, Kerstin Dautenhahn, on using robots as a tool for education and therapy for children with autism, which is sadly missing from the current book.
In sum, this book does not contain much of what one might classify as ‘traditional linguistics’, but it offers a wealth of fully operational studies and methods that may push the state-of-the-art in linguistics. The contributions never become too technical and should appeal to a broad, interdisciplinary audience.