The Structure of Multimodal Dialogue II ed. by M.M. Taylor, F. Néel, D.G. Bouwhuis (review)

342 letters in canada 2001 university of toronto quarterly, volume 72, number 1, winter 2002/3 similar, but compatible, goals. First, it is a reference grammar. But further, it is an introduction to linguistics. For example, the first chapter defines language and dialect; a typical reference grammar would name and discuss language and dialect but without background on how these differ. Likewise , in a traditional grammar, no setting would be provided in discussion of the sound system; the system would be presented under the assumption that readers know the basic parameters of articulatory phonetics. In NRG, on the other hand, discussion of the Nishnaabemwin sound system is placed in a larger setting detailing how sounds are made, what the vocal tract looks like, etc. My sense is that embedding the grammar within a general introduction to linguistics does not detract from the grammar, and that linguists will find the book useful even though it provides background that they have. The audience of teachers engaged in the promotion and preservation of Nishnaabemwin will be greatly aided by the presentation. NRG differs from standard reference grammars in another, less dramatic, way. Example sentences come from texts, and often background is given to help the user understand the subtlety of use of a construction. One of the many highlights of NRG is the wealth of examples, with each grammatical claim being amply illustrated. NRG, clearly a work of tremendous love, represents many years of work, and is a major contribution to Algonquian linguistics. In the preface, Valentine states what probably anyone who has devoted years of his or her life to writing a grammar feels: >Writing a grammar is a profoundly humbling experience. Languages are almost unbelievably complex and represent the richest traditions that we as human beings possess. What remotely compares with them, whether we are considering the massively intricate and fluid physical gestures involved in the articulation of sounds, or the systems behind the thousands of possibilities of distinct expression in the verb system of a language such as Nishnaabemwin?= Having dabbled in Algonquian languages, I have frequently been frustrated. Even trivial things like abbreviations often make no sense to me. One of the many accomplishments of NRG is to demystify the Algonquian literature. As a linguist, I now find myself far better able to think about Algonquian languages. Valentine is to be congratulated for the accessible way in which he leads the reader through this fascinating language; the field is indeed fortunate to have scholars of his calibre. (KEREN RICE) M.M. Taylor, F. Néel, and D.G. Bouwhuis, editors. The Structure of Multimodal Dialogue II John Benjamins 2000. xviii, 522. US $125.00 Face-to-face human communication involves more than the interchange of vocal signals. In normal circumstances, people consciously or subcon- humanities 343 university of toronto quarterly, volume 72, number 1, winter 2002/3 sciously use various other modes instead of or in parallel with voice in order to communicate. These may include gestures, facial expressions, or body language. They may also make use of other media to facilitate communication, e.g., pictures, graphs, or recordings. With more advances in speech recognition, discourse analysis, and human-computer interaction (HCI), the idea of simulating natural conversation involving all these various modes and media between humans and computers is becoming less and less far fetched. This has led to the emergence of a new and exciting interdisciplinary enterprise that brings together researchers from linguistics, psychology, cognitive science, philosophy, and several fields of computer science including speech technology, AI, software engineering, and HCI. The book under discussion is the fruit of the second workshop on the Structure of Multimodal Dialogue held in Aquafredda di Maratea, Italy, in 1991. It has certainly taken the editors a very long time to compile this volume, especially since it deals with such a rapidly evolving discipline. Nevertheless, it is worth the wait. The book comprises twenty-eight chapters/papers divided into a prologue and two parts. The prologue (three chapters) discusses the structure of dialogue, the history of HCI, and the phases that it has gone through since the advent of computers. Part 1 (eleven chapters) presents some theoretical models and strategies taken...